Browse Articles

Fixation-related potentials reveal that confusing program code elicits a late frontal positivity

Abstract As software pervades more and more areas of our professional and personal lives, there is an ever-increasing need to maintain software and for programmers to efficiently write and understand program code. In the first study of its kind, we analyze fixation-related potentials (FRPs) to explore the online processing of program code patterns that are confusing to programmers, but not to the computer (so-called atoms of confusion ), and their underlying neurocognitive mechanisms in an ecologically valid setting. Relative to clean counterparts in program code without an atom of confusion, confusing code elicits a late frontal positivity of about 400 to 700 ms after first looking at the atom of confusion. This frontal positivity resembles an event-related potential (ERP) component found during natural language processing that is elicited by unexpected but plausible words in sentence context. Thus, we suggest that the brain engages similar neurocognitive mechanisms in response to unexpected and informative inputs in program code and in natural language. In both domains, these inputs update a comprehender’s situation model, which is essential for information extraction from a quickly unfolding input. Our results have far-reaching implications for programming and pave the way for interdisciplinary collaborations between software engineering and psycholinguistics.

Nar1 binds the cytosolic iron–sulfur cluster assembly targeting complex via bipartite interactions

ASCENT-04: Analysis of efficacy by biomarker subgroups with sacituzumab govitecan (SG) + pembrolizumab (pembro) vs chemotherapy (chemo) + pembro in participants (pts) with previously untreated PD-L1+ metastatic triple-negative breast cancer (mTNBC).

1013 Background: In the phase 3 ASCENT-04/KEYNOTE-D19 study (NCT05382286) SG + pembro showed significant and clinically meaningful progression-free survival (PFS) improvement vs chemo + pembro in pts with previously untreated PD-L1+ mTNBC. SG + pembro exhibited a manageable safety profile consistent with prior studies for each agent. We present preplanned exploratory efficacy analyses in ASCENT-04 by biomarker subgroups. Methods: 443 pts received (1:1 ratio) SG + pembro in 21-day cycles or chemo (gemcitabine + carboplatin; taxane) + pembro. From centrally tested fresh or archival tumor samples, Trop-2 expression was determined by immunohistochemistry (IHC), tumor BRCA (tBRCA) status by whole exome sequencing, and HER2 expression by in situ hybridization (ISH) + IHC. Pts were grouped by Trop-2 expression quartiles, tBRCA status (wild-type [WT] or mutant [mut] in BRCA1/BRCA2 /both), and HER2 status (IHC 0 vs Low [IHC 1+ or IHC 2+/ISH–]). Biomarker subgroups were analyzed descriptively for association with PFS by blinded independent central review (BICR); other outcomes by biomarker status are forthcoming. Results: Median Trop-2 H-score was 280, ranges: quartile 1 (Q1) 0-224, Q2 225-279, Q3 280-298, Q4 299-300. SG + pembro demonstrated longer PFS by BICR in Q3/Q4 vs Q1/Q2, and benefit was observed for SG + pembro vs chemo + pembro in all quartiles (Table). Hazard ratios (HR; 95% confidence interval [CI]) were: Q1, 0.81 (0.48-1.36); Q2, 0.73 (0.44-1.22); Q3, 0.46 (0.27-0.80); Q4, 0.57 (0.33-0.99). Proportion of pts with tBRCA mutations (~20%) was similar between treatment groups. Longer PFS was observed with SG + pembro with HR (95% CI) of 0.67 (0.49-0.91) in the tBRCA WT and 0.88 (0.45-1.74) in the tBRCA mut subgroups. PFS was longer with SG + pembro vs chemo + pembro in the HER2 IHC 0 and HER2 low subgroups, HR (95% CI) 0.69 (0.46-1.04) and 0.67 (0.48-0.92), respectively. Conclusions: SG + pembro demonstrated longer PFS vs chemo + pembro across all Trop-2, tBRCA, and HER2 subgroups. These analyses strengthen support for the significant, clinically meaningful benefit of SG + pembro as first-line treatment for mTNBC across subgroups. Clinical trial information: NCT05382286 . Efficacy, BICR N Median PFS (95% CI), mo Biomarker Subgroup SG + pembro Chemo + pembro SG + pembro Chemo + pembro HR (95% CI) Trop-2(n = 400) Q1 48 47 9.3(7.4-19.4) 9.0(6.0-10.9) 0.81(0.48-1.36) Q2 50 50 9.6(7.3-16.7) 7.4(6.9-9.7) 0.73(0.44-1.22) Q3 55 50 13.5(9.3-NR) 8.4(5.6-10.8) 0.46(0.27-0.80) Q4 51 49 16.6(8.1-NR) 9.2(5.5-11.3) 0.57(0.33-0.99) tBRCA(n = 332) WT 130 131 9.6(7.6-16.7) 7.4(6.9-9.2) 0.67(0.49-0.91) Mut 39 32 16.6(7.5-NR) 12.9(7.1-NR) 0.88(0.45-1.74) HER2(n = 435) IHC 0 84 82 16.6(9.1-21.2) 9.0(7.2-10.8) 0.69(0.46-1.04) Low 133 136 11.2(9.1-16.6) 7.7(7.0-9.4) 0.67(0.48-0.92)

Breast cancer screening and detection after the 2024 USPSTF recommendation for women aged 40–49.

11126 Background: In April 2024, the US Preventive Services Task Force (USPSTF) expanded its biennial screening mammography recommendations to include women aged 40–49, citing rising breast cancer incidence in this age group and persistent disparities in outcomes. How this expansion has influenced diagnosis and stage at diagnosis in real-world practice is not yet well characterized. Methods: Using a subset of de-identified electronic health record data from Truveta, we identified women aged 40–74 who underwent screening mammography between 2018 and 2025, had no prior breast cancer diagnosis and who had evidence of outpatient care 6-18 months before and 1-12 months after screening. For patients with multiple eligible screening mammograms, only the first was retained for analysis. Analyses were restricted to health systems with available cancer stage data. April 30, 2024, was used as the cut point (pre- vs. post-guideline). Outcomes included: (1) screening uptake by age group, (2) breast cancers diagnosed within 180 days of screening, (3) stage at diagnosis (early [AJCC 0–I] vs. later [II–IV]). Diagnosis and stage were derived from clinical notes and categorized as early- (AJCC stage 0–I) versus later-stage (II–IV). Logistic regression, adjusted for race, ethnicity, rural–urban residence, and family history, was used to evaluate differences by age group and guideline period. Results: Among 985,687 women aged 40-74 who received a screening mammogram, 880,685 (89.3%) occurred pre-guideline and 105,002 (10.7%) occurred post-guideline. Women aged 40–49 comprised a substantially larger share of screenings post-guideline (53,227/105,002, 50.7% vs. 262,330/880,685, 29.8% pre-guideline p<.001). The likelihood of a breast cancer diagnosis within 180 days did not differ between the pre- and post-guideline periods for women aged 40–49 (OR 1.03, 95% CI 0.79–1.34) or 50–74 (OR 0.87, 95% CI 0.75–1.01). Among women diagnosed within 180 days with cancer stage information available (n = 3,866), the likelihood of early- versus later-stage diagnosis did not differ between the pre- and post-guideline periods for women aged 40–49 (early-stage: 63.6% vs. 68.3%; OR 1.26, 95% CI 0.72–2.21) or 50–74 (OR 0.91, 95% CI 0.67–1.24). Conclusions: Following the April 2024 USPSTF guideline expansion, screening mammography uptake among women aged 40–49 increased substantially, however the likelihood of a breast cancer diagnosis within 180 days did not significantly differ between the pre- and post-guideline periods. Stage at diagnosis did not differ significantly in the short post-guideline period; longer follow-up is needed to assess whether earlier detection translates to stage-shift benefits over time. These findings provide early real-world evidence to inform the ongoing risk-benefit discussion surrounding breast cancer screening in younger women.

Clinical outcomes of venetoclax with and without azole antifungals in acute myeloid leukemia: A large real-world cohort study.

e18540 Background: Venetoclax is widely used in the treatment of acute myeloid leukemia (AML) and is frequently co-administered with azole antifungals for prophylaxis or treatment of fungal infections. Azoles are strong CYP3A inhibitors and increase venetoclax exposure, potentially exacerbating toxicity and affecting outcomes. However, real-world data evaluating the clinical impact of this combination on survival, infectious complications, and hematologic toxicity remain limited. Methods: We conducted a retrospective cohort study using the TriNetX Global Collaborative Network, including adult patients with AML treated with venetoclax. Patients were stratified into two cohorts: venetoclax monotherapy and venetoclax plus azole antifungals (itraconazole, voriconazole, or posaconazole). Propensity score matching (1:1) was performed to balance baseline demographics and comorbidities. Outcomes included all-cause mortality, sepsis with or without septic shock, invasive fungal infections (IFI), severe neutropenia (≤0.5 ×10³/µL), and severe thrombocytopenia (≤50 ×10³/µL). Outcomes were assessed at 1 month, with analyses at 3 and 6 months using risk analyses and Kaplan–Meier survival models. Results: After matching, 12,984 patients were included in each cohort. At 1 month, mortality was lower in the venetoclax-only cohort compared with venetoclax plus azoles (7.5% vs 8.9%; risk difference −1.4%, 95% CI −2.1% to −0.7%; p < 0.001; hazard ratio 0.83, 95% CI 0.77–0.91). This survival advantage persisted at both 3 and 6 months, with hazard ratios favoring venetoclax monotherapy. There was no difference in sepsis ± septic shock at 1 month (4.8% vs 5.1%; p = 0.24), and this association remained non-significant at 3 and 6 months. The incidence of IFI was lower in the venetoclax-only cohort at 1 month (2.4% vs 3.3%; risk difference −0.9%; p < 0.001; 0.73), with this difference remaining statistically significant at 3 and 6 months. In contrast, hematologic toxicity was more frequent with azole co-administration: severe neutropenia occurred in 60.5% vs 57.0% (p < 0.001) and severe thrombocytopenia in 72.0% vs 69.0% (p < 0.001) for venetoclax plus azoles versus venetoclax alone. Conclusions: In this large real-world cohort of patients with acute myeloid leukemia treated with venetoclax, concomitant azole antifungal use was associated with higher mortality and increased hematologic toxicity compared with venetoclax monotherapy, without a reduction in sepsis risk. Invasive fungal infections were less frequent in patients treated without azole antifungals; however, interpretation of this finding is limited by the observational nature of the study. These results highlight the clinical impact of venetoclax–azole drug–drug interactions and support the need for prospective studies to define optimal antifungal strategies in venetoclax-based regimens.

Turning silence into rhythm: Outcomes of a school-based <i>Cancer Olympiad</i> in India.

9016 Background: Late-stage presentation continues to account for nearly 70% of cancer diagnoses in low- and middle-income countries, exposing the gap between medical progress and public participation. Traditional awareness drives often falter in communities shaped by stigma, low literacy, and fear of cost. The Cancer Olympiad was conceived as an arts-based, school-centered intervention harnessing creativity, performance, and peer learning to transform knowledge into shared action. By positioning children as credible messengers, it aimed to bridge the emotional and educational distance between health information and household practice. Methods: A quasi-experimental, mixed-methods study engaged 25,000 students (grades 1–12) and 2,500 families across 125 urban and rural schools in Udaipur District, North India. Schools implemented structured creative modules—drawings, performances, debates, health exhibitions, reflective writing, and parent–child tasks—culminating in an interschool finale attended by educators, clinicians, and policymakers. Knowledge, attitude, and behavior (KAB) outcomes were measured at baseline and three months post-intervention using validated questionnaires adapted for regional literacy levels. Secondary endpoints included stigma reduction, screening and vaccination uptake, tobacco-cessation behavior, and community-engagement indices such as school-initiated health clubs and family advocacy actions. Qualitative interviews and focus groups explored cultural receptivity, emotional learning, and pathways of sustained engagement. Results: Knowledge of cancer risk factors and early-warning signs improved by 23 percentage points (p &lt; 0.001), while recognition of curability rose from 58% to 83%. Stigma scores fell by 0.35 SD, and 19% of participating families undertook at least one preventive action—screening, vaccination, quitting tobacco, or seeking medical evaluation—within six months. Primary-health-center screening visits increased by 17% compared with matched control blocks, and teacher-reported health-dialogue frequency in classrooms quadrupled. Qualitative analysis revealed recurring themes of “children as catalysts of credibility,” “art as emotional pedagogy,” and “community pride replacing fear.” Implementation fidelity exceeded 90%, school-level engagement remained high, and more than half the schools adopted the Olympiad as an annual, self-sustained event. Conclusions: The Cancer Olympiad demonstrates that child-led, arts-integrated education can align biomedical understanding with cultural identity, shifting awareness into collective prevention. When stories and performances enter homes, they humanize cancer knowledge and normalize early action. Embedding narrative and creativity within oncology outreach may extend medicine’s reach beyond clinics—turning silence into rhythm, and rhythm into a movement of lasting change.

Evaluating the prognostic performance of an AI-based model in clinically low-risk HR+/HER2− early breast cancer.

552 Background: Accurate prediction of recurrence risk in early HR+/HER2- breast cancer remains central to adjuvant treatment decision making. Clinically approved prognostic models largely rely on genomic features, yet their performance remains insufficient for optimal clinical stratification. To address these limitations, AI-based approaches have increasingly been developed to improve risk prediction. Ataraxis Breast (ATX) is a multimodal artificial intelligence test integrating clinical information with morphological information from H&amp;E-stained whole slide images for HR+/HER2- patients. Here, we evaluate ATX’s ability to prognosticate recurrence risk in an external cohort of clinically low risk early breast cancer patients from Dordrecht, the Netherlands. Methods: Clinical information and H&amp;E images were accessed for 887 patients with HR+/HER2- breast cancer. ATX scores were generated using a locked model and a prespecified cut-off was applied to classify patients into ATX high vs low groups. The primary endpoint was recurrence-free interval (RFI); recurrence-free probabilities were estimated for the two groups using a Kaplan-Meier estimator. Additionally, performance of ATX was assessed using C-index and hazard ratio (HR per 0.1 increase in ATX score). No patients from this dataset were used in the training of ATX. Results: The cohort of 887 patients was characterized by low clinical risk, where 631tumors were classified as T1 and 239 as T2, and the majority of patients were node-negative (N0: n = 639). Tumors were predominantly low to intermediate grade (grade 1: n = 296; grade 2: n = 496), with relatively few high-grade cases (grade 3: n = 88). Histological analysis identified invasive lobular carcinoma in 131 patients. The median tumor size was 15 mm (interquartile range, 11-22 mm). Accordingly, most patients (n = 773, 87%) were classified as ATX low risk. Patients in the ATX high risk group had lower estimated 5-year probability of meeting the RFI endpoint (0.88, 95% CI = 0.80-0.93) than patients classified as ATX low risk (0.98, 95% CI = 0.96-0.98). When modeled as a continuous variable, ATX demonstrated strong discriminatory performance for recurrence risk (C-index = 0.740, 95% CI, 0.68-0.80) and was associated with a significantly increased hazard of an RFI-contributing event (HR = 2.68, 95% CI, 2.09-3.45; p &lt; 0.001). To account for clinical confounding, we fitted a multivariate Cox proportional hazards model adjusting for T stage, N stage, grade, age at diagnosis, adjuvant endocrine therapy, adjuvant chemotherapy, and tumor size. We found that ATX score remained significant (HR = 1.99, 95% CI = 1.38-2.86, p &lt; 0.001) and tumor size was the only other variable that was statistically significant (p &lt; 0.05). Conclusions: ATX accurately prognosticates recurrence risk in a clinically low risk group of HR+/HER2- breast cancer patients from an external validation cohort.

Mitochondrial size as a potential factor in colorectal cancer metastasis.

e15546 Background: The accelerated progression of colorectal cancer, with frequent recurrences and distant metastases, highlights the need for targeted attention to endogenous structures such as mitochondria. Their ability to actively translocate within environments and tissues is combined with powerful signaling regulation of structural and metabolic processes during tumor growth. Modern technologies make it possible to determine some physical characteristics of mitochondria to understand their role in metastasis processes. The aim of the study was to investigate the perimeter size of mitochondria isolated from rectal adenocarcinoma tissue and conditionally healthy tissue along the tumor resection line in patients of both sexes, depending on the presence or absence of metastases. Methods: Samples were obtained from 44 patients (median age: 66 years; range: 58–73) who underwent surgery for stage T2-3N0M0, G2-grade rectal adenocarcinoma. Tumor fragments and intestinal tissue from the resection margin were fixed in a formaldehyde/glutaraldehyde solution. Standard tissue processing methods were used, and ultrathin sections were examined using a Jeol JEM-1011 transmission electron microscope (TEM). Mitochondrial perimeter was measured on electron micrographs using FiJi software. Statistical analysis of the results was performed using the Statistica 10.0 software package. Results: A comprehensive ultrastructural study and statistical analysis of the perimeter size of mitochondria isolated from rectal adenocarcinoma in patients of both sexes was conducted. The results indicated that there were no statistically significant intergroup differences in mitochondrial size, categorized as conditionally small (0-1 μm), medium (1-2 μm) and large ( &gt; 2 μm), depending on the presence or absence of metastases. In contrast to the tumor, in conditionally healthy tissue along the rectal resection line, the proportion of mitochondria with a perimeter of ˃2 μm, which were characterized by the most pronounced signs of dysfunction, increased by 3.6 times in women and 4.8 times in men with metastases compared to the indicators in all patients without metastases (p &lt; 0.05). Conclusions: The significance of differences in the perimeter indices of mitochondria isolated from conditionally healthy tissue at the resection line of rectal adenocarcinoma, regardless of the gender differences of patients, indicated the possibility of a morphofunctional relationship between processes occurring at the ultrastructural mitochondrial level and the metastasis of rectal cancer. Translocation of mitochondria from rectal adenocarcinoma into layers of conditionally healthy submucosal tissue at the level of the tumor resection line indicates the dangerous potential of the motor and regulatory activity of mitochondria of malignant neoplasms in tumor progression.

Development and internal validation of a clinical prognostic model for SMARCB1-deficient renal medullary carcinoma (RMC).

e16535 Background: RMC is a rare and highly aggressive malignancy that predominantly affects young individuals with sickle hemoglobinopathies. Due to its rarity, no validated prognostic models exist to guide clinical decision-making. We aimed to develop and internally validate a prognostic model for patients with RMC using routinely available clinical and laboratory variables. Methods: We analyzed a single-institution cohort of patients with RMC. Candidate variables were identified based on biological plausibility and represented in a directed acyclic graph. Missing data were handled via multiple imputation by chained equations ( m = 50). Variables were ranked using elastic net Cox regression ( α = 0.5, λ = λ min ) with stability selection derived from 100 bootstrap resamples per imputed dataset. Predictors achieving &gt; 60% selection frequency were retained for a nested parsimony analysis. Bootstrap internal validation ( B = 2000) estimated optimism-corrected discrimination and calibration performance, with the calibration slope applied as global shrinkage to penalize final coefficients for overfitting. Clinical utility was evaluated via decision curve analysis (DCA) at 18- and 24-month horizons. Results: Among 180 patients (median age 29; IQR 23–37), 142 deaths occurred with a median follow-up of 12.4 months. The final model included four predictors (Table). The optimism-corrected c-index was 0.667 (optimism = 0.009), and the calibration slope was 0.964, indicating minimal overfitting. Calibration plots showed excellent predicted-observed agreement at 12, 18, 24, and 36 months. DCA showed net benefit across threshold probabilities of 10%–75% at 18 months and 10%–80% at 24 months compared to default strategies. In sensitivity analysis, inclusion of TP53 mutation status improved discrimination ( c -index 0.686) and model fit ( ΔAIC −10.6). An interactive web-based risk calculator was developed to facilitate clinical implementation. Conclusions: This study presents the first internally validated prognostic model for RMC, demonstrating acceptable discrimination and excellent calibration using routinely available variables. By stratifying patients into distinct risk groups, this model may inform treatment intensity—guiding escalation to multi-agent regimens or sequential chemotherapy cycling in high-risk patients, while identifying lower-risk patients who may benefit from de-escalation strategies such as metastasis-directed therapy for oligometastatic disease or consideration of cytoreductive nephrectomy. External validation is warranted. Variable Time Ratio 95% CI p value ECOG performance status 0.64 0.53–0.77 &lt;0.001 Distal (non-retroperitoneal) lymph node metastases 0.69 0.52–0.92 0.011 Neutrophil-to-lymphocyte ratio 0.98 0.94–1.02 0.25 Corrected calcium (mg/dL) 0.77 0.56–1.06 0.10 Sensitivity Analysis Only TP53 Mutation 0.63 0.36-1-10 0.10

Real-world outcomes and prognostic risk stratification of neoadjuvant immunotherapy for stage III lung squamous cell carcinoma.

8047 Background: While neoadjuvant immunotherapy (nIO) is the standard of care for resectable NSCLC, its efficacy in "potentially resectable" (PR) stage III lung squamous cell carcinoma (LUSC)—specifically patients with T4 invasion or multi-station N2 disease—remains underrepresented in clinical trials. We evaluated whether nIO could facilitate surgical conversion in PR patients comparable to initially resectable (IR) candidates and constructed a multidimensional risk stratification system to optimize decision-making. Methods: We retrospectively analyzed patients with stage III LUSC treated with neoadjuvant PD-1/PD-L1 inhibitors plus chemotherapy (January 2019–December 2024). The cohort was stratified into IR and PR groups, with PR defined by complex anatomy (T4 invasion or multi-station N2). Endpoints included surgical conversion, major pathological response (MPR), disease-free survival (DFS), and overall survival (OS). Baseline predictors were identified via logistic regression. A pathological risk score (pRS) was developed using Cox regression based on independent risk factors for DFS in the R0 resection cohort. Results: Of 210 evaluable patients, 170 (81.0%) underwent surgery with a 96.5% R0 resection rate. The PR cohort (n = 128) achieved surgical conversion rates comparable to the IR cohort (n = 82) (82.9% vs. 90.7%; P = .19). Notably, PR patients demonstrated superior nodal downstaging (74.5% vs. 58.8%; P = .047) compared with IR patients. Long-term survival outcomes showed no statistically significant difference between PR and IR groups (DFS: HR, 1.25; 95% CI, 0.77-2.04; OS: HR, 1.41; 95% CI, 0.76-2.62). Multivariable analysis identified baseline NLR ≥ 2.75 and CYFRA 21-1 ≥ 6.0 ng/mL as independent predictors of poor therapeutic benefit. The constructed pRS system (integrating non-MPR status, vascular, and pleural invasion) effectively stratified postoperative recurrence risk (P &lt; .0001; 1-year AUC, 0.73). Analysis of treatment failure (n = 40) revealed that baseline fibrinogen &gt; 3.5 g/L was associated with disease progression, while grade 3–5 pneumonitis was a primary cause of surgical dropout in patients with high PD-L1 expression. Conclusions: In this real-world cohort, nIO facilitated high R0 resection rates in complex stage III LUSC, allowing potentially resectable patients to achieve survival outcomes equivalent to initially resectable candidates. Nodal clearance appears to be a key driver of benefit in the PR subgroup. The novel pRS system and biomarkers (NLR, CYFRA 21-1, fibrinogen) provide a robust framework for patient selection and postoperative management.

Determinants of metastatic phenotypes at diagnosis in prostate cancer: A population-based analysis of organotropism, disparities, and clinical implications.

e17113 Background: Metastatic prostate cancer is biologically heterogeneous, yet metastatic sites are often grouped together in clinical decision-making. Methods: Using the National Cancer Database (2016–2022), 56,112 patients with metastatic prostate cancer at diagnosis were identified. Patients were classified into four mutually exclusive metastatic phenotypes: bone-only, distant lymph node (LN)-only, visceral organ-only (liver, lung, or brain), and multi-site disease. Multinomial logistic regression was used to estimate adjusted odds ratios (ORs) for each metastatic phenotype relative to bone-only disease. Results: Bone-only metastasis was the most common phenotype (64.5%), followed by multi-site (26.4%), distant LN-only (6.0%), and visceral organ-only disease (3.1%). Small cell/neuroendocrine carcinoma was strongly associated with visceral organ-only metastasis (OR 13.7, 95% CI 10.5–17.9) and multi-site disease (OR 6.4, 95% CI 5.3–7.8) compared with acinar adenocarcinoma. Each additional year of age was associated with lower odds of distant LN-only disease (OR 0.98 per year) but higher odds of visceral organ-only metastasis (OR 1.01 per year). Black race (OR 1.20, 95% CI 1.13–1.26) and Medicaid insurance (OR 1.18, 95% CI 1.09–1.27) were independently associated with higher odds of multi-site disease. Hispanic ethnicity was associated with increased odds of visceral organ-only (OR 1.22, 95% CI 1.01–1.48) and multi-site metastases (OR 1.20, 95% CI 1.11–1.30). Conclusions: Distinct metastatic phenotypes at prostate cancer diagnosis are strongly associated with tumor histology, age, and sociodemographic factors. These findings highlight biologic organotropism in aggressive histologic subtypes and persistent disparities in metastatic burden. Adjusted odds ratios for metastatic phenotypes relative to bone-only disease. Predictor Distant LN-only OR (95% CI) Visceral organ-only OR (95% CI) Multi-site OR (95% CI) Age (per year) 0.98 (0.98–0.98)* 1.01 (1.00–1.02)* 0.99 (0.98–0.99)* Black vs White 1.05 (0.95–1.16) 1.10 (0.96–1.26) 1.20 (1.13–1.26)* Hispanic vs Non-Hispanic 1.04 (0.90–1.20) 1.22 (1.01–1.48)* 1.20 (1.11–1.30)* Medicaid vs Private 0.67 (0.57–0.78)* 1.02 (0.83–1.26) 1.18 (1.09–1.27)* Charlson ≥3 vs 0 0.87 (0.74–1.02) 1.20 (1.00–1.44)* 1.11 (1.03–1.20)* Small cell / neuroendocrine 2.83 (2.00–4.01)* 13.69 (10.50–17.85)* 6.40 (5.29–7.75)* Adenocarcinoma w/ NE differentiation 2.70 (1.62–4.48)* 3.37 (1.78–6.36)* 3.20 (2.35–4.36)* Undifferentiated / anaplastic ------ 98.88 (21.35–457.86)* 10.96 (2.36–50.82)* Odds ratios derived from multivariable multinomial logistic regression. Bone-only metastasis served as the reference outcome. Models adjusted for demographics, insurance status, comorbidity burden, tumor grade, histology, facility type, and year of diagnosis. *p&lt;0.05.

Financial toxicity and health-related social needs in adolescent and young adult cancer patients and their caregivers: Insights from an LMIC population.

e23180 Background: Adolescents and young adults with cancer experience substantial financial toxicity (FT) and health-related social needs (HRSN), particularly in low and middle-income countries. These hardships also extend to their caregivers, yet quantitative evidence is limited. This study aimed to assess the extent and determinants of FT and HRSN among AYA cancer patients and their caregivers and their correlation. Methods: A cross-sectional study was conducted among consecutively enrolled AYA cancer patients at a tertiary care cancer hospital. Financial toxicity was assessed using the Comprehensive Score for Financial Toxicity (COST-FACIT) questionnaire and health-related social needs using the AHC-HRSN core instrument. The sociodemographic data and responses were collected. Moderate-severe FT was defined as COST scores &lt; 13. Multivariable logistic regression was used to identify independent predictors of FT, adjusting for socioeconomic factors. Results: Among 399 participants enrolled between August and October, 2025, the mean COST score recorded was 13.39 ± 6.11. 40.9% participants reported mild while 55.1% participants reported moderate to severe financial toxicity. Female gender (p = 0.039), rural household(p = 0.003), and level of education (p = 0.005) showed significant association with COST toxicity categories. Internal consistency of COST was excellent (α = 0.838). At least one unmet social need was reported by 62.4% of participants, with food insecurity (41.4%) and housing instability (29.6%) being most prevalent. All the domains of health-related social needs showed significant association with financial toxicity (p &lt; 0.05). Multivariable analysis when adjusting for socio-economic factors found that cumulative needs were independently associated with moderate–severe FT (adjusted OR 2.00, CI 1.54-2.59, p &lt; 0.001), while other factors were not significant after adjustment. On sensitivity analyses, unmet social needs remained significantly associated with lower COST scores (B = −2.20, p &lt; 0.001). A moderate inverse correlation was found with COST scores and number of unmet needs (r = –0.328, p &lt; 0.001), indicating a dose dependent effect. Furthermore, it was found that 98.2% of the caregivers used to stay at the hospital, for a median of 10 days, providing appreciable support. Conclusions: It was found that financial toxicity and health-related social needs were highly significant in this cohort, confirming the hypothesis. Unmet health-related social needs are the dominant independent predictor of clinically significant financial toxicity among AYA cancer patients in India, exceeding the influence of traditional sociodemographic factors. These findings emphasize the critical need for integrated financial and social support assessment for adolescents and young adults and their caregivers. Future research work is required to evaluate long-term outcomes in lower- and middle-income countries.

PKCI–YAP1–CCL7 axis and tumor-associated macrophage–mediated anti–PD-1 resistance in lung adenocarcinoma.

e20619 Background: Immune checkpoint blockade targeting PD-1 has improved outcomes in non-small cell lung cancer (NSCLC), yet most patients fail to achieve durable responses. Tumor-intrinsic oncogenic signaling is increasingly recognized as a key driver of immunosuppressive tumor immune microenvironment (TIME) formation and therapeutic resistance. We investigated whether protein kinase Cι (PKCι), an oncogenic atypical PKC highly expressed in NSCLC, mediates resistance to anti-PD-1 therapy through regulation of myeloid cell recruitment. Methods: PKCι expression was examined in human NSCLC specimens and correlated with immune infiltration and clinical response to immune checkpoint inhibitors. Murine lung adenocarcinoma models with genetic knockdown or overexpression of PKCι, YAP1, or CCL7 were established. Tumor growth under anti-PD-1 therapy was evaluated in immunocompetent mice. Immune cell populations were characterized by flow cytometry and immunohistochemistry. Macrophage migration was assessed using transwell and 3D tumor-spheroid assays. RNA sequencing, qPCR, ELISA, and chromatin immunoprecipitation were performed to define PKCι downstream signaling. The clinically approved drug auranofin was tested as a pharmacologic PKCι inhibitor, alone or combined with anti-PD-1 therapy, in orthotopic and subcutaneous lung tumor models. Results: High PKCι expression in NSCLC correlated with poor response to PD-1 blockade, increased tumor-associated macrophage (TAM) infiltration, and reduced CD8⁺ T-cell abundance. Genetic inhibition of PKCι significantly enhanced anti-PD-1 efficacy in vivo and selectively reduced TAM accumulation without substantially affecting other suppressive immune subsets. Mechanistically, PKCι activated YAP1/TEAD-dependent transcription of the chemokine CCL7, which was required for macrophage recruitment and establishment of an immunosuppressive TIME. Knockdown of CCL7 or YAP1 abrogated TAM migration and restored sensitivity to PD-1 blockade, whereas re-expression of PKCι or CCL7 reinstated TAM infiltration and therapeutic resistance. Auranofin suppressed PKCι-YAP1-CCL7 signaling, reduced TAM recruitment, increased infiltration and activation of cytotoxic CD8⁺ T cells, and markedly improved antitumor efficacy when combined with anti-PD-1 therapy. CD8⁺ T-cell depletion abolished the therapeutic benefit of the combination regimen. Conclusions: The PKCι-YAP1-CCL7 axis is a critical driver of TAM-mediated immune suppression and resistance to PD-1 blockade in lung adenocarcinoma. Pharmacologic targeting of PKCι with the repurposed drug auranofin reprograms the TIME and sensitizes tumors to anti-PD-1 therapy, providing a mechanistic rationale for a clinically translatable combination strategy in NSCLC.

Using the NCCN distress thermometer to evaluate clinically elevated levels of distress among caregivers of HSCT patients.

12093 Background: Caregivers of patients receiving an allogeneic hematopoietic stem cell transplant (HSCT) play a vital role in the care and recovery of patients. They must be present with the patient 24/7 for up to 100 days post-transplant, and they often report levels of distress higher than patients. Importantly, caregiver and patient distress are interdependent. Thus, knowing when to deploy appropriate resources and support to a distressed caregiver is essential. To date, the NCCN Distress Thermometer (DT) is a validated, single-item clinical measure consistently used in the oncology setting to assess and intervene upon patient distress. If such a tool was available for HSCT caregivers, oncology providers could assess and provide support for those in heightened distress. To date, limited research has been conducted in this area. This study examines the relationship of DT with self-reported depression and anxiety in a large sample of HSCT caregivers with the goal of proposing a potential clinical cut-point of elevated distress on the DT. Methods: Caregivers completed the DT, Center for Depression Studies Depression Scale (CESD), and Generalized Anxiety Disorder 7 (GAD7) 1-2 weeks pre-transplant as part of a larger study. Spearman correlation coefficients evaluated the association of the DT with CESD and GAD7. Logistic regression evaluated DT as a predictor of clinically elevated levels of depression (CESD &gt;16) and anxiety (GAD7 &gt;10). DT cut points (e.g., ≥4) as predictors of clinically elevated depression and anxiety were evaluated using sensitivity, specificity, and area under the curve (AUC). Results: Caregivers (N = 287) were 73% female, 75% spouse/partner of the patient, and ranged in age from 22-87 years (M = 58.2, SD = 13.4). The DT ranged from 0-10 with M = 4.5 (SD = 2.6). The CESD ranged from 0-51 with M = 13.3 (SD = 10.3) and 36% scored &gt;16. The GAD7 ranged from 0-21 with M = 6.4 (SD = 5.5) and 24% scored &gt;10. The DT was very strongly correlated ( p ’s &lt; .0001) with the CESD ( r s = .69) and GAD7 ( r s = .76). DT score significantly predicted elevated depression (OR = 1.92, 95% CI[1.64-2.24]) and elevated anxiety (OR = 2.30, 95% CI[1.84-2.86]). Table 1 presents sensitivity, specificity, and AUC for DT cut points of 4-7. Conclusions: Findings suggest that although a cut-point of 4 on the DT has high specificity and a cut-point of 7 has high sensitivity, a cut-point of 6 indicates more balanced specificity and sensitivity when considering both depression and anxiety. In the clinical setting, providing the single-item DT as a measure of caregiver distress should allow for a brief and useful assessment indicating a greater need to provide support resources. DT CESD GAD7 cut point: N (%) Sensitivity Specificity AUC Sensitivity Specificity AUC 4: 173 (60%) .60 .95 .77 .52 .99 .75 5: 148 (52%) .69 .88 .78 .61 .93 .77 6: 114 (40%) .80 .75 .78 .74 .85 .80 7: 80 (28%) .89 .58 .73 .86 .73 .80

Proof of concept for machine learning–based models to predict response to anticancer treatments.

e13002 Background: Selecting the most effective anticancer drugs remains challenging. Comprehensive molecular profiling identifies alterations actionable with targeted agents, but genomic and clinical data may also predict benefit across different drug classes. We developed machine learning–based models (Support Vector Machines - SVM and Random Forest - RF) to predict treatment outcomes using real-world clinical, pathological, genomic, and prior-therapy data at the treatment line level. Genomic alterations were summarized at the pathway level. Methods: We retrospectively analyzed 46 homogeneous patients with breast cancer who received PARP inhibitors (PARPi) discussed at the IEO Molecular Tumor Board between 2019 and 2023. Outcomes were progression-free survival (PFS) and objective response rate (ORR). As a quality control measure, models were trained to predict overall survival in first-line setting to assess data reliability. The most important predictors were expected (hormone receptor (HR), Ki-67 and age), thus supporting data reliability. We then trained outcome models for the four most represented treatment classes: PARPi, taxanes (Tx), endocrine therapies (ET), and pyrimidine analogues (Pyr). Results: RF models outperformed linear SVMs, particularly for PFS prediction, suggesting non linear interactions among predictors. Overall discrimination was modest (Table 1), reflecting limited sample size and the scarcity of true negative controls, especially for targeted treatments. Prior treatments emerged as the strongest predictor: longer PFS of previous platinum therapy (pPFS) was among the most important features to predict longer PFS and better ORR for PARPi, but similar results were obtained for all the treatment classes. Features such as Ki-67, performance status (PS), age, HR status and treatment history added predictive value, while genomic pathway alterations provided weaker but detectable signals. Conclusions: Our study demonstrates the feasibility of integrating data to predict treatment benefit and prioritize therapeutic options. Although current limited performance, the approach highlights the dominant yet still underestimated importance of prior treatment response on subsequent therapies’ outcome prediction, establishing a foundation for a methodological refinement, expansion across tumors, and external validation. It could evolve into a clinical decision support tool optimizing both personalized and standard treatments, which is especially relevant for low-income countries without access to expensive drugs. Drug PFS(RMSE) ORR(ROC) Top featuresORR Direction RF SVM RF SVM PARPi46 9 11.5 0.63 0.59 RAS wt + pPFS platinum + BRCA1, BRCA2 and PALB2 wt - Lobular - ET35 14.4 15.9 0.48 0.45 PGR% + pPFS ET + RAS altered - NOTCH wt - Tx35 8.2 10.6 0.75 0.58 pPFS anthracyclines + pPFS TROP2 ADC + PS - pPFS PARPi - Pyr34 15.5 22.8 0.6 0.58 pPFS alkylating + Lobular + Ki-67 - pPFS HER2 ADC -

Androgen receptor (AR) gene dynamics in ctDNA post-abiraterone/enzalutamide: Longitudinal study in mCRPC with germline results.

5076 Background: Approximately 10–15% of men with metastatic prostate cancer harbor germline mutations. Circulating tumor DNA (ctDNA) provides a non-invasive approach to identifying AR mutations as resistance mechanism after abiraterone and enzalutamide initiation. In this study, we report longitudinal ctDNA profiling in patients with metastatic castration-resistant prostate cancer (mCRPC) who had germline testing. Methods: A retrospective review was conducted at Tulane Cancer Center to identify mCRPC patients with serial ctDNA after starting abiraterone and/or enzalutamide and germline test results. Germline testing was performed using a multi-gene cancer panel from Invitae, covering 50–86 genes, while somatich alterations in ctDNA were assessed through Guardant360, which included 70–83 genes. Linear mixed-effects models (LMMs) with random intercepts were used to model AR gene allele fraction and AR amplification copy number as continuous outcomes. Fixed effects included serial ctdna timepoints and germline gene pathogenicity, and random effects accounted for repeated measure using patient-specific intercepts. Cox Proportional Model was used to analyze the time from initiation of ARPi to AR amp and/or AR mutation in ctDNA. Analyses were performed using the R version 4.4.2. Results: From 2015-2025, 341 mCRPC patients with germline test results received an average of 2.96 ctDNA tests post abiraterone or enzalutamide. In this cohort, 47 patients had germline pathogenic mutations (14%) and 294 patients had germline non-pathogenic mutations (86%). Among those with pathogenic germline findings, there were 29 patients with mutation in homologous recombination repair genes (62%) and 18 patients with mutation in other genes. Overall, 55.7% patient in this cohort developed either AR amplification or AR pathogenic mutation in ctDNA. In linear mixed models, germline pathogenicity and pathogenic HRR mutations were not associated with AR copy number, AR allele frequency, or occurrence of somatic AR alterations. Serial ctDNA timepoints following baseline assessment are associated with a significant increase in AR amplifications copy numbers in ctDNA(p=0.003). AR amplification copy number increases by 0.12 for each subsequent timepoint from baseline. Only 318 patients were included in survival analysis due to missing ARPi start date. Time to ctDNA AR amplification or mutation after 1 st ARPi initiation was shorter in mCRPC patients without pathogenic HRR germline mutations compared to those with pathogenic HRR mutations (median: 38 vs 58 months, p = 0.049). Conclusions: Later serial timepoints correlated with a significant increase in AR amplifications copy numbers but not correlated with highest AR mutant allele frequency. mCRPC patients without germline HRR mutations will find ctDNA AR amplification or mutation sooner than those with germline HRR mutations.

Severe cognitive effects of avapritinib: Towards a mechanistic understanding.

e24164 Background: Approximately 5-10% of Gastrointestinal Stromal Tumors (GIST) harbor a PDGFRα D842V mutation. For patients with advanced disease, avapritinib is currently the only approved therapy. However, emerging real-world data raises concerns about severe neurocognitive adverse events during and after avapritinib treatment. We hypothesize that avapritinib crosses the blood-brain barrier (BBB) and induces microvascular injury by affecting pericytes through PDGFRβ inhibition. Here, we report real-world patient data alongside complementary in vivo mouse studies designed to investigate this mechanism. Methods: We retrospectively reviewed neurocognitive and neurovascular events among Dutch GIST patients treated with avapritinib between August 2017 and December 2025, graded according to CTCAE version 5.0. In parallel, we conducted mouse experiments to (1) quantify brain penetration of avapritinib using LC-MS, (2) assess long-term effects with monthly MRI and histopathology, and (3) evaluate acute pericyte detachment following avapritinib exposure using electrophysiology and functional ultrafast ultrasound imaging to assess cerebral blood flow. This latter model was additionally used to explore the potential neuroprotective effect of the investigational agent C381. Results: Sixteen patients received avapritinib for a median duration of 442 days (range 27-1354). Memory impairment (12 patients; 75%), dullness (9; 56%), and aphasia (6; 38%) were the most frequently observed neurocognitive events, occurring after a median of 246, 96, and 682 days from treatment initiation, respectively. Eight patients experienced grade ≥3 neurocognitive events: four without evidence of cerebrovascular accident (CVA), three with a CVA, and one who developed grade 3 neurocognitive impairment followed by a CVA. Brain MRI frequently showed signs of vascular disease; but the absence of baseline and follow-up imaging in most patients limited causal inference. In mice, the avapritinib brain-to-plasma ratio was 0.69, indicating a modest BBB penetration. Interim analyses from the long-term exposure model (week 20) revealed no major MRI differences between control and high-dose groups. Results of histopathological analyses after long-term exposure, as well as effects on cerebral blood flow, are expected shortly. Conclusions: Severe neurocognitive events are common in patients treated with avapritinib. Clinical and radiological data suggest micro- and macrovascular brain injury. In vivo pharmacokinetic data confirm that avapritinib crosses the BBB. Ongoing in vivo experiments aim to elucidate whether avapritinib induces pericyte-mediated microvascular injury as a mechanistic basis for these neurocognitive events and to explore potential neuroprotective strategies. Final results are expected in early 2026.

A phase 1b/2 study of QLS1304, a novel, highly potent, and selective dual inhibitor of KAT6A/6B, plus endocrine therapy (ET) in patients (pts) with estrogen receptor–positive (ER+), human epidermal growth factor receptor 2–negative (HER2−) breast cancer (BC).

TPS1144 Background: QLS1304 is a novel, highly-potent, and selective dual inhibitor of KAT6A/6B. KAT6A acetylates histone H3 at the 23rd lysine residue, thereby modulating gene transcription, and implicated in ER+/HER2- BC. QLS1304 showed potent efficacy in standard of care resistant, KAT6A-high and ESR1 -mutant ER+/HER2- BC patient-derived xenograft model, whether administered as a monotherapy or combination therapy ( Cancer Res (2025) 85 (8_Supplement_1): 457 ). In the ongoing phase 1 trial, QLS1304 monotherapy showed an acceptable safety profile and potential anti-tumor activity in pts with BC. This study aimed to evaluate the safety and preliminary efficacy of QLS1304 + ET combination (fulvestrant or QLC1401, a novel oral selective ER degrader) ± CDK4/6 inhibitor (CDK4/6i) in pts with ER+/HER2- locally advanced or metastatic (LA/M) BC. Methods: This multicenter, open-label, phase 1b/2 trial consists of phase 1b combination dose-escalation and phase 2 dose-expansion. The phase Ib study will enroll pts with LA/M ER+/HER2- BC, who have received prior ET plus CDK4/6i, utilizing the Rolling-Six design. There are 2 cohorts: Cohort 1A QLS1304 + fulvestrant and Cohort 1B QLS1304 + QLC1401. Three dose levels of QLS1304 are planned, which are 4 mg/d, 8 mg/d, and 12 mg/d. The dose of QLC1401 is 150 mg, which is the recommended phase 2 dose (RP2D) of monotherapy. Both QLS1304 and QLC1401 are orally administered in a 28-day cycle. A total of 12~18 pts per cohort are planned to be enrolled, which might be adjusted based on the occurrence of dose-limiting toxicities (DLTs) during the first cycle. The primary endpoints are safety, tolerability, and RP2D of QLS1304 combined with fulvestrant/QLC1401. In phase 2, 1~2 dose levels of QLS1304 will be explored in 5 cohorts. In Cohort 2A (QLS1304 + fulvestrant) and Cohort 2B (QLS1304 + QLC1401), pts with LA/M ER+/HER2- BC who received prior ET plus CDK4/6i will be enrolled. In Cohort 2C (QLS1304 + CDK4/6i + letrozole), pts with LA/M ER+/HER2- BC with no prior therapy will be enrolled. In Cohort 2D (QLS1304 + CDK4/6i + fulvestrant), adjuvant ET-resistant or ET+CDK4/6i pretreated LA/M ER+/HER2- BC pts will be enrolled in 2 subgroups. In Cohort 2E (QLS1304 + CDK4/6i + QLC1401), adjuvant ET-resistant, treatment naïve, or ET+CDK4/6i pretreated LA/M ER+/HER2- BC pts will be enrolled in 3 subgroups. Doublet combination will be firstly explored, which will support the dose selection of triplet combination. In triplet cohorts 2C–2E, 6 pts will be enrolled in a safety run-in to assess the preliminary safety profile in each cohort. Expansion will be started after safety confirmation. A total of 316~366 pts will be enrolled. The primary endpoint of phase 2 is ORR per RECIST v1.1. Patient enrollment is currently ongoing. Clinical trial information: NCT07235176 .

Circulating tumor DNA clearance of adjuvant chemotherapy in localized colorectal cancer using an ultrasensitive structural variant–based assay.

3625 Background: High-risk localized colorectal cancer (CRC) remains prognostically unfavorable. Adjuvant chemotherapy (ACT) is administered to patients with poor-risk features to eradicate molecular residual disease (MRD) and improve outcome. Despite ACT selection, some patients experience recurrence and others receive unnecessary treatment. Circulating tumor DNA (ctDNA) has emerged as a strong prognostic biomarker that could improve selection for ACT, but detection in the MRD setting is dependent on high sensitivity. In this study, we use an ultrasensitive assay to investigate ACT clearance in localized CRC patients included in the CITCCA trial and its impact on outcome. Methods: Patients with stage I–III CRC were enrolled in a prospective Swedish cohort study between 2020-2024. Standard of care treatment was given, and patients were followed with longitudinal ctDNA before and after surgery at 4-6 weeks, 3 and 6 months, 1 and 2 years. The clinical landmark (CLM) was defined as the 4-6 week postoperative test prior to initiation of ACT. The first sample after definitive treatment (FDT) was defined as the first test after treatment termination. ACT clearance was defined as a positive ctDNA test at either CLM or at CLM and 3 months, with subsequent negative ctDNA, in patients without recurrence. An ultrasensitive SV-based assay, Pathlight, was used for ctDNA analysis. The primary outcome was recurrence-free interval (RFI). Results: Of the 377 patients included, 135 patients (36%) received ACT with a median length of 5.2 months (range 0.1-7.4). Single agent fluoropyrimidine was given to 65 patients (48%) and 70 (52%) received an oxaliplatin-based doublet. Analyzable CLM samples were obtained from 106 patients (79%). Nineteen patients (18%) were ctDNA positive at CLM and received ACT. ctDNA positive patients who received ACT still showed a substantially higher risk of recurrence compared to ctDNA negative patients (HR, 38; 95% CI, 15–97). The same was seen in ctDNA-positive patients who did not receive ACT, i.e. clinically low-risk (HR, 35; 95% CI, 14–88). In the ctDNA positive cohort, patients who received ACT versus no ACT had a 2-year RFI of 58% (95% CI:39%-85%) versus 39% (95% CI:22%-69%). Persistent ACT clearance was seen in 8 patients, none of whom recurred. Transient clearance was seen in 6 patients, all of whom recurred. Five patients had persistently positive ctDNA despite ACT; all recurred. Eighty-seven patients (82%) were ctDNA negative at CLM but received ACT; 9 of these recurred (10%). Eighteen patients were ctDNA positive at CLM but did not receive ACT and 13 of these recurred. Conclusions: ctDNA is a strong prognostic biomarker and may be utilized to guide adjuvant treatment decisions in patients with localized CRC. Permanent clearance of ctDNA by administration of ACT is a favorable prognostic feature whereas ctDNA persistence after ACT predicts recurrence. Clinical trial information: NCT04726800 .

Outcomes after recurrence on neoadjuvant chemo-immunotherapy (NCIT) in patients (pts) with high-risk early-stage triple-negative breast cancer (eTNBC).

601 Background: NCIT improves survival in high-risk eTNBC; however, trial-based data show that most relapses after NCIT are early. Real-world data are lacking. Methods: This retrospective study included pts from the DFCI Multicenter TNBC registry with eTNBC or estrogen receptor (ER)-low (≤10%), HER2-negative breast cancer (BC) treated with NCIT who underwent surgery before 7/1/2025. Aims were to evaluate (1) patterns of relapse; (2) BC–specific event–free survival (BC-EFS) from the first NCIT dose to relapse, contralateral BC, or death; (3) first-line (1L) metastatic systemic treatment patterns, and (4) time to progression (TTP). Results: 220 pts were identified, with median age of 50.1 yrs (IQR: 40.5-60.8). Median follow-up (FU) was 32.7 months (mo) (IQR, 30.4–34.4). At last FU, 29 pts (13.2%) had relapsed (locoregional, n=2; distant, n=27). Among these, 86.2% had not experienced pathologic complete response. Among relapsed pts with PD-L1 assessment (n=16), 43.8% (n=7) had a Combined Positive Score (CPS) &lt;10; 56.2% (n=9) had a CPS ≥10. BC-EFS is shown in the Table. 1L therapy consisted of antibody-drug conjugate (ADC) in 51.7% (n=15), Poly ADP-ribose Polymerase inhibitor (PARPi) in 13.8% (n=4), chemotherapy in 13.8% (n=4), ADC+PARPi in 3.4% (n=1), CIT in 3.4% (n=1), and HER2-directed in 3.4% (n=1); 10.3% (n=3) died before 1L therapy. Median (m)TTP in 1L (n=26) was 7.7 mo (95% CI, 6.0–13.0) and detailed in the Table. Among pts with disease-free interval (DFI) ≤6 mo (n=8), mTTP was 5.1 mos (2.7–NR), with TTP rates of 50.0% (25.0–100.0) at 6 mo and 16.7% (2.9–95.3) at 12 mo. In those with DFI 6–12 mo (n=9), mTTP was 9.7 mo (7.4–NR); TTP rates were 88.9% (70.6–100.0) at 6 mo and 25.4% (7.7–83.8) at 12 mo. For DFI &gt;12 mo (n=9), mTTP was 8.6 mo (5.6–NR), with TTP rates of 71.4% (44.7–100.0) at 6 mo and 28.6% (8.9–92.2) at 12 mo. Among pts treated with 1L ADC (n=16), mTTP was 8.6 mo (7.4–NR); TTP rates were 77.9% (58.4–100.0) at 6 mo and 26.0% (9.9–68.3) at 12 mo. In non-ADC-treated pts (n=10), mTTP was 6.2 mo (3.78–NR), with TTP rates of 60.0% (36.2–99.5) at 6 mo and 20.0% (5.8–69.1) at 12 mo. Median overall survival (from diagnosis) among pts with relapse was 26.3 mo (95% CI, 25.0–NR). Conclusions: Relapse after NCIT was predominantly early with poor metastatic outcomes, underscoring an urgent need for new strategies in this population largely excluded from clinical trials. Any NCIT(n = 216) KEYNOTE-522(n = 176) 1L for metastatic TNBC(n = 26) Time from NCIT start (mo) BC-EFS % (95% CI) Event N BC-EFS % (95% CI) Event N Time from 1L start (mo) TTP survival % (95% CI) Event N 6 100.0(100.0, 100.0) 0 100.0(100.0, 100.0) 0 6 70.8(54.6, 91.7) 7 12 95.4(92.5, 98.4) 9 95.0(91.7, 98.4) 8 12 23.6(11.0, 50.5) 17 18 89.4(84.9, 94.0) 19 89.7(84.9, 94.8) 15 18 14.2(5.0, 40.3) 19 24 86.4(81.3, 91.8) 23 86.0(80.3, 92.2) 19 24 9.4(2.5, 35.2) 20 36 79.9(72.2, 88.3) 27 82.2(74.8, 90.3) 21