The development of a predictive model to improve risk stratification for patients after resection for lung cancer may improve treatment strategies and outcomes in addition to providing better prognostic estimates for patients, families, and physicians. The effectiveness of the model is influenced by practice patterns for surveillance after resection, the frequency and intensity of radiologic examination, and the aggressiveness of the confirmation of recurrence. In this study [1Williams B.A. Sugimura H. Endo C. et al.Predicting post-recurrence survival among completely resected nonsmall-cell lung cancer patients.Ann Thorac Surg. 2006; 81: 1021-1027Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar], fewer than 25% of the patients identified with recurrence were free of symptoms, and 21% of patients were not analyzed due to inadequate follow-up information. In a setting with more frequent clinical assessment, post-recurrence survival would likely be longer, and other prognostic factors might have been identified as significant. Although the prolonged postrecurrence survival may only represent lead-time bias, the factors identified that relate to the prognostic model may alter clinical practice and improve actual survival.For a prognostic model to influence patient care and outcome, there should be a treatment algorithm associated with the scoring system. Although this model is intriguing, the patients scored at moderate risk or high risk would likely have been accurately identified without this model, as the risk impact of performance status, symptoms, and multiple sites of recurrence are well described. Furthermore, these patients are generally considered poor candidates for therapy, regardless of the predicted length of survival. The ideal model would allow identification and stratification of a larger fraction of patients with early recurrence for which varying therapeutic strategies may be used, including targeted therapy. In addition, stratification would be improved by incorporating tumor-specific factors, such as genetic mutations, protein expression, and treatment sensitivity or resistance analysis. The development of a predictive model to improve risk stratification for patients after resection for lung cancer may improve treatment strategies and outcomes in addition to providing better prognostic estimates for patients, families, and physicians. The effectiveness of the model is influenced by practice patterns for surveillance after resection, the frequency and intensity of radiologic examination, and the aggressiveness of the confirmation of recurrence. In this study [1Williams B.A. Sugimura H. Endo C. et al.Predicting post-recurrence survival among completely resected nonsmall-cell lung cancer patients.Ann Thorac Surg. 2006; 81: 1021-1027Abstract Full Text Full Text PDF PubMed Scopus (76) Google Scholar], fewer than 25% of the patients identified with recurrence were free of symptoms, and 21% of patients were not analyzed due to inadequate follow-up information. In a setting with more frequent clinical assessment, post-recurrence survival would likely be longer, and other prognostic factors might have been identified as significant. Although the prolonged postrecurrence survival may only represent lead-time bias, the factors identified that relate to the prognostic model may alter clinical practice and improve actual survival. For a prognostic model to influence patient care and outcome, there should be a treatment algorithm associated with the scoring system. Although this model is intriguing, the patients scored at moderate risk or high risk would likely have been accurately identified without this model, as the risk impact of performance status, symptoms, and multiple sites of recurrence are well described. Furthermore, these patients are generally considered poor candidates for therapy, regardless of the predicted length of survival. The ideal model would allow identification and stratification of a larger fraction of patients with early recurrence for which varying therapeutic strategies may be used, including targeted therapy. In addition, stratification would be improved by incorporating tumor-specific factors, such as genetic mutations, protein expression, and treatment sensitivity or resistance analysis.
Read full abstract