The Effects of Androgen-Deprivation Therapy on MRI Delta Radiomics Features in a Prospective Radiotherapy Treatment Trial for Prostate Cancer

Abstract

The aim of this study is to evaluate the effects of androgen-deprivation therapy (ADT) on the variation of multiparametric (mp)MRI-radiomics features, defined as "delta radiomics", in patients from a prospective radiotherapy trial in which 48% of men received short term androgen deprivation therapy (STADT). Twenty-five patients with favorable to high-risk prostate cancer were enrolled on a Phase I trial of MRI-guided radiotherapy treatment trial. Median follow up is 63 mo, range 17 to 90. mpMRI was carried out pre-RT, 3 and 9 months after completion of RT and at 2.0-2.5 years after completion of all treatment. A high dose (12-14 Gy) was given on day 1 to a portion near the center of the mpMRI delineated GTV(s). The mpMRI exam consisted of T2 weighted, T1 non-contrast, T1 dynamic contrast-enhanced (DCE)-MRI, and diffusion weighted imaging (DWI) with the generation of an Apparent Diffusion Coefficient (ADC) map. Prostate, peripheral zone (PZ) and transition zone (TZ) were contoured in MIM. GTV(s) were translated from the planning CT to all 4 mpMRI. Thirty variables (mean±SD) were extracted from GTV, normal appearing (NA)-PZ and NA-TZ on T2w and ADC, as well as the perfusion parameters Ktrans, Kep and ve on DCE-MRI. Baseline variables were compared between patients with and without ADT using Student t-test. Area-under-the-curve (s) (AUCs) for the variables over the baseline, 3 and 9 mo were compared between the +/-ADT groups using Mann-Whitney test. Two patients received only diagnostic mpMRI and were excluded from the study. From the reminding 23 patients, 14 had 3 longitudinal mpMRI studies and 9 had 4. The data was acquired on 3T Discovery MR750 (n = 45) (GE, Waukesha, WI), 3T MR Magnetom Trio (n = 7), Skyra (23) and 1.5T Symphony (n = 3) (Siemens, Erlagen, Germany). Twelve patients received STADT. There were no significant differences in the means of baseline radiomics variables, including the volumes of prostate, PZ and TZ between the (+/-ADT) groups. Similarly, there was no statistical difference between the AUC of the GTV on ADC and perfusion parameters. ADC values increased significantly on average 30%, 44% and 47% after treatment. ve also increased significantly on average 108%, 91% and 124% from baseline. Because of the different scanners, T2-weighted intensities were not compared. The study indicates that ADT does not affect the overall delta radiomics variables before and after radiotherapy. Prostate cancer has long natural history with relatively low number of failures. We demonstrate and +/-ADT patients can be analyzed together, allowing for increased number of events. The ultimate goal of the project is to identify delta radiomics signature that relates very early response to patient outcome such that therapeutic adjustments may be made to improve long term outcome. While 17% increase in ADC before and after ADT have been previously reported, the large magnitude of changes (∼40 %) here indicate that the radiation effects drive the observed variations.