Feasibility of a Novel Recurrence Algorithm to Improve NGS Testing in a Cohort of ER+/HER2- Metastatic Breast Cancer Patients in Kaiser Permanente Northern California

Molecularly targeted treatment of patients with recurrent ER+/HER2- metastatic breast cancer is an area of active investigation and clinical trials. One area for improvement in this domain is ensuring that all patients who have metastatic recurrence undergo NGS (Next-Generation Sequencing) testing. A 2-year learning collaborative with ASCO at Kaiser Permanente Northern California (KPNC) aims to develop a novel algorithm to detect breast cancer recurrence on a contemporary cohort of patients from 21 cancer centers throughout KPNC.

We identified 23,291 patients with early stage (I-III) ER+/HER2- breast cancer between 1/1/1996 and 12/31/2024 and alive in 2025. Recurrence was identified using large language models (LLM) based on pathology reports and diagnoses assigned by oncologists. Data on orders for tissue or liquid NGS testing were obtained.

Of the 23,291 eligible patients, 305 met additional criteria for having metastatic or local recurrence of prior breast cancer and 199 (65.3%) had NGS/tempus ordered. The median age at recurrence was 67 years (IQR: 57–78) and 54.8% of patients were White.

We demonstrated feasibility of an algorithm to identify ER+/HER2- patients with breast cancer who have recurred. These patients could potentially benefit from NGS testing to help select for molecularly targeted treatment and clinical trials. Our goal from a quality improvement standpoint is to understand barriers and facilitators to testing and bolster testing rates by 10% or higher.