Model dosimetry.

Given its particularly disappointing results, RTOG 0617 has been milked for all its worth in providing useful data outside its primary endpoint. Its long term results indicated that higher radiation doses to central structures led to worse survival outcomes, particularly the volume of heart receiving <5 and <30 Gy. To delve into more granular dosimetry details, this post-hoc dosimetric analysis first developed a predictive model of survival based on data from 306 enrollees. Among 191(!) variables, the stand-outs were mean lung dose, minimum dose to the hottest 45% of the atria, mean dose to the hottest 55% of the pericardium, and mean dose to the hottest 5% of the ventricles. That alone is interesting, but what’s really newsworthy is its ability to predict survival pretty spot on in the validation cohort of another 131 enrollees (see ensemble model in figure 2). TBL: The authors suggest exploring the utilization of these combined dosimetric thresholds in standard treatment planning for advanced non-small cell lung cancer. | Thor, Clin Cancer Res 2020