Development of a patient-centered aggregate score to predict survival after lung resection for non-small cell lung...

The objective of this analysis was to develop a survival aggregate score (SAS), including objective and subjective patient-based parameters, and assess its prognostic role after major anatomic resection for non-small cell lung cancer.

A total of 245 patients underwent major lung resections for non-small cell lung cancer with preoperative evaluation of quality of life (Short-Form 36v2 survey) and complete follow-up. The Cox multivariable regression and bootstrap analyses were used to identify prognostic factors of overall servival, which were weighted to construct the scoring system and summed to generate the SAS.

Cox regression analysis showed that the factors negatively associated with overall survival and used to construct the score were 36-item short-form health survey physical component summary score less than 50 (hazard ratio [HR], 1.7; P = .008), aged older than 70 years (HR, 1.9; P = .002), and carbon monoxide lung diffusion capacity less than 70% (HR, 1.7; P = .01). Patients were grouped into 4 risk classes according to their SAS. The 5-year overall survival was 78% in class SAS0, 59% in class SAS1, 42% in class SAS2, and 14% in class SAS3 (log-rank test, P < .0001). SAS maintained its association with overall survival in patients with stages pT1 (log-rank test, P = .01), pT2 (log-rank test, P = .02), or pT3-4 (log-rank test, P = .001), and in those with stages pN0 (log-rank test, P = .0005) or pN1-2 (log-rank test, P = .02). The 5-year cancer-specific survival was 83% in class SAS0, 71% in class SAS1, 63% in class SAS2, and 17% in class SAS3 (log-rank test, P < .0001).

This system may be used to refine stratification of prognosis for clinical and research purposes.