Whodunnit.

Top Line: To what extent is radiation toxicity attributable to the treating physician?

The Study: In yet another secondary analysis of a recent trial comparing neoadjuvant vs adjuvant sequencing of short-course ADT and prostate radiation, the authors of this study sought to compare toxicity among patients treated by individual radiation oncologists on trial. The >400 patients received 76 Gy in 38 fractions using 6-field 3D conformal planning and fiducial guidance, and the initial 56 Gy in 28 fractions included the proximal seminal vesicles. Margins were standard on protocol. So, off the bat we can hypothesize that any differences in toxicity could be 1) bad luck from lack of statistical power, 2) underlying patient characteristics, or 3) variation in physician target delineation or planning. Patients were clustered according to their treating physician for analysis. At baseline there were numerous significant differences in baseline patient characteristics. Sadly, none of those analyzed characteristics included things like baseline urinary or bowel function, prostate size, or target volume and dosimetric data. In addition, there were pretty big discrepancies in the number of patients enrolled by each physician. For instance, 45% of analyzed patients were treated by a single physician while the top two physicians accounted for >60% of analyzed patients. Lastly, the rates of late GI (2.5%), GU (3.5%), and overall toxicity (5.9%) were low. With these limitations, the authors did find a significant difference among physicians for overall grade 3+ toxicity events. However, there was no difference in toxicity among physicians when GI and GU toxicity were analyzed individually.

TBL: This study demonstrates the methodological difficulty of attributing toxicity risk to individual treating physicians in radiation oncology. | Roy, Pract Radiat Oncol 2020