Physiological and radiological criteria are both used to identify candidates for LVRS. This study compares the predictive value of these screening techniques among patients with homogeneous (Ho) and heterogeneous (He) emphysema. Preoperative inspiratory lung conductance (G(Li)) during spontaneous breathing and quantitative radioisotope V/Q scan (QVQS) results were available for 48 of 50 patients undergoing bilateral LVRS for emphysema. Ho disease (n = 21) was defined by QVQS as an upper/lower perfusion ratio (ULPR) between 0.75 and1.25. G(Li) correlated with 6-mo improvement in FEV(1) (DeltaFEV(1)-6) (r = 0.53, p < 0.001) for the entire cohort, and for patients with both Ho (n = 21, r = 0.56, p = 0.015) and He disease (n = 27, r = 0.46, p = 0.017). ULPR correlated less well with DeltaFEV(1)-6 (n = 48, r = -0.38; p = 0.008) for the cohort, and was significantly correlated with outcomes only in the subgroup of patients with He disease (r = -0.40, p = 0.04). Multivariate regression demonstrated that by combining G(Li) and ULPR criteria, 33% of the DeltaFEV(1)-6 response could be accounted for. We conclude that both physiological and radiological criteria help identify appropriate candidates for LVRS. G(Li) best identifies patients with Ho emphysema who may benefit from surgery, but would be excluded on the basis of strictly radiological criteria. ULPR helps identify patients with He disease that improves with surgery, despite unfavorable G(Li).