Reply to M.C. Chamberlain et al

Le résumé fourni par la source

An Accurate and Facile Diagnosis-Specific Tool to Estimate Survival for Patients With Brain Metastases." 2We not only largely agree with their comments but have included them in the article.As we explicitly stated in our article, our analysis was retrospective and has the inherent weaknesses of any retrospective study.We remind readers that the Graded Prospective Assessment (GPA) was originally designed from prospective data from four Radiation Therapy Oncology Group (RTOG) trials, and then was later refined to be diagnosis specific. [3]4][5] Chamberlain and Silbergeld 1 suggest that the survival by treatment in Table 2 of our article is problematic.Point three in our discussion section states: "Indeed, the survival by treatment data shown in Table 2 is certainly fraught with selection bias and should not be blindly applied or expected.Nonetheless, these data reflect patterns of care for patients with brain metastases over the last quarter century." 2(p424) Furthermore, the multivariate Cox regression analysis that we performed for Table 2 was stratified by institution; hence, the treatment differences reflected the average differences across 11 institutions that may have different institutional policies with respect to modality of therapy.Therefore, the result is more generalizable than it would be if the analysis was based on just one institution.We also agree with the comment by Chamberlain and Silbergeld 1 that a retrospective study does not supplant a prospective study, specifically RTOG 9508. 6 We (P.W.S., L.E.G., and M.P.M.) are authors of that study and are familiar with its findings.Furthermore, that study is currently undergoing reanalysis to evaluate whether its conclusions might vary if post stratified by the Diagnosis-Specific Graded Prognostic Assessment (DS-GPA) rather than the Recursive Partitioning Analysis.We would disagree, however, with the contention of Chamberlain and Silbergeld 1 that retrospective data provide "little guidance with respect to the best treatment of BM for specific clinical scenarios."The DS-GPA can aid, but certainly not replace, clinical judgment when a clinician, patient, and family are making treatment choices.Finally, we agree that the DS-GPA is an improvement over the RTOGRecursivePartitioningAnalysisclassificationandthattheDS-GPA should be validated in prospective trials.That validation is already underway.The DS-GPA has been adopted for purposes of stratification in two prospective, randomized, RTOG clinical trials (1118 and 1119).

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.