Neoadjuvant chemotherapy (NAC) is considered for patients with cN0 triple-negative (TN) or HER2-positive breast cancer ≥1 cm. If the mammogram shows primarily calcifications spanning ≥1 cm, the appropriateness of NAC is unclear due to uncertainty regarding invasive tumor size. We hypothesized that these patients largely do not have invasion ≥1 cm and should undergo upfront surgery. Women with TN/HER2-positive invasive breast cancer undergoing upfront surgery with a mammogram showing calcifications ≥1 cm were identified in our database. Patients with mass ≥1 cm on ultrasound were excluded. Imaging findings were compared with pathologic tumor size. Of 149 patients, 83

BACKGROUND:Despite the paucity of outcome data, axillary lymph node dissection (ALND) is increasingly being omitted in patients with positive sentinel lymph nodes after neoadjuvant chemotherapy, particularly in those with low-volume residual disease. We investigated oncological outcomes in patients with breast cancer and residual micrometastases in the sentinel lymph nodes treated with or without ALND. METHODS:OPBC-07/microNAC was a retrospective cohort study, using data obtained from the institutional databases of 84 cancer centres in 30 countries. Patients aged 18 years or older with clinical T1-4, N0-3 breast cancer at diagnosis treated with neoadjuvant chemotherapy followed by surgery between Jan 1, 2013, and May 31, 2023, who were found to have residual micrometastases (metastasis measuring >0·2 mm or >200 cells, not exceeding 2·0 mm in size) on frozen section or on final paraffin sections as determined by sentinel lymph node biopsy, targeted axillary dissection (sentinel lymph node biopsy with single or dual-tracer mapping plus image-guided localisation of the initially biopsy-proven and clipped node), or the marking axillary lymph nodes with radioactive iodine seeds (MARI) procedure were eligible for inclusion. The primary endpoint was the 5-year rate of any axillary recurrence (isolated or combined with local or distant recurrence) stratified by type of axillary surgery. Given the median follow-up, here we report 3-year rates and exploratory 5-year estimates. This study was registered with ClinicalTrials.gov, NCT06529302. FINDINGS:1585 female patients with ypN1mi disease were analysed, of whom 804 (50·7%) underwent ALND and 781 (49·3%) did not. Of 1585 women, 238 (15·0%) self-identified as Asian, 65 (4·1%) as Black, 200 (12·6%) as Hispanic, 968 (61·1%) as White, and 114 (7·2%) as unknown race and ethnicity. 925 (58·4%) of 1585 women had cT2 tumours, 1054 (66·5%) were node positive, and 1267 (79·9%) received nodal radiotherapy. The median follow-up was 3·1 years (IQR 1·8-5·2). The 3-year rate of any axillary recurrence (isolated or combined with local or distant recurrence) for the entire cohort was 2·0% (95% CI 1·3-2·9), with no statistical difference identified by extent of axillary surgery. However, patients with triple-negative disease who did not receive ALND had significantly higher rates of any axillary recurrence than women treated with ALND (8·7% [95% CI 4·4-15·0] vs 2·4% [95% CI 0·7-6·5], p=0·018). On multivariable analysis, triple-negative breast cancer (hazard ratio 3·83 [95% CI 1·72-8·52]) and omission of nodal radiotherapy (2·62 [1·19-5·73]) but not omission of ALND (0·86 [0·37-2·00]) were independently associated with an increased risk of any axillary recurrence. INTERPRETATION:Overall, these results do not support ALND for all patients with ypN1mi on sentinel lymph node biopsy treated with nodal radiotherapy; however, tumour biology should be taken into account when considering ALND omission. FUNDING:US National Institutes of Health, National Cancer Institute.

e12746 Background: Adenoid cystic carcinoma (AdCC) of the breast is a rare ( < 0.1%) form of breast cancer with two variants, solid basaloid (SB-AdCC) and classic (C-AdCC). Due to its rarity, limited outcome data exists. We aimed to describe clinicopathological features and long-term outcomes in patients with SB-AdCC and C-AdCC. Methods: Patients diagnosed with stage I-III AdCC between 2007-2024 were retrospectively identified from an institutional database. Patient characteristics were compared between groups using Fisher’s exact test and Wilcoxon rank sum test. Recurrence free survival (RFS) was estimated using the Kaplan-Meier method and compared between groups using log-rank test. Results: 77 patients were included. The median age was 58 (IQR 48,69). Most patients self-identified as White (n = 64, 86%) and not Hispanic (n = 65, 89%). Most patients (n = 55, 71%) had SB-AdCC histology and 22 patients (29%) had C-AdCC. Median tumor size was 17 mm (IQR 13,26), with no significant size difference between SB-AdCC and C-AdCC (p = 0.5). Most patients (n =44, 57%) presented with clinical stage I disease and 34 (50%) patients had grade 3 disease, with a higher proportion of high-grade disease in patients with solid basaloid subtype (p < 0.001). Eight patients (10%) were clinically node-positive at presentation. Nearly all tumors were ER-/HER2- (88%, n = 68). Most patients underwent breast conserving therapy (70%, n = 52). Forty percent (n = 31) of patients received systemic treatment. Patients with Sb-AdCC were more likely to receive systemic therapy, 53% vs 9% (p<0.001). Among patients who received systemic therapy (n = 31), the majority (n = 19, 61%) received anthracycline–based chemotherapy, 9 (29%) received taxane-based therapy, and 3 (10%) received CMF. Few patients overall (n = 10, 13%) received immunotherapy. All patients who received neoadjuvant systemic therapy (n = 10, 14%) had Sb-AdCC histology, and none achieved pCR. Overall, 67% (n = 47) of patients received radiation, 42 (89%) after lumpectomy and 5 (11%) after mastectomy. Median RFS was 7.2 years; 18 years for C-AdCC and 5.6 years for Sb-AdCC (p = 0.014). There was no significant difference in median RFS by receipt of systemic therapy in patients with Sb-AdCC (p = 0.8). Conclusions: Despite the use of standard systemic therapy, patients with Sb-AdCC have much worse oncological outcomes compared to C-AdCC. Alternative systemic therapy options are urgently needed. Patient age and disease-related factors by AdCC variant. Overall N=771 1 Classic AdCC N=221 1 Solid Basaloid AdCC N=551 1 p-value 2 Age 58 (48,69) 60 (50,70) 56 (48,68) 0.4 Subtype ER-/HER2- 68 (88%) 17 (77%) 51 (93%) 0.11 ER+/HER2- 9 (12%) 5 (23%) 4 (7.3%) Tumor Size (mm) 17 (13, 26) 18 (14, 26) 16 (11, 25) 0.5 Grade 1 6 (8.8%) 5 (26%) 1 (2.0%) <0.001 2 28 (41%) 13 (68%) 15 (31%) 3 34 (50%) 1 (5.3) 33 (67%) 1 Median (IQR); n (%). 2 Wilcoxon rank sum test; Fisher’s exact test.

Objective: This study prospectively compared lymphedema rates with perometer, tape measure, and bioimpedance spectroscopy (L-Dex) in patients with breast cancer undergoing axillary lymph node dissection (ALND). Background: Perometer, tape measure, and L-Dex are all used for diagnosing lymphedema, although lymphedema rates may vary between these techniques. Methods: From November 2016 to March 2020, patients who underwent ALND had arm measurements performed preoperatively and biannually with perometer, tape measure, and L-Dex. A lymphedema diagnosis was made if arm volume increased by ≥10% by perometer or tape measure, or if L-Dex increased ≥6.5 units from baseline. We assessed concordance between the different measurement techniques using scatter plots, Pearson correlations, and Cohen’s Kappa statistic. Results: We included 281 patients who underwent ALND and had at least one follow-up (median, 2.3 years; >1100 follow-up measurements). Two-year lymphedema rates were 21.9% with perometer, 34.6% with tape measure, and 67.5% with L-Dex. Evaluating all timepoints, we found moderate correlation between perometer and tape measure (R=0.67, P <0.001), perometer and L-Dex (R=0.66, P <0.001) and tape measure and L-Dex (R=0.57, P <0.001). Diagnostic agreement was weak between perometer and L-Dex (k=0.29, P <0.001), and tape measure and L-Dex (k=0.28, P <0.001); and moderate between perometer and tape measure (k=0.51, P <0.001). Conclusion: In our prospective ALND cohort, perometer, tape measure, and L-Dex demonstrated moderate correlation between measurements, and weak-to-moderate diagnostic agreement. Lymphedema rates varied between measurement techniques, with the highest rates observed with L-Dex, raising concern for overdiagnosis.

Nipple-sparing mastectomy (NSM) confers similarly excellent patient-reported outcomes (PROs) as skin-sparing mastectomy; however, the relative weight of individual patient- and treatment-related factors on satisfaction with breasts and psychosocial well-being following NSM is unclear. We assessed predictors of PROs following NSM. Patients undergoing NSM between April 2018 and July 2021 at a single institution were included in a prospective examination. Routinely collected preoperative and postoperative BREAST-Q responses were recorded. Univariable and multivariable linear regression identified predictors of satisfaction with breasts and psychosocial well-being. 333 patients underwent NSM; median age was 43 years (interquartile range [IQR] 37–49 years). Of patients, 86

Electronic patient-reported outcomes (ePROs) are used postoperatively to detect complications through real-time symptom monitoring. This study examines whether alerts triggered through the “Recovery Tracker” (RT), an ePRO system, predict 30-day re-admission or re-operation after lumpectomy. We retrospectively reviewed breast cancer patients who underwent lumpectomy at a single institution between August 2018 and May 2024. Patients who completed RT surveys on postoperative days 1–5 were included. Symptom alerts categorized as red (urgent) and yellow (less urgent) were analyzed using generalized additive and univariable logistic regression models. Among 8723 included patients, 2552 (29

BACKGROUND:Same-day discharge after mastectomy (SDM) is safe for patients meeting narrow eligibility criteria. Our institution embarked on a feasibility study for an SDM pathway with expanded eligibility criteria. STUDY DESIGN:An SDM pathway was created to ensure proper education and expectations. Data were prospectively collected for all patients who underwent mastectomy from February 2022 to May 2023. Demographics, treatments, and complications were compared for patients who underwent SDM with those who stayed overnight. Patient-reported outcomes were collected via Recovery Tracker, an electronic platform that allows for daily postoperative patient inputs. RESULTS:Of 1,664 patients, 1,580 patients (95%) stayed overnight; 84 patients (5%) elected for SDM, including 59 initially planned for SDM and 30 who converted from overnight stays to SDM. Five planned for SDM opted to stay overnight. Compared with overnight-stay patients, SDM patients were significantly older (61 vs 51 years, p < 0.001), more likely to undergo unilateral mastectomy (70% vs 46%, p < 0.001), and less likely to undergo reconstruction (32% vs 74%, p < 0.001). SDM patients lived significantly farther away from the hospital (34 vs 28 miles, p = 0.027). There was no significant difference in the number of complications experienced within 2 days after surgery; SDM patients experienced more complications within 30 days. There were no emergency room visits, transfers, or admissions. According to the Recovery Tracker, 3 SDM patients requested a nurse telephone call. CONCLUSIONS:SDM is safe even with expanded criteria for patient eligibility. Selection or inclusion should not be limited by age, surgery, or distance from the hospital. Instead, patient motivation and clear expectation setting by clinical teams are relevant considerations.

577 Background: African compared to European genetic ancestry is associated with lower prevalence of clinically actionable alterations despite comparable overall prevalence of driver alterations by ancestry group. This likely contributes to breast cancer (BC) disparity. A limitation of prior ancestry studies is lack of neighborhood-level disadvantage (ND) data, a contributor to BC disparity. The objective of this study was to investigate associations between ND and somatic mutations in a real-world BC cohort. Methods: Somatic mutations from primary and metastatic BC samples and genetic ancestry data were identified from MSK IMPACT (FDA-authorized sequencing panel). Samples sequenced between 2015-2025 were annotated and linked with census-tract level socioeconomic status (Yost Index 1-100; higher reflects increasing ND). Multivariate logistic regression analyzed associations between ND and somatic mutations, adjusting for covariates (FDR p<0.05). Results: 6703 samples (62% primary, 38% metastatic) were analyzed. Mean age was 60 (SD 13). Mean Yost was 26 (SD 24). 67% had ER+/HER2- disease, 9.7% had ER+/HER2+, 5.1% had ER-/HER2+, and 19% had ER-/HER2-. Adjusting for age, BMI, ancestry, and subtype, patients from ND had lower odds of CDH1 (OR 0.99, 95% CI 0.98- 0.99, p<0.001), BRCA1 (OR 0.96, 95% CI 0.94-0.99, p=0.021), and AKT1 (OR 0.98, 95% CI 0.97- 0.99, p=0.006) somatic mutations in metastatic samples. Conclusions: We identify novel associations between ND and somatic mutations. Notably, patients with ND had lower odds of actionable BRCA1 and AKT1 mutations. This highlights the importance of ensuring patients of diverse ancestry and ND have equitable inclusion in clinical sequencing workflows to improve identification of actionable mutations in all populations. Genetic ancestry (GA) and somatic mutational frequency by Yost. TotalN=6703 1 Yost 1-20(Most Advantaged)N=3601 1 Yost 21-40N=1539 1 Yost 41-60N=826 1 Yost 61-80N=477 1 Yost 81-100(Most Disadvantaged) N=260 1 p-value 2 European GA* 0.75(0.36) 0.83(0.30) 0.72(0.38) 0.64(0.39) 0.58(0.39) 0.42(0.35) <0.001 African GA* 0.12(0.27) 0.04(0.16) 0.15(0.30) 0.21(0.34) 0.25(0.36) 0.44(0.38) <0.001 East Asian GA* 0.05(0.21) 0.05(0.21) 0.06(0.22) 0.06(0.21) 0.07(0.23) 0.04(0.17) <0.001 South Asian GA* 0.06(0.16) 0.06(0.17) 0.05(0.15) 0.06(0.17) 0.04(0.12) 0.03(0.08) <0.001 Native American GA* 0.03(0.10) 0.01(0.07) 0.03(0.10) 0.04(0.12) 0.06(0.16) 0.08(0.17) <0.001 Mutational Frequency, metastases N=2536 1 N=1403 1 N=584 1 N=311 1 N=165 1 N=73 1 p-value 2 ^ CDH1 341(13%) 220(16%) 66(11%) 26(8.4%) 18(11%) 11(15%) <0.001 AKT1 127(5.0%) 74(5.3%) 34(5.8%) 12(3.9%) 4(2.4%) 3(4.1%) 0.006 BRCA1 21(0.8%) 13(0.9%) 7(1.2%) 1(0.3%) 0(0%) 0(0%) 0.021 *Calculated using >3,000 common SNP markers. ^MVA, control for age, BMI, ancestry, subtype. 1 Mean (SD), n (%). 2 Pearson’s Chi 2 test; Kruskal-Wallis rank sum test; Fisher’s exact test.

PURPOSE:Tumor size is an established and independent risk factor for locoregional recurrence (LRR) and distant metastasis. More recently, the recurrence score (RS) calculated from a 21-gene expression assay (Oncotype DX, Exact Sciences) has been shown to correlate with LRR and distant metastasis. This study sought to evaluate the impact of RS on recurrence risks for small tumors (≤1 cm) with RS ≥26 to similar-sized tumors with RS <26. METHODS AND MATERIALS:Between 2008 and 2020, 231 patients with breast cancers measuring ≤1 cm and high RS of ≥26 were retrospectively identified. These were compared with a control cohort of 1747 breast cancers that were ≤1 cm with RS <26 over the same time interval. Rates of LRR and invasive recurrence (IR)-free interval were estimated using the Kaplan-Meier method, and multivariable Cox regression was conducted. RESULTS:Patients with RS ≥26 had significantly worse time to LRR (P < .001) and IR than patients with RS <26 (P < .001). The LRR-free probabilities for RS ≥26 and RS<26 patients are 94% versus 99% at 5 years and 88% versus 96% at 10 years. Similarly, the IR-free probabilities were 94% versus 99% at 5 years and 87% versus 95% at 10 years. In multivariable analyses, RS was independently associated with LRR and IR. CONCLUSIONS:Our findings suggest that patients with small tumors and high RS are at a higher risk for LRR than patients with ≤1 cm breast cancers with low RS, despite the best available standard-of-care treatment. These findings may have important implications for the tailoring of locoregional treatment strategies because they relate to omission of radiation therapy and selection of radiation therapy modality.

OBJECTIVE:To evaluate the association between living in disadvantaged neighborhoods in New York City (NYC) with tumor grade, a clinical proxy for proliferation and tumor aggressiveness, and breast cancer-specific survival (BCSS). BACKGROUND:Neighborhood disadvantage (ND) is associated with shorter BCSS, independent of individual-level, tumor, and treatment characteristics, highlighting unmeasured factors associated with this survival disparity. METHODS:Women with stage I to III breast cancer (BCa) living in NYC treated at Memorial Sloan Kettering Cancer Center from 2013 to 2024 were included. ND was stratified using the Area Deprivation Index (ADI). The median ADI for the cohort was 3 and was used as the cutoff between neighborhood advantage (NA, ADI 1-3) and ND (ADI 4-10). Multivariable logistic regression and Cox proportional hazards modeling, controlling for individual, tumor, and treatment factors, were used to determine the association between ND and tumor grade and BCSS, respectively. RESULTS:Five thousand four hundred fifty-two women with BCa were included. Three thousand four hundred seventy-nine (64%) lived in NA and 1973 (36%) in ND. On multivariable analysis, ND had higher odds of poorly versus well/moderately differentiated tumors (adjusted odds ratio: 1.23, CI: 1.03-1.48), independent of age, race/ethnicity, insurance, body mass index, smoking/alcohol, stage, and subtype. ND was also associated with shorter BCSS (adjusted hazard ratio: 1.56, CI: 1.05-2.38). CONCLUSIONS:Women living in ND in NYC were more likely to present with poorly differentiated tumors and have shorter BCSS. These findings merit further inquiry and lay the foundation for future translational studies to externally validate the mechanisms by which ND "gets under the skin" to impact aggressive BCa tumor biology, and ultimately survival.

Background: Adding pembrolizumab to neoadjuvant chemotherapy (NAC+P) for stage II-III triple-negative breast cancer (TNBC) improves pathologic complete response (pCR) rates and event-free survival compared to neoadjuvant chemotherapy (NAC) alone. Recurrence patterns and factors associated with disease-free survival (DFS) in a real-world cohort are not known. We compared recurrence rates and assessed factors associated with DFS in patients receiving NAC+P versus NAC for early-stage TNBC. Methods: This retrospective analysis compared two consecutive cohorts of patients with stage II-III TNBC: 344 treated with NAC+P from 6/1/21-4/30/24 based on the KEYNOTE-522 regimen and 513 treated with NAC from 7/1/09-3/15/19. Demographic, clinicopathologic characteristics, and recurrence data were collected. Tumor infiltrating lymphocytes (TILs) were scored on pre-treatment core needle biopsy for NAC+P only. Primary outcomes were local recurrence (LR), locoregional recurrence (LRR), and distant recurrence (DR). Secondary outcomes were DFS and factors associated with DFS. Kaplan-Meier time-to-event analyses assessed for DFS. Univariate and multivariable Cox regression models assessed for factors associated with DFS. Results: In the NAC+P group, most patients presented with cT2 (70%), cN0 (53%) disease at a median age of 52 years (IQR 41, 62). In the NAC group, most presented with cT2 (48%) or cT3/4 (28%), cN1 (52%) disease at a median age of 51 years (IQR 41, 60). Rates of breast, nodal, and overall pCR (58% vs 36%, p<0.001) were higher with NAC+P. At 1-year, crude LR rates were not different between NAC+P and NAC: 2.3% vs 4.3%, p=0.13. However, 1-year LRR and DR rates were lower with NAC+P: LRR (0.6% vs 4%, p=0.01) and DR (8% vs 12%, p=0.03). Thirty-one NAC+P patients had DR. The most common sites of first progression were liver (29%), lung (29%), or brain (26%). Among 121 NAC patients with DR, the most common sites of first progression were multiorgan (50%), lung (19%), or liver (12%). Of 61 patients with multiorgan DR, 36 (59%) had brain metastases.At a median follow-up of 1.2 years (IQR 0.8, 1.6), the 1-year disease recurrence rate for NAC+P was 10% (95% CI: 6%-13%). With a median follow-up of 6 years (IQR 3.3, 7.7), the 5-year disease recurrence rate for NAC was 24% (95% CI: 20%-28%). Factors associated with worse DFS for NAC+P included extreme breast density (HR 3.78, p=0.004) and cN1 disease (HR 2.68, p=0.04). On univariate analysis only, cN2/3 disease and comprehensive RT were associated with worse DFS, while lower cT stage, immune-related adverse events, breast pCR, nodal pCR, and overall pCR were associated with improved DFS. Using a 10% cutoff, stromal TILs on pre-treatment core needle biopsy were not associated with DFS. Factors associated with improved DFS for NAC were cT2 vs cT3/T4 (HR 0.56, p=0.01), ACT-based regimen (HR 0.61, p=0.02), and overall pCR (HR 0.32, p<0.001), while lymphovascular invasion on biopsy (HR 2.1, p<0.001) was associated with worse DFS. Univariate associations for worse DFS included cN1 and cN2/3 disease. Conclusions: Compared to NAC, NAC+P resulted in improved crude 1-year rates of LRR and DR, along with numerical improvements in DFS with shorter follow-up. Factors associated with DFS (including pCR) were not similar across cohorts and isolated brain metastases as a site of first progression were more common in patients treated with NAC+P who developed DR. Citation Format: Lauren Perry, Varadan Sevilimedu, Natalia Polidorio, Marisa Lazarus, Giacomo Montagna, Nour Abuhadra, Ellen Yang, Rachel Han, Hannah Y. Wen, Lior Braunstein, Monica Morrow, George Plitas, Stephanie Downs-Canner. Patterns and Predictors of Recurrence After Neoadjuvant Chemo-Immunotherapy in Early-Stage Triple-Negative Breast Cancer [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2024; 2024 Dec 10-13; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(12 Suppl):Abstract nr P5-09-17.

Introduction: Neoadjuvant chemo-immunotherapy with pembrolizumab (NAC+P) has become the standard care for stage II-III triple-negative breast cancer (TNBC) following the KEYNOTE-522 (KN522) trial, which showed improved pathologic complete response (pCR) rates and event-free survival compared to chemotherapy alone (NAC). We aimed to describe pCR rates by age and treatment in TNBC patients receiving either NAC+P or NAC alone at our institution. Methods: Women aged ≥65 years with stage I-III TNBC treated with NAC+P from 6/2021 to 9/2023 were compared to patients aged <65 years treated with NAC+P and to patients ≥65 who received NAC alone from 03/2010 to 9/2019. Patients with metaplastic breast cancer and neoadjuvant radiotherapy were excluded. Clinicopathological characteristics and pCR rates (ypT0/is pN0) - were compared between groups using Chi-square tests. Logistic regression for univariate (UVA) and multivariable analysis (MVA) was used to evaluate the association between age, treatment, and pCR. Results: 481 patients were included: 69 NAC+P ≥65 years (median age 69, IQR: 67-73); 338 NAC+P <65 years (median age 48, IQR: 38-56); and 74 NAC ≥65 years (median age 69, IQR: 66-73). Compared to younger patients, older patients treated with NAC+P were more likely to have public insurance (87% vs 13%, p<.001), lobular and other non-ductal histologies (6% vs 1.2% and 10% vs 5%, respectively, p= .009), cT1 and cT4 stages (22% vs 11% and 13% vs 3.9%, respectively, p< .001) and cN+ disease at presentation (58% vs 45%, p=.04). In both age groups, 81% received the KN522 backbone regimen with carboplatin, and the remaining mostly received NAC with cyclophosphamide, anthracycline, and taxane. There was no difference in treatment completion rate - defined as the completion of all preplanned neoadjuvant therapy cycles regardless of interruptions, regimen changes, or dose reductions - and in immune-related adverse events (33%) between age groups. However, older patients had a longer time from NAC+P to surgery (38 vs 32 days, p= .002). Breast surgery was similar between age groups, but axillary dissection (ALND) was more common in older patients (32% vs 19%, p= .023). The pCR rates did not differ between older and younger patients (49% vs 58%, p=.2). On MVA, high tumor grade (OR 13.9, 95% CI 4.54-61.7, p< .001) was positively associated with pCR, whereas other histologies (OR 0.13, 95% CI 0.04-0.39, p< .001), cT4 stage (OR 0.10, 95% CI 0.02-0.36, p<.001) and cN+(OR 0.60, 95% CI 0.38-0.95, p= .029) were less likely to facilitate pCR.In patients ≥65 years, those receiving NAC+P, compared to NAC alone, more frequently had lobular and other histologies (6% vs 0% and 10% vs 4.1%, respectively, p= .028), intermediate tumor grade (15% vs 1.4%, p=0.003), and cT1 and cT2 stages (22% vs 16% and 61% vs 46%, respectively, p= .045).NAC+P patients more frequently received carboplatin (81% vs 6.8%, p<.001). There was no difference in NAC completion and time to treatment start, but the NAC+P group had longer time to surgery (38 vs. 30 days, p = 0.039). More NAC patients underwent ALND (56% vs 32%, p= .005), with no difference in cN+ disease and breast surgery. The pCR rates were significantly higher in the NAC+P group (49%) compared to NAC (23%, p=.001). On UVA, NAC+P significantly improved overall pCR (OR 3.26 p=.001). However, on MVA, after adjusting for tumor histology, grade, stage, neoadjuvant carboplatin, and completion rate, the use of immunotherapy was not independently associated with pCR. Conclusions: Within the NAC+P cohort, pCR rates did not differ by age, and no significant increase in irAEs was seen in elderly patients, though we acknowledge a possible selection bias toward healthier elderly patients. Compared to NAC only, NAC+P showed potential for improving pCR in older individuals and could be a viable treatment option for elderly TNBC patients who are likely to tolerate it. Citation Format: Natalia Polidorio, Lauren M. Perry, Srinivasa Sevilimedu Veeravalli, Giacomo Montagna, Nour Abuhadra, Diana Lake, Mark Robson, Monica Morrow, Stephanie Downs-Canner, Iris Zhi. Outcomes in Elderly Patients with Triple-Negative Breast Cancer Receiving Neoadjuvant Chemo-immunotherapy and Chemotherapy Alone [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2024; 2024 Dec 10-13; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2025;31(12 Suppl):Abstract nr P5-11-16.

Randomized trials established the safety of postmastectomy radiation/regional nodal irradiation (PMRT) as an alternative to axillary lymph node dissection (ALND) for cN0 patients undergoing upfront mastectomy with one or two positive sentinel lymph nodes (+SLNs). In January 2022, the authors adopted a policy omitting routine frozen section for cT1-3N0 patients undergoing upfront mastectomy, and this study sought to examine the impact of this policy on axillary treatment. Consecutive patients with cT1-3N0 breast cancer who underwent upfront mastectomy from January 2022 to July 2023 were identified. For patients with three or more +SLNs and those with one or two +SLNs not meeting institutional PMRT criteria, ALND was indicated. This study evaluated ALND/PMRT rates after adoption of this policy. In this study, 623 patients with cT1-3N0 breast cancer had upfront mastectomy with sentinel lymph node biopsy. Overall, 4.7

557 Background: Second opinions and multidisciplinary reviews (MDR) are known to alter the management of cancer patients in a substantial percentage of cases, yet these resources are not universally available. We previously presented more than 400 anonymized complex cancer scenarios (cases) to MDR by experts in five major tumor types by video conference (2020-2021, www.precisca.com) and tracked treatment recommendations. Here, we compare MDR panel recommendations to those generated by 3 different foundational AI models. Methods: 50 complex cancer scenarios in breast and lung cancer that were previously adjudicated by MDR were analyzed by three different foundational AI models (OpenAI’s ChatGPT 4.5, Anthropic’s Claude Opus 4 and Google’s Gemini Ultra) using PrecisCa’s proprietary method for prompts. The AI recommendations from each system were scored on a scale of 1-5 (5 being the highest) for completeness, reasoning, clarity, menu of options, recency, and relevance as compared to the MDR panel recommendations. The maximum possible competence score was 30 points per scenario and an aggregate score of 750 for each tumor type. Final AI recommendations were also compared with current National Comprehensive Cancer Network (NCCN) guidelines for serious omissions. Comparison in reverse (additional AI options that the experts missed) was not performed due to interval changes in treatment recommendations over the past 4 years. Results: Patient characteristics and aggregate concordance and competence scores are listed in the table. While there was some variation in the concordance and competence rates of the three AI models, there was excellent concordance with the expert opinion for all of them. Discordant cases were reviewed and primarily involved minor differences that would not have altered the patients’ management significantly. Conclusions: This study demonstrates a high degree of concordance between three leading AI models and expert panel decisions for common yet complex breast and lung cancer clinical scenarios. These findings indicate that it is not premature to incorporate AI as a decision support tool in daily practice with human experts’ review. Updated data with additional cases and tumor types will be presented. Patient characteristics and aggregate competence score (n=50). Characteristic Breast Cancer (n=25) Lung Cancer (n=25) Median Age (Range) 55 (27-83) 70 (53-87) Stage I 5 (20%) 3 (12%) II 6 (24%) 1 (4%) III 2 (8%) 5 (20%) IV 12 (48%) 16 (64%) Histology Ductal Carcinoma 24 (96%) N/A Lobular Carcinoma 1 (4%) N/A Non-Small Cell N/A 22 (88%) Small Cell N/A 3 (12%) Aggregate / Median Competence Score (Range) OpenAI 4.5 649 / 25 (23-27) 626 / 25 (21-29) Claude Opus 4 708 / 27 (24-30) 694 / 27 (23-30) Gemini Ultra 720 / 28 (26-30) 727 / 28 (26-30)