Interobserver variability in delineation of target volumes in head and neck cancer

Overview of artificial intelligence-based applications in radiotherapy: Recommendations for implementation and quality assurance

“…Accurate segmentation of OARs and target volumes is the starting point of radiation treatment planning. Manual segmentation is a time-consuming task with high intra and interobserver variability both within [34] and across [35] radiotherapy centers. Recently, convolutional neural network (CNN)-based auto-segmentation models have been shown to improve consistency and efficiency of this process [36,37].…”

Quality of Life, Toxicity and Unmet Needs in Nasopharyngeal Cancer Survivors

“…While the above studies have highlighted a number of different strategies focusing on reducing toxicity, two large sources of heterogeneity in the treatment planning process which may significantly impact toxicity are variations in target volume contouring and what constitutes an acceptable radiotherapy treatment plan (83)(84)(85). While the pioneering work of Peters et al eloquently demonstrated the detrimental impact of poorly compliant radiotherapy treatment plans on long-term survival in HNC patients (86), a less obvious sin is that of overcontouring, which while resulting in locoregional control, may labor survivors with unnecessary morbidity.…”

Radiotherapy Quality Assurance for Head and Neck Squamous Cell Carcinoma

“…However, in 2010, Rasch et al reported considerable heterogeneity in CTV delineation among Dutch radiation oncologists, despite the publication of guidelines on CTV delineation by Gregoire et al (31). Furthermore, in 2017, van der Veen et al found large discrepancies in the selection of prophylactic nodal levels and CTVs delineated among Belgian centers (14/22) (37,38), illustrating that continued efforts are required in training and education to improve standardization.…”