Click here to

Session: Novel Strategies Using Existing Imaging Technology for Planning, Delivery and Toxicity Analyses [Return to Session]

A Non-Parametric Analysis to Identify High-Risk Vs. Low-Risk Anatomic Regions Associated with Reduced Overall Survival in the RTOG 0617 Clinical Trial Data

J Zhu1*, M Thor2*, A Apte2, J Oh2, A Rimner2, J Deasy2, A Tannenbaum1, (1) Stony Brook University, Stony Brook, NY, (2) Memorial Sloan Kettering Cancer Center, New York, NY


TH-F-TRACK 4-1 (Thursday, 7/29/2021) 4:30 PM - 5:30 PM [Eastern Time (GMT-4)]

Purpose: Radiation dose to the cardio-pulmonary system has been identified as a new key detriment for risk of death in locally-advanced non-small cell lung cancer (LA-NSCLC). However, primarily one-dimensional dose metrics of pre-defined structures have typically been used for deducing this association. Here, we introduce an approach based on a metric called the unbalanced optimal mass transport (OMT) distance, which uses neither pre-selection of sub-regions nor normalization (total mass need not be preserved). The method ranked risk of death in the RTOG 0617 based on the distance.

Methods: The 409 patients retrieved from The Cancer Imaging Archive with evaluable data were randomly split into training and validation subsets. Multi-step B spline image registration was used to deform CT scans and the associated dose distributions to a chosen patient. Pairwise unbalanced OMT distances of all patients were efficiently computed through a newly proposed approximation method via vector-valued OMT, which measured the dose spread pattern and total dose amount differences. The distances were translated into risk scores by C-indices. Risk status in validation data was measured as distances to the training reference patients. Lastly, high-risk and low-risk regions were deduced according to the 50 riskiest (or least risky, respectively) reference patients. The locations of the sub-regions were qualitatively compared to those of eight deep learning auto-segmented cardio-pulmonary substructures.

Results: The identified riskiest region overlapped with the base of the heart, including the ascending aorta, pulmonary artery and the superior part of the left atrium. In contrast, the ‘safe’ distributions mostly avoided all cardio-pulmonary substructures except for parts of the superior vena cava and right pulmonary artery.

Conclusion: This OMT-based approach provides a tool to identify regions associated with survival without requiring pre-defined contours. For the RTOG 0617 dataset, base of the heart was identified as a critical region that influenced overall survival.

Funding Support, Disclosures, and Conflict of Interest: This study was supported by AFOSR grants (FA9550-17-1-0435, FA9550-20-1-0029), NIH grant (R01-AG048769), MSK Cancer Center Support Grant/Core Grant (P30 CA008748), and a grant from Breast Cancer Research Foundation (BCRF-17-193).



    Dose Response, Radiation Therapy, Radiosensitivity


    IM- Radiation Dose and Risk: General (Most Aspects)

    Contact Email