The Impact of Physician Productivity Metrics and Patient Tumor Biology on Receipt of RT in Elderly Patients with ER+ T1N0 Breast Cancer in a Tertiary Cancer Center

Abstract

Among elderly patients with early-stage invasive ER+ breast cancer, multiple randomized clinical trials show adjuvant radiation (RT) after lumpectomy does not reduce rates of survival or mastectomy (CALGB 9343, PRIME II). However, in the US, rates of adjuvant RT in this population have decreased modestly, from 79% to 75% (Soulos, JCO 2012). To understand barriers and facilitators to de-implementation of RT in these patients, we hypothesized that the enactment of a “new start” productivity metric for radiation oncologists would increase the rate of RT receipt among patients who were seen in consult by a radiation oncologist and/or increase the use of partial breast irradiation (PBI). We reviewed patients at a tertiary care center over the age of 70 treated with lumpectomy for ER+ pT1N0 from 2015-2018. A productivity metric based on number of “new starts” for patients treated with RT was implemented in January 2017; patients were divided into pre- and post- metric implementation for primary analysis. Chi squared analyses were performed to evaluate patterns of RT receipt before and after the new start metric. Secondarily, chi squared analyses and binary multivariate logistic regression including age, HER2 status, multifocality, and LVI were performed on all patients referred for RT to investigate biologic predictors of receipt of RT. 501 patients met inclusion criteria, and 35% received adjuvant RT. 188 patients (37%) were not referred for RT consult and were excluded. Among 313 patients referred, the rate of RT was 55% prior to metric implementation versus 59% after (NS). There was no difference in RT use among high-volume or low-volume providers, or treatment at main campus versus the regional setting. There was no difference in use of PBI (20% vs 20%), hypofractionation (73% vs 71%), or conventional fractionation (1% vs 0%). Secondarily, univariate analysis of patients referred for RT trended toward LVI increasing rates of RT compared to no LVI (65% vs 53%, respectively, p=0.09); on multivariate analysis, only HER2 amplification was associated with receipt of RT (p=0.014). Review of physician notes indicated multifocality as a risk factor driving RT recommendation, but this was not consistent across providers. The implementation of productivity metrics did not significantly impact use of RT in a tertiary care center that is below the national average of RT use in elderly patients with early-stage ER+ breast cancer. Leadership priorities and centralized peer review may overshadow financial incentives in this context; these implementation strategies should be evaluated outside the tertiary setting. Secondarily, HER2 amplification, which was not available in the era of CALGB 9343, is infrequent (5%) but drives use of RT when present. With data suggesting favorable outcomes in early-stage ER+HER2+ patients, investigation of radiation de-escalation in this cohort is warranted.