Abstract Background. Advancements in cancer medicine have resulted in prolonged survival for the majority of patients. Personalized care pathways which account for clinical needs of this growing population of cancer survivors—and which address oncology workforce shortages—are urgently needed. Rapid identification of low-risk cancer survivors who could be cared for in other settings is a critical element of personalized care. Here we present final algorithm performance on risk-stratification from an ongoing clinical trial evaluating a primary care physician-led cancer survivorship clinic for breast cancer survivors. We captured data on real-time clinical risk stratification of early-stage patients between 6-36 months post-treatment. Methods. With oncology stakeholder input, we developed a screening algorithm to identify low-risk breast cancer survivors from the electronic medical record based on data from pathology, treatment, and utilization records. The algorithm identified patients meeting study eligibility: adult stage 0-IIb breast cancer patients diagnosed with first primary cancer, excluding those with treatments indicative of high-risk or metastatic disease, ongoing ovarian suppression, neoadjuvant therapy, or enrollment in other cancer clinical trials. Next, the treating oncologist was asked to confirm or deny patient eligibility; if denied, we tracked and categorized the reason provided for ineligibility. We describe: 1) characteristics and proportions of patients identified by the algorithm; 2) a breakdown of patients confirmed or denied eligibility; and 3) oncologists’ reasons for ineligibility. Results. The algorithm identified 1186 patients. Of those, 91 were flagged by oncology as not followed (e.g., consult only), and 204 were not categorized. Of the remaining 890 patients, 716 (81%) were categorized as eligible. Mean age was 62 years (SD: 11.4, Table 1). There were significant differences by stage, with a higher proportion of stage 0 in the eligible group (24% vs. 18%) and a lower proportion of stage II (24% vs. 35%). There was also a significant difference by HER2 status, with a greater proportion of HER2+ patients categorized as eligible (21% vs. 16%). There were significant differences by race/ethnicity, with a higher proportion of Asian and Black patients categorized eligible (14% vs. 8%, and 28% vs. 17%, respectively) and fewer Hispanic and White patients categorized eligible (26% vs. 36%, and 30% vs. 37%, respectively). There were not significant differences by ER/PR, surgery, or endocrine therapy. Reasons for ineligibility included: suspicious for recurrence (6%), new primary disease (3%), considered high-risk based on genetic/tumor factors (9%), co-occurring heme/oncology disorder (9%), treatment-related concerns (e.g., patient declined chemotherapy) (7%), complex case/condition (11%), patient preference per conversation with oncologist (32%), other (19%), and lost to follow-up (4%). Conclusions. Leveraging technology to support survivorship in the primary care setting is critical. Our findings demonstrate that consensus-based risk algorithms with oncologist review can effectively perform identification of low-risk patients who may be appropriately transitioned to other settings for survivorship care. These patients may benefit from primary care-led survivorship with a focus on healthy lifestyle, managing comorbid conditions, and psychosocial care. Eligible/Ineligible per oncologist by patient demographics and cancer characteristics Citation Format: Farah Brasfield, Aiyu Chen, Eric Haupt, Lindsay Lyons, Talar Habeshian, Corrine Munoz-Plaza, Ernest Shen, Huong Nguyen, Michael Gould, Alexander Ferreira, Allen Joo, Seyed Monemian, Christy Lai, Yung-Mee Park, Hollis Lee, Steven Steinberg, Patricia Ganz, Erin Hahn. Technology-enabled identification of low-risk breast cancer survivors for personalized survivorship care [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO2-03-02.
Read full abstract