Feasibility of Ultrashort Echo Time MR Imaging for the Detection of Subclinical Achilles Enthesitis in Psoriasis and Psoriatic Arthritis Authors: Takeshi Fukuda, Yoshinori Umezawa, Mami Momose, Akihiko Asahina, Akira Ogihara, Ryuichi Itou, Hisashi Kitagawa, Katsutoshi Murata, Stefan Sommer Published: 2026 ABSTRACT AbstractSubclinical inflammation (SI) is clinically relevant in psoriasis (PsO) and psoriatic arthritis (PsA), but quantitative imaging biomarkers are lacking. We evaluated the feasibility of ultrashort echo time (UTE) bicomponent MRI for quantifying Achilles enthesis SI. Conventional MRI showed no group differences, whereas short T2* (T2*s) values were significantly elevated in PsO/PsA compared with healthy volunteers. These findings support the feasibility of UTE bicomponent analysis, as a promising tool for evaluating SI and warrant further investigation.Keywords: inflammation, magnetic resonance imaging, psoriasis, psoriatic arthritis Introduction Subclinical inflammation (SI) is defined as the presence of imaging abnormalities with or without arthralgia in the absence of clinical arthritis.1 Importantly, SI has been reported to predict arthritis development in psoriasis (PsO) patients,2 and the European Alliance of Associations for Rheumatology (EULAR) task force emphasized that SI is one of the key elements of subclinical psoriatic arthritis (PsA) and can be considered a short-term predictor of PsA onset.1 Furthermore, recognition of SI may have therapeutic implications, as the presence of risk factors for PsA development could influence drug choice in PsO, favoring systemic agents effective in both skin and joint disease.1 In established PsA, ultrasound studies have shown that SI is highly prevalent in patients with early oligoarthritis and often reclassifies them as having polyarthritis, which carries a worse prognosis and may warrant escalation to more intensive therapy.3 Despite its clinical relevance and prior reports using ultrasound and MRI, no previous study has attempted to quantitatively assess SI, particularly at the asymptomatic stages.4 Ultrashort echo time (UTE) imaging overcomes the limitation of conventional MRI in obtaining signal from the short-T2 tissues such as tendons and ligament entheses. Beyond morphological imaging, UTE bicomponent analysis quantifies short T2* (T2*s, representing bound water which reflects collagen integrity), long T2* (T2*l, representing free water which reflects edema), and fast fraction (FF, the proportion of bound water).5 Previous UTE bicomponent analyses have shown feasibility for detecting very early stage of abnormality like Achilles or patellar tendinopathy.5,6 However, it has never been applied to quantify SI. We therefore conducted a prospective pilot study to evaluate the feasibility of UTE bicomponent analysis for quantifying SI. Based on previous reports identifying the Achilles tendon enthesis as a frequent site of SI,7 we selected the Achilles tendon enthesis as the target site. Materials and Methods The study was approved by the Jikei University Ethics Committee (approval number: 33-167[10784]), and written informed consent was obtained from all participants. We prospectively enrolled PsO or PsA patients fulfilling the ClASsification of Psoriatic ARthritis (CASPAR)8 criteria but without clinical Achilles tendon enthesitis, and healthy volunteers (HV). PsA was diagnosed by experienced dermatology specialists at the Jikei University Hospital based on comprehensive clinical evaluation. Exclusion criteria included prior Achilles tendon injury, history of local steroid injection to the Achilles tendon, or local infection involving the tendon. All participants underwent MRI of the right ankle using a 3T scanner (MAGNETOM Vida; Siemens Healthineers, Forchheim, Germany) with sagittal T2-weighted images (T2WI) and short-tau inversion recovery (STIR). Additionally, a 3D double-echo research application UTE sequence was acquired with a total of 12 echo times (TE = 0.04, 0.23, 0.47, 0.77, 1.18, 1.80, 2.88, 4.73, 7.07, 9.61, 12.26, and 15.0 ms). ROIs were manually placed at the anatomically defined fibrocartilaginous portion of the Achilles tendon enthesis by 2 independent board-certified radiologists with 5 and 12 years of experience in musculoskeletal imaging. Each reader performed the ROI placement independently once. The ROIs were positioned at the tendon insertion near the calcaneus, covering approximately the deep half of the tendon thickness on the axial plane while avoiding the calcaneal cortex to minimize partial volume effects (Fig. 1). ROI placement was standardized across all subjects to encompass a comparable anatomical region of the enthesis. High-signal areas on the early-echo subtraction images were used as an additional guide to identify the fibrocartilaginous enthesis. Early-echo subtraction images were generated by subtracting a later-echo image (TE = 2.88 ms) from the first-echo image (TE = 0.04 ms). Because signals from short T2 tissue decay rapidly with increasing TE, this subtraction enhances signal originating from short T2 components such as the fibrocartilaginous enthesis as previously reported.9 These images were used solely to assist anatomical localization of the enthesis. For bicomponent analysis, a bi-exponential fitting model was applied to the signal decay curve obtained from multiple echo times. This model separates a rapidly decaying T2*s component, representing bound water associated with collagen, and a slow decaying T2*l component, representing free water (Fig. 2). For conventional MRI, signal intensity ratios (referred to as T2 ratio and STIR ratio, respectively) were calculated as the signal of identical enthesis ROI divided by the signal of flexor hallucis longus ROI showing homogeneous muscle signal. In addition, qualitative assessment was performed for T2WI, STIR, and UTE subtraction images using 4-point Likert scales. Different scoring systems were applied for conventional MRI sequences and UTE subtraction images because the normal fibrocartilaginous enthesis typically appears hyperintense on UTE subtraction images. The detailed scoring criteria are summarized in Table 1, and mean scores between 2 independent readers were used for analysis. Statistical analyses were performed using SPSS softwares (IBM, Armonk, NY, USA). Continuous variables are presented as median and interquartile range (IQR). Because of the small sample size and the exploratory nature of this pilot study, group comparisons were performed using the Mann–Whitney U test. P-values < 0.05 were considered statistically significant. To assess interobserver reproducibility of ROI placement and qualitative values, the Dice similarity coefficient (DSC) and intraclass correlation coefficient (ICC) were calculated. DSC and ICC were interpreted as follows: values < 0.40 indicated poor agreement, 0.40–0.59 moderate agreement, 0.60–0.79 good agreement, and ≥ 0.80 excellent agreement. ROI placement for the Achilles tendon enthesis. (a) Sagittal UTE subtraction image demonstrating the ROI placement (yellow dotted line) at the calcaneal insertion of the Achilles tendon, highlighting the fibrocartilaginous enthesis. (b) Axial image showing the ROI placed at the enthesis, covering approximately the deep half of the tendon thickness (yellow dotted line).UTE, ultrashort echo time. Schematic illustration of biexponential T2* decay. The signal decay curve consists of a rapidly decaying component (fast decay) and a slowly decaying component (slow decay), corresponding to short and long T2* component, respectively. Table 1Visual scoring systems for qualitative assessmentScoreT2WI/STIRUTE subtraction image0Normal appearance without increased signal intensity or thickeningHigh signal confined to the expected fibrocartilaginous enthesis1Mild focal signal increase without definite tendon thickeningSlightly increased conspicuity of enthesis signal but remaining localized2Moderate signal increase with mild tendon thickeningDefinite expansion of the high-signal area with extension into the perientheseal region3Marked signal increase with clear tendon thickeningMarked expansion of the high-signal area involving a broad perientheseal regionSTIR, short-tau inversion recovery; T2WI, T2-weighted images; UTE, ultrashort echo time. Visual scoring systems for qualitative assessment STIR, short-tau inversion recovery; T2WI, T2-weighted images; UTE, ultrashort echo time. Results We finally enrolled 6 PsO and 5 PsA patients, and 11 HV. There were no significant differences between PsO/PsA and HV groups in age (median [IQR], 49.0 [42.0–59.0] vs 40.0 [29.0–60.0] years; P = 0.365) or sex distribution (9 men and 2 women in each group; P = 1.000) (Table 2). Body mass index (BMI) was higher in PsO/PsA group than in HV (23.9 [22.5–27.4] vs 22.2 [20.4–23.0]; P =0.040). The disease duration of PsO was 15.0 [7.0–24.0] years in PsO/PsA group, and among them, the median duration of arthritis in PsA patients was 2.0 [0.8–7.5] years. Eight of the 11 patients were previously receiving either biologic agents or phosphodiesterase-4 inhibitors. As values below the limit of detection (0.04 mg/dL) were replaced with 0.02 mg/dL, the C-reactive protein level in the PsO/PsA group was 0.09 [0.02–0.26] mg/dL. Table 2Comparison of clinical characteristics and MRI parameters between PsO/PsA and healthy volunteersPsO/PsAHealthy volunteersP valueAge (years)49.0 [42.0–59.0]40.0 [29.0–60.0]0.365Sex (M:F)9:29:21.000BMI (kg/m2)23.9 [22.5–27.4]22.2 [20.4–23.0]0.040T2 ratio0.11 [0.09–0.13]0.11 [0.08–0.17]0.699STIR ratio0.10 [0.09–0.10]0.08 [0.07–0.10]0.270short T2* (ms)1.88 [1.60–2.22]1.52 [1.19–1.70]0.005long T2* (ms)24.5 [21.2–27.1]23.5 [18.5–26.9]0.332fast fraction0.83 [0.78–0.85]0.82 [0.78–0.84]0.699Values are presented as median [interquartile range]. Group comparisons were performed using the Mann–Whitney U test for continuous variables and Fisher’s exact test for sex distribution.BMI, body mass index; PsA, psoriatic arthritis; PsO, psoriasis; STIR, short-tau inversion recovery. Comparison of clinical characteristics and MRI parameters between PsO/PsA and healthy volunteers Values are presented as median [interquartile range]. Group comparisons were performed using the Mann–Whitney U test for continuous variables and Fisher’s exact test for sex distribution. BMI, body mass index; PsA, psoriatic arthritis; PsO, psoriasis; STIR, short-tau inversion recovery. On T2WI and STIR, signal intensity ratios showed no significant differences (PsO/PsA vs. HV; T2 ratio: 0.11 [0.09–0.13] vs 0.11 [0.08–0.17], P = 0.699; STIR ratio: 0.10 [0.09–0.10] vs 0.08 [0.07–0.10], P = 0.270). In contrast, bicomponent analysis revealed significantly elevated T2*s values in the PsO/PsA group vs. HV (1.88 ms [1.60–2.22] vs 1.52 ms [1.19–1.70], P = 0.005). T2*l values (24.5 ms [21.2–27.1] vs 23.5 ms [18.5–26.9], P = 0.332), and FF (0.83 [0.78–0.85] vs 0.82 [0.78–0.84], P = 0.699) did not differ significantly. In the exploratory subgroup analysis between PsO and PsA, no significant differences were found in any of the UTE parameters (T2*s, P = 1.000; T2*l, P = 0.329; FF, P = 0.429). In qualitative analysis, scores on UTE subtraction images did not show significant differences between PsO/PsA and HV (0.0 [0.0–1.0] vs 0.5 [0.0–0.5], P = 0.507). No abnormalities were detected on T2WI or STIR in either group, and all qualitative scores were 0 in both groups. Representative cases are shown in Fig. 3, illustrating that the Achilles enthesis appeared indistinguishable on routine protocol, whereas UTE subtraction images demonstrated high signal intensity in the PsO/PsA group patient. The high-signal area on subtraction images may reflect tissue alterations, potentially related to SI, and may be associated with increased T2*s values. Representative MRI of the Achilles tendon enthesis in a healthy volunteer and an asymptomatic PsA patient. (a–c) Forty-one-year-old healthy male: sagittal T2WI (a), STIR (b), and UTE subtraction image (c). (d–f) Forty-nine-year-old male with PsA without clinical Achilles enthesitis: sagittal T2WI (d), STIR (e), and UTE subtraction image (f). Conventional sequences (T2WI and STIR) show no abnormal signal in either case. In this representative PsA case, the UTE subtraction image demonstrates a more conspicuous high signal area at the enthesis (f, arrow) than in the healthy volunteer. This finding may reflect subtle tissue alteration. The measured short T2* value was 1.40 ms in the healthy volunteer and 2.22 ms in the PsA patient.PsA, psoriatic arthritis; STIR, short-tau inversion recovery; T2WI, T2-weighted images; UTE, ultrashort echo time. Interobserver agreement for ROI placement was good, with a mean DSC of 0.73 ± 0.08. In addition, ICC demonstrated excellent interobserver agreement for T2*s (0.94; 95% confidence interval (CI): 0.88–0.98), T2*l (0.93; 95% CI: 0.85–0.97), and FF (0.89; 95% CI: 0.77–0.95). Interobserver agreement for the T2 ratio was excellent (0.89; 95% CI: 0.75–0.95) and good for the STIR ratio (0.68; 95% CI: 0.36–0.85). Discussion This feasibility study demonstrated that UTE bicomponent MRI detects subtle Achilles tendon abnormalities in PsO and asymptomatic PsA patients, undetectable by conventional MRI. Conventional MRI and ultrasound primarily detect inflammatory changes such as edema, hyperemia, or structural thickening, which may represent relatively advanced stages of enthesitis. In contrast, UTE bicomponent MRI enables direct assessment of tendon microstructure, allowing detection of matrix alterations before overt inflammatory changes become apparent. Specifically, elevated T2*s values likely reflect early collagen matrix disorganization as previous UTE bicomponent MRI studies have demonstrated that T2*s component is particularly sensitive to early tendon degeneration and microstructural collagen abnormalities, even in cases without apparent signal changes on conventional MRI.5, 6 For example, Kijowski et al. reported that patients with patellar tendinopathy, including those without morphologic abnormalities on conventional MRI, showed significantly increased T2*s values and reduced FF, suggesting disruption of the collagen fiber network at a microscopic level.6 Similarly, Juras et al. demonstrated that T2*s component discriminated degenerated Achilles tendons from healthy tendons robustly and showed a strong correlation with clinical tendon scores, supporting its role as a sensitive marker of early tendinopathy.5 In line with these observations, our study demonstrated that only the T2*s component was significantly prolonged in PsO and asymptomatic PsA patients compared with HV, whereas the T2*l component showed no significant difference. Notably, this alteration was not reflected in the qualitative visual assessment of UTE subtraction images. This may be because UTE subtraction images are generated by simple subtraction of 2 echo images and do not directly represent the biexponential parameters derived from a quantitative model. Our finding may suggest that subclinical Achilles enthesis involvement in PsO/PsA is characterized by early collagen matrix disorganization rather than increased free water content. This interpretation is supported by previous biochemical and histopathological evidence demonstrating increased collagen turnover and degradation in PsA, which may precede overt inflammatory changes and contribute to early loss of tissue integrity.10 Our findings also align with previous UTE research by Moazamian et al., who reported elevated UTE-T1 values and reduced macromolecular proton fraction (MMF) in the Achilles tendon and enthesis of symptomatic PsA patients compared with controls, suggesting tissue degeneration and matrix alterations.11 However, this is the first study to apply bicomponent analysis, enabling separation of bound and free water contributions, in Achilles enthesis particularly in the absence of clinically apparent inflammation. Our study expands on previous work by showing that collagen integrity changes, quantified as increased T2*s, are already detectable at the subclinical stage. This also provides novel evidence that SI may be directly linked to early collagen disorganization. Because this was a pilot study, limitations include the small sample size and cross-sectional design. A formal a priori sample-size calculation was not performed, and the number of participants was determined pragmatically based on the number of eligible participants during the study period. Importantly, these limitations preclude establishing a causal relationship between the observed UTE-derived abnormalities and SI related to PsO/PsA. Furthermore, the imaging findings are not specific and may also reflect other factors such as mechanical loading or subclinical degeneration. Nevertheless, these exploratory findings support the feasibility of UTE bicomponent MRI for detecting and quantifying SI in future clinical and research applications. It would also be of interest for future studies to determine whether the subtle abnormalities detected by UTE bicomponent MRI can be identified by ultrasound. Although BMI difference was modest with no participants meeting criteria for obesity, mean BMI was significantly higher in the PsO/PsA group than in HV. Hence, potential influence of BMI on elevated T2*s cannot be completely excluded. However, higher BMI is a well-known characteristic of PsO populations,12 and our cohort reflects a representative real-world PsO population rather than a selection bias. Moreover, unlike age, which has been reported to correlate with T2*s,13 no direct association between BMI and T2*s has been established to date. As age did not differ between groups in the present study, the impact of patient characteristics on the calculated UTE parameters is considered minimal. Although prior imaging studies have reported that SI can be detected even in patients receiving biologic therapy,2 the inclusion of treated patients in the present cohort may have led to an underestimation of the differences observed between the PsO/PsA group and HV. Future studies with larger, longitudinal cohorts are warranted to clarify the independent effects of BMI and other potential confounders and to establish cutoff values predictive of PsA onset or disease progression. In conclusion, biomarkers derived from UTE bicomponent analysis such as T2*s may represent a promising tool for the evaluation of SI and warrant further investigation. Funding Statement This work was supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI (Grant Number JP21K15792). Footnotes Conflicts of Interest: KM and SS are employees of Siemens Healthineers. Their involvement in this study was limited to MRI protocol development and optimization. They had no access to the study data, were not involved in data acquisition, analysis, or interpretation, and had no influence on the study results or conclusions. All other authors declare that they have no conflict of interest. References