Correlating the extent of [⁶⁸Ga]-PSMA PET positive disease with pre-scan PSA in prostate cancer

Objectives To determine whether pre-scan PSA level predicts disease extent as diagnosed on whole body [⁶⁸Ga]-PSMA (HBED) PET, in patients for restaging of biochemically relapsed prostate cancer after local therapy.

Methods Patients with history of prostate cancer and PSA > 0.20ng/mL were scanned. Following injection of 100-150MBq of [⁶⁸Ga]-PSMA, PET/CT scans were acquired at 60 minutes +/- 2 hours (delayed) on a ToF PET/CT. Images were dual reported by two nuclear medicine specialists with access to information including Gleason score, treatment, PSA, CT and bone scan. Scan findings were extracted from the finalised reports and classified in accordance to AJCC. Each patient was assigned to one of six categories. Disease limited to prostate, prostate bed or seminal vesicle was local. Nodal disease within the true pelvis was locoregional. Distant disease was categorised as distant node, skeletal only, or skeletal plus other (any non-skeletal site).

Results 47 patients were available for analysis (age 48-94 years, median 73. PSA 0.21-57, median 4.9). 26 (55%) had prior radical prostatectomy. Gleason score was available for 43 (91%), of which 1 was GS5, 2 were GS6, 24 were GS7, 3 were GS8 and 13 were GS9. Delayed imaging was performed in 36 (77%). There were 42 (89%) positive (PSA 0.48-56) and 5 (11%) negative studies (PSA 0.21-1.9). 10 had local, 7 had locoregional and 25 had distant disease. All patients with PSA ≥ 2 had positive studies (33), compared to 9/14 (64%) patients with PSA < 2. Correlation between PSA and disease extent was not significant (one-way ANOVA p=0.15).

Conclusions For restaging of biochemically relapsed prostate cancer, pre-scan PSA correlates poorly with extent of PSMA positive disease. Low PSA does not predict local disease. Referral bias may explain the low mean PSA in patients with bone limited disease. All patients with PSA ≥ 2ng/mL had positive studies.