Background: Guidelines recommend the use of genomic assays to aid decision making regarding the use of adjuvant chemotherapy (CT) for hormone receptor-positive, HER2-negative (HR+/HER2-) early breast cancer (EBC). Recently, the RSClin tool, which integrates the 21-gene recurrence score (RS) and clinicopathologic features, was developed to better estimate CT benefit. Here, we studied patients with RSClin/RS discordance to determine if patients (pts) should be treated according to RS or RSClin. Methods: This was a retrospective cohort study of pts from the National Cancer Database who were diagnosed with HR+/HER2- node negative EBC from 2010-2020 who underwent RS testing and received adjuvant endocrine therapy with or without CT. Pts were classified as low RS ( < 26) or high RS (≥ 26) and low ( < 3%) versus high (≥ 3%) predicted CT benefit per RSClin. The primary outcome was overall survival (OS) benefit associated with CT administration for pts with discordant RS/RSClin, measured as hazard ratio [HR] for receipt of CT. Hazard ratios were adjusted for age and comorbidity index, and inverse probability of treatment weighting was performed for age, comorbidity index, race, ethnicity, and insurance status. Results: 262,748 pts with EBC were included, with a median follow-up of 58 months. Median age was 60 years; 82% were White, 7% Black, 5% Hispanic, and 4% Asian/Pacific Islander. Median tumor size was 1.5 cm, and median RS was 15. The 31,955 cases with concordant high RS/RSClin had an OS benefit with CT (HR: 0.71, 95% CI: 0.67 - 0.75, p < 0.001), while the 206,631 cases with concordant low RS/RSClin did not have a survival benefit (HR: 1.00, 95% CI: 0.95 - 1.04, p = 0.85, Table). High RS with low RSClin was seen in 435 cases – all with grade 1 tumors, with RS no greater than 28, and tumor size under 1.3cm. No significant CT benefit was seen in this small subset of pts (HR 1.57, 95% CI 0.75 – 3.26, p = 0.23). In 23,727 pts with low RS but high RSClin, a CT benefit was seen (HR 0.73, 95% CI 0.67 – 0.78, p < 0.001). These pts had high-intermediate RS (median 23, interquartile range 20 – 24) with other higher risk features. Similar results were seen in older (age > 50) and younger (age ≤ 50) subgroups – although no pts ≤ 50 were in the low RS/high RSClin group. Conclusions: RSClin may identify a small number of grade 1 tumors with high RS for which CT is not needed, and may help guide appropriate use of CT for EBC in the sizeable group with intermediate RS.