Predictive factors of 18F-choline PET/CT positivity in patients with prostate cancer recurrence after radiation therapy: is the impact of PSA nadir underestimated?

BACKGROUND The objective of this study is to explore the impact of PSA nadirs on detection rates of prostate cancer (PCa) recurrence with 18F-choline (CH) PET/CT after external beam radiation therapy (EBRT). METHODS In this retrospective study, data were collected from 54 patients with suspicion of PCa biochemical recurrence after EBRT (28 patients treated initially with EBRT and 26 as salvage therapy in the absence of PSA decrease after initial treatment), who underwent 18F-CH PET/CT between 2010 and 2015. PSA nadir and trigger PSA were collected from patient files. Relative PSA was calculated by subtracting the nadir from the trigger PSA. RESULTS Median PSA nadir was 0.31 (0.01-13.31) ng/mL, trigger PSA was 7.85 (0.47-111.60) ng/mL, and relative PSA was 6.05 (0.24-104.59) ng/mL. Overall, 40 (74%) PET/CT scans were positive: recurrence was local and/or regional in 29 patients, distant in 15 and combined both in four, with no association between PSA values and sites of recurrence. In univariate analysis, trigger (p = 0.015) and relative (p = 0.0005) PSA values and PSA velocity (p = 0.01) were significantly linked to positive PET/CT, but PSA nadir was not. In subgroup analysis, these significant differences were only found in the salvage EBRT group. Akaike Information Criterion multivariate model comparison found that relative PSA was a better predictor of positive PET/CT than trigger PSA (PSAt). 18F-CH PET/CT detection rates increased with trigger and relative PSA: 0% (0/4 patients), 71% (5/7 patients), and 81% (35/43 patients) for PSAt <2 ng/mL, 2≤ PSAt ≤4 ng/mL, and PSAt >4 ng/mL, respectively, and 14% (1/7 patients), 50% (5/10 patients), and 92% (34/37 patients) when relative PSA was taken into account instead of trigger PSA, with seven (13%) patients changing subgroups. CONCLUSIONS We found a high overall detection rate and an increase in detection rates proportional to trigger and relative PSAs. Although relative PSA, taking into account PSA nadir, was a better predictive factor of PET/CT positivity in univariate analysis, this was most noticeable for high PSAs. For low PSAs, trigger PSA remains most relevant. Larger series with intermediate PSA values need to be studied to fully apprehend nadir impact.