1177P - Variations of discordance of the clinical TNM stage with the pathological TNM stage between Japanese designated cancer hospitals

Date 30 September 2012
Event ESMO Congress 2012
Session Poster presentation II
Topics Imaging, Diagnosis and Staging
Presenter Fumiaki Nakamura
Authors F. Nakamura1, T. Higashi2, Y. Emori3, H. Nishimoto3
  • 1Healthcare Epidemiology, Kyoto University Graduate School of Medicine and Public Health, 606-8501 - Kyoto/JP
  • 2Department Of Public Health/health Policy, The University of Tokyo Graduate School of Medicine, Tokyo/JP
  • 3Cancer Surveillance Division, National Cancer Center, 1040045 - Tokyo/JP



Previous studies showed that TNM stages that were clinically determined before surgery were not often concordant with pathological TNM stages. However, no previous studies have examined variations of discordance of the clinical TNM stage with the pathological TNM stage among hospitals. We aimed to examine the discordance of the clinical and pathological stages among Japanese designated cancer hospitals using compiled data from the hospital-based cancer registry submitted from 286 designated cancer care hospitals in Japan.


The registry data had UICC TNM stages before and after surgery for stomach, colorectal, lung and breast cancer patients treated in these hospitals. We excluded patients who received adjuvant chemotherapy or radiotherapy, patients who received care from facilities with less than 10 patients, male breast cancer and patients whose stages were unknown from the analysis. We also calculated discordance of stages that could have theoretically resulted in changes in surgical procedures or treatment choice. Discordance was also calculated after excluding stage IV patients as sensitivity analysis.


In total, 145,726 patients (38,692 stomach, 47,304 colorectal, 19,643 lung and 40,096 breast cancer patients) were included in the analysis. Patients' mean age and clinical TNM stage varied across hospitals. The facility-level discordance between clinical and pathological stages ranged from 4% to 52% in stomach cancer, 3% to 57% in colorectal cancer, 0% to 50% in lung cancer and 1% to 45% in breast cancer. TNM stages before surgery were lower than after surgery in more than half of the cases. The multi-level logistic regression model showed that variance in facilities existed after adjustment for age, sex and clinical TNM stage. Discordance of stages that could lead to changes in the surgical procedures or treatment choice ranged from 4% to 58% in stomach cancer, 4% to 63% in colorectal cancer, 0% to 5% in lung cancer, and 5% to 52% in breast cancer. Sensitivity analysis yielded similar results.


We can confirm variations of discordance of TNM stages before and after surgery.


All authors have declared no conflicts of interest.