ESTRO 2024 - Abstract Book

S1478

Clinical - Lower GI

ESTRO 2024

may be reduced with the aid of a radiologist, possibly improving clinical outcome. 2 A pilot study was performed to investigate the benefit of additional GTV delineation by an expert radiologist for LRRC.

Material/Methods:

14 LRRC cases treated with neoadjuvant CRT were retrospectively selected. Participants received clinical information, a planning CT and a matched MRI and/or PET-CT, as available. First, 8 expert radiologists delineated the GTV on CT. A median delineation per case was constructed and defined as the “gold standard”. Second, 12 ROs delineated GTV on CT for all cases, with the gold standard available in 7. The case mix was alternated so that no ROs delineated a case twice and so that each case was delineated by a similar amount of ROs with and without the aid of the gold standard (GTV+ and GTV- respectively). IOV was calculated for radiologists in reference to the gold standard. To compare RO's IOV between GTV+ and GTV-, variation was analysed in reference to a median delineation per group (i.e. median of GTV+ delineations and median of GTV- delineations). IOV was analysed with the Dice Similarity Coefficient (DSC), Surface Dice (SDSC) and Hausdorff Distance (98th percentile) (HD98%). 3,4 IOV was categorized as (1) Red: High variation (SDSC/DSC <0.6 and HD98>20mm), (2) Yellow: Average variation (0.6≤SDSC/DSC<0.8 and 10mm

Results:

A total of 269 GTV delineations were returned (112 radiology and 157 RO (9-12 per case)). Patient characteristics and overall IOV are shown in table 1.