Patterns of Care and Data Quality in a National Registry of Black and White Patients with Merkel Cell Carcinoma

Purpose/Objective(s)

Merkel Cell Carcinoma (MCC) is a rare cutaneous cancer most commonly affecting White patients. Less is known regarding presentation, treatment, and quality of data reporting for Black patients with MCC. We sought to identify the characteristics of MCC, and to compare patterns of treatment, data reporting quality, and as a secondary outcome – all-cause mortality – by race.

Materials/Methods

We queried the National Cancer Database for all patients with MCC and coded White or Black race from 2006 – 2017. Race and treatment associations were assessed using multivariable models (logistic, proportional odds logistic, and baseline category logistic regression, adjusting for factors including age, sex, stage, immune status, surgical margins, and facility case volume). Definitive resection (DR, excluding excisional biopsy), radiation (RT), and chemotherapy (CT) were considered as time-varying predictors in the Cox regression. Multiple imputation was used to mitigate missing data bias. Statistical significance was defined as p < 0.05.

Results

34,503 patients with MCC were included (2,566 Black patients). Black patients were younger (median age 52, interquartile range (IQR) 37-65) compared to White patients (median age 72, IQR 59-81, p<.0001). Black patients had higher rates of immunosuppression (28%) compared to White patients (14%, p=.0062). Black patients were more likely to be diagnosed at higher stage (proportional OR=1.41, 95% CI 1.25 – 1.59), and with larger tumor size category (proportional OR 1.99; CI 1.77 – 2.24) relative to White patients. No significant differences were noted by race across receipt of DR, though Black patients did have longer time from diagnosis to DR in our cohort. Black patients were less likely to receive adjuvant RT or CT relative to White patients, adjusting for receipt of DR, stage, tumor size, tumor subsite, margin status and other relevant factors. Regarding data reporting quality (adjusting for facility characteristics and treatment received), Black patients were more likely to have missing cancer stage (OR=1.69, CI 1.52-1.88) or missing nodal information (OR=2.10, CI 1.84-2.40). Black patients had decreased adjusted risk of all-cause mortality in our cohort (HR 0.73, 0.65 – 0.81).

Conclusion

Important baseline patient, tumor, and treatment related differences by race were demonstrated in our cohort. A significant difference in quality of data reporting by race was determined in a large national cancer registry. Given the importance of registry analyses for rare cancers, continued efforts are needed to ensure complete data coding for all patients. Continued efforts by clinicians to recognize MCC as a cancer which can occur in Black patients, and to increase enrollment of underserved populations on MCC clinical trials are paramount to ensuring patients with MCC continue to receive optimal care.