The work presented by Agritelley, et al., in which I served as senior author, attempts to address the difficulties of differentiating benign bone uptake of PSMA from real sites of osseous metastatic disease related to prostate adenocarcinoma by providing the largest cohort to our knowledge of histopathologically verified PSMA positive lesions on 18F-DCFPyL PET/CT, and utilizing this tissue confirmation to retrospectively assess metrics of uptake to provide a pragmatic approach to reliably distinguish incidental benign bone lesions from osseous metastatic disease.
PSMA Positive Bone Lesions: A Significant and Persistent Clinical Concern
With the FDA approval of 18F-DCFPyL in 2021, the arrival of PSMA-targeted PET agents provided an impressive sensitivity and specificity for prostate cancer detection and has been consistently considered a game changer in Nuclear Medicine. Despite these successes, the sensitivity of PSMA PET agents has saddled the interpretation of these exams with the problematic IBL. Since binding of PSMA PET agents is not exclusive to prostate adenocarcinoma, PET uptake is often present in benign bone processes, particularly fibrous dysplasia and, to a lesser degree, hemangiomas and Paget disease of bone. Notably, the essential study by Mainta et al.1 found that up to half of bone lesions detected on 68Ga-PSMA-11 PET/CT are indeterminate, highlighting the concerns and frustrations that IBLs present for clinical interpretation and management directions, thus undermining the diagnostic precision of PSMA PET.
The consequences of mischaracterization of an IBL as osseous metastatic disease can be severe. A false-positive diagnosis of osseous metastases could lead to incorrect staging and possibly expose the patient to unnecessary systemic or localized radiation therapy. Further, efficacious surgical approaches may be abandoned. Previously published PSMA PET reporting systems (e.g., PSMA-RADS,2 PROMISE V2,3 amongst others) provide a structured interpretation approach for PSMA-positive lesions but, by design, they have left the issue of equivocal lesions unresolved. To alleviate this issue in the hope of providing a way forward, we have hopefully provided a step forward by providing a pathologically validated approach for navigating through the assessment of confounding PSMA-positive bone lesions.
A Firm Step Forward Built on a Pathologic Foundation
We feel the importance of this study is founded on our methodological rigor of only using pathologically confirmed PSMA-positive lesions for our analysis of PET metrics. Prior investigations prioritized imaging follow-up and imaging characteristics with only sparse pathologic sampling. An essential 2023 paper by Phelps et al.,4 which is currently the only comparable analysis of IBLs on 18F-DCFPyL, provided tissue confirmation in only 8 of 98 lesions. By contrast, our study evaluated 42 lesions confirmed by biopsy to be either a benign bone process or osseous metastatic prostate cancer, with diagnosis established with a median of just 24 days after PET/CT, minimizing the risk that new disease could confound interpretation.
Going Beyond SUVmax
Our study's central contribution was to provide quantitative SUV-based thresholds with meaningful predictive accuracy given the pathologic confirmation of the lesion. Further, we evaluated not only IBL SUVmax, the most evaluated metric in the literature. but SUVmean, SUVpeak, and ratios of lesion uptake-to-background organ uptake, which all provided a statistically significant predictive value.
When threshold values were applied, the results were clinically relevant. Notably, an SUVmax of 8 or greater predicted malignancy with 88% accuracy. Possibly even more relevant, our findings of an IBL/liver SUVmax ratio of 1 or greater, meaning lesion uptake exceeded background liver uptake, performed comparably to SUVmax thresholds, suggesting a possible generalizable lesion-to-liver background ratio. Notably, the IBL/Liver ratio provides an elegant and familiar solution, which has been successful in PET staging systems (e.g., Deauville score), since it is based on a simple assessment of the liver uptake relative to the lesion, which requires no intensive image post-processing and can be routinely and quickly performed.
Further, our results align directly with PROMISE V2 scoring, in which IBL uptake exceeds liver uptake, equating to either a score of 2 or 3, which both carry a high suspicion for prostate cancer, lending external consistency to the proposed thresholds and is promising for general use across scanners and PET agents.
Regarding SUVmax, when compared to Phelps et al.,4 the SUVmax threshold of 8 proposed here is modestly higher than their threshold of 5 for 18F-DCFPyL but more closely aligns with the thresholds proposed for other PSMA tracers (~7.2 for 18F-PSMA-1007 and greater than 6.2 for 68Ga-PSMA-11). This convergence of SUV metrics across PET agents is reassuring; however, caution in interpretation is necessary given that more studies are necessary for direct PSMA PET agent comparisons given known differences in pharmacokinetics, background organ uptake, and scanner-dependent parameters.
Clinical Context Will Always Be Essential
SUV metrics shouldn’t be used in a vacuum, and both clinicians and image interpreters should take clinical cues into account. Notably, our study identified two independent predictors of prostate metastases to the bone. First, multiple bone lesions were strongly associated with malignancy in both univariable analysis and multivariable modeling when evaluated in conjunction with SUVmax. A solution that feels intuitive given that widespread osseous involvement is a hallmark of metastatic disease, while benign processes are limited to a remote and focal lesion with an expected anatomic pattern, such as the lateral rib. Second, biochemical recurrence was also predictive, underscoring the need that PSMA PET interpretation of an IBL carries a substantially different probability of malignancy depending on whether it arises in the context of a rising serum PSA post-treatment versus initial staging in a patient without known recurrence. Incorporating clinical disease state into the interpretive framework is not merely sensible; it suggests it is supported by the statistical evidence.
Limitations and the Road Ahead
While we hope this is a promising step forward to solving the IBL conundrum, I must emphasize our study's limitations. While this is the largest amount of histologically confirmed PSMA-positive lesions in the IBL literature, the sample size of 42 is modest. Further subgroup analyses, especially the anatomic lesion location, are underpowered. Also, selection bias is inherent to biopsy, given the clinical concern and location of the lesion likely shifts true positive lesion data to higher SUV lesions, while avoiding lesions difficult to biopsy (i.e., the ribs).
Further, we evaluated a single PSMA tracer (18F-DCFPyL) given that it was the first widely available PSMA PET agent, likely contributing to more biopsies given the clinical context and uncertainty of results of a novel PET exam. Also, while our health network encourages standardization of scanner technology and protocols, it is difficult to control for all parameters unless scanning on a single machine. Thus, the generalizability of our results should be approached cautiously given the advances in scanners and the increasing array of commercial PSMA PET agents. Steps toward prospective, multicenter validation, including other PSMA PET agents, would be necessary to confirm our findings and provide evidence to support extending our results.
Conclusion
Our study represents a single, and hopefully meaningful, step to resolving the frustrating challenges related to IBLs in PSMA PET interpretation. By anchoring its findings to histopathologic standards, we have been able to devise a practical system to differentiate prostate cancer bone metastases from benign bone processing, which can be easily incorporated into radiologists’ previously existing reporting frameworks. The frustration related to the IBL has been a relatively new challenge; however, continued study of the PET metrics in conjunction with clinical context will hopefully push the current frustrations of the IBL into the column of resolved issues.
The author would like to disclose that he is a consultant for Siemens Healthineers.
Written by: Ryan Avery, MD, Department of Radiology, Division of Nuclear Medicine, Northwestern University Feinberg School of Medicine, Chicago, Illinois
References:
- Mainta IC, Neroladaki A, Wolf NB, Benamran D, Boudabbous S, Zilli T, Garibotto V. [68Ga]Ga-PSMA-11 PET and Prostate Cancer Bone Metastases: Diagnostic Performance of Available Standardized Criteria. J Nucl Med. 2024 Sep 3;65(9):1376-1382. doi: 10.2967/jnumed.124.267899. PMID: 39117453.
- Cheng L, Jin C, Zhang H, Sun Z, Fan J, Yang T, Mao F, Fu G, Wang Q, Niu X, Tao W. PSMA-RADS 2.0: clinical validation and technical considerations for prostate-specific membrane antigen positron emission tomography/computed tomography image interpretation. Quant Imaging Med Surg. 2026 Feb 1;16(2):172. doi: 10.21037/qims-2025-1928. Epub 2026 Jan 23. PMID: 41669484; PMCID: PMC12883463.
- Seifert R, Emmett L, Rowe SP, Herrmann K, Hadaschik B, Calais J, Giesel FL, Reiter R, Maurer T, Heck M, Gafita A, Morris MJ, Fanti S, Weber WA, Hope TA, Hofman MS, Fendler WP, Eiber M. Second Version of the Prostate Cancer Molecular Imaging Standardized Evaluation Framework Including Response Evaluation for Clinical Trials (PROMISE V2). Eur Urol. 2023 May;83(5):405-412. doi: 10.1016/j.eururo.2023.02.002. Epub 2023 Mar 17. PMID: 36935345.
- Phelps TE, Harmon SA, Mena E, Lindenberg L, Shih JH, Citrin DE, Pinto PA, Wood BJ, Dahut WL, Gulley JL, Madan RA, Choyke PL, Turkbey B. Predicting Outcomes of Indeterminate Bone Lesions on 18F-DCFPyL PSMA PET/CT Scans in the Setting of High-Risk Primary or Recurrent Prostate Cancer. J Nucl Med. 2023 Mar;64(3):395-401. doi: 10.2967/jnumed.122.264334. Epub 2022 Oct 20. PMID: 36265908; PMCID: PMC11927076.