• Citer

Standardizing mediastinal nodal CTV delineation in Stage III NSCLC: results of a two-phase dummy run

par Charlier, Florian ;Remouchamps, Vincent;Lambrecht, Maarten;Geets, Xavier;Lievens, Yolande;Moretti, Luigi
Référence Radiotherapy and oncology, 133, page (S405)
Publication Publié, 2019-04-01

Article révisé par les pairs

• Accès en ligne
• Détails
• Contenu
• Statistiques

Résumé :

[Purpose or Objective] Lymph node delineation in stage III non-small cell lung cancer (NSCLC) remains the subject of discussion as the inclusion or omission of a lymph node or station can have an important impact on both oncological outcome and toxicity. The Belgian College for Physicians in Radiation Oncology aims to improve the quality of radiotherapy in Belgium. Within this framework the Project on Cancer of the Lung (ProCaLu) focuses on standardizing delineation for locally advanced NSCLC. A two-phase contouring dummy-run on a single test case was performed to analyze the baseline interobserver variability on target definition and delineation and evaluate the effects of a national consensus before the start of a peer-review program for routine clinical practice. [Material and Methods] All 25 radiotherapy centers in Belgium and Luxembourg were invited to take part in a delineation dummy-run by sending contours from at least one radiation oncologist (RO) dedicated to lung cancer treatments. The case consisted of a squamous cell carcinoma of the right upper lobe with nodal extension to station 4R staged cT1bN2M0 (TNM 7). The results of chest CT, PET/CT and endoscopic samplings were provided to participants and the planning CT was made available through a secure transfer platform. RO's were asked to delineate and upload back the GTV and CTV for the nodal disease. More than a year later and after the formulation of a national consensus based on current ESTRO guidelines (inclusion of only involved nodes with a 5 mm CTV margin) the case was resent to all centers with the same purpose. On all received contours, a description of the TV definition and delineation was obtained through visual inspection. Using an open-source software (3D Slicer with SlicerRT), the DICE Similarity Index and Hausdorff distances were analyzed to compare the delineations. To allow an evaluation of uniformity with these metrics, independent GTV and CTV contours by the first author were used as reference. [Results] At the first phase, 16 contours sent by 14 RT centers were analyzed. Important variations were due to the inclusion of a negative node in GTV (n