Lack of consensus on post-operative radiotherapy (PORT) fields used in non-small cell lung cancer (NSCLC)

Abstract

7658 Background: Survival in patients with a completely resected stage III-N2 NSCLC may be improved by PORT. LungART is a phase III trial for patients with resected N2 disease who will be randomized to between PORT and no PORT, irrespective of the use of chemotherapy. LungArt target volumes are based upon the post-surgical findings in each patient, and on the surgical literature. A ‘dummy run’ was performed to study PORT volumes used internationally and to evaluate the LungART protocol. Methods: A CD-rom based contouring program containing the planning CT scans of 2 patients (post-lobectomy and pneumonectomy, respectively) were mailed to 17 thoracic radiation oncologists from 4 continents. Relevant clinical, radiological and pathological reports were provided in order to allow clinicans to contour their ‘routine’ clinical target volumes (CTV). Contours were saved onto a CD-rom and mailed to the study center. The LungART protocol was then provided and ‘protocol’ CTVs were generated. Contours of each case were transferred to a template CT dataset of that patient, and nodal CTVs were jointly generated by two clinicians using published definitions. An optimised treatment plan was generated for all CTV's. Results: For each case, a total of 10 routine CTVs were received. Both routine and protocol CTVs were provided for each case by 6 clinicians. The routine CTVs varied up to 3-fold between clinicians and this was significantly decreased by use of the protocol (F test, p=0.00003 and p=0.00002 for cases 1 and 2, respectively). Coverage of nodal stations 2R and 4R improved when CTVs were defined using the protocol. Risks of lung toxicity, as defined by the lung volumes treated to 20 Gy and 5 Gy, varied between the smallest and second largest routine CTVs. For the post-lobectomy case, the V20 (range) was 11.8–19.3% and a V5 of 33.8–57.9%. The corresponding V20 and V5 values were 1.5–2.1% and 31.6–39.7% for the post-pneumonectomy case. Both cardiac volumes receiving 45 Gy and spinal doses were similar for both CTVs. Conclusions: The variations in PORT volumes generated by experts led to important differences in the risk of radiation-induced lung toxicity. Standardized target definitions are needed in order to optimize the survival benefits from PORT, and the LungArt protocol may serve this purpose. No significant financial relationships to disclose.