Tumor Heterogeneity and Permeability as Measured on the CT Component of PET/CT Predict Survival in Patients with Non–Small Cell Lung Cancer

Abstract

Purpose: We prospectively examined the role of tumor textural heterogeneity on positron emission tomography/computed tomography (PET/CT) in predicting survival compared with other clinical and imaging parameters in patients with non–small cell lung cancer (NSCLC). Experimental Design: The feasibility study consisted of 56 assessed consecutive patients with NSCLC (32 males, 24 females; mean age 67 ± 9.7 years) who underwent combined fluorodeoxyglucose (FDG) PET/CT. The validation study population consisted of 66 prospectively recruited consecutive consenting patients with NSCLC (37 males, 29 females; mean age, 67.5 ± 7.8 years) who successfully underwent combined FDG PET/CT-dynamic contrast-enhanced (DCE) CT. Images were used to derive tumoral PET/CT textural heterogeneity, DCE CT permeability, and FDG uptake (SUV_max). The mean follow-up periods were 22.6 ± 13.3 months and 28.5± 13.2 months for the feasibility and validation studies, respectively. Optimum threshold was determined for clinical stage and each of the above biomarkers (where available) from the feasibility study population. Kaplan–Meier analysis was used to assess the ability of the biomarkers to predict survival in the validation study. Cox regression determined survival factor independence. Results: Univariate analysis revealed that tumor CT-derived heterogeneity (P < 0.001), PET-derived heterogeneity (P = 0.003), CT-derived permeability (P = 0.002), and stage (P < 0·001) were all significant survival predictors. The thresholds used in this study were derived from a previously conducted feasibility study. Tumor SUV_max did not predict survival. Using multivariable analysis, tumor CT textural heterogeneity (P = 0.021), stage (P = 0.001), and permeability (P < 0.001) were independent survival predictors. These predictors were independent of patient treatment. Conclusions: Tumor stage and CT-derived textural heterogeneity were the best predictors of survival in NSCLC. The use of CT-derived textural heterogeneity should assist the management of many patients with NSCLC. Clin Cancer Res; 19(13); 3591–9. ©2013 AACR.