REAL-WORLD TREATMENT PATTERNS AND CLINICAL OUTCOMES IN PREVIOUSLY UNTREATED LOCALLY ADVANCED OR METASTATIC TRIPLE-NEGATIVE BREAST CANCER, BY PD-L1 STATUS: A NORSTELLALINQ US LINKED CLAIMS-EHR STUDY
Author(s)
Becky Hollenberg, MPH1, Allison Perry, PhD1, Rahul Das, PhD1, Karan Bhanot, PhD1, ilan behm, MPH2.
1Norstella, New York, NY, USA, 2Norstella, Englewood, CO, USA.
1Norstella, New York, NY, USA, 2Norstella, Englewood, CO, USA.
OBJECTIVES: To characterize first-line treatment patterns and real-world clinical outcomes among adults with previously untreated locally advanced inoperable or metastatic triple-negative breast cancer (la/m TNBC) in the United States, stratified by PD-L1 status.
METHODS: Retrospective study using NorstellaLinQ’s US real-world linked electronic health record (EHR), open-claims, and closed-claims data (January 2018-December 2024). From 4,437,456 patients with breast cancer, adults with la/m TNBC initiating systemic therapy (index 2020-2023) were identified. PD-L1 status was derived from structured laboratory data and NLP-assisted abstraction of clinical notes. Treatment patterns by line of therapy (LOT), real-world time on treatment (rwTOT), time to next treatment (rwTTNT), and overall survival (rwOS) were assessed using Kaplan-Meier methods.
RESULTS: The final cohort comprised 724 patients (PD-L1 positive 398; negative 163; remainder unknown/undocumented); 99.6% were female, mean age 59 years, and 70% had ECOG 0-1. Common metastatic sites were lymph nodes (34%), bone (21%), lung (20%), and liver (16%). First-line treatment differed substantially by PD-L1 status: 61% of PD-L1-positive patients received chemotherapy plus immunotherapy, versus 74% of PD-L1-negative patients receiving chemotherapy alone. Overall, 59% initiated second-line and 34% third-line therapy. Antibody-drug conjugates were used predominantly after first line, accounting for 19% of second-line and 27% of third-line regimens. Median rwTOT remained under 3 months across all lines (2.6, 2.8, 2.9 months), and median rwTTNT was 6.5 and 7.2 months. Median rwOS was not reached during available follow-up (108 deaths).
CONCLUSIONS: In real-world la/m TNBC, first-line treatment varied by PD-L1 status, but time on treatment was short and many patients proceeded to subsequent lines, with antibody-drug conjugates increasingly used later. Despite biomarker-guided first-line selection, treatment duration remained consistently short across all lines, highlighting substantial unmet need. Linked claims and EHR data, supplemented with PD-L1 ascertainment from clinical notes, enabled characterization of biomarker-defined treatment patterns difficult to evaluate using structured data alone.
METHODS: Retrospective study using NorstellaLinQ’s US real-world linked electronic health record (EHR), open-claims, and closed-claims data (January 2018-December 2024). From 4,437,456 patients with breast cancer, adults with la/m TNBC initiating systemic therapy (index 2020-2023) were identified. PD-L1 status was derived from structured laboratory data and NLP-assisted abstraction of clinical notes. Treatment patterns by line of therapy (LOT), real-world time on treatment (rwTOT), time to next treatment (rwTTNT), and overall survival (rwOS) were assessed using Kaplan-Meier methods.
RESULTS: The final cohort comprised 724 patients (PD-L1 positive 398; negative 163; remainder unknown/undocumented); 99.6% were female, mean age 59 years, and 70% had ECOG 0-1. Common metastatic sites were lymph nodes (34%), bone (21%), lung (20%), and liver (16%). First-line treatment differed substantially by PD-L1 status: 61% of PD-L1-positive patients received chemotherapy plus immunotherapy, versus 74% of PD-L1-negative patients receiving chemotherapy alone. Overall, 59% initiated second-line and 34% third-line therapy. Antibody-drug conjugates were used predominantly after first line, accounting for 19% of second-line and 27% of third-line regimens. Median rwTOT remained under 3 months across all lines (2.6, 2.8, 2.9 months), and median rwTTNT was 6.5 and 7.2 months. Median rwOS was not reached during available follow-up (108 deaths).
CONCLUSIONS: In real-world la/m TNBC, first-line treatment varied by PD-L1 status, but time on treatment was short and many patients proceeded to subsequent lines, with antibody-drug conjugates increasingly used later. Despite biomarker-guided first-line selection, treatment duration remained consistently short across all lines, highlighting substantial unmet need. Linked claims and EHR data, supplemented with PD-L1 ascertainment from clinical notes, enabled characterization of biomarker-defined treatment patterns difficult to evaluate using structured data alone.
Conference/Value in Health Info
2026-11, ISPOR Europe 2026, Vienna, Austria
Value in Health, Volume 29, Issue 12S
Code
CO201
Topic
Clinical Outcomes
Topic Subcategory
Comparative Effectiveness or Efficacy, Relating Intermediate to Long-term Outcomes
Disease
Oncology, Personalized & Precision Medicine