A new formula for predicting liver metastasis in patients with colorectal cancer: Immunohistochemical analysis of a large series of 439 surgically resected cases

Objective: The purpose of this study was to establish a new formula predicting liver metastasis in patients with colorectal cancer (CRC). Methods: Nine previously reported predictive markers for liver metastasis and/or prognosis (COX-2, dysadherin, E-cadherin, β-catenin, Ki-67, p53, laminin5γ2, matrilysin and MUC-1) were immunohistochemically investigated in 439 consecutive patients with CRC. We tried to determine the combination of molecules which best predicted liver metastasis. A formula for predicting liver metastasis was constructed using a training cohort comprising 150 cases, and applied to a validation cohort comprising 190 cases and another comprising 99 cases from an outside hospital. Results: A combination of dysadherin, E-cadherin and matrilysin was identified to be best for predicting liver metastasis (area under the curve value, 0.807). The predictive formula: 3× dysadherin score [0 for low expression (≤50% of tumor cells positive) or 1 for high expression (>50%)] + 4× E-cadherin score [0 for preserved (>80% of tumor cells positive) or 1 for reduced (≤80%)] + 2× matrilysin score [0 for low expression (≤30% of tumor cells positive) or 1 for high expression (>30%)] was able to discriminate patients with liver metastasis in the training cohort with a sensitivity of 85.7% and a specificity of 58.9%. The discriminative capacity of the formula was validated in the first cohort with a sensitivity of 87.0% and a specificity of 66.5%, and in the second cohort with a sensitivity of 80% and a specificity of 60.0%. Conclusions: We have established a formula for predicting liver metastasis in patients with CRC, and confirmed that it has a high sensitivity potentially useful for clinical application.