IS THE CHEST RADIOGRAPH A RELIABLE TOOL IN THE ASSESSMENT OF TUMOR RESPONSE AFTER RADIOTHERAPY IN NONSMALL CELL LUNG-CARCINOMA

Background: The purpose of this study was to evaluate whether the ches t radiograph is a reliable tool to assess response to radiotherapy. Ma terials and Methods: Pre- and post-treatment chest radiographs and com puted tomographs (CT) of 63 patients with nonsmall cell lung cancer (N SCLC) treated by radiotherapy were reviewed by four observers with reg ard to suitability for tumor measurement, and response. Suitability fo r tumor measurement was expressed as the number of measurable diameter s. In addition, the consequences to clinical outcome were studied by s urvival analysis. Results: The CT turned out to be more suited for tum or measurement before as well as after radiotherapy, resulting in an i ncrease of the number of measurable cases. The number of measurable ca ses with CT was 52 (83%) as compared to 28 (44%) with chest radiograph y. Especially in case of centrally localized tumors, the presence of a n atelectasis, or squamous cell carcinoma, CT contributed to a higher rate of measurable cases. The interobserver agreement with regard to r esponse using chest radiograph was good (mean kappa = 0.74). In 25 of 28 cases (89%) measurable with CT as well as with chest radiograph, re sponse was equally classified. When CT was used, the median survival o f the responders was 14.2 months as compared to 6.8 months of the nonr esponders. When chest radiograph was used, the median survival of thes e groups was 12.0 and 6.6 months respectively, which was not significa ntly different when response was assessed by CT. Conclusion: We conclu de that CT is more suited for tumor measurement because more measurabl e lesions can be found and more evaluable lesions on chest radiograph become measurable on CT. The chest radiograph does have a valuable rol e to play in those lesions that are measurable because of the good int erobserver agreement with regard to the response classification, the h igh overall agreement between CT and chest radiograph in case of measu rable cases, and the lack of important differences with regard to surv ival. (C) 1998 Elsevier Science Inc.