Comparison of investigator-delineated gross tumor volumes and quality assurance in pancreatic cancer: Analysis of the pretrial benchmark case for the SCALOP trial.

BACKGROUND AND PURPOSE To evaluate the variation in investigator-delineated volumes and assess plans from the radiotherapy trial quality assurance (RTTQA) program of SCALOP, a phase II trial in locally advanced pancreatic cancer. MATERIALS AND METHODS Participating investigators (n=25) outlined a pre-trial benchmark case as per RT protocol, and the accuracy of investigators' GTV (iGTV) and PTV (iPTV) was evaluated, against the trials team-defined gold standard GTV (gsGTV) and PTV (gsPTV), using both qualitative and geometric analyses. The median Jaccard Conformity Index (JCI) and Geographical Miss Index (GMI) were calculated. Participating RT centers also submitted a radiotherapy plan for this benchmark case, which was centrally reviewed against protocol-defined constraints. RESULTS Twenty-five investigator-defined contours were evaluated. The median JCI and GMI of iGTVs were 0.57 (IQR: 0.51-0.65) and 0.26 (IQR: 0.15-0.40). For iPTVs, these were 0.75 (IQR: 0.71-0.79) and 0.14 (IQR: 0.11-0.22) respectively. Qualitative analysis showed largest variation at the tumor edges and failure to recognize a peri-pancreatic lymph node. There were no major protocol deviations in RT planning, but three minor PTV coverage deviations were identified. . CONCLUSIONS SCALOP demonstrated considerable variation in iGTV delineation. RTTQA workshops and real-time central review of delineations are needed in future trials.