Abstract
PURPOSE: To evaluate differences in clinical outcomes, treatments received, recurrence, and sociodemographic characteristics in patients with triple-negative breast cancer (TNBC) classified as invasive lobular carcinoma (TNBC-ILC) or invasive carcinoma of no special type (TNBC-NST). METHODS: Using national registry data, we conducted a retrospective, population-based cohort study of 6449 women diagnosed with primary TNBC (stratified by histological subtype) in Sweden (2007-2021). Clinical and treatment data were analyzed using descriptive statistics, logistic regression, machine learning (Boruta/XGBoost), and Cox proportional hazards models adjusted for patient age, tumor size, grade, nodal status, comorbidities, and receipt of adjuvant chemotherapy (ACT). RESULTS: TNBC-ILC accounted for 2.7% of all TNBC cases and affected older patients (median age 70 vs 62 years). Compared to TNBC-NST, TNBC-ILC had lower Ki-67, fewer high-grade tumors, higher T stage, and greater socioeconomic vulnerability. Machine learning identified age and post-operative tumor size as key predictive features of TNBC-ILC. ACT was administered to 40% of TNBC-ILC versus 59% of TNBC-NST cases (P < 0.001), with a survival benefit observed only in TNBC-NST. TNBC-ILC patients aged 50-64 years were less likely to receive ACT. Despite lower proliferative activity, TNBC-ILC was associated with worse overall (OS; adj-HR 1.39, 95% CI 1.04-1.86) and disease-specific survival (DSS; adj-HR 1.98, 95% CI 1.41-2.79), particularly in patients ≥ 50 years of age. TNBC-ILC patients ≥ 75 years had the poorest 5-year survival (DSS 55%; OS 42%). CONCLUSIONS: TNBC-ILC is a distinct subgroup with older age, lower grade and Ki-67, undertreatment, and poorer survival, emphasizing the need for age- and subtype-specific treatment strategies.