162P - MR imaging biomarkers profiles in patients with prostate cancer treated with androgen deprivation therapy

Background

Radiomics has emerged as a potential tool to predict clinical outcomes in prostate cancer (Pca ). MRI has a fundamental role in staging and radiotherapy planning, being affected in the case of receiving prior androgen deprivation treatment (ADT). The aim of this study was to determine imaging biomarker profiles that allow to identify patients with a worse prognosis and a higher risk of biochemical relapse (BQR).

Methods

A single-center, retrospective, observational study was designed. Baseline MRIs (T2w, DWI and/or DCE sequences) from PCa patients treated with ADT and radiotherapy were included. Using Quibim’s QP-Prostate ® solution, subregions (Transition + Central Zone [TZ + CZ], Peripheral zone [PZ] and seminal vesicles [SV]) were automatically segmented. For each subregion and for the whole prostate (WP), computed using the weighted biomarkers averaging, 105 radiomic features (shape and textures) from T2w, apparent diffusion coefficient (ADC) variables from DWI and perfusion parameters from DCE were extracted.

Results

A total of 128 T2w, 107 DWI and 62 DCE MRIs from 128 PCa patients (32 [25%] low or favourable intermediate risk [group 1] and 96 (75%) high or unfavourable intermediate risk [group 2]) were included, of whom 20 (16%) experienced BQR (18 [19%] from group 2). According to risk stratification, patients from group 1 presented more homogeneous textures in PZ and WP than group 2. Regarding BQR, the ADC mean was higher in the TZ in those who did not relapse, while patients with greater major axis length in the PZ, WP and SV were more likely to relapse. In SV and WP heterogenicity-related feature values were significantly higher in BQR. In group 2, diffusion parameters and shape variables in the SV were significantly different in those experiencing BQR.

Conclusions

Imaging biomarker profiles in this PCa population may help to better stratify patients at risk based on baseline MRIs. High-risk patients may be identified by the heterogeneity of textures. Diffusion and texture parameters may determine a higher probability of relapse.

Clinical trial identification

Editorial acknowledgement

Legal entity responsible for the study

The authors.

Funding

Has not received any funding.

Disclosure

All authors have declared no conflicts of interest.