Toward harmonized PROMISE PSMA scoring: a simple approach for correcting 18F-PSMA-1007 liver uptake based on 68Ga-PSMA-11 and 18F-DCFPyL reference values
In brief
A 0.56 liver-uptake correction reclassifies 38% of PSMA-1007 score-1 lesions
In 331 scans using three prostate-specific membrane antigen (PSMA) tracers, 18F-PSMA-1007 showed higher liver uptake than the other tracers, while spleen uptake was an inconsistent substitute. Applying a liver correction factor of 0.56 reclassified 38.2% of score-1 lesions as score 2 and brought score distributions closer to those of the other tracers; whether this improves clinical interpretation remains untested.
- Journal
- EJNMMI research (Q1)
- Published
- 15 August 2026
- Study design
- Prospective / inception cohort
- Evidence level
- Level 2, Moderate (CEBM 2b)
- Authors
- Reut Anconina, Miriam Priss, Esther Landau, Moshe Sason, Arnon Schwartz, Eli Atar, et al.
- PMID
- 42902110
- DOI
- 10.1186/s13550-026-01499-5
Why clinicians should know about it
- Picked for Radiology, Radiation Oncology, Nuclear Medicine, Medical Physics and Imaging (paper of the day, 23 October 2026): Harmonized PROMISE PSMA scoring corrects liver uptake
Abstract
BACKGROUND: The PROMISE framework standardizes PSMA PET interpretation using reference organs, particularly the liver. For hepatobiliary-excreted tracers (e.g., 18F-PSMA-1007), the spleen was proposed as an alternative reference, though without quantitative validation. We evaluated PSMA-ligand uptake patterns among 18F-PSMA-1007, 68Ga-PSMA-11, and 18F-DCFPyL to assess the validity of spleen-based referencing and develop a simple approach that may harmonize 18F-PSMA-1007 liver uptake with that of renally-excreted tracers. RESULTS: 331 consecutive PSMA PET-CT studies (n = 118/95/118 for 68Ga-PSMA-11/18F-PSMA-1007/18F-DCFPyL) were analyzed: SUVs were retrospectively measured in all studies in reference tissues (blood pool, liver, parotid), with SUVspleen also recorded for 18F-PSMA-1007. SUVliver differed markedly across tracers: 18F-PSMA-1007 (13.95 ± 4.71) > 68Ga-PSMA-11 (7.45 ± 4.59) ≈ 18F-DCFPyL (8.11 ± 1.78), p < 0.01. The elevated 18F-PSMA-1007 liver uptake expanded the PROMISE score-1 range (SUV range between SUVblood pool and SUVliver) and compressed that of score-2 (SUV range between SUVliver and SUVparotid). SUVspleen in this group (12.24 ± 5.67) showed only moderate correlation with SUVliver (r = 0.43) and exceeded SUVliver in 29% of cases, questioning the utility of spleen uptake as a consistent alternative to liver uptake-based cutoff. Linear correction using the inverse ratio of mean SUVliver (18F-PSMA-1007 versus other tracers; factor = 0.56, 95% CI 0.51-0.61) yielded SUVliver-corrected of 7.8 ± 2.63 in the 18F-PSMA-1007 group. Applying SUVspleen as a PSMA score 1-2 cutoff failed to resolve score distribution imbalance, upstaging only 11.8% of score-1 lesions to score-2. By contrast, the use of SUVliver-corrected (SUVliver × 0.56) as this cutoff harmonized the score distribution in the 18F-PSMA-1007 cohort relative to 68Ga-PSMA-11 and 18F-DCFPyL, leading to reclassification of 38.2% of score-1 lesions as score-2. CONCLUSION: In PET with 18F-PSMA-1007, PROMISE scoring based on unadjusted liver or spleen uptake systematically underestimates lesion scores. The simple linear correction approach proposed here offers a research avenue for enhancing the consistency of PSMA scoring for improved interpretive standardization. This study contains data obtained from a prospective Registry Trial: ClinicalTrials.gov, NCT03718260, registered October 24, 2018.
Abstract as published, via PubMed.
For healthcare professionals. The summary is generated by AI from the published abstract, and the evidence level is assigned automatically from the study design on the Oxford CEBM hierarchy. Neither is medical advice. Read the full paper before changing practice.