Racial disparities in non-recommendation of adjuvant chemotherapy in stage II-III ovarian cancer

Gynecol Oncol. 2021 Nov 13:S0090-8258(21)01553-5. doi: 10.1016/j.ygyno.2021.10.090. Online ahead of print.


OBJECTIVES: To identify patient factors associated with not receiving a recommendation for adjuvant chemotherapy after primary surgery for ovarian cancer.

METHODS: This retrospective cohort study used the National Cancer Database (NCDB) data from 2004 to 2015 to identify patients with stage II-III ovarian cancer who underwent primary surgery. Multivariate logistic regression analyses evaluated factors associated with notation in the NCDB that “chemotherapy was not recommended/administered because it was contraindicated due to patient risk factors (i.e., comorbid conditions, advanced age).” Survival data were assessed via Kaplan-Meier analyses.

RESULTS: Of the 48,245 patients who met the inclusion criteria, 522 (1.08%) did not receive adjuvant chemotherapy because it was determined to be contraindicated. In multivariate analyses, independent predictors for not receiving a recommendation for adjuvant chemotherapy were age ≥ 70 years old (adjusted odds ratio, aOR = 2.43, p < 0.0001), non-zero Charlson-Deyo comorbidity scores (score 1, aOR = 1.41, p = 0.002; score ≥ 2, aOR = 2.57, p < 0.0001), and Black race (aOR = 2.12, p < 0.0001). For Black patients, recommendation against adjuvant chemotherapy occurred at a younger median age (64.5 years vs. 72 years) and was associated with lower 5-year survival (25.9% vs. 40.3%, p < 0.0001).

CONCLUSIONS: Patients with ovarian cancer who underwent surgery but did not receive chemotherapy “because it was contraindicated due to patient risk factors” were older and had higher comorbidity scores. Even after controlling for these differences, Black patients were disproportionately not recommended for chemotherapy, which was associated with worse survival. Determining eligibility for adjuvant chemotherapy requires an individualized approach, and the possible influence of racial bias on risk estimation should be further investigated.

PMID:34785030 | DOI:10.1016/j.ygyno.2021.10.090