ABSTRACT
PURPOSE: Tissue inhibitor of metalloproteinase-1 (TIMP-1) is known to protect cells against apoptosis. We raised the hypothesis that elevated tumor tissue levels and thereby plasma levels of TIMP-1 would predict resistance to apoptosis-inducing chemotherapy. EXPERIMENTAL DESIGN: Ninety patients with metastatic colorectal cancer were included in the study. Plasma TIMP-1 and serum carcinoembryonic antigen (CEA) were measured in samples obtained before the first cycle of chemotherapy. RESULTS: Analysis of best objective response (complete or partial response versus stable or progressive disease) showed that patients with low plasma TIMP-1 had higher probability of obtaining an objective response [odds ratio (OR), 3.5; 95% confidence interval (95% CI), 1.4-8.5, P=0.007]. CEA treated as a continuous variable was also a statistically significant predictor of no response (OR, 1.3; 95% CI, 1.0-1.7, P=0.02, area under the curve 0.66) but much less so. Plasma TIMP-1 was the only significant covariate in a multivariable analysis of best objective response (OR, 3.6; 95% CI, 1.4-9.5; P=0.001). Plasma TIMP-1 scored as a continuous variable on the log scale (log(e)) was significantly associated with overall survival [OS; hazard ratio (HR), 3.8; 95% CI, 2.4-5.9; P<0.0001] and with time to progression (TTP; HR, 1.5; 95% CI, 1.0-2.3; P=0.048). Multivariable analysis showed that plasma TIMP-1 was significant for OS when including routine clinical baseline covariates (HR, 3.5; 95% CI, 2.1-5.8; P<0.0001). A multivariable analysis including TTP instead of OS showed that only plasma TIMP-1 was retained in the model (HR, 1.5). CEA was not significantly associated with TTP or OS when TIMP-1 was included in the model. CONCLUSION: This study shows that plasma TIMP-1 levels are significantly and independently associated with objective response, TTP, and OS in patients with metastatic colorectal cancer receiving combination chemotherapy.
