Lung cancer chemotherapy. Response-survival relationship depends on the method of chest tumor response evaluation

In a previous study we found that tumor responses as assessed by CT scan and fiberoptic bronchoscopy are sometimes discordant. We hypothesize that the response-survival relationship might vary according to the method of tumor response assessment. In a multivariate analysis of survival using the landmark method, we evaluated the prognostic significance of tumor response assessed by CT scan or fiberoptic bronchoscopy together with bronchial tumor location and histology of bronchial biopsies at restaging. A total of 133 lung cancer patients (50 small cell lung cancers and 83 non-small cell lung cancers) were entered in controlled chemotherapy trials and prospectively evaluated for chest tumor response by CT scan and fiberoptic bronchoscopy (FOB). Only 106 patients were fully evaluable for response by both methods. For these patients, a statistical concordance was observed between the two tests (kappa = 0.271; p < 0.001). There was a significant correlation between response and survival whatever the test used. However, only CT scan evaluation resulted in a classification showing that the more unfavorable the response stage was, the worse the survival became with no intersection between survival curves. Cox's hazard model demonstrated that CT-evaluated progression, proximal bronchial location at second FOB (intermediate, main bronchus or trachea) and positive histologic status at restaging were all prognostic determinants of poor survival. In conclusion, CT-evaluated response led to the best response-survival relationship as this method classified patients into four groups with different outcomes. Fiberoptic bronchoscopy should be avoided in patients who were found to have no endobronchial lesion during the pretreatment staging. For patients with pretreatment assessable endobronchial lesions, the decision of a second FOB depends on the results of CT restaging: FOB is probably unnecessary in patients for whom progression is disclosed by CT scan. In patients for whom CT scan discloses tumor response or stabilization, bronchial tumor location and histology of bronchial biopsies at second FOB are independent prognostic factors.