Interobserver delineation uncertainty in involved-node radiation therapy (INRT) for early-stage Hodgkin lymphoma: On behalf of the Radiotherapy Committee of the EORTC lymphoma group

Background and purpose: In early-stage classical Hodgkin lymphoma (HL) the target volume nowadays consists of the volume of the originally involved nodes. Delineation of this volume on a post-chemotherapy CT-scan is challenging. We report on the interobserver variability in target volume definition and its impact on resulting treatment plans.Materials and methods: Two representative cases were selected (1: male, stage IB, localization: left axilla; 2: female, stage IIB, localizations: mediastinum and bilateral neck). Eight experienced observers individually defined the clinical target volume (CTV) using involved-node radiotherapy (INRT) as defined by the EORTC-GELA guidelines for the H10 trial. A consensus contour was generated and the standard deviation computed. We investigated the overlap between observer and consensus contour [SOrensen-Dice coefficient (DSC)] and the magnitude of gross deviations between the surfaces of the observer and consensus contour (Hausdorff distance). 3D-conformal (3D-CRT) and intensity-modulated radiotherapy (IMRT) plans were calculated for each contour in order to investigate the impact of interobserver variability on each treatment modality. Similar target coverage was enforced for all plans.Results: The median CTV was 120cm(3) (IQR: 95-173cm(3)) for Case 1, and 255cm(3) (IQR: 183-293cm(3)) for Case 2. DSC values were generally high (>0.7), and Hausdorff distances were about 30mm. The SDs between all observer contours, providing an estimate of the systematic error associated with delineation uncertainty, ranged from 1.9 to 3.8mm (median: 3.2mm). Variations in mean dose resulting from different observer contours were small and were not higher in IMRT plans than in 3D-CRT plans.Conclusions: We observed considerable differences in target volume delineation, but the systematic delineation uncertainty of around 3mm is comparable to that reported in other tumour sites.This report is a first step towards calculating an evidence-based planning target volume margin for INRT in HL