Yun-Jau Chang1, Li-Ju Chen, Yao-Jen Chang, Kuo-Piao Chung, Mei-Shu Lai. 1. Department of General Surgery, Zhongxing Branch, Taipei City Hospital, Graduate Institute of Health Policy and Management, College of Public Health, National Taiwan University Hospital, No.17 Suchow Road,Taipei, Taiwan.
Abstract
PURPOSE: There is a controversy about whether lymph node yield can be used as a proxy of quality care for patient with colorectal cancer. We aim to use propensity score models to investigate the association between lymph node number and long-term survival for colorectal cancer patients. MATERIALS AND METHODS: Taiwan Cancer Database was employed to review all patients with newly diagnosed colorectal cancer from 2003 to 2005. Exclusion criteria included those patients with stage IV disease or without information of lymph node. Propensity score models (examined lymph node >12 or <12 as dependent variable) were applied to group of patients with Stage II or Stage III disease and primary end point was 5-year survival (and mortality). We also report results of Stage I-III for comparison. RESULTS: We identified 15,731 newly diagnosed colorectal cancers during study period, among which a total of 10,517 colorectal cancer patients treated at 32 hospitals fulfilled the inclusion criteria. Pathology reports of about 63 % (6658/10517) patients revealed lymph node retrieval >12. After propensity score matching, there were 2888, 1079, 1094 pairs recruited for Stage I-III, Stage II and Stage III, respectively. According to analysis of these matched pairs, the 5-year risk adjusted overall mortality were lower for lymph node examined ≥12 than <12 among Stage II (24.3% vs. 31.1%, p=0.012) and Stage I-III (20.8% vs. 23.6%, p=0.003), but insignificant for Stage III (40.2% vs. 45.6%, p=0.073). Similar situation happened with regard to disease-free and disease-specific mortality. CONCLUSION: For patients with colorectal cancer undergoing colorectal surgery, the quality metric of lymph node is associated with significantly better 5-year survival except for Stage III disease.
PURPOSE: There is a controversy about whether lymph node yield can be used as a proxy of quality care for patient with colorectal cancer. We aim to use propensity score models to investigate the association between lymph node number and long-term survival for colorectal cancerpatients. MATERIALS AND METHODS: Taiwan Cancer Database was employed to review all patients with newly diagnosed colorectal cancer from 2003 to 2005. Exclusion criteria included those patients with stage IV disease or without information of lymph node. Propensity score models (examined lymph node >12 or <12 as dependent variable) were applied to group of patients with Stage II or Stage III disease and primary end point was 5-year survival (and mortality). We also report results of Stage I-III for comparison. RESULTS: We identified 15,731 newly diagnosed colorectal cancers during study period, among which a total of 10,517 colorectal cancerpatients treated at 32 hospitals fulfilled the inclusion criteria. Pathology reports of about 63 % (6658/10517) patients revealed lymph node retrieval >12. After propensity score matching, there were 2888, 1079, 1094 pairs recruited for Stage I-III, Stage II and Stage III, respectively. According to analysis of these matched pairs, the 5-year risk adjusted overall mortality were lower for lymph node examined ≥12 than <12 among Stage II (24.3% vs. 31.1%, p=0.012) and Stage I-III (20.8% vs. 23.6%, p=0.003), but insignificant for Stage III (40.2% vs. 45.6%, p=0.073). Similar situation happened with regard to disease-free and disease-specific mortality. CONCLUSION: For patients with colorectal cancer undergoing colorectal surgery, the quality metric of lymph node is associated with significantly better 5-year survival except for Stage III disease.