Chemotherapy (CT) use by Oncotype DX (O-Dx) recurrence score (RS) in hormone receptor–positive, HER2-negative (HR+/HER2-) invasive lobular carcinoma (ILC): A contemporary National Cancer Database (NCDB) analysis.

Abstract

531 Background: ILC has a distinct biology and is considered less chemosensitive than invasive ductal carcinoma (IDC). O-Dx RS can guide adjuvant CT decisions for patients with HR+/HER2- breast cancer (BC); however, pivotal trials largely included patients with IDC. Patients with high RS have been shown to benefit from CT. Whether this genomic risk threshold is applied to patients with ILC in real-world practice is not well understood. We evaluated contemporary CT use by O-Dx RS and its association with overall survival (OS) in patients with ILC. Methods: Patients ≥18 years with HR+/HER2- ILC diagnosed from 2018-2022 who underwent surgery, had pT1-T3, pN0-N1 disease, and known O-Dx RS were identified in the NCDB. Baseline characteristics were compared by RS groups: 0-10 (low), 11-25 (intermediate), and >25 (high). Multivariable logistic regression models identified factors associated with CT use in the overall cohort and stratified by RS. We evaluated the association between CT use and OS within each RS group, adjusting for covariates using propensity score matching (by year of diagnosis, age, pT, pN). Results: A total of 30,393 patients with early-stage HR+/HER2- ILC and O-Dx score were identified. Among them, 21% had low, 72% intermediate, and 7% high RS. Overall, 9.2% received CT, including 2.4% of those with low RS, 5.5% of intermediate RS, and 66% of high RS. CT use did not change over time (p=0.47). On multivariable analysis in all patients, higher RS was the key determinant of CT use (RS 11-25 vs 0-10: aOR=2.98, 95%CI 2.49-3.58; RS >25 vs 0-10: aOR=333.4, 95%CI 266.8-416.7), while older age was associated with lower CT use (aOR=0.91, 95%CI 0.91-0.92). In multivariable models stratified by RS, clinicopathologic features, including younger age, larger tumor size, higher nodal stage, and higher grade, were associated with CT use in low and intermediate RS groups (i.e., pN1 aOR=18.64 for RS 0–10; aOR=7.37 for RS 11–25; both p<0.001), whereas in high RS group, age had similar effect; however, tumor size and nodal status had significant but attenuated effects. In propensity-matched cohorts within RS groups, CT use was not associated with different OS in low RS (5-year OS 98% vs 100%, p=0.29) or intermediate RS (5-year OS 97% vs 97%, p=0.68) groups; however, in high RS group, CT use was associated with improved OS (5-year OS 96% vs 90%; p=0.01). Conclusions: In clinical practice, CT use in patients with ILC is guided by O-Dx RS but influenced by age and clinicopathologic risk. CT use conferred no OS benefit in patients with low or intermediate RS but was associated with OS benefit in high RS. Despite this, one-third of patients with ILC with high RS did not receive CT. These findings support RS-guided adjuvant CT decision-making in ILC and suggest that increasing CT use in patients with high RS may improve outcomes.