Treatment patterns and oncologic outcomes in patients with triple-negative (TN) lobular breast cancer.

Abstract

605 Background: Triple-negative breast cancer (TNBC) with invasive lobular carcinoma (ILC) histology represents an exceedingly rare (<1% of breast cancers) and aggressive disease subtype with poorly described outcomes. In an effort to better understand treatment patterns and prognosis for these patients, we conducted an analysis using the National Cancer Database to assess treatments and outcomes in this high-risk, understudied population. Methods: We identified patients ≥18 years with stage I-III TNBC diagnosed between 2017-2022 who underwent mastectomy or lumpectomy with ILC or invasive ductal histology (IDC) (n=80,589). Of these patients, 69,940 were treated with chemo- or immunotherapy (CTx/IO). Subgroup analysis included patients treated with neoadjuvant chemotherapy (NACT) ± IO (n=47,822). Among this subgroup, trends in NACT±IO and pathologic complete response (pCR) rates by histology were assessed using the Cochran-Armitage test. Associations between histology and NACT use or pCR were evaluated with multivariable logistic regression. Impact of pCR on overall survival (OS) was examined via Kaplan-Meier estimates and multivariable Cox models with propensity score weighting and matching (1:1 by stage and NACT type). Results: Among 80,589 patients with TNBC (median age 58 years), 79,645 (98.8%) had IDC and 944 (1.2%) had ILC. Patients with ILC were older (66 vs 58 years) and less likely to receive CTx/IO compared to those with IDC (75.5% vs 86.9%). Among 69,940 patients treated with CTx/IO, those with ILC were older (median age 65 vs 56 years), had higher clinical tumor stage, and were less likely to receive NACT compared to those with IDC (61.9% vs 68.4%; p<0.001). Among 47,822 patients who received NACT ± IO, patients with ILC had significantly lower pCR rates (16.1% vs 40.2%; aOR=0.44, 95% CI 0.34-0.57). The addition of IO showed a trend toward improved pCR rates in ILC (aOR=1.69, 95%CI 0.95-3.01, p=0.07). In 29,330 matched patients, pCR was associated with improved 5-year OS (94% vs 78%, p<0.001). However, in 108 matched ILC patients, no survival difference was observed by pCR status (5-year OS: 66% with pCR vs 69% without pCR, p=0.54). In a multivariable Cox model with inverse probability weighting among patients who received NACT±IO, ILC was associated with increased mortality risk (aHR=1.34, 95%CI 1.05-1.70, p=0.02), after adjustment for pCR, IO, age, race, tumor size, and insurance status, among other factors. Conclusions: Patients with TN-ILC who received NACT had lower pCR rates and worse OS compared to those with TN-IDC. Although limited by small sample size, this is the largest analysis of its type in the current NACT and IO era. These data suggest that pCR in TN-ILC may lack the prognostic significance it has in TN-IDC. These findings highlight the need for larger studies and novel therapeutic approaches tailored to the distinct biology of this rare subtype.