In a study reported in Clinical Cancer Research, Gold and colleagues developed a risk model for recurrence of resected non–small cell lung cancer (NSCLC) based on protein expression and clinical variables.
The study involved analysis of samples from 370 patients with NSCLC resected between 2002 and 2005 at The University of Texas MD Anderson Cancer Center. Proteins examined were from pathways related to metabolism, DNA repair, inflammation, and growth factors. Protein expression was measured by immunohistochemistry, chosen because the technique is widely available and can be used on formalin-fixed paraffin-embedded samples. Overall, 63% of patients had stage I disease. Median follow-up was 5.3 years.
The multivariate Cox model for recurrence-free survival consisted of age (hazard ratio [HR] = 1.024 per year increase, P = .001), stage (HR = 1.765 for stage II and 2.676 for stage III vs stage I, P = .002 and < .001), positive membrane insulin receptor (HR = 1.442, P = .012; 5-year recurrence-free survival = 40% vs 48% for negative), cytoplasmic CXCR2 above the median (HR = 1.360, P = .038; 5-year recurrence-free survival = 41% vs 49% below the median), and elevated IGF1R (HR = 1.517 per 100 increase, P = .040) as significant predictors of shorter recurrence-free survival, and positive cytoplasmic pAMPK (HR = 0.648, P = .004; 5-year recurrence-free survival = 50% vs 33% for negative), positive cytoplasmic pmTOR (HR = 0.696, P = .029; 5-year recurrence-free survival = 49% vs 34% for negative ), positive cytoplasmic EpCAM (HR = 0.708, P = .024; 5-year recurrence-free survival = 49% vs 39% for negative) and higher membrane CASK (0.680 per 100 increase, P = .049) as significant predictors of longer recurrence-free survival.
The multivariate Cox model for overall survival consisted of age (HR = 1.028 per year increase, P = .001), stage (HR = 1.425 for stage II and 2.620 for stage III vs stage I, P = .087 and < .001), positive cytoplasmic pAMPK (HR = 0.669, P = .018; 5-year overall survival = 64% vs 52% for negative), positive cytoplasmic pmTOR (HR = 0.662, P = .026; 5-year overall survival = 65% vs 48% for negative), and positive cytoplasmic EpCAM (HR = 0.648, P = .012; 5-year overall survival = 66% vs 51% for negative) as significant predictors of longer overall survival and higher cytoplasmic CXCR2 (HR = 1.568, P = .007; 5-year overall survival = 55% vs 66% for lower), and higher nuclear FEN1 (HR = 1.424, P = .035; 5-year overall survival = 54% vs 68% for lower) as significant predictors of shorter overall survival.
The investigators concluded, “We have developed a comprehensive risk model predictive for recurrence in our large retrospective database, which is one of the largest reported series of resected NSCLC.” ■
Gold KA, et al: Clin Cancer Res. December 23, 2013 (early release online).
