PO-0782 TARGET DELINEATION IN RECTAL CANCER RADIOTHERAPY: RADIOBIOLOGICAL CONSEQUENCES OF ATLAS IMPLEMENTATION

Abstract

Purpose/Objective: Target volume delineation variability is a major concern for cooperative groups and multi-center clinical trials. The aim of this study is to ascertain the dosimetric and radiobiologic impact of using a consensus guideline target volume delineation atlas regarding the reduction of inter-observer target delineation variation. Materials and Methods: Using a representative case and target volume delineation instructions and a standardized case derived from a proposed SWOG IMRT rectal cancer clinical trial, gross tumor volume (GTV), and clinical target volumes (CTV) were contoured by 13 physician observers. Observers were then randomly assigned to receipt (atlas) or non-receipt (control) of a cooperative-group supported consensus guideline/atlas for anorectal cancers, then instructed to re-contour the same case. PTV margin expansion and plan optimization were performed using standardized parameters from the study protocol, using sequential dynamic arc IMRT. The calculation of tumor control probability (TCP), normal tissue complication probability (NTCP), complication-free tumor control probability (P+), biologically effective uniform dose and generalized equivalent uniform dose (gEUD) was performed for all the plans. Grouped Atlas/Control TCP values were compared using Wilcoxon rank sum test and Levene’s test of variance equality for the grouped cohort GTV, CTV, and PTV inter-observer TCP grand mean (TCP’; i.e. the mean of all interobserver TCP calculations) Results: Although the variation of P+ among the members of the atlasassisted group (0.8%) was larger than that of the control group (0.5%) initially, in the second delineation the reduction of the deviation was lower for the atlas-assisted group (0.7% against 1.0%). Atlas implementation resulted in mean±SD P+ of 85.7±0.8, 78.1±2.0, and 76.8±1.1 for GTV, CTV, and PTV respectively, while for the control group the equivalent values were 85.1±0.6, 78.1±1.3, 76.4±0.6. Reduction in TCP’ variation between plans was observed with atlas use for CTV and PTV ROIs, with a reduction in TCP’ difference between plans from 13.7% (pre-atlas) to <1% (post atlas) for CTV, and from 32.5% (pre-atlas) to 7.7% (post-atlas) [p<0.01]. For the control cohort, no statistically significant differential in grand mean difference was noted, with TCP’ values for CTV of 10.2 and 6.5%, and PTV of 23.6% and 28.2% on sequential recontouring. A <1% TCP’ difference was observed in TCP for both atlas or control cohorts, suggesting minimal influence of the atlas upon GTV definition. Conclusions: The use of a visual atlas and consensus treatment guidelines in the development of rectal cancer IMRT treatment plans resulted in at least 4-fold reduction in the inter-observer dependent radiobiologic variability in TCP’ for CTV and PTV volumes. The observed differences in tumor control probability for the CTV and PTV due to the use of the atlas are potentially clinically meaningful, and suggest routine atlas implementation is a key component for reducing potential confounders of outcome in conformal radiotherapy trials. PO-0783 CONE BEAM CT BASED MARGIN GENERATION: DO WE REQUIRE DIFFERENT MARGINS FOR SUB-REGIONS OF HEAD AND NECK? A. Budrukkar 1 , R. Krishnatry 1 , K. Rajput 2 , S. Yadadu 2 , B. George 2 , S. Chaudhari 2 , S. Ghosh Laskar 1 , V. Murthy 1 , T. Gupta 1 , J. Agarwal 1