e19036 Background: The TNM classification system is the global standard for staging solid tumors, but the ability of TNM to differentiate prognostic groups may vary between settings because of differences in case-mix and treatments. Little information is available about prognostication using TNM in safety-net health systems, which provide care for socioeconomically disadvantaged individuals and have limited resources. Therefore, we aimed to assess the ability of TNM to differentiate prognostic groups for colorectal cancer patients in an urban safety-net cancer center. Methods: We used institutional registry data from the JPS Oncology and Infusion Center (Fort Worth, TX), which is a Comprehensive Community Cancer Program. Our eligible population included adults who were diagnosed with first primary colorectal cancer between 2008 and 2017, had TNM pathologic stage I – IV, and received at least part of the first course treatment at JPS. Our outcome of interest was 5-year all-cause mortality and our exposure of interest was TNM pathologic stage. Patients were followed from cancer diagnosis until death, loss to follow up, or end of study (December 31, 2018), whichever occurred first. We used a pseudo-observation approach and generalized linear models with log link to estimate risk ratios (RR) and corresponding 95% confidence limits (CL) comparing 5-year mortality between TNM stages, where stage I was the reference category. Results: Our eligible population comprised 655 colorectal cancer survivors, of whom 85% were aged < 65 years, 54% were male, 60% were racial/ethnic minorities, and 65% were uninsured at time of diagnosis. Stage IV patients had 4.4 times higher risk of 5-year mortality compared with stage I (RR = 4.4, 95% CL: 2.9, 6.6). Stage III patients had 2.1 times higher risk of 5-year mortality compared with stage I (RR = 2.1, 95% CL: 1.3, 3.2). Stage II patients had 1.4 times higher risk of 5-year mortality compared with stage I (RR = 1.4, 95% CL: 0.81, 2.3). Conclusions: Our findings suggest that TNM stages do not adequately differentiate colorectal cancer prognostic groups in a socioeconomically disadvantaged population. Stage IV is distinct, but stages I – III have limited differentiation and notable overlap. The use of TNM stages for colorectal cancer prognostication may be misleading for socioeconomically disadvantaged individuals. Factors beyond TNM should be considered for better prognostication in this population.
Optimal Follow-up of Stage I Colorectal Cancer Patients
