Correcting diffusion artifacts in urinary oxygen tension monitoring Authors: Ali Ramezani, Natalie A. Silverton, Laurence M. Saint-Pierre, Guillaume L. Hoareau, Kai Kuck Published: 2026 ABSTRACT Urinary oxygen tension (PuO2) is a promising continuous biomarker of renal medullary oxygenation, but distal catheter measurements are distorted by transit delay and wall diffusion, especially during oliguria. We developed and validated a physics-based algorithm that reconstructs bladder PuO2 from the distal signal. We formulated complementary steady-state and transient diffusion-transport models that couple wall diffusion to a lumen mass balance and solve an inverse problem to estimate bladder PuO2. Performance was assessed on a bench-top catheter system with controlled flows and in eight swine undergoing hemorrhagic shock and resuscitation (68,788 paired samples). On the bench top, adjustment reduced mean bias relative to bladder PuO2 from − 10.9 to −0.64 mmHg, narrowed 95% limits of agreement from 39.1 to 15.8 mmHg, and improved correlation with ground truth (R2 0.53→0.94). In vivo, pooled bias decreased from + 20.9 ± 11.5 to −4.1 ± 13.5 mmHg and mean absolute error from 21.1 to 9.5 mmHg, with the largest gains at low urinary flow. Modeling transport delay and wall diffusion enables reconstruction of bladder PuO2 from distal catheter measurements, supporting reliable, real-time urinary hypoxia monitoring and broader application to other conduit-based analyte measurements.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1038/s41598-026-53461-z. Introduction Acute kidney injury (AKI) is a serious complication across medical and surgical settings, affecting up to 60% of intensive care unit patients and occurring in up to 42% of adult cardiac surgery patients1–4. Even mild AKI is associated with sharply increased mortality, intensive care unit length of stay, and healthcare costs5–7. Standard AKI diagnostic criteria, such as a rise in serum creatinine concentration as defined by the Kidney Disease: Improving Global Outcomes (KDIGO) guidelines, often take several hours or days to manifest7–9, hampering early intervention. Novel biomarkers (such as neutrophil gelatinase-associated lipocalin or cystatin C) can predict AKI sooner, but require intermittent manual sampling and lab processing9–11. There is a pressing need for a continuous, real-time, non-laboratory monitor of renal damage that can signal impending AKI, allowing clinicians to intervene before irreversible damage occurs. One key pathophysiological driver of AKI is inadequate renal oxygenation, particularly in the vulnerable renal medulla12. The outer medulla has a high metabolic demand and relatively low blood flow, making it prone to hypoxic injury during systemic insults like sepsis, major surgery, cardiopulmonary bypass, or heart failure13. Directly monitoring renal medullary oxygenation in patients is invasive and impractical clinically. However, the urinary tract presents an appealing possibility: when urine is first formed in the nephron and enters the collecting ducts, its urinary oxygen tension (PuO2) equilibrates with that of the surrounding medullary tissue. Thus, the oxygen tension of recently produced urine has been proposed as a surrogate window into the renal medullary oxygen partial pressure14,15. By the time urine reaches the bladder, some oxygen exchange may have occurred across the ureter, potentially confounding the measurement. Nevertheless, studies have shown that changes in bladder urinary oxygen tension track with changes in renal medullary oxygenation and hemodynamics13,15. Continuous urinary oxygen monitoring could, therefore, serve as a real-time indicator of renal medullary oxygen insufficiency, and thus of AKI risk. When urinary catheterization is already clinically indicated, adding a distal PuO2 sensor does not require additional instrumentation within the urinary tract and preserves the standard catheter workflow16. However, a significant challenge in translating urinary oxygen monitoring to routine clinical use lies in the dynamics of urine flow and transit time. If urine flow is very low or intermittent (oliguria), the urine may linger in the catheter for extended periods, during which the oxygen tension can drift due to ongoing diffusion17. In other words, the transit of urine from bladder to the sensor not only delays but also distorts the PuO2 readings, reducing their reliability as real-time indicators of kidney status. In prior cardiac-surgery data, bladder-to-distal urine transit times commonly ranged from approximately 10 to 30 min during low-flow states, estimated from catheter dead volume and measured urine output (~ 10–30 mL/h)18. Here, low-flow refers to states in which measured urine output is low enough to prolong catheter transit time, rather than to a visual assessment of bag filling. Such prolonged transit causes the corresponding O2 measurements to reflect renal O2 that occurred much earlier. Clearly, accounting for urine transit time is crucial to ensure that the measured PuO2 accurately reflects current renal medullary oxygenation. To overcome this limitation, our recent work has focused on computational solutions to estimate and correct for urine transit delays19. However, a second important factor is radial oxygen diffusion across the urinary catheter wall. Urinary catheters are made of oxygen-permeable polymers (e.g., silicone and latex); when urine flow is low, molecular oxygen from the surrounding tissues (intra-abdominal catheter segment) and room air (extracorporeal catheter segment) readily diffuses inward or outward, altering the PuO2. This diffusion bias is larger during periods of oliguria when clinicians most need a reliable renal oxygen signal, and it cannot be eliminated by simply discarding low-flow data. Unless wall-diffusion is adjusted, PuO2 will systematically and nonlinearly overestimate true medullary oxygenation. Accordingly, the objective of the present study was to develop and validate a mathematical model that quantifies and compensates for catheter-wall oxygen diffusion, allowing distal PuO2 readings to reflect bladder PuO2 and, consequently, medullary oxygen tension across a broad spectrum of urine flows. Building on our previously validated transit-time algorithm, we developed and validated diffusion-compensation models, and we hypothesized that they would reduce diffusion-induced error in both bench-top (in vitro) and large animal (in vivo) experiments19. Methods Mathematical modeling To adjust the distortion that PuO2 acquires while transiting from the bladder to the distal sensor, we developed two mathematical models based on the principles of mass transport. The first was an analytical steady-state model that provides a foundational, closed-form solution. The second was a more computationally intensive transient model that accounts for the dynamic nature of diffusion within the catheter wall. General assumptions Both models are based on a core set of physical assumptions: Laminar flow: The maximum urinary flow does not exceed 1300 mL/h. Under this condition, and for a catheter of an inner diameter of 4 mm, the Reynolds number (Re) is ≈ 115. Because Re < 2300, the flow in the catheter was therefore taken to be laminar under all operating conditions measurable by the sensor.Well-mixed lumen: At any given cross-section, the PuO2 is assumed to be radially uniform within the urine. This is a lumped-parameter assumption that neglects radial concentration gradients within the lumen itself. This assumption refers to radial uniformity within a cross-section, not axial uniformity along the catheter.Only radial diffusion: Oxygen transport occurs exclusively in the radial direction, with axial diffusion in both the urine and the wall material being considered negligible. This is justified because axial advection dominates axial diffusion over the catheter lengths and flow rates studied here. For example, for a 4 mm inner-diameter catheter at 1300 mL/h, urine traverses a 20 cm segment in approximately 7 s, whereas the corresponding axial diffusion length of oxygen is only about 0.1 mm. Even at 10 mL/h, the axial diffusion length is only about 1.3 mm, which remains much smaller than the catheter length.Constant environment and properties: The environmental oxygen pressures (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{air}$$\end{document}) and the material properties (diffusion coefficient \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{D}_{wall}\:$$\end{document}, Henry’s solubility coefficient H) are assumed to be constant along their respective segments and the catheter wall is treated as having a uniform effective diffusivity within each segment.Surface equilibrium: The oxygen concentration at any material surface is in equilibrium with the environment it is in contact with. Laminar flow: The maximum urinary flow does not exceed 1300 mL/h. Under this condition, and for a catheter of an inner diameter of 4 mm, the Reynolds number (Re) is ≈ 115. Because Re < 2300, the flow in the catheter was therefore taken to be laminar under all operating conditions measurable by the sensor. Laminar flow: The maximum urinary flow does not exceed 1300 mL/h. Under this condition, and for a catheter of an inner diameter of 4 mm, the Reynolds number (Re) is ≈ 115. Because Re < 2300, the flow in the catheter was therefore taken to be laminar under all operating conditions measurable by the sensor. Well-mixed lumen: At any given cross-section, the PuO2 is assumed to be radially uniform within the urine. This is a lumped-parameter assumption that neglects radial concentration gradients within the lumen itself. This assumption refers to radial uniformity within a cross-section, not axial uniformity along the catheter. Well-mixed lumen: At any given cross-section, the PuO2 is assumed to be radially uniform within the urine. This is a lumped-parameter assumption that neglects radial concentration gradients within the lumen itself. This assumption refers to radial uniformity within a cross-section, not axial uniformity along the catheter. Only radial diffusion: Oxygen transport occurs exclusively in the radial direction, with axial diffusion in both the urine and the wall material being considered negligible. This is justified because axial advection dominates axial diffusion over the catheter lengths and flow rates studied here. For example, for a 4 mm inner-diameter catheter at 1300 mL/h, urine traverses a 20 cm segment in approximately 7 s, whereas the corresponding axial diffusion length of oxygen is only about 0.1 mm. Even at 10 mL/h, the axial diffusion length is only about 1.3 mm, which remains much smaller than the catheter length. Only radial diffusion: Oxygen transport occurs exclusively in the radial direction, with axial diffusion in both the urine and the wall material being considered negligible. This is justified because axial advection dominates axial diffusion over the catheter lengths and flow rates studied here. For example, for a 4 mm inner-diameter catheter at 1300 mL/h, urine traverses a 20 cm segment in approximately 7 s, whereas the corresponding axial diffusion length of oxygen is only about 0.1 mm. Even at 10 mL/h, the axial diffusion length is only about 1.3 mm, which remains much smaller than the catheter length. Constant environment and properties: The environmental oxygen pressures (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{air}$$\end{document}) and the material properties (diffusion coefficient \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{D}_{wall}\:$$\end{document}, Henry’s solubility coefficient H) are assumed to be constant along their respective segments and the catheter wall is treated as having a uniform effective diffusivity within each segment. Constant environment and properties: The environmental oxygen pressures (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{air}$$\end{document}) and the material properties (diffusion coefficient \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{D}_{wall}\:$$\end{document}, Henry’s solubility coefficient H) are assumed to be constant along their respective segments and the catheter wall is treated as having a uniform effective diffusivity within each segment. Surface equilibrium: The oxygen concentration at any material surface is in equilibrium with the environment it is in contact with. Surface equilibrium: The oxygen concentration at any material surface is in equilibrium with the environment it is in contact with. Steady-state diffusion model The steady-state model provides an analytical solution by assuming that for any segment of urine, the concentration gradient within the catheter wall instantly reaches a stable state. Derivation of molar flow rate. Derivation of molar flow rate. Derivation of molar flow rate. The derivation begins with Fick’s first law for radial diffusion in a cylinder20:1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{J}_{r}\left(r\right)=-{D}_{wall}\frac{dC\left(r\right)}{dr}\:\end{array}$$\end{document} where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{J}_{r}\left(r\right)$$\end{document} is the radial molar flux of oxygen (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:mol\:{m}^{-2}\:{s}^{-1})$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{D}_{wall}$$\end{document} is the radial diffusion coefficient in the wall \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({m}^{2}\:{s}^{-1}\right)$$\end{document}, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:C\left(r\right)$$\end{document} is the oxygen concentration at the radius r \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(mol\:\:{m}^{-3}\right)$$\end{document}. The total steady-state molar flow rate of oxygen (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\dot{m}}_{ss}$$\end{document}​) across the wall over a cylindrical segment of length L is the product of flux and surface area. By separating variables and integrating across the wall from the inner radius \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({r}_{i}\right)$$\end{document} to the outer radius \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({r}_{o}\right)$$\end{document} of the wall, we solve for \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\dot{m}}_{ss}$$\end{document}:2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{\int\:}_{{C}_{l}}^{{C}_{env}}dC\left(r\right)=-\frac{{\dot{m}}_{ss}}{2\pi\:L{D}_{wall}}{\int\:}_{{r}_{i}}^{{r}_{o}}\frac{dr}{r}\:\end{array}$$\end{document} Where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{C}_{env}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{C}_{l}$$\end{document} are the concentrations in the environment and in the lumen, respectively \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(mol\:\:{m}^{-3}\right)$$\end{document}. This yields the expression for the steady-state molar flow rate:3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{\dot{m}}_{ss}=\frac{2\pi\:L{D}_{wall}}{\mathrm{ln}\left(\frac{{r}_{o}}{{r}_{i}}\right)}\left({C}_{env}-{C}_{l}\right)\end{array}$$\end{document} Where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{C}_{env}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{C}_{l}$$\end{document} are the oxygen concentrations in the environment and lumen, respectively \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(mol\:\:{m}^{-3}\right)$$\end{document}. We define \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\dot{m}}_{ss}$$\end{document} as inward molar flow (into the lumen) and take that as positive. Equation 2 is written with the wall’s outward normal, so the surface flux is negative when oxygen enters the lumen. Thus, when solving Eq. 2 for the inward flow \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\dot{m}}_{ss}$$\end{document}, the leading minus sign cancels, giving the positive expression in Eq. 3. b.Derivation of lumen concentration. b.Derivation of lumen concentration. Derivation of lumen concentration. A mass balance on a differential segment of urine is used, so axial inflow and outflow cancel; only radial wall flux contributes. With \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\dot{m}}_{ss}$$\end{document} defined as the net radial molar flow across the wall (positive into the lumen), the accumulation is:4\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{V}_{l}\frac{d{C}_{l}\left(t\right)}{dt}={\dot{m}}_{ss}\:\end{array}$$\end{document} where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{V}_{l}$$\end{document} is the lumen volume \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({m}^{3}\right)$$\end{document}. Substituting Eq. 3 into Eq. 4 yields a first-order ordinary differential equation (ODE):5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}\frac{d{C}_{l}\left(t\right)}{dt}={k}_{j}\left({C}_{env}-{C}_{l}\left(t\right)\right)\end{array}$$\end{document} Where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{k}_{j}$$\end{document} \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({s}^{-1}\right)$$\end{document} is the first-order rate constant for mass transfer for the segment j and is equal to:6\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{k}_{j}=\frac{2{D}_{wall,j}}{{r}_{i}^{2}\mathrm{ln}\left(\frac{{r}_{o}}{{r}_{i}}\right)}\:for\:j\in\:\{in,out\}\end{array}$$\end{document} c.Solution and inversion. c.Solution and inversion. Solution and inversion. Integrating this ODE over the transit time \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({{\Delta\:}t}_{j}\right)$$\end{document} of the urine in a segment of j from an initial concentration \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{C}_{inlet}$$\end{document} yields the solution for the outlet concentration, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{C}_{outlet}$$\end{document}:7\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{C}_{outle{t}_{j}}={C}_{en{v}_{j}}+\left({C}_{inle{t}_{j}}-{C}_{en{v}_{j}}\right){e}^{-{k}_{j}{{\Delta\:}t}_{j}},\:j\in\:\{in,out\}\end{array}$$\end{document} Where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{C}_{outle{t}_{j}}$$\end{document} is the concentration of oxygen at the end of the segment \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:j\:\left(mol\:\:{m}^{-3}\right)$$\end{document}. For two segments in the series:8\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{C}_{junction}={C}_{en{v}_{in}}+\left({C}_{bladder}-{C}_{en{v}_{in}}\right){e}^{-{k}_{in}{\Delta\:}{t}_{in}}\end{array}$$\end{document}9\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{C}_{distal}={C}_{en{v}_{out}}+\left({C}_{junction}-{C}_{en{v}_{out}}\right){e}^{-{k}_{out}{\Delta\:}{t}_{out}}\end{array}$$\end{document} Where the subscript "in" refers to the catheter segment inside the body, surrounded by tissue, and the subscript "out" refers to the catheter segment outside the body, exposed to ambient air, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{C}_{bladder}$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{C}_{junction}\:$$\end{document}and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{C}_{distal}$$\end{document} are the oxygen concentrations in the bladder, junction between in-body and out-body segments, and at the end of the urinary catheter, respectively \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(mol\:\:{m}^{-3}\right)$$\end{document} (Fig. 1). Given the measured distal concentration and environment values:10\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{C}_{junction}={C}_{en{v}_{out}}+\left({C}_{distal}-{C}_{en{v}_{out}}\right){e}^{{k}_{out}{{\Delta\:}t}_{out}}\end{array}$$\end{document}11\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{C}_{bladder}={C}_{en{v}_{in}}+\left({C}_{junction}-{C}_{en{v}_{in}}\right){e}^{{k}_{in}{\Delta\:}{t}_{in}}\end{array}$$\end{document} By converting concentration to oxygen partial pressure using Henry’s law ( C=HP; Appendix, Eq. 31), where H is the Henry’s law solubility constant, the concentration-domain solution can be written in the pressure domain. Assuming H is constant within each segment, H cancels algebraically, yielding Eq. (12)21. The Henry’s-law conversion and calibrated steady-state expression are summarized in the Appendix (Eqs. 31–34).12\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{\widehat{P}}_{bladder}={P}_{tissue}+\left({P}_{air}-{P}_{tissue}\right){e}^{{k}_{in}{\Delta\:}{t}_{in}}+\left({P}_{distal}-{P}_{air}\right){e}^{{k}_{in}{{\Delta\:}t}_{in}+{k}_{out}{{\Delta\:}t}_{out}}\end{array}$$\end{document} Where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\widehat{P}}_{bladder}$$\end{document} is the estimated bladder PuO2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(mmHg\right)$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document} is the measured distal catheter PuO2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(mmHg\right)$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document} is the ambient tissue oxygen partial pressure surrounding the in-body catheter segment \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(mmHg\right)$$\end{document}, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{air}$$\end{document} is the ambient room-air oxygen partial pressure surrounding the out-of-body catheter segment \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(mmHg\right)$$\end{document}. Fig. 1Schematic of the two catheter segments and notation. Urine flows from the bladder (left) to the distal end, where there is a sensor (right). Segment 1 (inside body) is surrounded by tissue with environmental partial pressure \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document}; Segment 2 (outside body) is exposed to room air with \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{air}$$\end{document}. The dashed line marks the body exit (segment junction). Oxygen partial pressures along the lumen are labeled \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\widehat{P}}_{bladder}$$\end{document}(bladder PuO2), \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{junction}$$\end{document}(after the in-body segment), and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}(sensor). Geometric radii \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{r}_{i},\:{r}_{o}$$\end{document} are shown for the catheter wall. Transit times are \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\Delta\:}{t}_{in}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\Delta\:}{t}_{out}$$\end{document} (corresponding to length \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{L}_{in}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{L}_{out}$$\end{document}). Schematic of the two catheter segments and notation. Urine flows from the bladder (left) to the distal end, where there is a sensor (right). Segment 1 (inside body) is surrounded by tissue with environmental partial pressure \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document}; Segment 2 (outside body) is exposed to room air with \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{air}$$\end{document}. The dashed line marks the body exit (segment junction). Oxygen partial pressures along the lumen are labeled \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\widehat{P}}_{bladder}$$\end{document}(bladder PuO2), \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{junction}$$\end{document}(after the in-body segment), and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}(sensor). Geometric radii \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{r}_{i},\:{r}_{o}$$\end{document} are shown for the catheter wall. Transit times are \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\Delta\:}{t}_{in}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\Delta\:}{t}_{out}$$\end{document} (corresponding to length \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{L}_{in}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{L}_{out}$$\end{document}). Transient diffusion model The primary limitation of the steady-state model is its assumption of instantaneous equilibrium within the catheter wall. This simplification posits that the concentration gradient within the polymer wall immediately reaches a stable, linear profile in response to a change in lumen concentration. In reality, this process requires finite time to transition to a new equilibrium when lumen or environmental oxygen concentrations change, rather than equilibrating instantaneously22. The steady-state model, therefore, does not capture the initial, higher-flux phase of diffusion that occurs while the wall transitions toward a new concentration equilibrium. This discrepancy becomes particularly significant during rapid changes in urine concentration or for short transit times, when the wall may never reach an actual steady state. Long transit times are clinically important because they increase signal delay and cumulative distortion, whereas short transit times are the regime in which the transient and steady-state wall models differ most, because the wall may not have time to approach steady state. The transient model below addresses this by explicitly solving for the time-dependent concentration profile within the wall. Governing equations. Governing equations. Governing equations. The model is described by a system of coupled differential equations: a partial differential equation (PDE) for diffusion within the catheter wall and an ODE for the mass balance in the well-mixed lumen. The governing equation for transient diffusion in the wall is Fick’s second law in cylindrical coordinates, expressed in terms of the partial pressure of oxygen using Henry’s Law23:13\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}\frac{\partial\:P}{\partial\:t}=D\left(\frac{{\partial\:}^{2}P}{\partial\:{r}^{2}}+\frac{1}{r}\frac{\partial\:P}{\partial\:r}\right),\:\:{r}_{i}\le\:r\le\:{r}_{o}\end{array}$$\end{document} The ODE for the lumen is derived from the principle of conservation of mass, where the rate of oxygen accumulation in the lumen volume equals the net radial molar flow rate from the wall. We write this balance for a moving urine element, so axial inflow and outflow cancel; equivalently, a fixed control volume would include an additional axial in-out term. For generality, we introduce an interface mapping parameter, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\chi\:$$\end{document}, where the pressure at the inner surface is defined as \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:P\left({r}_{i},\:t\right)=\chi\:{P}_{l}\left(t\right)$$\end{document}, and a dimensionless solubility ratio \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\rho\:={\alpha\:}_{wall}/{\alpha\:}_{lumen}$$\end{document}, where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\alpha\:$$\end{document} is Henry’s solubility coefficient relating concentration to partial pressure (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:C=\alpha\:P)$$\end{document}, yielding:14\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}\frac{d{P}_{l}\left(t\right)}{dt}={\left.\frac{2{D}_{wall}\rho\:}{{r}_{i}}\frac{\partial\:P(r,t)}{\partial\:r}\right|}_{r={r}_{i}}\end{array}$$\end{document} Where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}\left(t\right)$$\end{document} is the (well-mixed) lumen oxygen partial pressure and P(r,t) is the wall oxygen partial pressure field. This pressure-domain ODE/PDE is the analogue of the concentration-domain balance given earlier. In the present implementation, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\chi\:$$\end{document} was fixed at 1, so the inner wall pressure was taken equal to the lumen pressure at the interface, consistent with a local equilibrium assumption. b.Boundary and initial conditions. b.Boundary and initial conditions. Boundary and initial conditions. To solve the coupled system, a complete set of boundary and initial conditions must be defined for each segment: Outer wall: The partial pressure of oxygen at the outer wall is constant and equal to the environment: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:P({r}_{o},\:t)={P}_{env}$$\end{document}.Inner wall: The partial pressure of oxygen at the inner wall is equal to the dynamic lumen pressure: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:P\left({r}_{i},\:t\right)=\chi\:{P}_{l}\left(t\right)$$\end{document}. Equivalently, this corresponds to setting \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\chi\:$$\end{document} = 1 in the general formulation. Outer wall: The partial pressure of oxygen at the outer wall is constant and equal to the environment: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:P({r}_{o},\:t)={P}_{env}$$\end{document}. Outer wall: The partial pressure of oxygen at the outer wall is constant and equal to the environment: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:P({r}_{o},\:t)={P}_{env}$$\end{document}. Inner wall: The partial pressure of oxygen at the inner wall is equal to the dynamic lumen pressure: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:P\left({r}_{i},\:t\right)=\chi\:{P}_{l}\left(t\right)$$\end{document}. Inner wall: The partial pressure of oxygen at the inner wall is equal to the dynamic lumen pressure: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:P\left({r}_{i},\:t\right)=\chi\:{P}_{l}\left(t\right)$$\end{document}. Equivalently, this corresponds to setting \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\chi\:$$\end{document} = 1 in the general formulation. 3.Initial wall state: At the start of the segment’s simulation (t=0), the wall is assumed to be fully saturated with the environmental pressure: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:P\left(r,\:t=0\right)={P}_{env}$$\end{document}.4.Initial lumen state: At t=0, the lumen begins with the partial pressure of oxygen of the fluid entering the segment: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}\left(0\right)={P}_{{j}_{inlet}}$$\end{document}. 3.Initial wall state: At the start of the segment’s simulation (t=0), the wall is assumed to be fully saturated with the environmental pressure: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:P\left(r,\:t=0\right)={P}_{env}$$\end{document}. Initial wall state: At the start of the segment’s simulation (t=0), the wall is assumed to be fully saturated with the environmental pressure: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:P\left(r,\:t=0\right)={P}_{env}$$\end{document}. 4.Initial lumen state: At t=0, the lumen begins with the partial pressure of oxygen of the fluid entering the segment: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}\left(0\right)={P}_{{j}_{inlet}}$$\end{document}. Initial lumen state: At t=0, the lumen begins with the partial pressure of oxygen of the fluid entering the segment: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}\left(0\right)={P}_{{j}_{inlet}}$$\end{document}. c.Numerical solution using the finite difference method. c.Numerical solution using the finite difference method. Numerical solution using the finite difference method. The complexity of the coupled PDE-ODE system, particularly due to the dynamic boundary condition at the lumen-wall interface, precludes a closed-form analytical solution. Therefore, the system is solved numerically. An implicit Crank-Nicolson scheme was chosen for its unconditional stability24. The wall domain is discretized into M radial intervals of size \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\Delta\:}r$$\end{document} with nodes \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:z=0,\dots\:,M$$\end{document}, with cell centers \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{r}_{z}$$\end{document} and faces \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{r}_{z\pm\:\frac{1}{2}}$$\end{document}, and time is advanced in steps of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\Delta\:}t$$\end{document} with time levels n and n+1. The wall PDE is discretized using a conservative finite-volume approach. The Crank-Nicolson method averages the resulting spatial operator between the current time level \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left(n\right)$$\end{document} and the next (n+1), creating a tridiagonal system of linear equations for the unknown future wall pressure, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{z}^{n+1}$$\end{document}:15\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}-{a}_{z}{P}_{z-1}^{n+1}+{b}_{z}{P}_{z}^{n+1}-{c}_{z}{P}_{z+1}^{n+1}={d}_{z}^{\left(n\right)},\:for\:z=1,\dots\:,\:M-1\end{array}$$\end{document} Where the right-hand side vector, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{d}_{z}^{\left(n\right)}$$\end{document}, contains all known pressure values from the current time step, n. The coefficients are defined as:16\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{\gamma\:}_{z}=\frac{{D}_{wall}{\Delta\:}t}{2{r}_{z}{\Delta\:}{r}^{2}}\end{array}$$\end{document}17\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{a}_{z}={\gamma\:}_{z}{r}_{z-\frac{1}{2}}\end{array}$$\end{document}18\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{b}_{z}=1+{\gamma\:}_{z}\left({r}_{z+\frac{1}{2}}+{r}_{z-\frac{1}{2}}\right)\end{array}$$\end{document}19\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{c}_{z}={\gamma\:}_{z}{r}_{z+\frac{1}{2}}\end{array}$$\end{document}20\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{d}_{z}^{\left(n\right)}=\left(1-{\gamma\:}_{z}\left({r}_{z+\frac{1}{2}}+{r}_{z-\frac{1}{2}}\right)\right){P}_{z}^{n}+{\gamma\:}_{z}\left({r}_{z+\frac{1}{2}}{P}_{z+1}^{n}+{r}_{z-\frac{1}{2}}{P}_{z-1}^{n}\right)\end{array}$$\end{document} In vector form, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathbf{A}{\mathbf{P}}^{\mathrm{n}+1}={\mathbf{d}}^{\left(\mathrm{n}\right)}$$\end{document} with \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\mathbf{A}$$\end{document} tridiagonal and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\mathbf{P}}^{\mathrm{n}+1}{=[{P}_{1}^{n+1},\dots\:,\:{P}_{N}^{n+1}]}^{\mathrm{T}}$$\end{document}. This system is solved at each time step using the Thomas algorithm25,26. d.Lumen-wall coupling. d.Lumen-wall coupling. Lumen-wall coupling. A challenge in this implicit scheme is that the inner wall boundary condition, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{0}^{n+1}=\chi\:{P}_{l}^{n+1}$$\end{document}, depends on the lumen pressure at the new time step, which is also unknown. This implicit coupling is handled using a Picard iteration within each time step25,27. First, the lumen ODE (Eq. 15) is discretized using a second-order accurate one-sided finite difference to approximate the gradient at the inner wall. This results in the explicit update rule for the lumen pressure:21\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{P}_{l}^{n+1}={P}_{l}^{n}+\frac{2{D}_{wall}\rho\:{\Delta\:}t}{{r}_{i}}\left(\frac{-3{P}_{0}^{n+1}+4{P}_{1}^{n+1}-{P}_{2}^{n+1}}{2{\Delta\:}r}\right)\end{array}$$\end{document} Within each time step, we run a Picard iteration indexed by \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:m=0,\:1,\:2,\dots\:$$\end{document}. Start with m=0, and increment m after each sweep. Then, the following steps are performed: Guess: An initial guess is made for the new lumen pressure: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}^{\left(0\right)}\leftarrow\:\:{P}_{l}^{n}$$\end{document} (For the first substep of a segment, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}^{n}={P}_{inlet}$$\end{document}.Wall solve: The tridiagonal system for the wall (Eq. 15) is assembled and solved using the current guess to set the inner boundary condition: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{0}^{n+1}=\chi\:{P}_{l}^{\left(m\right)}$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{M}^{n+1}={P}_{env}$$\end{document}.Lumen update: The new wall pressures \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:({P}_{0}^{n+1},\:{P}_{1}^{n+1},\:{P}_{2}^{n+1})$$\end{document} are used in Eq. 21 to calculate an updated lumen pressure to a new value, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}^{(m+1)}$$\end{document}.Iteration and stop criterion: Repeat steps 2–3 and, after each sweep, check convergence. We declare convergence when the relative change in the lumen is below the tolerance \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:( \epsilon =0.5\%)$$\end{document}. For robustness, also impose a safety cap of at most \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{m}_{max}=10$$\end{document} Picard sweeps per time step. If the tolerance is met (or the cap is reached), set \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}^{n+1}\leftarrow\:\:{P}_{l}^{(m+1)}$$\end{document}, and then proceed to the next time step; use the wall profile from the last sweep as the initial wall state for the next substep. Guess: An initial guess is made for the new lumen pressure: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}^{\left(0\right)}\leftarrow\:\:{P}_{l}^{n}$$\end{document} (For the first substep of a segment, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}^{n}={P}_{inlet}$$\end{document}. Guess: An initial guess is made for the new lumen pressure: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}^{\left(0\right)}\leftarrow\:\:{P}_{l}^{n}$$\end{document} (For the first substep of a segment, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}^{n}={P}_{inlet}$$\end{document}. Wall solve: The tridiagonal system for the wall (Eq. 15) is assembled and solved using the current guess to set the inner boundary condition: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{0}^{n+1}=\chi\:{P}_{l}^{\left(m\right)}$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{M}^{n+1}={P}_{env}$$\end{document}. Wall solve: The tridiagonal system for the wall (Eq. 15) is assembled and solved using the current guess to set the inner boundary condition: \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{0}^{n+1}=\chi\:{P}_{l}^{\left(m\right)}$$\end{document}, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{M}^{n+1}={P}_{env}$$\end{document}. Lumen update: The new wall pressures \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:({P}_{0}^{n+1},\:{P}_{1}^{n+1},\:{P}_{2}^{n+1})$$\end{document} are used in Eq. 21 to calculate an updated lumen pressure to a new value, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}^{(m+1)}$$\end{document}. Lumen update: The new wall pressures \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:({P}_{0}^{n+1},\:{P}_{1}^{n+1},\:{P}_{2}^{n+1})$$\end{document} are used in Eq. 21 to calculate an updated lumen pressure to a new value, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}^{(m+1)}$$\end{document}. Iteration and stop criterion: Repeat steps 2–3 and, after each sweep, check convergence. We declare convergence when the relative change in the lumen is below the tolerance \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:( \epsilon =0.5\%)$$\end{document}. For robustness, also impose a safety cap of at most \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{m}_{max}=10$$\end{document} Picard sweeps per time step. If the tolerance is met (or the cap is reached), set \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}^{n+1}\leftarrow\:\:{P}_{l}^{(m+1)}$$\end{document}, and then proceed to the next time step; use the wall profile from the last sweep as the initial wall state for the next substep. Iteration and stop criterion: Repeat steps 2–3 and, after each sweep, check convergence. We declare convergence when the relative change in the lumen is below the tolerance \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:( \epsilon =0.5\%)$$\end{document}. For robustness, also impose a safety cap of at most \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{m}_{max}=10$$\end{document} Picard sweeps per time step. If the tolerance is met (or the cap is reached), set \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{l}^{n+1}\leftarrow\:\:{P}_{l}^{(m+1)}$$\end{document}, and then proceed to the next time step; use the wall profile from the last sweep as the initial wall state for the next substep. e.Inverse problem solution. e.Inverse problem solution. Inverse problem solution. To determine the bladder PuO2 (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}) from a measured distal PuO2 (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}), the finite difference Crank–Nicolson solver serves as a forward model within an iterative inverse problem solver. A root-finding algorithm (Brent’s method) adjusts an initial guess, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder,guess}$$\end{document}, and runs the forward simulation for the in-body and out-of-body segments in series until the predicted distal PuO2 matches the measured distal PuO2 (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}) to within a specified tolerance \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:({\Delta\:}P\approx\:0)$$\end{document}28. For fixed \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{t}_{in},\:{t}_{out},\:{P}_{tissue},\:{P}_{air}$$\end{document}, the forward map \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:g:\:{P}_{bladder}\mapsto\:{P}_{distal}$$\end{document} is continuous and increasing; so the inverse exists and is unique. A summary of the forward/inverse implementation is provided in the Appendix (Sect. 5 of the transient diffusion model). Model calibration The Crank-Nicolson solver constitutes the “forward model,” which maps an initial bladder PuO2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{(P}_{bladder})$$\end{document} to a predicted distal catheter PuO2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{(P}_{distal,\:model})$$\end{document} by running sequentially for the in-body and out-of-body segments. To estimate bladder PuO2 from a distal measurement, an inverse problem is solved using Brent’s method to find the value of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} for which the model output matches the measured \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{(P}_{distal})$$\end{document}28. Finally, a dimensionless calibration factor, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\Lambda\:}$$\end{document}, is applied to the model-derived bladder PuO2 estimate to account for residual device-level uncertainties and simplifying assumptions, including effective wall diffusivity, temperature-dependent solubility differences, and deviations from the well-mixed lumen assumption. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\Lambda\:}$$\end{document} was estimated once from the bench-top calibration data by minimizing the root-mean-square error between the model-derived bladder PuO2 estimate and the directly measured ground-truth bladder PuO2. The resulting value of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\Lambda\:}$$\end{document} was then fixed and applied unchanged to the in vivo porcine dataset. The physical rationale for \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\Lambda\:}$$\end{document} and the corrected diffusion model are summarized in the Appendix (Sect. 7 of steady-state diffusion model). Likewise, the solubility ratio, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\rho\:$$\end{document}, was treated as a system-level material property term in the pressure-domain formulation, not as a separately fitted physiological variable for each subject. Thus, neither \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\Lambda\:}\:$$\end{document}nor \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\rho\:\:$$\end{document}was varied between subjects in the present implementation.22\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\begin{array}{c}{P}_{bladder,\:corrected}=\varLambda\:\cdot \:{P}_{bladder,\:model}\end{array}$$\end{document} Real-time implementation For clinical applications, the adjustment algorithm must operate in real-time. The steady-state model, being a closed-form analytical expression, is inherently real-time capable, as its evaluation is computationally trivial. The transient model, however, requires solving an iterative inverse problem (Sect. 2.1.3e), which might be computationally intensive for direct real-time use in simple processors. In our calculations, a single forward simulation of both segments with the default settings (12 wall cells, 60 time steps per second, 2 Picard sweeps per time step, total residence of 2 s) costs ~ 1 million floating-point operations (flops). The Brent root finder, seeded with the steady-state results, typically needs ~ 3–8 such forward evaluations (~ 3–8 million flops); the worst-case cap is 35 evaluations (~ 30 million flops). To translate the accuracy of the transient model to a real-time environment, a surrogate model must be created. This involves running the high-fidelity transient model offline across a wide grid of plausible inputs \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:({P}_{distal},\:{t}_{in},\:{t}_{out})$$\end{document} to generate a comprehensive dataset. This dataset can then be used to create a computationally inexpensive surrogate, such as a multi-dimensional lookup table with linear interpolation, that can be deployed on a clinical monitor to provide instantaneous, adjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} values. Experimental setup: bench-top validation To validate the diffusion models in a controlled environment that mimics key aspects of the clinical scenario, a specialized bench-top apparatus was constructed (Fig. 2). The setup was designed to independently control the initial bladder partial pressure of oxygen (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}) in the simulated urine, the PO2 of the environment surrounding the catheter (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document}), and the flow rate through the catheter lumen. Fig. 2Schematic of the bench-top experimental apparatus for model validation. A clinical-grade urinary catheter is passed through a larger, temperature-controlled tube that simulates the ‘in-body’ tissue environment. The catheter tip is sealed within a smaller chamber representing the bladder. Water, serving as simulated urine, is supplied from a reservoir where its PO2 is precisely controlled by a mixture of oxygen and nitrogen gases regulated by mass flow controllers to provide \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}. A second reservoir and pump system circulates water of a known PO2 through the outer ‘tissue’ tube to provide \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document}. Oxygen and temperature sensors continuously measure the PO2 at four locations—right flask, simulated bladder (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder})$$\end{document}, simulated tissue (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document}), and distal catheter end (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document})—and a flow sensor measures volumetric flow. All data were recorded on a computer. Schematic of the bench-top experimental apparatus for model validation. A clinical-grade urinary catheter is passed through a larger, temperature-controlled tube that simulates the ‘in-body’ tissue environment. The catheter tip is sealed within a smaller chamber representing the bladder. Water, serving as simulated urine, is supplied from a reservoir where its PO2 is precisely controlled by a mixture of oxygen and nitrogen gases regulated by mass flow controllers to provide \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}. A second reservoir and pump system circulates water of a known PO2 through the outer ‘tissue’ tube to provide \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document}. Oxygen and temperature sensors continuously measure the PO2 at four locations—right flask, simulated bladder (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder})$$\end{document}, simulated tissue (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document}), and distal catheter end (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document})—and a flow sensor measures volumetric flow. All data were recorded on a computer. Apparatus and circuit The core of the apparatus consisted of a simulated in-body environment created from a polycarbonate tube, within which a smaller, sealed tube served as a “bladder” reservoir. A standard clinical-grade 20 Fr Bard catheter (C. R. Bard, Franklin Lakes, NJ, USA) was inserted into this system such that its proximal half was contained within the larger tube (the “tissue” environment) and its inflatable tip was sealed inside the smaller “bladder” tube. Two flasks containing water served as reservoirs for the simulated urine and tissue fluid. The PO2 and temperature of the water in each flask were controlled by bubbling a mixture of nitrogen and oxygen gas through it. The gas mixture for each flask was regulated by a dedicated mass-flow controller (Alicat, Tucson, AZ, USA). A diaphragm pump was used to circulate water from the “tissue” reservoir through the outer tube, creating a controlled environment around the in-body segment of the catheter. The “bladder” reservoir was connected to the catheter inlet. A valve placed at the distal end of the catheter was used to regulate the outflow, thereby controlling the flow rate of the simulated urine. Measurements and data acquisition A suite of optode-based oxygen and temperature sensors (OxyLite™ Pro with OxyLite NX sensors, Optronix, Oxford, UK) was used for continuous monitoring. Ground-truth bladder PO2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({P}_{bladder}\right)$$\end{document}: A sensor was placed directly inside the “bladder” tube to provide the true initial PO2 of the fluid entering the catheter.Tissue PO2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({P}_{tissue}\right)$$\end{document}: A sensor was placed in the outer tube to continuously monitor the environmental PO2 of the simulated tissue surrounding the in-body segment of the catheter. Thus, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{t}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{u}\mathrm{e}}$$\end{document} was measured directly in the bench-top system.Distal sensor PO2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({P}_{distal}\right)$$\end{document}: A sensor was placed in a T-junction at the distal end of the catheter to measure the final PO2 of the fluid after its transit, mimicking the measurement from a clinical device.Flow rate: A flow sensor was placed in-line after the distal PO2 sensor to provide continuous measurement of the fluid flow rate. Ground-truth bladder PO2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({P}_{bladder}\right)$$\end{document}: A sensor was placed directly inside the “bladder” tube to provide the true initial PO2 of the fluid entering the catheter. Ground-truth bladder PO2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({P}_{bladder}\right)$$\end{document}: A sensor was placed directly inside the “bladder” tube to provide the true initial PO2 of the fluid entering the catheter. Tissue PO2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({P}_{tissue}\right)$$\end{document}: A sensor was placed in the outer tube to continuously monitor the environmental PO2 of the simulated tissue surrounding the in-body segment of the catheter. Thus, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{t}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{u}\mathrm{e}}$$\end{document} was measured directly in the bench-top system. Tissue PO2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({P}_{tissue}\right)$$\end{document}: A sensor was placed in the outer tube to continuously monitor the environmental PO2 of the simulated tissue surrounding the in-body segment of the catheter. Thus, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{t}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{u}\mathrm{e}}$$\end{document} was measured directly in the bench-top system. Distal sensor PO2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({P}_{distal}\right)$$\end{document}: A sensor was placed in a T-junction at the distal end of the catheter to measure the final PO2 of the fluid after its transit, mimicking the measurement from a clinical device. Distal sensor PO2 \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\left({P}_{distal}\right)$$\end{document}: A sensor was placed in a T-junction at the distal end of the catheter to measure the final PO2 of the fluid after its transit, mimicking the measurement from a clinical device. Flow rate: A flow sensor was placed in-line after the distal PO2 sensor to provide continuous measurement of the fluid flow rate. Flow rate: A flow sensor was placed in-line after the distal PO2 sensor to provide continuous measurement of the fluid flow rate. All sensor data, including PO2 values, temperatures, and flow rates, were recorded continuously by a data acquisition system for subsequent analysis. Experimental protocol To test the models across a wide range of clinically relevant conditions, the experiment was designed to dynamically vary both the initial PO2 (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}) and the flow rate over several hours while maintaining the “in-body” segment at a physiologically relevant temperature (≈ 37 °C). The \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} was cyclically ramped between hyperoxic (PO2 > 100 mmHg) and low-oxygen (PO2 < 20 mmHg) states to simulate physiological fluctuations, with the low-oxygen condition serving as an experimental challenge state rather than a clinical definition of hypoxia. These bounds were selected to span a broad range of oxygenation states relevant to urinary PuO2 monitoring and to test the model under both low- and high-oxygen conditions, rather than to define clinical thresholds. Concurrently, the flow rate was modulated using the distal valve to create periods of high flow (> 400 mL/h) interspersed with periods of simulated oliguria, during which the flow was reduced to near-zero levels. This protocol ensured that the models were rigorously tested against a diverse and challenging dataset, including both high and low flow rates as well as rapid changes in oxygen partial pressure. In vivo validation in a porcine model To test the performance of the diffusion-adjustment models in a dynamic physiological environment, a porcine model of hemorrhagic shock and resuscitation was used. Animal model and ethical approval All experimental procedures were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Utah (protocol number 00002200; approved 07/26/2023). The study was conducted using castrated male or non-pregnant female American Yorkshire swine (Sus scrofa domesticus) weighing 50–60 kg. A total of nine animals were included in the study; one animal was excluded intraoperatively due to a fatal embolic event (air embolism) resulting in a final cohort of eight animals for analysis. All experimental procedures were conducted in accordance with relevant institutional and national guidelines and regulations, and the study is reported in compliance with the ARRIVE guidelines. Anesthesia and surgical preparation Animals (Sus scrofa, Premier BioSource, CA, USA) were fasted overnight with free access to water. Anesthesia was induced with an intramuscular injection of tiletamine/zolazepam (6–8 mg/kg). Following endotracheal intubation, animals were mechanically ventilated and maintained under anesthesia with 1.5-3.0% isoflurane mixed in oxygen. The inspired oxygen fraction was adjusted to maintain pulse oximetry readings between 95 and 98%, and the respiratory rate was varied to maintain an end-tidal CO2 level of 35–45 mmHg. Core body temperature was maintained at approximately 39 °C using warming blankets. Warmed isotonic crystalloids (Plasma-Lyte 148, Baxter Healthcare Corporation, Deerfield, IL, USA) were administered intravenously at 5 mL/kg/hour. A midline laparotomy was performed to remove the spleen to prevent autotransfusion during hemorrhage. Urinary oxygen monitoring To obtain continuous measurements of both the ground-truth \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} and the diffusion-altered \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}, two separate oxygen sensors were used. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\boldsymbol{P}}_{\boldsymbol{b}\boldsymbol{l}\boldsymbol{a}\boldsymbol{d}\boldsymbol{d}\boldsymbol{e}\boldsymbol{r}}\boldsymbol{}$$\end{document}: A latex urinary catheter (Bardex Foley Catheter, Covington, GA, USA) was secured directly into the bladder via a cystotomy without allowing air inside the bladder. An optode-based oxygen sensor (OxyLite™ NX, Optronix, Oxford, UK) was advanced through this catheter, and its tip was positioned ~ 1 cm proximal to the catheter’s drainage eyelets. This placement ensured the sensor measured the PuO2 within the bladder while avoiding direct contact with the bladder wall.\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\boldsymbol{P}}_{\boldsymbol{d}\boldsymbol{i}\boldsymbol{s}\boldsymbol{t}\boldsymbol{a}\boldsymbol{l}}\boldsymbol{}$$\end{document}: The urinary catheter was connected to a custom-built sensor housing containing an oxygen sensor (PreSens Precision Sensing GmbH, Regensburg, Germany) at its distal end, just upstream of the collection bag. This sensor measured the PuO2 after transit through the full length of the catheter. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\boldsymbol{P}}_{\boldsymbol{b}\boldsymbol{l}\boldsymbol{a}\boldsymbol{d}\boldsymbol{d}\boldsymbol{e}\boldsymbol{r}}\boldsymbol{}$$\end{document}: A latex urinary catheter (Bardex Foley Catheter, Covington, GA, USA) was secured directly into the bladder via a cystotomy without allowing air inside the bladder. An optode-based oxygen sensor (OxyLite™ NX, Optronix, Oxford, UK) was advanced through this catheter, and its tip was positioned ~ 1 cm proximal to the catheter’s drainage eyelets. This placement ensured the sensor measured the PuO2 within the bladder while avoiding direct contact with the bladder wall. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\boldsymbol{P}}_{\boldsymbol{b}\boldsymbol{l}\boldsymbol{a}\boldsymbol{d}\boldsymbol{d}\boldsymbol{e}\boldsymbol{r}}\boldsymbol{}$$\end{document}: A latex urinary catheter (Bardex Foley Catheter, Covington, GA, USA) was secured directly into the bladder via a cystotomy without allowing air inside the bladder. An optode-based oxygen sensor (OxyLite™ NX, Optronix, Oxford, UK) was advanced through this catheter, and its tip was positioned ~ 1 cm proximal to the catheter’s drainage eyelets. This placement ensured the sensor measured the PuO2 within the bladder while avoiding direct contact with the bladder wall. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\boldsymbol{P}}_{\boldsymbol{d}\boldsymbol{i}\boldsymbol{s}\boldsymbol{t}\boldsymbol{a}\boldsymbol{l}}\boldsymbol{}$$\end{document}: The urinary catheter was connected to a custom-built sensor housing containing an oxygen sensor (PreSens Precision Sensing GmbH, Regensburg, Germany) at its distal end, just upstream of the collection bag. This sensor measured the PuO2 after transit through the full length of the catheter. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\boldsymbol{P}}_{\boldsymbol{d}\boldsymbol{i}\boldsymbol{s}\boldsymbol{t}\boldsymbol{a}\boldsymbol{l}}\boldsymbol{}$$\end{document}: The urinary catheter was connected to a custom-built sensor housing containing an oxygen sensor (PreSens Precision Sensing GmbH, Regensburg, Germany) at its distal end, just upstream of the collection bag. This sensor measured the PuO2 after transit through the full length of the catheter. A LD20-2600B flow sensor (Sensirion AG, Stäfa, Switzerland; full-scale flow rate 16.6 mL/min, accuracy 5% of the measured value, repeatability 0.5% of the measured value) was placed in-line immediately after the distal PuO2 sensor. After completion of each experiment, the sensors were re-tested under room conditions to assess baseline drift; no substantial drift was observed. Experimental protocol: hemorrhage and resuscitation Following a stabilization period after instrumentation, animals were subjected to a controlled hemorrhage to induce renal hypoxia. Blood was removed from a femoral arterial line in staged increments (5 mL/kg, followed by two 10 mL/kg stages) to achieve approximately 25 mL/kg, corresponding to ~ 40% of the total blood volume. This was followed by a resuscitation phase, during which shed blood and isotonic crystalloids were administered to restore arterial blood pressure. This protocol was designed to generate a rich dataset with wide, dynamic variations in urine flow, renal perfusion, and urinary oxygen partial pressure for subsequent, post-hoc testing of the adjustment models. At the conclusion of the experimental protocol, animals were euthanized under deep anesthesia by intravenous administration of pentobarbital/phenytoin sodium solution (Euthasol; 390 mg/mL pentobarbital sodium, 50 mg/mL phenytoin sodium) at a dose of 0.22 mL/kg, in accordance with AVMA Guidelines for the Euthanasia of Animals. Death was confirmed by the absence of cardiac activity and respiratory movement prior to tissue collection. Data processing The transient model was implemented as the primary compensation algorithm because it explicitly resolves the time-dependent diffusion of oxygen across the catheter wall. To ensure comparison of corresponding urine element, the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{b}\mathrm{l}\mathrm{a}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{r}}$$\end{document} signal was shifted forward by the sample-specific transit time, so that each distal measurement was aligned with its corresponding upstream value and both signals were expressed on the distal-tip time axis19. At each measurement event, the goal of the solver was to recover the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} such that the simulated \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document} matched the observed end-of-catheter measurement. This inversion was framed as a root-finding problem, where the error function was defined as the difference between the measured and simulated \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document} values. To accelerate convergence, we adopted a hybrid initialization strategy. Specifically, the steady-state two-segment model was evaluated first to generate a physically informed estimate of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}. This value served as the initial seed for the transient solver. By seeding the solver in this way, the search interval for Brent’s method root-finding routine could be narrowed, reducing the number of expensive transient evaluations required. The solver maintained a convergence tolerance of ± 0.1 mmHg at the distal end. Throughout the analysis, ‘flow’ refers to the average flow derived from transit time, not instantaneous sensor reading. The catheter wall diffusion was discretized in the radial domain (15 nodes across the wall thickness) and advanced in time using a Crank–Nicolson scheme to maintain numerical stability. Lumen–wall coupling was enforced with Picard iterations (typically two per time step), and environmental boundary conditions were applied at the outer radius. Each residence interval was subdivided into 60 substeps per second of transit time, yielding a stable temporal resolution. To account for the experimental protocol, wall states were reset between measurement events; for clinical deployment, we instead propagated the wall state continuously across time, which avoids any dependence on event timing. This implementation strategy balanced physical rigor with computational efficiency: the steady-state seed prevented wasted iterations, while the transient model captured short-term diffusion dynamics. For the animal data post-processing, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document} was not measured directly and was instead prescribed as a fixed boundary condition of 40 mmHg to approximate the peri-catheter tissue oxygen environment29. For subgroup analysis of low-flow conditions, we examined an approximate KDIGO-based oliguria subgroup defined as flow ≤ 27.5 mL/h, corresponding to 0.5 mL/kg/h for the average body weight of the study animals8. Results Bench-top validation We evaluated the transient diffusion-compensation algorithm in a controlled apparatus that reproduced the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}, urine transit through the catheter, and ambient diffusion. The model was assessed against the ground-truth \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} reconstructed at the catheter tip time axis. Dynamic tracking of oxygen transients Unadjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document} exhibited variable amplitude distortion and temporal lag during transitions. However, model-adjusted values closely followed the ground-truth waveform across hypoxic and hyperoxic phases, preserving both amplitude and timing (Fig. 3). Fig. 3Time-series comparison of ground-truth \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}, unadjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}, adjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document}, plotted against elapsed time (minutes). Flow rate is overlaid as both a line plot and a shaded area in the background. Time-series comparison of ground-truth \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}, unadjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}, adjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document}, plotted against elapsed time (minutes). Flow rate is overlaid as both a line plot and a shaded area in the background. Reduction in bias and dispersion Bland–Altman analysis demonstrated a marked reduction in systematic error. After adjustment, the mean bias decreased from − 10.89 mmHg relative to the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} to -0.64 mmHg (94% reduction), and the 95% limits of agreement narrowed from \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\pm\:$$\end{document}39.12 mmHg to \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:\pm\:$$\end{document}15.78 mmHg (≈ 60% narrower) (Fig. 4). Overall error metrics reflected the same improvement: mean absolute error decreased from 18.81 mmHg to 6.38 mmHg, and RMSE from 22.74 mmHg to 8.06 mmHg (≈ 65% reduction). The relationship with ground truth strengthened substantially, with R2 increasing from 0.53 to 0.94. Error distributions shown as kernel density estimation (KDE) curves, shifted from a broad, biased profile to a compact, near-zero-centered unimodal profile (Fig. 5)30. Fig. 4Bland–Altman analysis of unadjusted (left) versus model-adjusted (right) measurements relative to \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}. Plots show mean bias and 95% limits of agreement. Bland–Altman analysis of unadjusted (left) versus model-adjusted (right) measurements relative to \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}. Plots show mean bias and 95% limits of agreement. Fig. 5Kernel-density error distributions for unadjusted (orange) versus adjusted (green) values. The adjusted curve is tighter and centered near zero (dotted line), indicating reduced bias and variance compared with the broad, skewed unadjusted distribution. Kernel-density error distributions for unadjusted (orange) versus adjusted (green) values. The adjusted curve is tighter and centered near zero (dotted line), indicating reduced bias and variance compared with the broad, skewed unadjusted distribution. Flow-dependent performance Because diffusion artifacts intensify at low flow, we stratified performance by flow band (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} vs. adjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}). RMSE decreased from 34.0 mmHg to 11.4 mmHg at < 50 mL/h (n = 1390, 67% reduction), from 22.9 mmHg to 7.1 mmHg at 50–150 mL/h (n = 1096, 69% reduction), and from 12.7 mmHg to 6.0 mmHg at > 150 mL/h (n = 2508, 53% reduction). In vivo validation in a porcine hemorrhagic shock model We evaluated the diffusion-compensation algorithm in eight anesthetized pigs undergoing controlled hemorrhage and resuscitation, where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{b}\mathrm{l}\mathrm{a}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{r}}$$\end{document} spanned 3.1–108.2 mmHg with a median of 26.9 mmHg (interquartile range 21.9–32.5 mmHg). Dual sensors recorded the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document} signals. The dataset comprised 68,788 paired samples across eight animals (min 4,599, max 11,801 per subject). A representative in vivo time-series example from one porcine experiment is shown in Fig. 6. During hemorrhage and resuscitation, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} varied dynamically over time, whereas the unadjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document} signal showed transport-related distortion, particularly during low-flow periods. The model-adjusted estimate more closely tracked the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} waveform, although divergence remained during intervals of unstable, very low, or near-zero urine flow when the inverse problem becomes less well conditioned. Fig. 6Representative in vivo time-series comparison of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}, unadjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}, model-adjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}, and urine flow in one porcine experiment. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} varied dynamically over time during hemorrhage and resuscitation, whereas the unadjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document} signal showed transport-related distortion during catheter transit. The model-adjusted estimate more closely followed the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} waveform, particularly during periods of changing oxygen tension, although larger discrepancies remained during intervals of unstable, very low, or near-zero urine flow. Flow rate is shown on the secondary y-axis as a light blue background trace. Representative in vivo time-series comparison of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}, unadjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}, model-adjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}, and urine flow in one porcine experiment. \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} varied dynamically over time during hemorrhage and resuscitation, whereas the unadjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document} signal showed transport-related distortion during catheter transit. The model-adjusted estimate more closely followed the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} waveform, particularly during periods of changing oxygen tension, although larger discrepancies remained during intervals of unstable, very low, or near-zero urine flow. Flow rate is shown on the secondary y-axis as a light blue background trace. Overall error and bias reduction Pooling all samples, unadjusted distal readings exhibited a significant positive bias of + 20.9 ± 11.5 mmHg (mean ± SD) relative to the bladder. After adjustment, the mean bias shifted to -4.1 ± 13.5 mmHg, representing an ≈ 80% reduction in |bias|. The magnitude of error likewise decreased: mean absolute error fell from 21.1 mmHg to 9.5 mmHg (≈ 55% reduction). Both improvements were highly significant (paired t-test and Wilcoxon signed-rank test, p < 1 × 10–16 for bias and |error|). Violin plots of the pooled data (Fig. 7) show the broad, right‑shifted raw distribution contracting into a tighter, near‑zero‑centered distribution after correction. Fig. 7Pooled bias and error distributions before and after adjustment. Violin plots comparing the unadjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document} measurements (orange) to adjusted values (green). Left: bias relative to \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}. Right: absolute error. Distributions are displayed as split violins overlaid with boxplots and jittered raw data. Adjustment reduced both bias and error, shifting the distributions toward zero. Statistical comparisons indicated highly significant improvements (p < 1 × 10− 16 for both metrics). Pooled bias and error distributions before and after adjustment. Violin plots comparing the unadjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document} measurements (orange) to adjusted values (green). Left: bias relative to \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}. Right: absolute error. Distributions are displayed as split violins overlaid with boxplots and jittered raw data. Adjustment reduced both bias and error, shifting the distributions toward zero. Statistical comparisons indicated highly significant improvements (p < 1 × 10− 16 for both metrics). Restoration of linearity and subject‑level agreement Per‑subject identity‑contour plots (Fig. 8) revealed poor and inconsistent alignment of unadjusted data with the identity line (attenuated slopes and upward displacement). After adjustment, the contours shifted toward identity with steeper slopes, a lower intercept, and a tighter spread. Mixed‑effects regression quantified these changes: relative to the adjusted condition, the unadjusted data showed a higher intercept ( ≈ + 33.3 mmHg) and a shallower slope (interaction − 0.339, all fixed effects p < 0.001). The pooled slopes were ≈ 0.15 for unadjusted and ≈ 0.49 for adjusted measurements, indicating substantial improvement toward restoration of one‑to‑one agreement with the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}. Fig. 8Subject-level identity plots of bladder vs. distal measurements. Kernel density contour plots for each animal illustrate the relationship between \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} (x-axis) and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document} (y-axis). Orange: unadjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}; green: adjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}; dashed line: identity (y = x). Regression lines (solid) demonstrate attenuation and offset in the unadjusted condition, with restoration toward identity after correction. Subject-level identity plots of bladder vs. distal measurements. Kernel density contour plots for each animal illustrate the relationship between \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} (x-axis) and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document} (y-axis). Orange: unadjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}; green: adjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document}; dashed line: identity (y = x). Regression lines (solid) demonstrate attenuation and offset in the unadjusted condition, with restoration toward identity after correction. Subject-level bias and error distributions Subject‑level summaries (Fig. 9) confirmed the pooled results: RMSE decreased in 7/8 subjects (the pooled RMSE decreased from 23.8 to 14.1 mmHg), and absolute bias decreased in 7/8. One animal (Subject 4) exhibited atypical behavior with a degraded identity relationship, despite reduced dispersion. As shown in Fig. 9, adjustment shifted the bias distributions toward zero and generally reduced the spread of absolute error across subjects. Most subjects showed lower median bias and lower interquartile error after adjustment, indicating improved agreement between \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{distal}$$\end{document} and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} across a range of physiologic states and urine flows. Fig. 9Subject-level distributions of bias and absolute error before and after adjustment. Split violin plots by subject, compare unadjusted (orange) and adjusted (green) values relative to \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}. Boxplots indicate medians and interquartile ranges; vertical bars denote 95% confidence intervals. Adjustment reduced bias and absolute error in most subjects, with the largest improvements observed in subjects with greater unadjusted distortion. Subject-level distributions of bias and absolute error before and after adjustment. Split violin plots by subject, compare unadjusted (orange) and adjusted (green) values relative to \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document}. Boxplots indicate medians and interquartile ranges; vertical bars denote 95% confidence intervals. Adjustment reduced bias and absolute error in most subjects, with the largest improvements observed in subjects with greater unadjusted distortion. Performance in hypoxic and oliguric subgroups We performed additional subgroup analyses for low \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{b}\mathrm{l}\mathrm{a}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{r}}$$\end{document} and low-flow conditions. First, for samples with reference \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{b}\mathrm{l}\mathrm{a}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{r}}$$\end{document} ≤ 15 mmHg, the adjusted signal showed substantially lower mean bias (8.6 vs. 38.7 mmHg) and lower mean absolute error (11.0 vs. 38.7 mmHg) than the unadjusted distal measurement31. Second, in an approximate KDIGO-based oliguria subgroup defined as flow ≤ 27.5 mL/h, corresponding to 0.5 mL/kg/h for the average animal weight in this study (55 kg), the adjusted signal reduced mean bias (-7.3 vs. 22.4 mmHg) and mean absolute error (12.3 vs. 24.3 mmHg) relative to the unadjusted distal measurement. Using a threshold of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{b}\mathrm{l}\mathrm{a}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{r}}$$\end{document} ≤ 15 mmHg to identify low \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{b}\mathrm{l}\mathrm{a}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{r}}$$\end{document} in the full cohort, the corrected signal achieved a sensitivity of 30.9% and specificity of 73.9%, whereas the unadjusted distal signal showed 0.0% sensitivity and 100.0% specificity. Within the approximate KDIGO-based oliguria subgroup, the corrected signal achieved a sensitivity of 34.5% and specificity of 60.6%, compared with 0.0% and 100.0% for the unadjusted signal. These results indicate that the adjustment improves recovery of low \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{b}\mathrm{l}\mathrm{a}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{r}}$$\end{document} states relative to the unadjusted distal measurement, although threshold-based classification performance remained limited under extreme low-flow conditions. Discussion The results support our hypothesis that a physics‑informed, transient diffusion–compensation algorithm can reconstruct \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} from a distal urinary catheter PuO2 measurement, substantially reducing error relative to unadjusted readings. Bench‑top validations first showed that the algorithm could recover upstream \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} under various controlled flow and transit conditions, thereby reducing the systematic diffusion bias, especially during low-flow conditions. In vivo, across eight pigs subjected to hemorrhagic shock and resuscitation (68,788 paired samples), the correction shifted the pooled mean bias from + 20.9 ± 11.5 mmHg to -4.1 ± 13.5 mmHg and reduced mean absolute error by ≈ 55%. Identity‑contour visualizations and a mixed‑effects model showed that adjustment moved the distal measurements closer to the identity line, with steeper slopes and lower intercepts than the unadjusted data. Notably, the most significant gains occurred during low‑flow (oliguric) epochs, when unadjusted distal values were most distorted, confirming that the algorithm preserves fidelity under the very conditions in which renal hypoxia is most likely to occur. At the lowest flow rates, however, some residual over-adjustment remained, likely because prolonged transit times drive \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{d}\mathrm{i}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{l}}$$\end{document} closer to environmental equilibrium, making the inverse estimate of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{b}\mathrm{l}\mathrm{a}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{r}}$$\end{document} more sensitive to boundary-condition and model-approximation uncertainty. In practical terms, this adjustment allows the distal catheter signal to track the bladder’s oxygen dynamics in real time, especially at low to moderate flows, whereas at extremely low or near-zero flow, uncertainty increases as the distal signal progressively approaches environmental equilibrium and the inverse problem becomes less well-conditioned. Additional subgroup analyses showed that the adjustment substantially reduced bias and absolute error in both hypoxic and approximate KDIGO-based low-flow conditions, although threshold-based classification performance remained modest, particularly at the lowest flow states. One subject (#4) represented an extreme low-flow regime, characterized by prolonged near‑zero urine flow, with intermittent large flow spikes (potentially from intraoperative manipulation or a sensor placement/malfunction). In this setting, very long transit times and highly variable flow likely reduced the stability of both transit-time estimation and inverse reconstruction, resulting in less consistent agreement compared with the rest of the cohort. Excluding this subject, pooled bias (all points) was 22.1 ± 10.8 mmHg (unadjusted) vs. -3.2 ± 10.9 mmHg (adjusted), an 85.5% reduction in |bias|; mean absolute error likewise tightened (≈ 22.1→≈8.8 mmHg), further strengthening agreement with bladder measurements. The distal signal conflates two transport phenomena: (i) transit delay, whereby older urine arrives at the distal tip, especially when flow is low, and (ii) radial oxygen exchange across the catheter wall. In the present study, this exchange is not measured directly but inferred from its effect on \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{d}\mathrm{i}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{l}}$$\end{document} under controlled flow and environmental conditions. These effects flatten the bladder–distal slopes and introduce a flow-dependent upward bias. The simple steady‑state model assumes a wall perpetually at equilibrium and therefore cannot capture the finite time required for wall saturation or recovery during rapid flow changes. Our advanced solver, an unconditionally stable Crank–Nicolson scheme coupled to a lumen mass balance, propagates the wall state in time and aligns the bladder reference to the distal time base. Both unadjusted \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{d}\mathrm{i}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{l}}$$\end{document} and adjusted estimates were evaluated on the same distal time axis; only the \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{b}\mathrm{l}\mathrm{a}\mathrm{d}\mathrm{d}\mathrm{e}\mathrm{r}}$$\end{document} reference was shifted to account for transit delay, so the comparison between unadjusted and adjusted signals was not affected by the alignment step. Clinical implications and other applications Transit time‑aware correction enables distal urinary PO2 sensing to serve as a monitor of bladder (urinary) oxygenation, which prior clinical studies have linked to AKI risk15. In perioperative and critical‑care settings (e.g., major vascular surgery, trauma, sepsis, cardiopulmonary bypass), the adjusted signal supports earlier recognition of renal hypoxia, interpretable trends across large flow swings, and pragmatic alarm thresholds. Prior clinical work has used a PuO2 ≤ 15 mmHg as a working definition of urinary hypoxia associated with postoperative AKI31; by removing transport artifacts, our approach makes such thresholds more reliable at the acute point of care and better suited as continuous endpoints for reno‑protective trials or protocolized resuscitation. These findings also have implications for next-generation urinary catheter design. Catheters intended for oxygen monitoring should minimize wall permeability and uncontrolled exchange with the environment, for example through lower-permeability materials, shorter extracorporeal segments, standardized wall thickness, or integrated distal sensor housings positioned close to the catheter outlet. Such design features could reduce diffusion artifacts at the hardware level and lessen the computational burden of correction. The framework generalizes to any conduit‑based measurement in which tubing transit and polymer interfaces blur dynamics. Candidate applications include distal measurements of other urinary analytes (PCO2, pH, metabolites), respiratory sampling lines and capnography adapters where long tubing damps transients, pleural/abdominal/ventricular drains, peritoneal dialysis catheters, microdialysis outlets, and sensors embedded in extracorporeal circuits or microfluidic point‑of‑care devices. In each case, explicitly modeling transport can turn a convenient but biased distal signal into a more faithful representation of the upstream compartment. Computational considerations for deployment Solving the inverse problem row‑by‑row is more computationally intensive than applying a static calibration, but it remains tractable offline. For real‑time integration into bedside monitors, the same algorithm can be delivered with a surrogate: precomputed lookup tables, compact polynomial maps, or neural emulators trained on the forward model across the physiological range. These surrogates retain interpretability while offering real-time inference and straightforward uncertainty quantification. Limitations and mitigations We model the lumen as laminar and well-mixed, and oxygen transport through the catheter wall as radially symmetric diffusion with fixed material properties. These simplifications are appropriate for standard urinary catheters and the flow ranges studied; however, unusual geometries, temperature gradients, aging, or multilayer walls could break these assumptions and introduce errors. To absorb residual device‑specific uncertainties (partition/solubility ratios, effective diffusivity, and geometric mismatches), we fit a single calibration factor, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{\Lambda\:}$$\end{document}, which improves accuracy but needs to be estimated under stable conditions. In the porcine analysis, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document} was prescribed rather than measured directly; the value of 40 mmHg was used as a physiologically plausible boundary condition based on peri-ureteral literature, but the peri-catheter tissue environment in the present preparation may differ. In an additional sensitivity analysis, repeating the porcine post-processing with \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document}= 30 mmHg instead of 40 mmHg did not materially alter the main findings, suggesting that the reconstruction is relatively insensitive to \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document} across this plausible range, although direct estimation of \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{tissue}$$\end{document} would still be preferable. At very low or zero flow, long transit times pull the distal signal toward ambient equilibrium, making the initial bladder value difficult to identify. To prevent over-correction, we constrain the inversion by bracketing \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{bladder}$$\end{document} in the Brent search and capping the Picard sweeps per time step. If the solver cannot meet the convergence tolerance or the measured transit time exceeds a predefined limit, the sample is flagged as invalid and yet displayed to the user. In clinical practice, distal monitoring is performed using our custom sensor positioned between the urinary catheter and the collection bag, and the bag should remain below bladder level, since bag height and accumulated urine volume may alter hydrostatic pressure, thereby affecting flow, transit time, and indirectly contributing to variation in \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\:{P}_{\mathrm{d}\mathrm{i}\mathrm{s}\mathrm{t}\mathrm{a}\mathrm{l}}$$\end{document} measurements. Finally, although the porcine model captures large hemodynamic excursions, prospective human studies across catheter sizes and materials are needed to establish generalizability. Future directions Prospective validation in surgical and ICU cohorts with prespecified decision thresholds is a priority. Real‑time implementation should pair the adjustment with confidence bands, flow‑aware alerts, and online parameter adaptation (estimating wall‑exchange and transit coefficients using routine perturbations such as voids or flushes). It is also important to establish how PuO2 monitoring can move beyond AKI risk prediction to actively guiding therapy—helping clinicians initiate, titrate, and evaluate reno-protective interventions. Extending the solver to multiplexed analytes and other catheterized systems will demonstrate the broader value of transport-aware corrections. Conclusion Explicitly modeling transport delay and wall diffusion converts a biased distal measurement into a physiologically faithful signal. This approach preserves the familiar clinical catheter workflow, achieves accuracy comparable to bladder sensing, and provides a broadly applicable framework for improving a wide range of conduit‑based monitoring systems. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1