Abstract

43 Background: Biochemical response to neoadjuvant androgen deprivation therapy (nADT) predicts relapse-free and overall survival for prostate cancer (PCa). Radiomic features change in response to nADT, showing potential as prognostic imaging biomarkers. Using multi-parametric MRI (mpMRI) before and after nADT, we aim to validate these changes, evaluate reproducibility and assess their association with previously described ADT-induced reductions in quantitative blood flow (BF) and blood volume (BV) imaging parameters. Methods: 20 patients with high-risk PCa received 4 mpMRI scans, 2 pre-nADT and 2 after 3 months therapy. T2-weighted sequences were used to define tumour regions-of-interest (ROI) which were subtracted from the prostate outline to form the benign prostate ROI. Established kinetic models were applied to dynamic contrast susceptibility MR sequences to derive changes in BF and BV in both ROI. Pyradiomics was used to extract homogeneity and energy features from both ROI. A paired student T-test compared feature values pre- and post-nADT. An inter-class correlation coefficient (ICC) calculated using mean values from consecutive scans assessed reproducibility. Feature changes and physiological parameters were correlated using Pearson’s coefficient. Results: In response to nADT, homogeneity and energy values significantly increased in benign prostate (p < 0.001) and decreased in tumour (p < 0.005). Both features exhibited high reproducibility pre- and post-nADT; all ICC ≥0.89. The reduction in tumour homogeneity and energy showed a positive association with the decline in BF and BV induced by ADT; all r 0.69-0.78, all p < 0.001. No correlation was seen in benign prostate. Conclusions: Energy and homogeneity radiomic features in benign and malignant prostate are reproducible and show significant reciprocal change in response to nADT. Validation of these changes and the strong association in tumour with ADT-induced physiological effects confirms their potential as prognostic biomarkers in high-risk PCa.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call