To assess the predictive role of multiparametric MR imaging (mpMRI) in the setting of salvage RT (sRT) for a presumed local failure after radical prostatectomy (RP).
This is a cross-sectional study including all patients treated with sRT at a single Institution from Feb 2014 to Sept 2016. All patients had failed biochemically after RP and underwent pre-sRT mpMRI (including a dynamic contrast-enhanced-DCE MRI) on a 3-T system with the evidence of a presumed local recurrence. All patients were treated with a homogeneous approach delivering 73.5 Gy and 69 Gy to the nodule and to the prostatic bed, respectively, in 30 fractions (IGRT-VMAT, SIB technique). Patients underwent a re-evaluation mpMRI at the time of the study. Patients without a complete response (CR), underwent repeated mpRM at 3-6 month-intervals until CR and/or a maximum of 3 repeated mpMRI. CR was defined as complete disappearance of the nodular area of early contrast enhancement in DCE-MRI or persistence of an area of contrast enhancement with a slower wash-in rate and/or a hypointense (more than muscles) appearance on T2-weighted sequence. On baseline DCE-MRI, the wash-in and wash-out were derived from the whole-volume time-intensity curve using home-made software as per Casciani et al, AJR, 2008. The two-sided Mann–Whitney rank test was used to compare the variables. Receiver operating characteristic (ROC) curves were performed for the most significant DCE-MRI parameters.
Thirty-seven patients were identified. The mean (SD) time interval between RP and sRT was 71.1 (58.1) mths. At sRT, mean (SD) serum PSA was 1.47 (1.58) ng/ml and the mean (SD) volume of the recurrence delineated on mpMR plus a 8 mm margin (rPTV) was 12.5 cc (6.3 cc). Eleven (29.7%) patients received also androgen deprivation (AD). All patients ended up showing a complete response, 28 (75.7%), 7 (18.9%) and 2 (5.4%) at the 1st, 2nd or 3rd re-eval mpMRI. The mean (SD) time from the end of sRT to CR was 16.5 (10.0) and 24.1 (5.8) mths for patients who responded at the 1st mpMRI (N=28) vs the slower ones (N=9), p= 0.0436. Compared to slower responders, patients achieving a CR at the 1st re-eval mpRM had significantly smaller rPTVs (9.4cc vs 13.1cc, p=0.0214) and a trend towards lower pre-sRT PSA values (1.21 vs 2.28, p=0.06). However, the stronger correlation was found for the wash-out parameter with a lower value highly predicting a delayed response (0.06 s-1 vs -0.99 s-1,continuum, p=0.0087). The sensitivity and specificity of the wash out parameter were 78.6% and 75%, respectively (area under ROC curve 0.808 p=0.0001, cut-off value ≥0.46 s-1). No correlation was found between time to response and AD (p= 0.3030).
sRT is highly effective in achieving a complete response at mpMRI. There is a strong correlation between the time to response and baseline wash-out features.
Author Disclosure: V. Landoni: None. A. Faiella: None. S. Marzi: None. A. Farneti: None. M.G. Petrongari: None. B. Saracino: None. L. Bertini: None. F. Spasiano: None. G. Sanguineti: None.
© 2018 Published by Elsevier Inc.