Defining MR-Based Parameters of Treatment Response to Immune-Modulatory Therapy for Grade Group 5 Prostate Cancer

Abstract

<h3>Purpose/Objective(s)</h3> A phase I/II study combining nivolumab with androgen deprivation therapy (ADT) and high dose brachytherapy (HDR) followed by external beam radiotherapy (EBRT) indicated that the regimen is well tolerated and effective with radiographic and pathologic evidence of increased antitumor activity. In the present study we sought to utilize MR-based imaging characteristics to define parameters of antitumor response to HDR+EBRT and nivolumab in patients with Grade Group (GG) 5 prostate cancer. <h3>Materials/Methods</h3> Eligible patients received ADT in combination with nivolumab and HDR followed by EBRT delivered using MR-guided radiotherapy. Data from 12 patients were analyzed using MRI from diagnosis, once-weekly treatment from fractions 5, 10, 15 and 25 and 3-month follow up. Images were fused and analyzed using an image analysis package. At time of enrollment the dominant tumor (ROI) was contoured and tracked at different timepoints. The change in T2 value was assessed using voxel data at diagnosis and follow up. The delta (δ) of average change was estimated between partial and complete responders from diagnosis and follow up. The same analysis was done on the treatment MR using steady-state coherent sequences. All data were normalized against the prostate values. <h3>Results</h3> Median follow up was 17 months. Complete radiographic response (CR) was seen in 6 out of 12 patients. A δ T2 was estimated between baseline and follow up T2 tumor ROI. Partial responders had higher δ of 62.5 as compared to complete responders with δ of 28. When compared longitudinally over 5 fractions, on average change in steady-state coherent sequence over time was higher in partial responders than complete responders. <h3>Conclusion</h3> Our preliminary results show a greater δ in T2 signal in partial responders consistent with continued tumor response on weekly imaging, while complete responders showed minimal change over time. This finding suggests that CR may derive a rapid and early response to nivolumab, ADT and HDR than the partial responders. Furthermore, tracking δ in T2 signal and steady-state coherent sequences may be a reliable tool for defining tumor response in real-time and needs further validation in a larger cohort. This study was registered under NCT03543189.