PSMA-RADS 2.0: clinical validation and technical considerations for prostate-specific membrane antigen positron emission tomography/computed tomography image interpretation

Introduction

Prostate-specific membrane antigen (PSMA) is a type II transmembrane glycoprotein possessing folate hydrolase and N-acetyl-alpha-linked acidic dipeptidase activities, first identified in 1987 (1). It demonstrates significantly elevated expression in aggressive prostate cancer (PCa) variants (2). The clinical adoption of PSMA-targeted radionuclide imaging has revolutionized PCa management through its superior diagnostic performance and therapeutic potential (3,4). Positron emission tomography/computed tomography (PET/CT) targeted imaging was used in clinical practice for staging, stratification, and treatment response evaluation of PSMA targeted radioligand therapy (5). Among current imaging options, PSMA PET/CT provides the most comprehensive snapshot of a patient’s tumor burden, and its accurate interpretation is pivotal for guiding subsequent prostate-cancer management (6-8).

PSMA Reporting and Data System (PSMA-RADS) 1.0 framework was the first scoring system based on molecular imaging PET/CT for PCa, and is widely used for staging PCa lesions (9-12). Although PSMA-RADS 1.0 has been increasingly adopted (13) with clear advantages, there are notable limitations, particularly for follow-up assessment of locally treated lesions. Recently, researchers have launched an improved set of categories, PSMA-RADS version 2.0, to improve characterization at the lesion level and facilitate optimal clinical decision-making. Lesions are classified into nine categories based on PSMA-RADS 2.0, including PSMA-RADS-1, PSMA-RADS-2, PSMA-RADS-3A-D, PSMA-RADS-4, PSMA-RADS-5 and PSMA-RADS-5T. Compared to the previous 1.0 version, PSMA-RADS 2.0 has simplified the definition of PSMA-RADS-1 for category 1 lesions and newly introduced PSMA-RADS 5T score, a new category specifically for post-treatment lesions (14,15). Although PSMA-RADS 2.0 has introduced significant improvements, its clinical applicability is still unclear. This study aims to evaluate the utility and reproducibility of PSMA-RADS 2.0 in clinical applications with intra- and inter-reader consistency in interpretation of PSMA PET/CT image and identify areas requiring further improvements. We present this article in accordance with the STROBE reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1928/rc).

Methods

Study population

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This retrospective study was approved by the Ethics Committee of Huaian No. 1 People’s Hospital (approval No. YX-2021-113-01). All the enrolled patients have signed the informed consent prior to the examination. From January 2021 to December 2022, PCa patients undergoing PSMA PET/CT were retrospectively analyzed in our institutional database. Exclusion criteria included: (I) histologically confirmed non-PCa primary tumors; (II) concomitant hematologic malignancies (lymphoma, multiple myeloma); (III) incomplete clinical data and poor image quality. Finally, a total of 162 PCa patients were retrospectively included in this study. The specific screening process of patients is illustrated in Figure 1.

Figure 1 Flowchart of patient screening and selection process. PCa, prostate cancer; PSMA PET/CT, prostate-specific membrane antigen positron emission tomography/computed tomography.

Imaging protocol

All scans were acquired using a Biograph mCT-X 4R PET/CT scanner (Siemens Healthineers, Erlangen, Germany) starting 60±5 minutes after intravenous administration of 192.4 MBq [interquartile range (IQR), 185–203.5 MBq] of ⁶⁸Ga-PSMA-11. The PET component employed a standardized acquisition protocol with 2 minutes per bed position, utilizing TrueX + TOF iterative reconstruction algorithm configured with 2 iterations and 21 subsets. Concurrently, the CT component operated at 120 kVp tube voltage with CARE Dose4D automatic exposure technology to optimize radiation dose based on patient anatomy. All imaging data were acquired in continuous bed motion mode with an axial field of view of 22.1 cm, ensuring complete anatomical coverage from skull base to mid-thigh. Images were analyzed using syngo.via VB40 (Siemens) and Hermes Hybrid Viewer 3.2.

Lesion inclusion

The demographic distribution of evaluable lesions among the 162 study participants is detailed in Table 1. A subset of 20 patients underwent serial PSMA PET/CT imaging for pre- and post-treatment comparisons. Through rigorous selection criteria, 109 discrete lesions were identified for comprehensive analysis, comprising six prostate primary/recurrent lesions, 36 lymph node metastases, 39 osseous lesions, seven hepatic metastases, and 21 extra-hepatic soft-tissue lesions. To ensure methodological rigor, we implemented a double-blind selection protocol for non-PSMA-RADS 5T lesions. Four blinded readers independently conducted duplicate evaluations of these lesions using standardized classification criteria, maintaining a minimum washout period of one month between assessments to mitigate recall bias.

First, we established consensus criteria to differentiate physiological uptake in sympathetic ganglia from pathological lymph node involvement, a distinction that proved particularly demanding for inexperienced readers (IRs). Second, we developed a unified methodology for post-treatment evaluation that combined the PSMA-RADS-5T framework with PERCIST 1.0 response criteria (complete metabolic response, partial metabolic response, stable metabolic disease, progressive metabolic disease) (16,17). This integration produced four clinically distinct subcategories: complete remission (PSMA-RADS-5T-CR), partial remission (PSMA-RADS-5T-PR), stable disease (PSMA-RADS-5T-SD), and progressive disease (PSMA-RADS-5T-PD). These subcategories were designed to enhance diagnostic precision in monitoring therapeutic outcomes while maintaining compatibility with established response evaluation paradigms.

Reading

Four board-certified nuclear medicine physicians performed independent interpretations of all PSMA PET/CT scans. The group comprised two experienced readers (ERs: E1 and E2) who each possessed over 5 years of specialized experience in PSMA PET/CT interpretation, alongside two IRs (IRs: I1 and I2) who had completed six months of structured training in molecular imaging interpretation. Before commencing lesion evaluations, all participants engaged in comprehensive training sessions that systematically compared PSMA-RADS versions 1.0 and 2.0 (11,14), with detailed comparative analyses recorded in Table 2. While maintaining consistency with version 1.0 scoring criteria for established parameters, we implemented two key standardization measures to address interpretive challenges in the updated classification system.

Statistical analysis

In this study, reliability was defined as the consistency of lesion evaluations between different observers (inter-reader agreement) and within the same observer across repeated assessments (intra-reader agreement). Following PSMA-RADS 2.0 classification into nine categories (1-5T), two ERs and two IRs independently evaluated 109 lesions twice with one month between evaluations. We implemented statistical analysis using Python 3.8’s pingouin library to calculate intra-class correlation coefficients (ICCs) with 95% confidence intervals (CIs), applying distinct models for different reliability assessments: a two-way random effects model (ICC: 1–2) for inter-reader agreement between ERs and IRs, and a one-way random effects model (ICC: 1–1) for intra-reader consistency.

The analysis workflow involved: (I) organizing evaluation data into a structured DataFrame with columns for lesion ID, reader ID, PSMA-RADS score, and assessment timepoint; (II) computing ICC values through pingouin’s intraclass_corr function. Interpretation followed established benchmarks: <0.50 (poor), 0.50–0.74 (moderate), 0.75–0.90 (good), >0.90 (excellent agreement) (18). Visualization components included confusion matrices generated with scikit-learn and seaborn-derived heatmaps to display rating distributions across diagnostic categories.

Results

Inter- and intra-reader agreements of the PSMA-RADS 2.0 score

PSMA-RADS 2.0 demonstrated excellent reliability with inter-reader ICC values of 0.930 (95% CI: 0.907–0.948) and 0.929 (95% CI: 0.906–0.948) across two independent readings by four observers. ERs demonstrated marginally higher agreement compared to IRs.

Analysis of anatomical subgroups revealed varying ICC ranges. Initial readings showed inter-reader reliability spanning from 0.817 (95% CI: 0.725–0.889) for osseous lesions to 1.000 (95% CI: 1.000–1.000) for hepatic, soft tissue, and prostatic lesions. Lymph node assessments showed moderate agreement (ICC: 0.797; 95% CI: 0.691–0.880). Subsequent readings maintained this pattern with improved ICC values for osseous and lymphatic lesions (0.831–0.823) and sustained perfect agreement for hepatic/prostatic lesions (Table 3).

Intra-reader reliability remained consistently high across all observers. ERs achieved ICC values of 0.954–0.985 (95% CI: 0.934–0.990), while IRs demonstrated ICC ranges of 0.920–0.956 (95% CI: 0.885–0.970). Detailed target lesion consistency data are presented in Table 4.

Visual analysis of confusion matrix heatmaps (Figures 2,3) revealed strong diagnostic concordance, evidenced by prominent diagonal patterns indicating high classification agreement between readers.

Figure 2 Inter-reader agreement analysis using confusion matrices for PSMA-RADS classification consistency between ERs and IRs. (A) Agreement between ER1 and ER2 in the first reading session (ICC =0.964). (B) Agreement between IR1 and IR2 in the first reading session (ICC =0.929). (C) Agreement between ER1 and ER2 in the second reading session (ICC =0.969). (D) Agreement between IR1 and IR2 in the second reading session (ICC =0.932). ER, experienced reader; ICC, intraclass correlation coefficient; IR, inexperienced reader; PSMA, prostate-specific membrane antigen; RADS, Reporting and Data System; TL, target lesion.

Figure 3 Intra-reader reliability assessment through confusion matrices showing PSMA-RADS scoring consistency across repeated evaluations by ERs and IRs. (A) Intrareader agreement for ER1 (ICC =0.985). (B) Intrareader agreement for ER2 (ICC =0.954). (C) Intrareader agreement for IR1 (ICC =0.920). (D) Intrareader agreement for IR2 (ICC =0.956). ER, experienced reader; ICC, intraclass correlation coefficient; IR, inexperienced reader; PSMA, prostate-specific membrane antigen; RADS, Reporting and Data System.

Post-treatment lesion analysis

Treatment response analysis of 20 patients (Table 5) revealed disease progression in 50% (n=10), stable disease in 10% (n=2), and complete remission in 10% (n=2). The remaining 30% (n=6) showed partial treatment response.

Head-to-head comparison between PSMA-RADS 1.0 and 2.0

To directly quantify the impact of the revision, one ER re-evaluated 55 randomly selected lesions first with PSMA-RADS 1.0 and immediately afterwards with version 2.0 under identical viewing conditions. Twenty percent (11/55) of the lesions received a different category assignment (Table S1).

Discussion

PSMA PET/CT has emerged as an essential tool for initial staging and treatment response evaluation in PCa, demonstrating superior lesion detection capabilities compared to conventional imaging modalities (19). While PSMA-RADS 1.0 has provided a foundation for PCa lesion assessment, its lack of dedicated post-treatment lesion categorization limits clinical utility for treatment plan optimization. Our study implemented PSMA-RADS 2.0 to evaluate target lesions, aiming to assess both the system’s reliability and identify potential areas for refinement.

Our analysis revealed that PSMA-RADS 2.0 demonstrates high practical utility and reproducibility. As evidenced by our tabular and graphical results, all readers maintained strong consistency in lesion staging, with even IRs achieving accurate classification for most lesions. The system showed good-to-excellent inter-reader and intra-reader consistency (ICC >0.75), including for lymph node metastasis interpretation, suggesting its effectiveness in reducing interpretive subjectivity. However, several limitations warrant consideration. First, the PSMA-RADS-3B classification for bone lesions risks PCa understaging and potential missed surgical interventions. The clinical implications of heterogeneous bone uptake patterns, particularly in rib lesions, remain unclear and require long-term follow-up data. Second, lymph node staging presented challenges, particularly for IRs, due to confounding factors including size thresholds, uptake intensity variations, and physiological uptake mimicry. We observed an IR misclassifying physiological sympathetic ganglion uptake as PSMA-RADS-3A lymph node involvement (Figure 4). This diagnostic pitfall has been previously documented by Rischpler et al. in ganglion-related misinterpretations (20). Third, we identified classification challenges specific to PSMA-RADS-3C and 3D categories. PSMA-RADS-3D lesions (CT-visible structural abnormalities without significant PSMA uptake) may represent aggressive PCa variants such as neuroendocrine tumors (Figure 5), where misclassification could delay appropriate management. Conversely, the inclusion of PSMA-avid non-prostatic malignancies (e.g., lung or pancreatic cancers) in PSMA-RADS-3C risks clinical misinterpretation and delayed PCa-specific treatment. Finally, the PSMA-RADS-5T classification’s broad scope appears problematic. Among our 20 patients with serial post-treatment scans, only two achieved complete remission while 50% showed disease progression. Current PSMA-RADS 2.0 criteria classify only regressing or resolving lesions as 5T (15), potentially overlooking progressive disease that necessitates urgent therapeutic modification (21). This limitation was exemplified by a 53-year-old post-prostatectomy patient with undetectable prostate-specific antigen (PSA) but histologically confirmed neuroendocrine transformation (22), highlighting the system’s inability to capture aggressive phenotype evolution.

Figure 4 ⁶⁸Ga-PSMA-11 PET/CT scan in a 69-year-old prostate cancer patient demonstrating physiological uptake in sympathetic ganglia (red arrows) despite stable PSA levels. MIP, maximum intensity projection; PET/CT, positron emission tomography/computed tomography; PSA, prostate-specific antigen; PSMA, prostate-specific membrane antigen.

Figure 5 ⁶⁸Ga-PSMA-11 PET/CT scan of a 53-year-old post-prostatectomy patient with undetectable PSA (<0.006 ng/mL), showing local recurrence (red arrows) with PSMA-negative characteristics. MIP, maximum intensity projection; PET/CT, positron emission tomography/computed tomography; PSA, prostate-specific antigen; PSMA, prostate-specific membrane antigen.

We propose subdividing PSMA-RADS-5T into four distinct categories: 5T-CR (complete remission), 5T-PR (partial remission), 5T-SD (stable disease), and 5T-PD (progressive disease). Alternatively, developing a dedicated treatment response scoring system analogous to lymphoma’s Deauville criteria could enhance clinical utility (23).

By consolidating 1A and 1B into a single Category 1 and adding the 5T classification, the system decreases reader hesitation and provides clearer management guidance. These changes also facilitate differentiation of treatment response from de-novo disease. While PSMA-RADS 2.0 demonstrates marked reliability and addresses crucial gaps in post-treatment evaluation, our findings underscore the need for continued refinement. Unlike Grawe et al.’s PSMA-RADS 1.0 application in mixed diagnostic populations (15), our focus on targeted lesion classification in treatment-naïve and treated PCa patients revealed higher overall scores, though this distributional shift does not compromise our primary validity assessment.

In conclusion, our study identifies key improvement areas in bone lesion classification, lymph node staging, and treatment response characterization. Future directions should prioritize category refinement and multicenter validation to establish standardized guidelines. Despite these challenges, PSMA-RADS 2.0 represents a significant advancement in post-treatment evaluation reproducibility, providing clinicians with an intuitive framework for treatment monitoring and adaptation. The system’s structured approach ultimately enhances diagnostic consistency and patient care quality, while its evolving nature ensures adaptability to emerging clinical needs.

There are several limitations in this study. First, inter-reader variability related to experience levels may affect interpretation consistency, though this should diminish with increased clinical adoption. Second, the moderate sample size necessitates validation through expanded multicenter studies. Third, limited interdisciplinary communication during the study period may have influenced clinical correlation accuracy, highlighting the importance of strengthened radiologist-clinician collaboration in future research.

Conclusions

Our analysis intentionally avoided direct version comparison due to preliminary evidence suggesting comparable staging performance. The introduction of PSMA-RADS-5T in version 2.0 significantly enhances post-treatment lesion characterization, though its current implementation lacks progression-specific guidance. This gap may lead to suboptimal management of advanced PCa cases. We anticipate the future research to develop more refined classification systems that address these limitations while maintaining PSMA-RADS 2.0’s strengths in standardization and reproducibility.

Acknowledgments

Thanks for the support from the Department of Urology, the Affiliated Huaian No.1 People’s Hospital of Nanjing Medical University.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1928/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1928/dss

Funding: Thanks for the support from the Department of Urology, the Affiliated Huaian No.1 People’s Hospital of Nanjing Medical University.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1928/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This retrospective study was approved by the Ethics Committee of Huaian No. 1 People’s Hospital (approval No. YX-2021-113-01). All the enrolled patients have signed the informed consent prior to the examination.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Davis MI, Bennett MJ, Thomas LM, Bjorkman PJ. Crystal structure of prostate-specific membrane antigen, a tumor marker and peptidase. Proc Natl Acad Sci U S A 2005;102:5981-6. [Crossref] [PubMed]
2. Wang F, Li Z, Feng X, Yang D, Lin M. Advances in PSMA-targeted therapy for prostate cancer. Prostate Cancer Prostatic Dis 2022;25:11-26. [Crossref] [PubMed]
3. Karimzadeh A, Heck M, Tauber R, Solaris E, Nekolla S, Knorr K, Haller B, D’Alessandria C, Weber WA, Eiber M, Rauscher I. The Impact of PSMA PET-Based Eligibility Criteria Used in the Prospective Phase II TheraP Trial in Metastatic Castration-Resistant Prostate Cancer Patients Undergoing Prostate-Specific Membrane Antigen-Targeted Radioligand Therapy. J Nucl Med 2023;64:1252-8. [Crossref] [PubMed]
4. Hartrampf PE, Serfling SE, Michalski K, Buck AK, Werner RA. PSMA PET/CT for Response Assessment of (177)Lu-PSMA Therapy. Semin Nucl Med 2024;54:69-76. [Crossref] [PubMed]
5. Esfahani SA, Morris MJ, Sartor O, Frydenberg M, Fanti S, Calais J, Vapiwala N. Standardized template for clinical reporting of PSMA PET/CT scans. Eur J Nucl Med Mol Imaging 2024;52:335-41. [Crossref] [PubMed]
6. Mari A, Cadenar A, Giudici S, Cianchi G, Albisinni S, Autorino R, Di Maida F, Gandaglia G, Mir MC, Valerio M, Marra G, Zattoni F, Bianchi L, Lombardo R, Shariat SF, Roupret M, Bauckneht M, Vaggelli L, De Nunzio C, Minervini A. A systematic review and meta-analysis to evaluate the diagnostic accuracy of PSMA PET/CT in the initial staging of prostate cancer. Prostate Cancer Prostatic Dis 2025;28:56-69. [Crossref] [PubMed]
7. Kluge K, Einspieler H, Haberl D, Spielvogel C, Amereller D, Egger G, Kramer G, Grubmüller B, Shariat S, Hacker M, Kenner L, Haug A. Comparison of discovery rates and prognostic utility of [68Ga]Ga-PSMA-11 PET/CT and circulating tumor DNA in prostate cancer-a cross-sectional study. Eur J Nucl Med Mol Imaging 2024;51:2833-2842.
8. Bauckneht M, Checcucci E, Cisero E, Rizzo A, Racca M, De Cillis S, Amparore D, De Luca S, Fiori C, Rescigno P, Porpiglia F. The prognostic role of next-generation imaging-driven upstaging in newly diagnosed prostate cancer patients. Eur J Nucl Med Mol Imaging 2024;51:864-70. [Crossref] [PubMed]
9. Choi SH, Lee SS, Park SH, Kim KM, Yu E, Park Y, Shin YM, Lee MG. LI-RADS Classification and Prognosis of Primary Liver Cancers at Gadoxetic Acid-enhanced MRI. Radiology 2019;290:388-97. [Crossref] [PubMed]
10. Nishino M. CT Volumetry for Lung-RADS Classification of Solid Nodules. Radiology 2020;297:685-6. [Crossref] [PubMed]
11. Rowe SP, Pienta KJ, Pomper MG, Gorin MA. Proposal for a Structured Reporting System for Prostate-Specific Membrane Antigen-Targeted PET Imaging: PSMA-RADS Version 1.0. J Nucl Med 2018;59:479-85. [Crossref] [PubMed]
12. Spak DA, Plaxco JS, Santiago L, Dryden MJ, Dogan BE. BI-RADS(®) fifth edition: A summary of changes. Diagn Interv Imaging 2017;98:179-90.
13. Rowe SP, Pienta KJ, Pomper MG, Gorin MA. PSMA-RADS Version 1.0: A Step Towards Standardizing the Interpretation and Reporting of PSMA-targeted PET Imaging Studies. Eur Urol 2018;73:485-7. [Crossref] [PubMed]
14. Werner RA, Hartrampf PE, Fendler WP, Serfling SE, Derlin T, Higuchi T, Pienta KJ, Gafita A, Hope TA, Pomper MG, Eiber M, Gorin MA, Rowe SP. Prostate-specific Membrane Antigen Reporting and Data System Version 2.0. Eur Urol 2023;84:491-502. [Crossref] [PubMed]
15. Grawe F, Blom F, Winkelmann M, Burgard C, Schmid-Tannwald C, Unterrainer LM, Sheikh GT, Pfitzinger PL, Kazmierczak P, Cyran CC, Ricke J, Stief CG, Bartenstein P, Ruebenthaler J, Fabritius MP, Geyer T. Reliability and practicability of PSMA-RADS 1.0 for structured reporting of PSMA-PET/CT scans in prostate cancer patients. Eur Radiol 2024;34:1157-66. [Crossref] [PubMed]
16. Fanti S, Goffin K, Hadaschik BA, Herrmann K, Maurer T, MacLennan S, Oprea-Lager DE, Oyen WJ, Rouvière O, Mottet N, Bjartell A. Consensus statements on PSMA PET/CT response assessment criteria in prostate cancer. Eur J Nucl Med Mol Imaging 2021;48:469-76. [Crossref] [PubMed]
17. Wahl RL, Jacene H, Kasamon Y, Lodge MA. From RECIST to PERCIST: Evolving Considerations for PET response criteria in solid tumors. J Nucl Med 2009;50:122S-50S. [Crossref] [PubMed]
18. Koo TK, Li MY. A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research. J Chiropr Med 2016;15:155-63. [Crossref] [PubMed]
19. Sonni I, Felker ER, Lenis AT, Sisk AE, Bahri S, Allen-Auerbach M, Armstrong WR, Suvannarerg V, Tubtawee T, Grogan T, Elashoff D, Eiber M, Raman SS, Czernin J, Reiter RE, Calais J. Head-to-Head Comparison of (68)Ga-PSMA-11 PET/CT and mpMRI with a Histopathology Gold Standard in the Detection, Intraprostatic Localization, and Determination of Local Extension of Primary Prostate Cancer: Results from a Prospective Single-Center Imaging Trial. J Nucl Med 2022;63:847-54. [Crossref] [PubMed]
20. Rischpler C, Beck TI, Okamoto S, Schlitter AM, Knorr K, Schwaiger M, Gschwend J, Maurer T, Meyer PT, Eiber M. (68)Ga-PSMA-HBED-CC Uptake in Cervical, Celiac, and Sacral Ganglia as an Important Pitfall in Prostate Cancer PET Imaging. J Nucl Med 2018;59:1406-11. [Crossref] [PubMed]
21. Ge R, Wang Z, Montironi R, Jiang Z, Cheng M, Santoni M, Huang K, Massari F, Lu X, Cimadamore A, Lopez-Beltran A, Cheng L. Epigenetic modulations and lineage plasticity in advanced prostate cancer. Ann Oncol 2020;31:470-9. [Crossref] [PubMed]
22. Wang Y, Wang Y, Ci X, Choi SYC, Crea F, Lin D, Wang Y. Molecular events in neuroendocrine prostate cancer development. Nat Rev Urol 2021;18:581-96. [Crossref] [PubMed]
23. Lopci E, Meignan M. Deauville score: the Phoenix rising from ashes. Eur J Nucl Med Mol Imaging 2019;46:1043-5. [Crossref] [PubMed]