Abstract P2-14-07: Omission of Sentinel Lymph Node Biopsy in Patients with Early Stage Breast Cancer: Looking Beyond the Choosing Wisely Guidelines for Age < 70

Abstract

Abstract Background: The Surgical Society of Oncology Choosing Wisely Campaign for breast cancer advocates against the routine use of sentinel lymph node biopsy (SLNB) for women ≥ 70 years with early stage estrogen positive (ER+), clinically node negative (cN0) disease, given the low likelihood of axillary involvement and axillary recurrence risk, absence of survival benefit and greater reliance on genomic testing for therapeutic decisions. We hypothesize that this practice may be extended to a younger cohort of patients. In this proof-of-concept feasibility study, we first sought to determine the incidence of node positive (N+) disease in our health system using natural language understanding (NLU) technology to extract relevant data from the electronic medical record (EMR). NLU of the clinical narrative has been proven to aid clinical decision support by extracting relevant information and can populate clinical databases to facilitate optimal population management strategies. The advantage of NLU over a cancer registry is the speed and efficiency of data extraction for a large number of patients in real time, plus the capture of data points not conventionally included in a registry. Methods: All patients with early stage ER+, cN0 breast cancer who had SLNB from January 2015-December 2017 were identified in an integrated academic health network comprised of 15 hospitals in Western Pennsylvania. Patitent clinical data were abstracted from the EMR using Realyze Intelligence™ NLU technology. The Realyze NLU pipeline uses a combination of machine learning algorithms and standard terminologies to create a breast cancer patient model that includes genomic, phenotypic, and clinical data. The pipeline gathers information from all data sources – structured and unstructured – and normalizes the information to create a complete model of patient clinical criteria. Realyze Information Models use clinical data formatting flexible enough to represent clinical disorders on a concept level as well as the encounter, patient, and population levels. A breast cancer model with focus on the lymph node identification, pathological as well as clinical tumor and node classification were developed and mapped to standard terminology. A Semantic Reasoning layer is provided by different mechanisms including a rule-based layer to render answers to the questions posed in this hypothesis. NLU performance was validated by manually verifying key clinical variables (i.e., clinical stage, pathologic stage, and nodal positivity) on a subset of patients. Statistical analysis to determine any difference in N+ rates by age was performed using Chi-square testing with significance set at p < 0.05. Results: We identified 602 pts with early stage ER+, cN0 breast cancer over this period who underwent SLNB. Average age was 59.6 years old. As a whole group, there was an increase in N+ rates as the stage increased (Table 1). When comparing incidence of N+ disease stratified by age (< 70 or >70), there was no difference in N+ rates across all stages. In addition, equally low rates of SLN positivity were seen for patients specifically with stage T1a and T1b disease. Conclusions: These data suggest that the Choosing Wisely recommendation to omit SLNB may be extended to a younger cohort of pts with ER+, cN0 disease, specifically those with stage T1a or T1b tumors. With low rates of N+ disease, and less reliance on axillary stage for treatment decision making, the harms of surgical axillary staging may outweigh the benefits. Future validation is needed with a larger sample size. Table 1 Citation Format: Neil Carleton, Gilan Saadawi, Steffi Oesterreich, ADRIAN V. LEE, Emilia Diego. Omission of Sentinel Lymph Node Biopsy in Patients with Early Stage Breast Cancer: Looking Beyond the Choosing Wisely Guidelines for Age < 70 [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P2-14-07.