ReIGNITE RT Boost: An International Study Testing the Feasibility, Accuracy and Reliability of Using Restriction Spectrum Imaging (RSI) MRI to Guide Radiotherapy Target Volume Delineation for Prostate Cancer Tumor Boost

Purpose/Objective(s)

Recently the FLAME clinical trial showed adding a radiation boost to the visible tumor improves prostate cancer control without additional toxicity. In FLAME, a radiologist assisted in each tumor delineation. Even for radiologists, there is substantial dependence on reader experience for prostate MRI. Radiation oncologists' ability to reliably identify prostate tumors is critical to wide adoption and patients benefitting from tumor radiotherapy boost. We have developed a quantitative diffusion MRI metric for prostate cancer based on the restriction spectrum imaging (RSI) framework, called the RSI restriction score (RSI_rs). RSI_rs has been shown to improve accuracy over guideline-recommended standard ADC maps. RSI_rs maps can be overlaid on conventional T2-weighted MRI sequences and used to identify prostate cancer. We hypothesize that contouring on RSI_rs maps will improve the accuracy and reliability of prostate tumor delineation.

Materials/Methods

Radiation oncologists (participants) from multiple institutions contoured prostate tumors on 20 patient cases. An expert radiologist contoured each tumor to establish ground truth. Participants were either board certified or a resident who had completed a clinical genitourinary cancer rotation. Participants were provided the patient's age, PSA level, biopsy result and radiologist's report (e.g., "PI-RADS 5 lesion in right lateral peripheral zone at the base"). Participants delineated volumes on either standard MRI alone (50% of cases; included T2-weighted, DWI, ADC) or standard MRI with RSI_rs map. Accuracy was quantified by comparing to the expert volume by distance to agreement, Dice coefficient, and Jaccard coefficient. Reliability was measured as variation across participants by the Volume Ratio (VR) method.

Results

In this interim analysis of an ongoing study, 126 target volumes have been analyzed from 6 participants. Comparing participant and expert volumes, use of RSI_rs maps significantly improved the mean distance to agreement (3.87mm vs. 1.68mm, p<0.01), Dice coefficient (0.45 vs. 0.65, p<0.001), and Jaccard coefficient (0.32 vs. 0.50, p<0.001). Target volume variability was also significantly improved; VR was reduced from 19.23 to 5.90 (p<0.05). Notably, there were 9 complete misses when participants contoured on standard MRI alone but no misses when RSI_rs maps were provided. 63 global participants have been enrolled in this study and additional data collection and analysis are in process.

Conclusion

The addition of RSI_rs maps to standard MRI improves the reliability and accuracy of target delineation for prostate tumor boost. RSI_rs maps have the potential to increase the feasibility of widely implementing the benefits of focal tumor boost for patients undergoing prostate radiotherapy.