A hybrid curve-fitting method for radiation dosimetry evaluation and simultaneous multi-organ biodistribution assessment in total-body dynamic PET/CT: application to a novel PD-L1-targeted tracer Authors: Yunze Xie, Wenxin Tang, Lichao Zhang, Guobing Liu, Hongcheng Shi Published: 2025 ABSTRACT AbstractObjectiveThis study explored a hybrid curve-fitting method optimized for radiation dosimetry evaluation in total-body dynamic positron emission tomography/computed tomography (PET/CT) imaging, compared to conventional methods reliant on multi-time-point static acquisition protocols, and evaluated the radiation dosimetry results and organ biodistribution of [68Ga]Ga-NOTA-SNA002.MethodsA total of 16 patients with solid tumors underwent a 60-min dynamic PET acquisition immediately after administration of [68Ga]Ga-NOTA-SNA002, followed by a 20-min static PET scan at about 120 min post-injection. Volumes of interest (VOIs) were manually delineated on the CT images for subsequent radiation dosimetry estimation and biodistribution analysis. Two datasets were created to generate organ time-activity curves (TACs): (1) two PET frames reconstructed at 15–35 min and 40–60 min from the dynamic PET dataset, and one frame from static PET acquisition, thereby simulating a conventional multi-time-point static protocol; (2) 55 frames from the dynamic PET dataset and a frame from static PET. To calculate the time-integrated activity coefficients (TIACs) for the tracer in source organs, two methods were used: a routine method (RM) that fitted Dataset 1 by using a conventional bi-exponential curve fitting approach, and a hybrid method (HM) that fitted Dataset 2 by combining rectangular integration for the early phase of TACs with exponential fitting for the later phase. Absorbed and effective radiation doses were subsequently estimated using OLINDA/EXM version 1.1.ResultsThe [68Ga]Ga-NOTA-SNA002 primarily accumulated in the urinary system, with relatively low overall uptake levels in other source organs. Analysis of tumor-to-organ standardized uptake value (SUV) ratios revealed that all ratios reached their highest values at 120 min post-injection. The TIACs derived from the HM were consistently higher than those obtained using the RM method, while preserving a similar rank order of distribution across organs. In both methods, the kidneys, liver, and lungs exhibited the largest TIACs. The bladder wall, kidneys, and spleen received the highest absorbed doses. The total effective dosimetry calculated with RM and HM were 20.47 ± 3.08 µSv/MBq and 36.33 ± 6.18 µSv/MBq, respectively.ConclusionThe radiation dosimetry results obtained from both the RM and the HM consistently support the safety and feasibility of total-body PET/CT imaging using [68Ga]Ga-NOTA-SNA002. Meanwhile, the HM provides theoretically more accurate results compared to the RM, and is therefore recommended for radiation dosimetry evaluation of other novel tracers.Supplementary InformationThe online version contains supplementary material available at 10.1186/s40658-025-00828-2.Keywords: Programmed death-ligand 1 (PDL-1), Positron emission tomography (PET), Dynamic imaging, Internal dosimetry Introduction Programmed Cell Death-1 (PD-1) / Programmed Cell Death-Ligand 1 (PD-L1) inhibitors, a class of immune checkpoint inhibitors, serve as a first-line treatment for many cancers [1]. Unlike traditional invasive methods such as tissue sampling followed by immunohistochemical assessment of PD-L1 expression, positron emission tomography/computed tomography (PET/CT) imaging with PD-L1–targeted tracers enables non-invasive, real-time assessment of PD-L1 expression, which is a key factor in determining the efficacy of PD-1/PD-L1 checkpoint inhibitor treatment [2, 3]. [68Ga]Ga-NOTA-SNA002 is a novel single-domain antibody-based PET tracer that targets PD-L1. As one of the new radiotracers, radiation dosimetry evaluation is essential for clinical translation. The Medical Internal Radiation Dose (MIRD) formalism, the most widely used approach for organ-level dosimetry, divides the absorbed dose estimation process into two steps: time-integrated activity estimation and absorbed dose calculation [4, 5]. The former step relies on acquiring a sufficient number of measurements to adequately characterize tracer kinetics within source organs and generate time-activity curves (TACs). An appropriate fitting model is then utilized to generate time-integrated activity coefficients (TIACs) [6, 7]. The latter step utilizes time-integrated activity coefficients as input and applies dose factors (S-values) derived from the standard phantoms to estimate absorbed radiation doses [8]. These dose factors are also available in software such as OLINDA/EXM, IDAC, and MIRDCalc [9]. In recent years, additional methods for absorbed dose calculation have been developed, including dose-point kernel convolution and Monte Carlo radiation transport simulation. It is worth pointing out that regardless of the calculation method employed, the accuracy of TIAC estimation directly influences the reliability of the resulting absorbed dose values. However, conventional PET/CT with axial view of view(AFOV) less than 35 cm acquisition protocols typically consist of static acquisitions at discrete time points, in which each acquisition yields only a single measurement per organ. Conducting multiple discrete scans imposes considerable patient burden due to prolonged examination times and repeated imaging sessions. This limitation restricts the acquisition of high-frequency time points. Consequently, TACs generated under routine clinical protocols are often derived from a limited set of measurements, which compromises the accuracy of TIAC calculation [6, 10]. Furthermore, different organs can exhibit different kinetic phases and/or different effective half-lives. The choice of the acquisition time points in the past usually prioritized a more accurate TACs estimation of the most irradiated organs and/or organs at risk, neglecting the assessment of other potentially affected organs [4]. The emergence of total-body PET/CT with 194 cm axial field of view (AFOV), with its large coverage, superior sensitivity and detection efficiency, enables dynamic imaging of the entire body [11]. It offers a significant opportunity to improve dosimetry evaluation for radiotracers. Simultaneous total-body imaging enables comprehensive assessment of tracer biodistribution and temporal changes, overcoming the temporal inconsistencies of conventional multi-bed short-axial PET scans [12]. In addition, unlike conventional static acquisition protocol that yield only a single measurement per scan, dynamic total-body PET acquisition preserves quantitative accuracy (even with frame durations as short as 1 s) while providing an abundance of temporal data points derived from appropriate reconstruction schemes [13, 14]. Although previous studies have utilized large field of view, dynamic total body PET imaging for radiation dosimetry evaluation of various tracers [12, 15–17], it remains a challenge whether curve-fitting and dosimetry calculation methods should be modified based on the characteristics of dynamic datasets. Therefore, the aim of this study is to explore a hybrid curve-fitting method for robust estimation of radiation dosimetry that is suitable for total-body dynamic PET imaging, and to evaluate the radiation dosimetry and biodistribution of [68Ga]Ga-NOTA-SNA002 in human for assessing its safety as a potential radiotracer. Materials and methods Study design and patients This study was part of a prospective, open-label clinical trial that aimed to investigate the safety, distribution and dosimetry of a novel PET tracer, [68Ga]Ga-NOTA-SNA002. This trial was registered at www.chinadrugtrials.org.cn (registry number: CTR20232251), and was approved by the Ethics Committee of Zhongshan Hospital, Fudan University (approval number: 2023-091R). Inclusion criteria for this study were: patients with pathologically confirmed solid tumors, good general condition with an Eastern Cooperative Oncology Group (ECOG) performance status of 0–2, ability to tolerate the examination, and no other major diseases. PET/CT image acquisition All PET/CT scans were obtained on a total-body PET/CT scanner with an AFOV of 194-cm (uEXPLORER, United Imaging Healthcare, Shanghai, China). Synthesis and purification of tracer [68Ga]Ga-NOTA-SNA002 were carried out as reported previously [18, 19]. Injected activities ranged from 37 to 200 MBq, and patients were assigned to four mass dose groups (0.1 mg, 0.3 mg, 0.5 mg, and 0.9 mg). A total-body low-dose CT (10 mAs, 120 kV) was initially performed for attenuation correction (AC). Then, a 60-min duration dynamic PET scan was started immediately after intravenous injection [68Ga]Ga-NOTA-SNA002 via a vein near the ankle. Next, diagnostic CT was performed using the following parameters: modulated tube current and tube voltage of 120 kV for the body, while 258 mAs and 120 kV for the head; collimation, 64 × 0.5 mm; pitch, 1.0; and matrix, 512 × 512. In addition, 20-min static PET with another ACCT was obtained at 120 min post-injection (p.i.). The dynamic PET raw data were divided into 55 frames: 36 × 5 s and 19 × 180 s [11, 13, 14, 20]. In addition, two sets of PET images were reconstructed over 15–35 min and 40–60 min intervals post-injection, simulating static PET acquisitions performed at “30 min” and “60 min” after injection, respectively. All PET images were reconstructed using ordered subset expectation maximization algorithm with the following parameters: time of flight and point spread function modeling, 2 iterations and 20 subsets, matrix 192 × 192, slice thickness 2.886 mm, pixel size 3.125 × 3.125 × 2.886 mm3 with a Gaussian post-filter (3 mm), and CT-based attenuation and scatter correction [13, 14, 20]. Time activity curves and biodistribution All images were transferred to PMOD software (version 3.2, Zürich, Switzerland) for subsequent image processing. Registration of PET and CT images was performed using mutual information maximization [21]. Two PET datasets were generated for time-activity curve (TAC) derivation: the first consisted of the three abovementioned PET images obtained at approximately “30 min”, “60 min” and 120 min p.i, respectively. Thus, the Dataset 1 provided three data points and simulating a conventional multi-time-point static imaging protocol; the second was a composite 56-frame dataset, including 55 frames reconstructed from the dynamic PET acquisition and one static frame acquired at 120 min p.i. A schematic diagram of the acquisition and reconstruction workflow is presented in Fig. 1. Schematic diagram illustrating the PET acquisition protocol, image reconstruction workflow, and the composition of the two PET datasets Then, TACs were obtained by delineating volumes of interest (VOI) around tumor and 16 source organs: brain, thyroid, heart contents, lungs, stomach, liver, spleen, gallbladder contents, kidneys, pancreas, small intestine, left colon, right colon, red marrow, prostate/uterus, and urinary bladder contents. VOI segmentation was performed manually on a slice-by-slice basis by two experienced nuclear medicine physicians. Initial organ contours were delineated on the corresponding CT images and subsequently refined on the PET images to minimize spillover effects and partial-volume effects through appropriate boundary adjustments. For the analysis of tumor-to-organ standardized uptake value (SUV) ratios, specific VOIs were defined as follows [22]: the blood pool VOI was generated by placing the largest possible regions of interest (ROIs) across 10 consecutive axial slices within the lumen of the descending aorta, carefully avoiding the vessel wall; the muscle VOI was defined by placing five circular ROIs (25 mm in diameter) across consecutive slices of the bilateral gluteus maximus muscles; and the background VOI was placed in the normal tissue contralateral to the tumor, matched as closely as possible in volume to the tumor VOI. The time-integrated activity coefficients calculation The time-integrated activity coefficients (TIACs) were obtained using both a routine method (RM) and a hybrid method (HM). The RM used a conventional bi-exponential fitting approach applying multi-time-point static acquisition protocols dataset (Dataset 1) as conducted in previous studies [4, 23, 24]. The HM employed a hybrid curve fitting strategy involving segmented fitting of TACs, which was applied to Dataset 2. Specifically, rectangular integration was employed during the initial phase of the TACs, until a clear exponential trend was observed. Once an exponential pattern was observed, the method transitioned to exponential curve fitting. The transition point was typically set at the TAC peak. To assess the goodness of fit among mono-exponential, bi-exponential, and tri-exponential models, the corrected Akaike Information Criterion (AICc) was used, as described in previous studies [25]. The model with the lowest AICc is selected. The ΔAICc was calculated as , where i represents the AICc value for the i-th model, and AICcmin represents the smallest AICc value among all compared models. The TIAC of red bone marrow was estimated by manually delineating the marrow region within the 2nd to 4th lumbar vertebrae and assuming that red marrow activity concentration in 2nd to 4th lumbar vertebrae represented total-body red marrow and that 2nd to 4th lumbar vertebrae contains 6.7% of total red marrow [26]. Notably, red bone marrow absorbed dose in the present study was estimated using quantitative PET imaging to derive specific bone marrow uptake [27, 28], rather than blood-based methods. In addition, for Dataset 1, a voiding bladder model was employed for dosimetry evaluation of urinary bladder according to previous study [29]. Bladder TIACs for Dataset 2 were calculated using trapezoidal integration. Radiation dosimetry estimation Based on the framework outlined by the MIRD Committee, the TIACs results from different fitting methods were imported into OLINDA/EXM software (version 1.1, Vanderbilt University) to calculate the absorbed doses and the effective doses, using standard adult phantoms [30]. Statistical analysis Quantitative values are expressed as means ± standard deviation. Curve fitting was performed using PMOD software. Statistical analyses were performed using MATLAB R2023b. Differences in TIACs between the two methods were assessed using a paired t-test. Results Patient characteristics and PET imaging Based on the inclusion criteria, a total of 16 patients were enrolled in this study. Of these, one patient was initially evaluated as having non-small cell lung cancer during the screening phase but was later confirmed by biopsy to have pulmonary metastasis from gastroesophageal junction adenocarcinoma. Patient characteristics and administered activity of [68Ga]Ga-NOTA-SNA002 (radiochemical purity > 94%) are summarized in Table 1. The tracer injection was well tolerated by all of the 16 patients. No adverse events or immediate side effects related to tracer administration were observed. Table 1Patients characteristicsPatient (no.)Age (year)SexHeight (cm)Weight (kg)Dose (mg)Injected activity (MBq)Diagnosis171Male175.059.00.173.44Gastric adenocarcinoma with lung metastasis264Male157.047.00.150.01Non-small cell lung cancer358Male168.162.00.1143.56Non-small cell lung cancer455Male171.861.90.3149.14Non-small cell lung cancer566Male174.077.00.382.12Non-small cell lung cancer656Female159.552.20.3141.11Non-small cell lung cancer774Male165.072.00.5126.73Non-small cell lung cancer871Male161.655.60.5127.61Non-small cell lung cancer958Male170.569.40.5144.56Non-small cell lung cancer1073Female146.647.40.5111.36Non-small cell lung cancer1169Female150.956.10.5122.94Non-small cell lung cancer1270Male151.454.50.9103.27Non-small cell lung cancer1361Male171.069.90.9117.74Non-small cell lung cancer1460Male169.079.00.998.95Non-small cell lung cancer1566Male154.939.90.9100.83Non-small cell lung cancer1675Male158.055.20.9107.62Non-small cell lung cancer Patients characteristics Time activity curves and biodistribution Representative TACs of delineated source organs in a typical patient are shown in Fig. 2a. Aggregated TACs from the remaining participants are provided in supplementary file. In brief, except for the urinary bladder contents TAC, which showed a gradual increase followed by a slow decline, the TACs of all other organs exhibited a rapid rise to peak values shortly after tracer administration, followed by a continuous decline and stabilization at relatively low levels. Example of time-activity curves (TACs) for source organs in Subject No. 1 (a), and comparison of liver TACs fitted using two different methods (b) In the context of biodistribution, the kidneys and urinary bladder demonstrated the highest SUVmean values. Specifically, in the 15–35 min PET image, the SUVmean were 14.50 ± 1.75 for the kidneys and 58.67 ± 31.87 for the bladder; in the 40–60 min image, 13.89 ± 2.46 and 84.58 ± 46.64, respectively; and in the 120 min image, 17.27 ± 4.87 and 27.08 ± 15.66, respectively. Among other organs, SUVmean showed a decreasing trend over time, and spleen maintained relatively high SUVmean across all three time points. These findings were consistent with preclinical biodistribution studies in animals. Detailed biodistribution data of [68Ga]Ga-NOTA-SNA002 are presented in Table 2. Table 2Pooled tumor-to-organ SUV ratio and SUVmean at 3 time points after [68Ga]Ga-NOTA-SNA002 administrationParameters15–35 min PET image40–60 min PET image120 min PET image SUV mean Tumor1.64 ± 0.461.30 ± 0.450.94 ± 0.55Background0.62 ± 0.210.41 ± 0.160.22 ± 0.09Brain0.13 ± 0.020.07 ± 0.020.03 ± 0.01Thyroid1.90 ± 0.401.12 ± 0.260.51 ± 0.14Blood3.27 ± 0.561.64 ± 0.390.67 ± 0.22Heart2.63 ± 0.481.49 ± 0.330.66 ± 0.20Lung0.57 ± 0.130.39 ± 0.100.23 ± 0.07Stomach0.99 ± 0.500.81 ± 0.370.51 ± 0.23Liver1.82 ± 0.331.14 ± 0.250.62 ± 0.19Spleen3.58 ± 0.652.68 ± 0.611.48 ± 0.37Gallbladder0.90 ± 0.390.73 ± 0.300.40 ± 0.10Kidneys14.50 ± 1.7513.89 ± 2.4617.27 ± 4.87Pancreas1.85 ± 0.351.16 ± 0.260.54 ± 0.18Intestine1.32 ± 0.230.99 ± 0.210.55 ± 0.14Left colon0.82 ± 0.280.70 ± 0.230.39 ± 0.11Right colon0.82 ± 0.230.67 ± 0.190.39 ± 0.14Red marrow1.04 ± 0.270.72 ± 0.240.42 ± 0.19Prostate1.56 ± 0.491.17 ± 0.680.92 ± 0.56Uterus (n = 3)1.38 ± 0.280.99 ± 0.160.91 ± 0.12Urinary bladder58.67 ± 31.8784.58 ± 46.6427.08 ± 15.66Muscle0.57 ± 0.090.44 ± 0.070.26 ± 0.06Whole body0.23 ± 0.210.30 ± 0.160.17 ± 0.09SUVmax – to – SUVmean ratioTumor to background5.47 ± 2.466.16 ± 3.158.18 ± 3.82Tumor to lung5.64 ± 2.036.10 ± 2.517.83 ± 3.52Tumor to blood0.95 ± 0.251.42 ± 0.472.66 ± 1.38Tumor to liver1.69 ± 0.442.02 ± 0.702.86 ± 1.69Tumor to muscle5.44 ± 1.285.21 ± 1.756.84 ± 3.78 Pooled tumor-to-organ SUV ratio and SUVmean at 3 time points after [68Ga]Ga-NOTA-SNA002 administration Additionally, analysis of tumor-to-organ SUV ratios revealed that all ratios reached their highest values at 120 min post-injection. A representative example is shown in Fig. 3. Total-body PET/CT images at three time points from subject No. 7 with non-small cell lung cancer following administration of [68Ga]Ga-NOTA-SNA002. Radiotracer activity is evident in the blood pool at early time points, with consistently high physiological uptake observed in the spleen, kidneys, and urinary bladder. Tumor uptake in the lung is clearly visible, with corresponding SUVmax values of 4.7 (a), 4.5 (b), and 5.2 (c), respectively Time-integrated activity coefficients In the HM, the transition time points from rectangular to mono-exponential fitting occurred very early for all source organs. The latest mean transition time was observed in the kidneys (5.1536 ± 1.4501 min). To determine the optimal exponential fitting model of HM, the goodness of fit was evaluated by selecting the exponential fitting model with the minimum corrected Akaike information criterion (AICc) value. Among all 16 source organs and whole body fitted using HM, mono-exponential fitting generally was generally associated with the highest ΔAICc values. Six organs, including the gallbladder contents, pancreas, left colon, right colon, prostate, and uterus, exhibited the smallest average ΔAICc for bi-exponential fitting. For the remaining organs, tri-exponential fitting yielded the smallest average ΔAICc. The TIACs of source organs and the ΔAICc obtained from different models are summarized in supplementary file 1. The TIACs of source organs and whole body using the best-fitting exponential model from HM are shown in Table 3. Table 3The TIAC results calculated using the RM and the HMOrganRM (h)HM (h) p Brain0.002 ± 0.0010.002 ± 0.0010.435Thyroid0.000 ± 0.0000.000 ± 0.000< 0.001Heart contents0.023 ± 0.0070.018 ± 0.005< 0.001Lungs0.025 ± 0.0050.020 ± 0.004< 0.001Stomach0.007 ± 0.0030.005 ± 0.002< 0.001Liver0.040 ± 0.0070.031 ± 0.005< 0.001Spleen0.009 ± 0.0040.007 ± 0.003< 0.001Gallbladder contents0.000 ± 0.0000.000 ± 0.000< 0.001Kidneys0.131 ± 0.0270.115 ± 0.025< 0.001Pancreas0.001 ± 0.0000.001 ± 0.000< 0.001Small intestine0.021 ± 0.0090.018 ± 0.008< 0.001Left colon0.004 ± 0.0020.003 ± 0.001< 0.001Right colon0.005 ± 0.0020.004 ± 0.002< 0.001Red marrow0.022 ± 0.0140.018 ± 0.0050.211Prostate0.000 ± 0.0000.000 ± 0.0000.093Uterus0.000 ± 0.0000.000 ± 0.0000.043Urinary bladder contents0.365 ± 0.0920.317 ± 0.081< 0.001Whole body1.507 ± 0.0531.154 ± 0.058< 0.001the TIAC of urinary bladder contents was derive using trapezoidal integration applied to Dataset 1. The p values indicate statistical comparisons between the TIAC results obtained using the routine method (RM) and the hybrid method (HM) The TIAC results calculated using the RM and the HM the TIAC of urinary bladder contents was derive using trapezoidal integration applied to Dataset 1. The p values indicate statistical comparisons between the TIAC results obtained using the routine method (RM) and the hybrid method (HM) For the TIAC results from both RM and HM, a consistent pattern emerged that the TIACs derived from the HM were consistently higher than those obtained using the RM method, while preserving a similar rank order of distribution across organs. Statistical analysis revealed that the TIAC results obtained using the two methods differed significantly in almost all source organs. Notably, the whole body TIAC, which reflects the total cumulative tracer exposure integrated over time, was substantially greater with the HM (1.507 ± 0.053 h) than with the RM (1.154 ± 0.058 h). Radiation dosimetry The absorbed doses and effective doses for all source organs are summarized in Table 4. The calculated absorbed doses from two methods showed a similar rank order of distribution, with the highest absorbed dose observed in the bladder wall, kidneys, and spleen. For RM, the results were 181.49 ± 51.55 µGy/MBq, 181.19 ± 39.20 µGy/MBq, and 23.14 ± 9.58 µGy/MBq, respectively. For HM, the absorbed doses were 452.94 ± 114.34 µGy/MBq, 208.19 ± 42.78 µGy/MBq, and 28.76 ± 11.52 µGy/MBq. Consistent with the TIAC findings, the HM yielded higher absorbed doses in nearly all target organs compared with the RM. The total effective dose calculated using the RM was 20.47 ± 3.08 µSv/MBq, while that calculated using the HM was a higher value of 36.33 ± 6.18 µSv/MBq. Table 4[68Ga]Ga-NOTA-SNA002 dosimetry summary of mean absorbed and effective doses using OLINDA/EXM 1.1RMHMTarget organAbsorbed dose (µGy/MBq)Effective dose (µSv/MBq)Absorbed dose (µGy/MBq)Effective dose (µSv/MBq)Adrenals8.11 ± 1.250.03 ± 0.0111.62 ± 1.520.06 ± 0.01Brain1.47 ± 0.330.01 ± 0.001.90 ± 0.420.01 ± 0.00Breasts4.88 ± 0.920.24 ± 0.057.53 ± 1.150.38 ± 0.06Gallbladder wall8.09 ± 1.390.00 ± 0.0014.61 ± 10.990.00 ± 0.00Left colon wall12.46 ± 2.251.50 ± 0.2719.59 ± 2.822.35 ± 0.34Small intestine16.76 ± 4.870.06 ± 0.0322.81 ± 5.710.11 ± 0.03Stomach wall11.31 ± 2.551.36 ± 0.3116.21 ± 3.371.95 ± 0.40Right colon wall10.90 ± 2.210.04 ± 0.0215.86 ± 2.810.08 ± 0.01Heart wall15.61 ± 3.680.00 ± 0.0021.66 ± 4.500.00 ± 0.00Kidneys181.19 ± 39.203.03 ± 2.13208.19 ± 42.781.04 ± 0.21Liver11.18 ± 2.420.56 ± 0.1214.81 ± 3.020.74 ± 0.15Lungs11.34 ± 2.201.36 ± 0.2714.90 ± 3.011.79 ± 0.36Muscle5.91 ± 0.910.02 ± 0.019.53 ± 1.130.05 ± 0.01Ovaries9.79 ± 0.791.96 ± 0.1616.87 ± 1.243.37 ± 0.24Pancreas7.64 ± 1.830.03 ± 0.0110.20 ± 2.230.05 ± 0.01Red marrow8.17 ± 1.340.98 ± 0.1611.71 ± 2.681.40 ± 0.32Osteogenic cells9.81 ± 2.040.10 ± 0.0214.76 ± 3.300.15 ± 0.03Skin4.78 ± 0.840.05 ± 0.017.58 ± 1.030.08 ± 0.01Spleen23.14 ± 9.580.08 ± 0.0528.76 ± 11.520.14 ± 0.06Testes6.02 ± 0.600.00 ± 0.0010.67 ± 0.580.00 ± 0.00Thymus5.63 ± 1.060.02 ± 0.018.65 ± 1.310.04 ± 0.01Thyroid5.62 ± 1.800.28 ± 0.097.32 ± 2.180.37 ± 0.11Urinary bladder wall181.49 ± 51.559.08 ± 2.59452.94 ± 114.3422.64 ± 5.71Uterus/prostate8.25 ± 1.290.03 ± 0.0116.05 ± 2.740.08 ± 0.01Total body7.37 ± 1.0520.47 ± 3.0811.31 ± 1.3336.33 ± 6.18 [68Ga]Ga-NOTA-SNA002 dosimetry summary of mean absorbed and effective doses using OLINDA/EXM 1.1 Discussion In this study, the TACs of [68Ga]Ga-NOTA-SNA002 obtained from most source organs consistently demonstrated early peak times and short equilibrium phases. Following a single administration, the radiotracer primarily accumulated in the urinary system, with relatively low overall uptake levels in other source organs. Throughout the entire trial, no side effects or adverse events were reported within the 120 min examination. By applying differing processing strategies to the same imaging dataset, we generated two distinct datasets. Dataset 1 was constructed to simulate the conventional multi-time-point static imaging protocol. Given that tracer activity at the 2-hour imaging time point was reduced due to biological clearance and physical decay of [68Ga]Ga-NOTA-SNA002, a 20-min acquisition duration was employed to ensure sufficient counts and reliable quantification. Thereby, composite images from the 1-hour dynamic acquisition were reconstructed over identical 20-min intervals (15–35 min and 40–60 min post-injection) to accurately emulate the image quality and quantitative characteristics of static acquisitions. Dataset 2 consisted of 55 frames reconstructed from the 60-min dynamic acquisition combined with the single 20-min static frame acquired at 120 min post-injection, resulting in a total of 56 temporal data points. For dynamic image reconstruction, a reconstruction scheme of 36 × 5 s + 19 × 180 s was adopted. This selection was informed by prior investigations demonstrating that [68Ga]Ga-NOTA-SNA002 exhibits no distinct plateau phase, with peak uptake occurring predominantly at early time points [19]. The high temporal resolution in the initial phase was therefore implemented to prevent omission of the activity peak. Furthermore, owing to the exceptionally high detection efficiency of the total-body PET/CT equipment, images reconstructed using OSEM maintained quantitative quality even at frame durations as short as 1 s [13], and acquisitions of 30–45 s were shown to yield images meeting the requirements for diagnosis [31]. Balancing image quality, quantitative accuracy, and accumulated institutional experience, the selected framing scheme was deemed optimal [20]. To calculate TIACs, the RM adhered to established MIRD formalism, which was widely accepted for both imaging protocol design and data-fitting strategy selection, offering a convenient and efficient means of obtaining dosimetry estimates [4, 6, 30, 32]. Specifically, source organ TACs were fitted with bi-exponential functions using the three discrete time points, whereas a voiding model was used for the urinary bladder. Recent advancements in PET/CT equipment have made simultaneous total-body dynamic acquisition protocols increasingly feasible. Such protocols enable continuous monitoring of tracer distribution across all organs, capturing the full range of pharmacokinetic processes. Prior work from our group, however, showed that direct application of exponential fitting to sparsely sampled data introduced substantial bias. The fewer the data points included, the larger the bias introduced. Moreover, even when exponential fitting was applied across the entire TAC derived from high-temporal-resolution dynamic data, fitting accuracy was compromised. The rapid early distribution phase could influence the fitting process, resulting in curves that deviate substantially from physiological kinetics [14]. Owing to its nature as a single-domain antibody-based imaging agent, [68Ga]Ga-NOTA-SNA002 exhibits numerous favorable physicochemical properties, including small size, high stability, rapid tumor uptake, and efficient normal-tissue clearance [18, 19]. The equilibrium phase was found to be very brief in most organs. Consequently, a 60-min dynamic acquisition window was sufficient to encompass the transition points in source organs. This kinetic characteristic permitted the application of a hybrid fitting strategy in the HM, in which rectangular integration was applied to the initial phase of the TAC and exponential fitting was subsequently employed for the later phase. Due to the sufficient number of early time points from the 1-hour dynamic acquisition data, rectangular integration enables HM to accurately capture blood flow and tracer biodistribution in source organs, providing more comprehensive and authentic information compared to RM. Similarly, for the exponential phase, the dense time sampling allowed for HM selecting the optimal individualized fitting model for each organ by comparing AICc values from different models [25]. Notably, both bi-exponential and tri-exponential models yielded optimal fits, consistent with the complex pharmacokinetics of tracer clearance. Another key distinction between HM and RM lies in their extrapolation performance. As demonstrated in the magnified view of Fig. 2b, TACs derived from 60-min dynamic PET data revealed that nearly all organs exhibited a distinct uptake and distribution phase, characterized by an initial increase from a low baseline to a peak within the first few min post-injection. In contrast, curves generated using the RM markedly underestimated cumulative activity during this initial phase relative to the high-temporal-resolution dynamic data. Furthermore, although both methods rely on fitted equations for extrapolation, the HM, supported by considerably greater kinetic sampling density, yielded markedly more stable and biologically consistent extrapolations. Notably, there lacked an objective standard to definitively the superiority of TIAC estimates between the RM and the HM. When judged against the high-temporal-resolution dynamic measurements, however, the HM results likely yielded more physiologically plausible and exhibited greater explanatory power. In this context, whole body TIAC can provides particular interpretive value. Unlike individual organ kinetics, whole body TIAC is influenced predominantly by physical decay and detector sensitivity rather than biological clearance when no voiding or excretion occurs during the acquisition period. In the absence of excretion, the theoretical value of the whole body TIAC is equal to total body cumulated activity divided by the administered activity [33]. According to the radioactive decay law, this value equals the physical half-life divided by ln(2). For a tracer labeled with 68Ga, this theoretical upper limit value is approximately 1.64 h. Given that [68Ga]Ga-NOTA-SNA002 is almost exclusively excreted via the urinary tract and participants voided only once between the dynamic and delayed static acquisitions, the observed whole body TIAC should be slightly below this theoretical maximum. From this benchmark, whole body TIAC results derived from the dynamic protocol (1.507 ± 0.053 h) demonstrated clear superiority over conventional multi-time-point static acquisition protocol (1.154 ± 0.058 h). Furthermore, across all source organs, TIACs derived from the HM were consistently higher than those obtained with the RM. This systematic difference suggests that the contribution of the early blood-flow-dominated distribution phase may have been substantially underestimated in conventional acquisition protocol. Radiation dosimetry calculations were performed using standard phantoms implemented in OLINDA/EXM version 1.1, based on the calculated TIACs. In the radiation dosimetry results, both RM and HM yielded similar rank orders of organ doses. The bladder wall, kidneys, and spleen were consistently identified as the top three organs receiving the highest absorbed doses. The whole body effective dose calculated by the RM was 20.47 ± 3.08 µSv/MBq, while that calculated by the HM was 36.33 ± 6.18 µSv/MBq. These effective dose levels remained well below limits for individual organ doses [34, 35], were similar to that reported for [68Ga]Ga-WL12 (18.50 ± 4.07 µSv/MBq) [10], and were substantially lower than the human-projected effective doses for tracers that have not undergone extensive clinical evaluation, such as [89Zr]Zr-DFO-anti-PD-L1 (131 µSv/MBq) [36] and [124I]I-durvalumab (186 µSv/MBq) [37]. There were several limitations in this study. First, no dedicated motion correction was applied during the dynamic acquisition, and organ segmentation relied entirely on manual VOI delineation rather than artificial intelligence-based automated methods. Although contours were carefully adjusted to minimize motion-related blurring and spillover effects, this approach substantially increased operator workload and introduced inter-observer variability that potentially affected reproducibility. In addition, the heterogeneity of tumor types among patients represented another potential source of variability in tracer biodistribution and kinetics. Furthermore, due to the absence of a notable plateau phase in most source organ TACs, the highest point was generally chosen as the transition point for switching from rectangular integration to exponential fitting. However, it is not a universally applicable choice. Especially for radiotracers exhibiting markedly slower clearance or longer effective half-lives, a 60-min dynamic acquisition performed immediately after administration or even a delayed dynamic acquisition may still have been insufficient to capture the transition points that characterize the later phase. This reflects the need to carefully consider whether the choice of transition point aligns with the physiological kinetics when applying the HM to other tracers. Moreover, the HM approach was challenging to apply to certain organs whose TACs do not exhibit a clear exponential trend, such as the urinary bladder in this study. Unlike RM, which utilized a voiding model incorporating physiological factors such as variable bladder volume, urine entry rates, and voiding schedules to allow for more accurate radiation dose estimations [38], the HM used trapezoidal integration to determine TIAC of the urinary bladder contents [39]. This simplification was deemed acceptable based on prior institutional experience indicating that absorbed dose to the urinary bladder wall is predominantly contributed by the first two voiding cycles post-injection [14]. The 60-min dynamic acquisition effectively captured tracer accumulation in the bladder prior to the initial void, whereas the 120-min static acquisition provided activity concentration immediately before the second void. Given the relatively rapid biological clearance of [68Ga]Ga-NOTA-SNA002 and the short physical half-life of 68Ga, omission of explicit voiding intervals in the integration process introduced limited bias. Under these conditions, trapezoidal integration was considered an acceptable approximation for urinary bladder TIAC estimation in this specific tracer cohort. Despite these limitations, the continuous 60-minute dynamic acquisition performed without prior voiding offers a distinct advantage in assessing bladder dosimetry. The dynamic acquisition partially reflects the subject-specific physiological clearance of the tracer via the urinary system. In contrast to conventional voiding models, continuous dynamic imaging may offer a more physiologically accurate assessment of bladder wall radiation exposure. Moreover, due to the high sensitivity and detection efficiency of the long-axial PET equipment, the total-body dynamic acquisition protocol used in this study can also be applied to pharmacokinetics evaluation of new drugs. This approach enables the use of imaging-derived pharmacokinetic parameters to replace the traditional method based on multiple invasive blood samplings. These are research areas currently being explored by our team. Conclusion This study revealed that the radiation dosimetry results from RM and HM, along with biodistribution findings, consistently demonstrated the safety and feasibility of PET/CT imaging with [68Ga]Ga-NOTA-SNA002. Meanwhile, the HM provides theoretically more accurate results compared to the RM, and is therefore recommended for radiation dosimetry evaluation of other novel tracers. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements Not applicable. Abbreviations Positron emission tomography Computed tomography Programmed death-ligand 1 Time-integrated activity coefficients Routine method Optimized method Standardized uptake value Axial field of view Time-activity curves Volume of interest Regions of interest Attenuation correction Post-injection The corrected akaike information criterion Author contributions Material preparation, data collection and analysis were performed by Yunze Xie. The first draft of the manuscript was written by Yunze Xie and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Funding This study was funded by the Shanghai 2024 “Explorer” program (grant number: 24TS1411100), the Innovative Medical Device Application Demonstration Program of Shanghai Municipal Commission of Economy and Informatization (grant number: 23SHS01200 and 23SHS01200-01), the Key Program of Ministry of Industry and Information Technology of China (grant number: CEIEC-2022-ZM02-0219), and the Chinese National Key Clinical Specialty Program (grant number: YWP2022-007). None of these funding bodies have been associated with the design of the study, or collection, analysis, and interpretation of data, or in writing the manuscript. Data availability The data that support the findings of this study are available from the corresponding author upon reasonable request. Declarations Ethics approval and consent to participate This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of Zhongshan Hospital, Fudan University (approval number: 2023091R). Written informed consent was signed for the use of the data from all the subjects. Consent for publication The authors affirm that human research participants provided informed consent for publication of the images in Figures 1 and 3. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Yunze Xie, Wenxin Tang and Lichao Zhang have contributed equally to this work. Contributor Information Guobing Liu, Email: liu.guobing@zs-hospital.sh.cn. Hongcheng Shi, Email: shi.hongcheng@zs-hospital.sh.cn. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement The data that support the findings of this study are available from the corresponding author upon reasonable request.