Top Line: What are sources of variation in contouring mediastinal target volumes when treating stage III NSCLC?
The Study: ProCaLung is a Belgian project designed to evaluate the quality of mediastinal nodal target volume delineation for stage III NSCLC. Contouring was performed in a two step process of target identification and then delineation. A study by Peeters, et al. in 2016 established a useful algorithm for determining when to include mediastinal nodes in the GTV based on PET/CT and EBUS findings. The algorithm essentially recommends covering PET+ nodes regardless of EBUS results, and then relying on EBUS to determine whether or not to include PET- nodes. In their meta-analysis, EBUS of PET- nodes reduced the false negative rate by a relative ~80% for both enlarged and normal sized nodes. Without EBUS, enlarged nodes (short axis ≥1cm) are considered positive. The ESTRO-ACROP guidelines from 2018 were then used as a basis for mediastinal target volume delineation. These guidelines offer two methods for nodal CTV delineation. First, the entire nodal station including a 5-8mm expansion of the GTV can be delineated. Second, a geometric expansion of 5-8mm from the nodal GTV can be used with no additional coverage of the involved station. These guidelines also allow for elective inclusion of the ipsilateral hilum and stations adjacent to the involved stations. As you can imagine, the latter option may introduce significant variation in practice. In ProCaLung, ESTRO-ACROP-based CTV delineation was restricted to a 5mm GTV expansion (not the entire station) and no elective inclusion of adjacent stations. Quantitative and qualitative analysis of mediastinal GTV and CTV contours identified 3 persistent sources of variation. First was the inclusion of small neighboring nodes, i.e. those ≤5mm from a positive node. While the guidelines recommended including these in the CTV, 40% of participants included them in the GTV. The other two sources were frequent inclusion of the pericardium and superior aortic recess as well as adjacent, uninvolved blood vessels. The authors noted a lack of clear “crop-to” mediastinal structures in contouring guidelines as a reason why isotropic expansions often extend into the heart and vessels.
TBL: Clear contouring guidelines that specify how to handle common adjacent non-tumor structures and qualitative peer review are important ways to reduce variation in mediastinal target volume delineation | Charlier, Radiother Oncol 2021