To determine whether pulmonary function tests (PFT) predict survival in lung cancer (LC).
150 patients diagnosed with LC were stratified into performance groups based on initial predicted PFTS (FEV1%, FVC%, and DLCO%). The groups for FEV1% and FVC% were: <40%, 40–59%, 60–79% and 80–100% and for DLCO% were: <50%, 50–74%, and 75%-100%. DLCO% were available for 113 patients. Survival times were compared using Kaplan-Meier (KM) product-moment curves and log-rank tests with SAS® Version 8, statistical procedure PROC LIFETEST. Multivariate analyses predicting survival were then performed stratifying across cancer staging and testing (log rank tests) for covariates: age, PFTS, ECOG score, and weight loss.
Only DLCO was significantly associated with predicted survival (p=0.043) [Table
Kaplan-Meier survival based on PFTSPFTNMedian Survival in Days, 95% CI (Range)X2 Valuep ValueFEV1 %3.470.324<4013201 (139–601)40–5937328 (202–468)60–7948376 (273–502)80–10052532 (416–808)FVC%4.4310.219<406166 (115–291)40–5936288 (201–450)60–7943400 (259–636)80–10065489 (416–640)DLCO6.290.043<5043262 (201–480)50–7443446 (334–561)75–10027882 (273–934), figure].For multivariate analysis, the KM curves across cancer stages were not equal, Χ28 = 27.67, p = <0.001. When stage was included, none of the other covariates were predictive of survival (DLCO % showed a trend towards significance, p = 0.081).CONCLUSIONS: Although baseline DLCO % initially appeared significant, multivariate analysis revealed that when cancer stage was included, no other variable added to predicted lung cancer survival.
A larger group of patients is needed to determine whether DLCO influences survival in addition to lung cancer stage.
R.L. Gross, None.