Browse Articles

Second-line therapeutic strategies following resistance to EGFR-TKIs in advanced EGFR-mutant NSCLC: A systematic review and network meta-analysis.

e20728 Background: After progression on EGFR tyrosine kinase inhibitors (EGFR-TKIs), multiple second-line options exist for targeting all-comers. However, their comparative efficacy, safety, and subgroup-specific characteristics of these therapies remain incompletely defined, complicating individualized treatment selection. Methods: We performed a PRISMA-compliant network meta-analysis of 16 randomized controlled trials (RCTs) (n = 4909) evaluating second-line systemic therapies for EGFRm advanced NSCLC after EGFR-TKI failure. Treatments were categorized as platinum–pemetrexed chemotherapy (Chemo), immunotherapy +Chemo (IC), anti-VEGF+Chemo (VC), IO+anti-VEGF+Chemo (IVC), Amivantamab+Chemo (AC), Amivantamab+Lazertinib+Chemo (ALC), HER3-directed antibody–drug conjugate (HER3-ADC), TROP2-directed ADC (TROP2-ADC), docetaxel and ICI monotherapy. Co-primary endpoints were progression-free survival (PFS) and overall survival (OS). Secondary endpoint was grade ≥3 treatment-related adverse events (TRAEs). Results: Compared with Chemo, most intensified strategies improved PFS. The largest PFS benefits were observed with ALC (HR 0.44, 95% CI 0.35–0.56), AC (HR 0.48, 95% CI 0.36–0.64), IVC (HR 0.52, 95% CI 0.45–0.60), and TROP2-ADC (HR 0.49, 95% CI 0.39–0.62). HER3-ADC and IC yielded more modest PFS gains (both HR 0.77), whereas ICI monotherapy was inferior (HR 1.92, 95% CI 1.27–2.90). OS effects were generally modest and homogeneous; TROP2-ADC was the only biomarker-agnostic regimen with a clear OS advantage versus Chemo (HR 0.60, 95% CI 0.44–0.82), while HER3-ADC remained OS-neutral despite PFS/ORR improvements. Rank analyses consistently placed IVC, Amivantamab-based regimens, and TROP2-ADC in the top efficacy tier. An efficacy–toxicity gradient was evident: grade ≥3 TRAEs were lowest with Chemo/ICI monotherapy, intermediate with VC and ADCs, and highest with Amivantamab-based and IO+anti-VEGF-based regimens. In prespecified PFS subgroup analyses, IVC showed particularly strong and stable efficacy in metastatic populations (brain and liver): brain metastases HR 0.38 (95% CI 0.29–0.52; I² = 0) and liver metastases HR 0.63 (95% CI 0.49–0.81; I² = 0). Across molecular subgroups, Amivantamab-based regimens and IVC remained consistently favorable in 19Del/L858R and post–3-generation TKI settings, while TROP2-ADC preserved benefit in T790M-positive disease. Conclusions: After EGFR-TKI failure, IVC demonstrates particularly strong and stable PFS benefit in patients with brain or liver metastases, supporting phenotype-guided selection. TROP2-ADC provides a biomarker-agnostic therapeutic option with demonstrable OS improvement. These findings provide a basis for second-line treatment decisions and support the conduct of head-to-head trials among high-value regimens.

HRS-4642 combined with adebrelimab in previously treated patients with metastatic pancreatic ductal adenocarcinoma.

4190 Background: Nearly 90% pancreatic ductal adenocarcinoma (PDAC) tumors were driven by RAS mutations, including KRAS G12D (~40%). The high prevalence of KRAS G12D mutations is thought to play a vital role in the universally immunosuppressive tumor microenvironment of PDAC. HRS-4642 is a highly selective KRAS G12D inhibitor. Here, we report preliminary results of HRS-4642 combined with adebrelimab (an anti-PD-L1 antibody) in patients (pts) with KRAS G12D-mutant PDAC. Methods: The study comprised two parts. Phase 1 (Dose escalation): A “3+3” design was used to determine the recommended phase 2 dose (RP2D) of HRS-4642 (4 dose levels: 300, 400, 500mg, qw; and 500mg D1/1200mg D8, q3w) combined with fixed-dose adebrelimab (1200mg iv, Day 1, q3w). Phase 2 (Dose expansion): Simon's two-stage optimal design was applied, requiring at least five objective responses. The primary endpoints were safety and RP2D (phase 1), and objective response rate (ORR; phase 2). The key secondary endpoints included progression-free survival (PFS), overall survival (OS), and disease control rate (DCR). Results: As of Dec 31, 2025, 48 pts were enrolled, with 87.5% having received ≥2 prior lines of therapy. No dose-limiting toxicity (DLT) occurred during phase 1, and the RP2D was determined as HRS-4642 (500mg D1/1200mg D8, q3w) with adebrelimab. Treatment-related adverse events (TRAEs) occurred in 47 pts (97.9%). The most common grade 3/4 TRAEs included lipase increased (16.7%), gamma-glutamyl transpeptidase increased (12.5%), and blood cholesterol increased (8.3%). No treatment related deaths occurred. Among 43 pts with post-baseline tumor assessment, the confirmed ORR was 41.9% and DCR was 83.7%, with a median mPFS of 5.6 months; the median mOS was not reached. Notably, the RP2D group (n=33) achieved a confirmed ORR of 48.5% and a DCR of 90.9%, successfully meeting the primary endpoint. Conclusions: HRS-4642 combined with adebrelimab showed a manageable safety profile and promising anti-tumor activity in metastatic PDAC harboring KRAS G12D mutation. Clinical trial information: NCT06427239 . Efficacy Evaluation. All (n=43) RP2D (n=33) Objective response, n (%) 18 (41.9) 16 (48.5) Best response, n (%) Partial Response 18 (41.9) 16 (48.5) Stable Disease 18 (41.9) 14 (42.4) Progressive Disease 7 (16.3) 3 (9.1) Disease control rate, n (%) 36 (83.7) 30 (90.9) Median progression-free survival, months (95%CI) 5.6 (4.0-7.3) 5.9 (4.2-7.9) Median overall survival, months (95%CI) NR 11.5 (9.2-NR)

MDX-124, a first-in-class annexin-A1 targeting antibody, in patients with locally advanced or metastatic solid malignancies: Data supporting RP2D from the dose escalation stage of a first-in-human phase Ib trial.

e15112 Background: Annexin-A1 overexpression correlates with tumor proliferation and poor prognosis in cancer making it a novel therapeutic target. MDX-124 is a first-in-class monoclonal antibody that targets annexin-A1. In preclinical models, MDX-124 significantly inhibits tumor growth and metastasis, induces antibody-dependent cellular cytotoxicity and has demonstrated synergy with several chemotherapeutics. Here we report safety and efficacy data from a first-in-human phase Ib study of MDX-124 in patients with advanced or metastatic solid malignancies. Methods: This modular, multi-arm, phase Ib study of MDX-124 is enrolling adult patients with solid tumors likely to overexpress annexin-A1. Module 1 is a single agent dose escalation using a BOIN (Bayesian Optimal Interval) design with an expansion cohort in which MDX-124 is administered Q2W at doses ranging from 1–30 mg/kg with a 21-day DLT period. The primary objective is to determine the recommended phase 2 dose (RP2D) of MDX-124 as monotherapy. Secondary objectives are to evaluate safety, tolerability, pharmacokinetics, and preliminary anti-tumor activity of MDX-124. Responses are evaluated according to RECIST v1.1. Adverse events are graded using CTCAE v5.0. Tumor and immunological biomarkers are exploratory endpoints. Module 2 will assess MDX-124 as monotherapy and in combination with SOC therapies in indication-specific arms using a standard 3+3 design. Results: At data cut-off, 21 patients received MDX-124 at doses ranging from 1–30 mg/kg (mean age 60 years, 10 primary tumor types and median of 3 prior treatment lines). A disease control rate of 55% (6/11) was observed in evaluable patients, including 3/4 cholangiocarcinoma patients with one confirmed partial response. No grade 3-4 treatment related adverse events (TRAEs) or DLTs were reported. TRAEs were reported in 62% patients with fatigue and nausea being most common (33%). PK data show dose proportional increase in serum concentration of MDX-124 across 1-30 mg/kg cohorts. Conclusions: Data indicate promising safety, tolerability, and anti-tumor activity of MDX-124. The RP2D for MDX-124 monotherapy was determined to be 20 mg/kg Q2W. Module 2 is currently enrolling. Clinical trial information: ISRCTN78740398.

Graphene goes dental: Unveiling a new era in implants, regeneration, and therapeutics

Spatially Strengthened Ion‐Dipole Electrolyte Enables High‐Temperature Anode‐Free Sodium Batteries

ABSTRACT Practical anode‐free sodium batteries (AFSBs) require high‐temperature adaptability, e.g., stable operation at 45°C. However, current AFSBs are usually confined to room/low temperatures, and pouch cell‐level AFSBs capable of stable cycling >25°C have rarely been reported. Here we report high‐temperature AFSBs via spatially strengthened ion‐dipole electrolyte chemistry. Specifically, a novel solvent, namely, ethyl tetrahydrofurfuryl ether (ETFE) is designed. The reversely anchored rigid cyclic head endows ETFE molecule with weakened steric hindrance effect and stronger cyclic ethereal O─Na + interaction accordingly, hence spatially strengthening overall ion‐dipole interaction. Consequently, less vulnerable free solvents, ameliorated electrolyte decomposition, and formation of stable electrode/electrolyte interphases enable highly reversible Na plating/stripping behaviors at elevated temperatures. Further, an ampere hour (Ah)‐level pouch cell capable of 200 cycles at 45°C with a capacity retention of 89.1% is realized even after initial 300‐cycle ageing at 25°C, featuring the first long‐term HT evaluation toward pouch cell‐level AFSBs. This work should expedite the practicability of AFSBs.

Geodemographic trends and disparities in liver cancer and sepsis-related mortality among adults aged ≥ 45 years in the United States, 1999-2024: A CDC WONDER analysis.

e16163 Background: In the United States, Liver Cancer (LC) is emerging as one of six major cancers with increasing incidence rates. LC is linked to increased sepsis risk through weakened immune response, and chronic alcohol use further predisposes to progression of the disease and weak immune system. The combined burden of these conditions is poorly comprehended and is associated with increasing mortality. To overcome this gap, this study analyses national mortality trends in relation to geodemographic variables. Methods: To analyze mortality associated with LC (C22) and sepsis (A40-A41), we utilized the CDC WONDER database, to extract geo-demographic data among adults aged ≥45 years. Age-Adjusted Mortality rates (AAMRs) were standardized per 1 million. Joinpoint regression software evaluated Average Annual Percentage Change (AAPC) with 95% confidence intervals (CI). Results: LC and Sepsis accounted for 26,320 deaths over the study period, with AAMRs rising between 1999 and 2024 (AAPC: 5.1*; 95% CI: 4.9-5.3). Despite lower AAMRs, females exhibited a more pronounced increase in mortality relative to males (AAPC: 5.2*; 95% CI: 5.0-5.4). Older adults (65-85+ years) exhibited the most significant rise in mortality (AAPC: 5.3; 95% CI: 4.7-5.9), as compared to adults (44-65 years). Mortality rose among all racial groups, with the sharp increase observed in Non-Hispanic (NH) White populations (AAPC: 4.7*; 95% CI: 3.4-6.1). Similarly, among the census regions, most significant increase was observed in the Midwest (AAPC: 5.8*; 95% CI: 5.3-6.2). Urban-rural disparities were apparent, with steeper increases in nonmetropolitan regions. Conclusions: Our analysis revealed distinct patterns in LC and sepsis-related mortality. Increased risk among older adults, females, and NH White population, predominantly in rural areas in the Midwest census region, justifies the need for focused screening of premalignant liver diseases and enhanced treatment aligned with AASLD guidelines. These trends also establish the need for strict adherence to antibiotic stewardship, Fast blood culture and sensitivity tests for early infection control, and subsequent research aimed at optimizing disease management. Trends in liver cancer– and sepsis-related mortality in U.S. Variable AAMR 1999 AAMR 2024 AAPC (95% CI) Overall 4.2 12.9 5.1*(4.9-5.3) Gender:FemaleMale 2.96 8.917.7 5.2*(5.0-5.4)4.9*(4.6-5.2) Age:45-64 years65-85+ years 2.27.9 625 4.4*(4.0-4.8)5.3*(4.7-5.9) Race:HispanicNH AsianNH WhiteNH Black 11.410.53.95.7 19.919.111.918 3.5*(2.9-4.1)1.1*(0.5-1.7)4.7*(3.4-6.1)4.3*(3.8-4.8) Census region:NortheastMidwestSouthWest 4.33.14.35.5 10.610.613.815.4 3.7*(3.2-4.2)5.8*(5.3-6.2)5.5*(5.1-5.9)4.7*(4.4-5.1) Urbanization (1999-2020):MetroNonmetro 1999 4.62.9 2020 10.59.6 4.1*(2.8-5.4)6.4*(5.7-7.1) *Indicates p< 0.05

Real-world treatment patterns and mortality associated with GLP1-RAs in breast cancer patients.

629 Background: GLP1 Receptor Agonists (GLP1) are widely used for type 2 diabetes (DM2) and obesity in adults. However, data on modern utilization and associated health outcomes, including all-cause mortality, in breast cancer (BC) patients is limited. We evaluated these endpoints in a BC population from a large US national healthcare database. Methods: Using Truven MarketScan, a commercial insurance claims database which includes over 270 million patients, we extracted eligible invasive BC patients diagnosed between 2014 and 2023 who were ≥ 18 years old, completed definitive BC surgery (index date), received GLP1 for at least 3 continuous months, and had available mortality data. Overall survival (OS) was defined as the time from BC surgery to death from any cause; patients alive at the time of analysis or lost to follow-up were censored at last contact date. Propensity score matching (1:1 ratio) was used to match patients who received GLP1 with those who did not, based on clinical prognostic factors (age; sex; deyo comorbidity index (CI); receipt of immunotherapy, chemotherapy, adjuvant endocrine therapy; DM2; weight category; stroke; myocardial infarction (MI)/coronary artery disease (CAD); heart failure); Kaplan-Meier estimates and log-rank tests were used to compare OS between groups and hazard ratios (HR) were estimated using multivariable Cox proportional hazards models. Results: In 137,493 BC patients, 7,249 (5.3%) received a GLP1 and over half (n=4,019; 55.4%) initiated it after BC surgery. The median follow up time in GLP1 users was 32 (IQR 15-57) months. In GLP1 users the median age was 59 (IQR 53-64) years; 5,860 (80.8%) had DM2 and 1,825 (25.2%) had a CI ≥3; in those with weight data available (n=4,687), 4,609 (98.3%) were in the overweight or obese category. A total of 3,828 (52.8%) received GLP1 concurrent with anti-cancer systemic therapy (chemotherapy, immunotherapy, or adjuvant endocrine therapy). The median duration of GLP1 use was 2.1 (IQR 0.8-6.1) years. GLP1 use increased over time (3.2% in 2014, 5.2% in 2018, 7.2% in 2023). Semaglutide was the most common agent prescribed across all years (35.8%) and exhibited the greatest annual increase in use after 2018. In the propensity score matched cohort, OS was significantly higher in GLP1 users vs non-users (n=7,248 per group): 5 year OS 95.8% GLP1 vs. 89.5% non-users (adjusted HR 0.41, 95% CI: 0.34-0.49, p<0.001). Conclusions: We describe one of the largest observational studies on real-world GLP1 prescription patterns and associated mortality in a BC cohort. Use of these agents is increasing over time in BC patients. Our research supports other emerging epidemiologic data suggesting a potential all-cause mortality benefit with these incretin mimetics in this population. Prospective studies are warranted to clarify the biological mechanisms underlying this association, including metabolic health implications and effects on BC tumor biology.

Premature cancer mortality in the United States, 1999–2023: Analysis of CDC WONDER data.

11042 Background: Premature cancer deaths (defined as mortality in adults aged 20-69 years) represent significant years of potential life lost (YPLL) and substantial disease burden, yet comprehensive national trends remain insufficiently characterized. Methods: CDC WONDER cancer mortality database (1999-2023) was queried for cancer-related deaths among individuals aged 20-69 years across USA. Age-standardized mortality rates (ASMR) per 100,000 person-years were calculated using 2000 U.S. Census standard population. Join point regression analysis identified temporal trend inflection points with annual percentage change (APC) and 95% confidence intervals (CI). Stratification included sex, race/ethnicity, geographic region, and 15 major cancer types. Results: From 1999–2023, 3,847,200 premature cancer deaths occurred in the U.S. (mean 153,888/year). Annual deaths increased 12.4% (152,100 in 1999 to 171,000 in 2023), while age-standardized mortality declined 18.6% (ASMR 87.3 to 71.1 per 100,000). Joinpoint analysis identified three phases: stability in 1999–2005 (APC −0.42%), accelerated decline in 2005–2015 (APC −2.18%), and slower improvement in 2015–2023 (APC −0.73%). Males had higher mortality than females in 1999 (ASMR 105.4 vs 69.2), with declines in both sexes (APC −1.89% males; −2.34% females), narrowing the male–female gap from 52.2% to 37.8% by 2023. Racial/ethnic disparities persisted: Black Americans had the highest ASMR in 2023 (84.3), 18.6% higher than Whites (71.1), with slower declines (APC −1.23% vs −2.07%); Hispanics/Latinos had the lowest ASMR (52.8) and fastest declines (APC −3.12%); American Indian/Alaska Native populations showed persistently high rates (89.2) with minimal change (APC −0.18%). In 2023, leading contributors were lung (38,200 deaths; ASMR 17.5; APC −3.78%), breast (23,100; 10.5; −2.45%), colorectal (11,800; 5.3; −3.12%), pancreatic (8,400; 3.8; −0.28%), and ovarian cancer (6,200; 2.8; −1.86%), together accounting for 58.2% of premature deaths. Cervical cancer declined sharply (APC −4.23%), while melanoma increased (APC +2.15%). Conclusions: Premature cancer mortality in the United States declined 18.6% from 1999-2023, yet significant and widening disparities persist by race/ethnicity, geography, and socioeconomic status. Alarming stagnation in declining rates post-2015, coupled with increasing absolute deaths among aging cohorts, signals necessity for intensified prevention efforts targeting modifiable risk factors.

Neoadjuvant disitamab vedotin plus carboplatin in HER2-expressing advanced ovarian cancer: A prospective, single-arm, phase II trial.

5556 Background: Standard neoadjuvant paclitaxel-carboplatin for advanced ovarian cancer causes significant neurotoxicity impacting quality of life. Disitamab vedotin (RC48), an anti-HER2 antibody-drug conjugate, demonstrated efficacy and favorable tolerability in HER2-expressing solid tumors. We evaluated RC48 plus carboplatin as neoadjuvant therapy in HER2-expressing advanced ovarian cancer. Methods: This single-arm phase II trial (NCT06574763) enrolled patients with newly diagnosed FIGO stage III-IV high-grade serous ovarian, fallopian tube, or primary peritoneal cancer expressing HER2 (IHC 1+ or ERBB2 amplification) who were ineligible for primary cytoreductive surgery. Treatment comprised carboplatin (AUC 5, day 1, Q3W) alone in cycle 1, then RC48 (2.5 mg/kg, day 1, Q3W) plus carboplatin for cycles 2-4. A 6-patient safety run-in determined optimal RC48 dosing. Interval debulking surgery (IDS) was performed 14-28 days after cycle 4. Primary endpoint: complete resection rate (CRR, defined as no macroscopic residual disease at IDS). Secondary endpoints: pathological complete response (pCR), objective response rate (ORR per RECIST v1.1), progression-free survival, overall survival, and safety (CTCAE v5.0). Results: From October1st, 2024 to December 31st, 2025, 33 patients were enrolled.; 25 completed neoadjuvant therapy and underwent IDS. Among 24 evaluable patients (1 excluded for mixed histology meeting exclusion criteria), median age was 57 years (range 42-71); all had stage IIIC (n=18, 75%) or IVB (n=6, 25%) disease. The regimen achieved exceptional surgical outcomes: CRR 91.6% (22/24), with 2 patients having minimal residual disease <1 cm. ORR was 87.5% (21/24). pCR rate was 4.2% (1/24). Robust tumor regression was observed: median residual viable tumor in primary lesions 22% (range 0-80%), in omental implants 44% (range 0-80%). With median follow-up of 5.0 months (range 2.3-10.6), no disease recurrence occurred. Grade ≥3 treatment-related adverse events occurred in 29.2% of patients, primarily hematological toxicities (neutropenia, thrombocytopenia) that resolved to ≤grade 2 within one week with supportive care. Critically, no patient (0/27, 0%) experienced peripheral neuropathy of any grade. Conclusions: Neoadjuvant disitamab vedotin plus carboplatin demonstrated exceptional surgical outcomes with 91.6% complete resection rate and complete elimination of neurotoxicity in HER2-expressing advanced ovarian cancer. The regimen induced substantial pathological regression with a favorable safety profile. These results support RC48-carboplatin as a promising neurotoxicity-sparing alternative to standard paclitaxel-based therapy. Updated survival data will be presented. Clinical trial information: NCT06574763 .

A phase 3 randomized, double-blind, placebo-controlled study of a pasritamig plus best supportive care in metastatic castration-resistant prostate cancer.

TPS5136 Background: Metastatic castration-resistant prostate cancer (mCRPC) remains an incurable disease with high morbidity and a median overall survival of 2 years. Human kallikrein 2 (KLK2) is highly and specifically expressed in normal and malignant prostate tissue, including late stage mCRPC. Pasritamig is a humanized, IgG1-based bispecific antibody that targets KLK2-expressing cells via CD3 engagement, inducing T cell-mediated targeted cytotoxicity. In a first-in-human study in participants with mCRPC refractory to standard therapies, pasritamig demonstrated a favorable safety profile (< 5% grade 3 or higher treatment-related adverse events and < 10% cytokine release syndrome, all grade 1) and encouraging efficacy at the recommended phase 2 dose with every 6-week (Q6W) outpatient dosing. Methods: KLK2-comPAS is a double-blind, randomized, global phase 3 study to evaluate the safety and efficacy of pasritamig in participants with mCRPC who have previously received all available and suitable life-prolonging therapies including androgen receptor pathway inhibitor, taxane chemotherapy, radioligand therapy (RLT), and poly(ADP-ribose) polymerase inhibitors (for patients with BRCA mutations). Other key inclusion criteria include ECOG performance status 0–2, PSA ≥2 ng/mL, adequate organ function, and ongoing androgen deprivation therapy or prior orchiectomy. Key exclusion criteria include visceral metastases, prior KLK2- or CD3-directed therapies, and active autoimmune disease requiring systemic immunosuppression. Approximately 663 patients will be randomized in a 2:1 ratio to receive pasritamig plus best supportive care (BSC) or placebo plus BSC, stratified by prior PSMA-targeted RLT, number of prior taxanes, and ECOG performance status. BSC may include palliative external beam radiation, low dose steroids, pain medications, bone protective agents, and needed palliative procedures. Pasritamig is administered outpatient at a target dose of 300 mg IV Q6W with two step-up doses of 3.5 mg IV on cycle 1 day 1 and 18 mg IV on cycle 1 day 8. Pre-medications (dexamethasone, diphenhydramine, acetaminophen) will be administered for step-up and first target dose. Participants will receive treatment until confirmed radiographic progression or unequivocal clinical progression, intolerable toxicity, withdrawal, or death. The primary endpoint is overall survival. Key secondary endpoints include radiographic progression-free survival (rPFS), PFS (including clinical progression and unconfirmed bone progression), time to symptomatic progression, and time to skeletal-related events. Enrollment commenced in September 2025 and remains ongoing. Clinical trial information: NCT07164443 .

Impact of early intracranial hemorrhage on survival and critical care utilization in acute promyelocytic leukemia: A population-based propensity-matched study.

e18539 Background: Intracranial hemorrhage (ICH) remains a leading cause of early death in acute promyelocytic leukemia (APL), but the extent to which excess risk persists among survivors, and the longer-term critical care burden in this population, are not well known. Methods: Using the TriNetX multi institutional network (2014-2025),we compared APL with ICH within 14 days of initial diagnosis of APL versus APL without ICH. Cohorts were 1:1 propensity-score matched (PSM) for key clinical and laboratory variables. Co-primary outcomes were all-cause mortality and ICU admission-rate (ICUAR; critical care procedure-based definition) at 30 days, 90 days, 1 year, and 5 years, analyzed with risk estimates and Kaplan–Meier/Cox proportional hazard models. Landmark analyses were performed among patients alive at 30 days (n=425/group), 6 months (n=350/group), and 1 year (n=332/group) to delineate the long-term impact on survivors. Results: We identified 16,301 patients with APL, including 789 patients with early ICH (4.8%) and 15,512 without ICH. Before matching, median age was 55.1 (range 37.0-73.2) versus 54.9 years (range 34.9-74.9), and female sex comprised 49% versus 48% of cohorts, respectively. After 1:1 PSM, 776 patients were included per group. Estimated mortality was higher in the early-ICH cohort versus no-ICH at 30 days (32% vs 6%; HR 6.1), 90 days (41% vs 11%; HR 5.0), 1 year (54% vs 25%; HR 3.1), and 5 years (64% vs 39%; HR 2.5). ICUAR was likewise higher at 30 days (42% vs 7%; HR 5.10), 90 days (45% vs 10%; HR 5.9), 1 year (48% vs 16%; HR 4.5), and 5 years (51% vs 21%; HR 3.9) (all p<0.001). Among 30-day survivors, estimated 1-year mortality remained higher (21% vs 11%) and ICUAR remained higher (39% vs 12%). Among 6-month survivors, 5-year mortality was similar (16 vs 15%), but ICUAR remained higher (40% vs 14%). Among 1-year survivors, 5-year mortality was lower in the prior-ICH cohort (10% vs 17%) consistent with probable survivorship bias, but ICUAR remained higher (38.0% vs 16.0%). Conclusions: Early ICH in APL is associated with significant early mortality and persistent increase in ICU utilization beyond the first year. As survivorship landmarks advance, the effect of ICH on mortality attenuates, but the critical care burden persists indicating ongoing morbidity among ICH survivors. Our study highlights the need to optimize early diagnosis and tailored management as well as downstream supportive care. Analysis/Window Matched n per group KM Estimated mortality (%) (ICH vs no ICH) ICU admission (%) (ICH vs no ICH) Diagnosis to 30 days 776 32 vs 6 42 vs 7 Diagnosis to 90 days 776 41 vs 11 45 vs 10 Diagnosis to 1 year 776 54 vs 25 48 vs 16 Diagnosis to 5 years 776 64 vs 39 51 vs 21 Landmark analyses Alive at 30 days: 1-year rate 425 21 vs 11 39 vs 12 Alive at 30 days: 5-year rate 425 35 vs 27 43 vs 16 Alive at 6 months: 5-year rate 350 16 vs 15 40 vs 14 Alive at 1 year: 5-year rate 332 10 vs 17 38 vs 16

Pre-onset increases in neutrophil-to-lymphocyte ratio as an identifier of mild and severe immune checkpoint inhibitor–related pneumonitis in lung cancer.

e24188 Background: Immune checkpoint inhibitor (ICI)–related pneumonitis is an uncommon but potentially life-threatening toxicity. Reliable biomarkers that predict pneumonitis prior to clinical onset are lacking. We evaluated whether dynamic changes in the neutrophil-to-lymphocyte ratio (NLR) preceding pneumonitis onset are associated with severity. Methods: Lung cancer patients treated with ICI-based therapy between October 2014 and December 2025 were retrospectively identified from an institutional pharmacy database. Pneumonitis events were graded per CTCAE v6.0 and categorized as mild (grade 1–2) or severe (grade ≥3). Serial complete blood counts were used to calculate baseline NLR and longitudinal pre-onset NLR metrics, including maximum NLR, absolute and percent change, and NLR slopes over 30- and 60-day intervals. Analyses were conducted at the pneumonitis-event level. Group comparisons used Wilcoxon rank-sum tests. Discriminatory performance was assessed using receiver operating characteristic (ROC) analysis, and associations with severe pneumonitis were evaluated using logistic regression. An optimal NLR slope threshold was identified by ROC analysis to define pre-event inflammatory acceleration. Results: Among 450 ICI-treated patients, 31 patients experienced 38 pneumonitis events (21 mild, 17 severe). Tumor types: adenocarcinoma/squamous cell carcinoma/small cell lung cancer/other malignancies 55%/26%/13%/6%. Median age at diagnosis: 72 years (IQR 65–76). Male: 55%. Race/ethnicity: Non-Hispanic White/non-White 77%/23%. Baseline NLR did not differ between mild and severe pneumonitis (median 3.99 vs 4.76; p = 0.21). In contrast, severe pneumonitis was associated with higher maximum NLR, greater absolute and percent increases in NLR, and steeper pre-event NLR slopes (all p < 0.01). The 30-day pre-event NLR slope was significantly higher in severe compared with mild pneumonitis (median 0.22 vs 0.02 units/day; p = 0.001). ROC analysis demonstrated good discrimination for severe pneumonitis (AUC 0.80) and identified a slope threshold of 0.04 units/day, which detected pre-event NLR acceleration in 94.1% of severe pneumonitis with a median of 15.5 days (IQR 37.3) and in 52.4% of mild pneumonitis with a median of 16.0 days (IQR 28.5) prior to clinical diagnosis. Each 0.1-unit/day increase in the 30-day NLR slope was associated with increased odds of severe pneumonitis (OR 1.68, 95% CI 1.15–3.18). Conclusions: Short-term pre-onset increases in NLR, particularly 30-day slopes, identify patients at increased risk for severe ICI-related pneumonitis. Longitudinal NLR monitoring may support earlier risk stratification and proactive clinical management, warranting prospective validation.

Associations between indoor and outdoor air pollution worry and awareness and acceptability of multi-cancer early detection testing: Evidence from a nationally representative US survey.

e22521 Background: Public concern regarding the carcinogenic effects of air pollution has intensified, shaping perceptions of cancer risk and prevention. Multi-Cancer Early Detection (MCED) blood tests, which aim to detect multiple malignancies from a single blood draw, represent a promising frontier in early detection. However, it remains unclear whether environmental worry particularly perceived health threat from indoor and outdoor air pollution translates into greater awareness of MCED testing or greater perceived value of these emerging technologies. Methods: We analyzed nationally representative data from U.S. adults aged ≥18 years using the 2024 Health Information National Trends Survey (HINTS 7; N = 6,597), restricted to respondents with complete data on air pollution worry and MCED outcomes. The primary exposures were self-reported worry that indoor and outdoor air pollution would harm health, categorized as little/none versus some/a lot. Outcomes included awareness of MCED tests and perceived value/acceptability of MCED testing. Weighted multivariable logistic regression models were fitted for each outcome, adjusting for demographic and socioeconomic characteristics and behavioral covariates. Results: The weighted sample was 48.9% female; 48.2% were aged ≥50 years; 60.8% were Non-Hispanic White, 17.0% Hispanic, and 11.0% Non-Hispanic Black. Overall, 37.0% reported that indoor air pollution was some/a lot harmful to their health, and 50.0% reported similar worry about outdoor air pollution. MCED awareness was low (16.8%), while perceived value was high (73.5%). Greater indoor air pollution worry (OR 0.89; 95% CI 0.69–1.16; p=0.38) and outdoor air pollution worry (OR 0.95; 95% CI 0.75–1.19; p=0.62) were not associated with MCED awareness. In contrast, both indoor (OR 1.50; 95% CI 1.13–2.00; p=0.006) and outdoor air pollution worry (OR 1.48; 95% CI 1.19–1.85; p=0.001) were associated with higher perceived value of MCED testing. Conclusions: In a nationally representative U.S. sample, air pollution worry was associated with greater perceived value of MCED testing but not greater awareness, suggesting that environmental risk concern may increase receptivity to novel screening technologies even when knowledge remains limited.

Efficacy of sonrotoclax vs pirtobrutinib in post–Bruton tyrosine kinase inhibitor (BTKi) relapsed/refractory (R/R) mantle cell lymphoma (MCL): An indirect comparison.

e19050 Background: Sonrotoclax (BGB-11417-201) and pirtobrutinib (BRUIN MCL trial; NCT03740529) have demonstrated efficacy in separate single-arm post-BTKi R/R MCL studies. In the absence of head-to-head randomized trials, a matching-adjusted indirect comparison (MAIC) was conducted to compare the efficacy of both therapies in a post-BTKi R/R MCL setting. Methods: An unanchored MAIC compared individual patient-level data from the sonrotoclax trial (n=103 [efficacy set]; median OS follow-up, 15.2 months) with aggregate data from the pirtobrutinib trial (n=120 [US Food and Drug Administration efficacy cohort]; median OS follow-up, 9.3 months). Matching covariates were selected based on literature review and clinical expert input. The base-case model maximized covariate adjustment while maintaining an effective sample size (ESS) >50. Sensitivity analyses with adequate sample size (ESS >50) and alternative covariates were conducted to validate the base-case model results. Outcomes included overall response rate (ORR), duration of response (DOR), progression-free survival (PFS), and overall survival (OS). Weighted Cox proportional hazards and logistic regression models were used for time-to-event and binary endpoints, respectively. Hazard ratios (HR) and odds ratios (OR) with 95% confidence intervals (95% CI) were reported. Results: In the base-case analysis, sonrotoclax showed a nonsignificant, numerically higher ORR (OR, 1.54; 95% CI, 0.80-2.97), as well as a numerically longer DOR (HR, 0.75; 95% CI, 0.37-1.54), versus pirtobrutinib. Consistent numerical improvements in the sonrotoclax cohort were observed for PFS (HR, 0.73; 95% CI, 0.48-1.11) and OS (HR, 0.66; 95% CI, 0.37-1.15). Sensitivity analyses using alternative covariate sets generated broadly consistent results (Table). Conclusions: Based on the data currently available, the MAIC findings suggest a nonsignificant trend of potential difference in efficacy outcomes between sonrotoclax and pirtobrutinib. Future studies with longer follow-up are warranted. MAIC results comparing efficacy outcomes with sonrotoclax vs pirtobrutinib in post-BTKi R/R MCL. Model ESS n (%) ORR-IRCOR (95% CI) P -value DOR-IRCHR (95% CI) P -value PFS-IRCHR (95% CI) P -value OSHR (95% CI) P -value Base case a-c 55 (53) 1.54(0.80-2.97).20 0.75(0.37-1.54).43 0.73(0.48-1.11).14 0.66(0.37-1.15).14 Sensitivity analysis 1 a,b 97 (94) 1.15(0.67-1.97).62 0.77(0.44-1.36).37 0.90(0.64-1.25).52 0.92(0.59-1.44).71 Sensitivity analysis 2 a 102 (99) 1.14(0.67-1.94).64 0.76(0.43-1.32).32 0.89(0.64-1.23).48 0.92(0.59-1.43).70 Covariates include: a sMIPI, no. of prior therapy lines; b PD to last BTKi; and c blastoid MCL. IRC, independent review committee; PD, progressive disease; sMIPI, simplified Mantle Cell Lymphoma International Prognostic Index.

Palliative chemotherapy for pancreatic cancer in an integrated healthcare system.

e16371 Background: Palliative systemic therapy for pancreatic ductal adenocarcinoma (PDAC) aims to extend survival and alleviate symptoms, but real-world benefits can vary across patient subgroups, particularly older adults. We evaluated overall survival (OS) associated with receipt of palliative chemotherapy within an integrated healthcare system. Methods: We conducted a retrospective analysis of patients with PDAC treated in an integrated health network. Patients were categorized by receipt of any palliative chemotherapy versus no chemotherapy. The primary endpoint was mean OS from the index date of palliative intent (or comparable reference date for non-treated patients). A prespecified subgroup analysis assessed outcomes among patients aged ≥70 years. Statistical comparisons were performed using two-sided tests with significance set at p < 0.001. Results: In the overall cohort, palliative chemotherapy was associated with longer survival compared with no chemotherapy (mean OS 12.9 months vs 5.2 months, p < 0.001). In the subgroup aged ≥70 years, the pattern differed: patients who received palliative chemotherapy had shorter mean OS compared with those who did not (3.2 months vs 9.8 months, p < 0.001). These findings suggest potential age-related differences in treatment effect or selection factors in clinical practice. Conclusions: Within this integrated healthcare system, palliative chemotherapy for PDAC was associated with improved mean OS overall, but not among patients aged ≥70 years, in whom shorter survival was observed with treatment. The contrasting outcomes underscore the need for careful patient selection, optimization of supportive care, and consideration of geriatric assessment when making palliative treatment decisions for older adults. Future work should incorporate performance status, comorbidity burden, regimen intensity, and goals-of-care discussions to clarify drivers of benefit and harm in this population.

Comprehensive genomic profiling of patients with cancer of unknown primary (CUP).

e15188 Background: Carcinoma of unknown primary (CUP) remains a clinical entity with limited standardized treatment options. However, recent studies have demonstrated that broad next-generation sequencing (NGS) panels provide clinically meaningful insights, enabling identification of actionable genomic alterations, refinement of tissue-of-origin hypotheses, and improved therapeutic stratification. A substantial proportion of CUP tumors harbor potentially targetable alterations, and a subset exhibits high tumor mutational burden (TMB-H) and/or high microsatellite instability (MSI-H) status. In this study, we assessed the clinical utility of comprehensive genomic profiling panels in CUP cases. Methods: Overall, 103 patients were analysed, including 74 tissue-based and 29 liquid biopsy cases. Matched leukocyte DNA was used to exclude clonal hematopoiesis–related variants. Targeted-capture NGS analysis was performed using two CE-IVD GenePlus assays covering 1021 cancer-related genes and 38 fusion genes, with integrated assessment of TMB and MSI. Sequencing was carried out on an MGI sequencing platform (DNBSEQ-T7). Results: NGS analysis revealed that 29% (30/103) of patients derived clinical benefit from pan-cancer biomarkers. Specifically, BRAF V600 mutations were detected in 5 cases, indicating sensitivity to BRAF/MEK inhibitors. TMB ≥10 muts/Mb was observed in 19 tissue samples, while TMB ≥16 muts/Mb was identified in 4 liquid biopsy samples, both suggesting eligibility for immunotherapy. Additionally, 2 cases were also MSI-H, supporting potential response to immune checkpoint inhibitors. Genomic alterations with potential relevance for off-label targeted therapy were identified in 30% of patients. The most frequently mutated gene was KRAS (18 pts), indicating potential sensitivity to KRAS/MEK inhibitors. Additional alterations were observed in PIK3CA/PTEN, ERBB2, IDH, FGFR2 (fusion), and homologous recombination deficiency–related genes (BRCA2, ATM, BAP1), suggesting possible responsiveness to PI3K/AKT/mTOR inhibitors, ERBB2-directed therapies, IDH, FGFR and PARP inhibitors, respectively. Finally, 29% of the patients were considered eligible for clinical trial enrolment, based on their molecular profiles. Conclusions: Therapeutically relevant molecular profiling results were identified in 59% of CUP patients. These findings support the integration of molecular profiling into routine CUP evaluation, facilitating biomarker-driven precision oncology strategies aimed at improving patient outcomes.

Outcomes of atypical <i>EGFR</i> mutations in early-stage <i>EGFR</i> -mutated non-small cell lung cancer (NSCLC).

8034 Background: Adjuvant osimertinib (osi) improves survival outcomes in patients (pts) with early-stage EGFR -mutated NSCLC harboring the classical EGFR exon 19 deletions (ex19del) or L858R mutation after surgery or chemoradiotherapy. Pts with early-stage NSCLC harboring atypical EGFR mutations were excluded from the respective clinical trials. Outcomes in this pt population remain unknown. Methods: We conducted a multi-institutional retrospective analysis of pts with stage IA-IIIC EGFR -mutated NSCLC (2010-2024) who were classified into two groups: those with tumors harboring classical EGFR mutations (c- EGFR : ex19del or L858R) or atypical EGFR mutations (a- EGFR : non-ex19del or L858R). Kaplan-Meier analyses compared disease-free survival (DFS), central nervous system (CNS) DFS, and overall survival (OS). Multivariable Cox regression adjusted for disease stage, definitive therapy, and adjuvant therapy. Results: Among 417 pts, 293 had tumors harboring c- EGFR and 124 had a- EGFR mutations. Baseline characteristics were similar between groups (Table). Median follow-up was 4 years. Among pts not treated with adjuvant osi (n = 346), a- EGFR was associated with a significantly lower 3-year DFS compared to c- EGFR (52.5% vs 71.2%; HR 1.71, 95% CI 1.21-2.42; P = 0.003). Compared to c- EGFR , a- EGFR had higher rates of recurrence to the ipsilateral lung (24% vs 14%) and bone (12% vs 6%). A- EGFR was associated with a significantly lower 3-year CNS DFS (81.3% vs 87.1%; P = 0.002) and 5-year OS (77.7% vs 85.6%; P = 0.007) compared to c- EGFR . Among pts with a- EGFR mutations, EGFR exon 20 insertions (ex20ins) were associated with a significantly reduced 3-year DFS compared to other alterations (35.9% vs 68.2%; HR 2.04, 95% CI 1.10-3.77; P = 0.013), and with a lower 3-year CNS DFS (70.9% vs 92.2%; P = 0.028) and trend towards lower 5-year OS (71.3% vs 84.4%; P = 0.141). Among pts who underwent surgery for stage IB-IIIA NSCLC (n = 209), adjuvant osi improved 3-year DFS among c- EGFR (84.2% with osi vs 63.4% without osi) whereas a- EGFR had a lower 3-year DFS (36.4%; P &lt; 0.001), which was driven by EGFR ex20ins compared to other alterations (19.9% vs 59.2%; P = 0.003). Conclusions: Atypical EGFR mutations are associated with inferior survival outcomes compared to classical EGFR mutations in early-stage NSCLC. EGFR ex20ins appear to drive poor outcomes among the atypical EGFR mutations. Novel adjuvant therapy strategies are warranted to improve definitive treatment of early-stage atypical EGFR -mutated NSCLC. Classical EGFR , N=293 (%) Atypical EGFR , N=124 (%) Ex19del / L858R 46 / 54 - Ex20ins / G719X / L861Q / Other - 49 / 26 / 17 / 8 Stage IA / IB / II / III 35 / 23 / 17 / 25 38 / 14 / 23 / 25 Surgery / XRT / CRT 81 / 8 / 8 82 / 8 / 7 Neoadjuvant Chemo / EGFR TKI 4 / 3 7 / 2 Adjuvant Chemo / EGFR TKI / Osi 26 / 25 / 23 26 / 8 / 4 3-year DFS* 71.2 52.5 3-year CNS DFS* 87.1 81.3 5-year OS* 85.6 77.7 *In non-adjuvant osi cohort (n=346). Statistically significant difference.

First-in-class PD-1/IL-2 ^α-bias bispecific antibody IBI363 (TAK-928) in patients (pts) with advanced immunotherapy-resistant non–small cell lung cancer (NSCLC): Updated results from a phase I study.

2618 Background: IBI363 is a first-in-class PD-1/IL-2 α-bias bispecific antibody fusion protein designed to block the PD-1/PD-L1 pathway and simultaneously activate the IL-2 pathway. IBI363 selectively expands and rejuvenates exhausted tumor-specific T cells by cis-activating IL-2 receptors. It has shown a manageable safety profile and encouraging efficacy in NSCLC (No. 8509, 2025 ASCO). The updated findings are reported here with a longer follow-up. Methods: Pts with advanced NSCLC who had progressed on standard therapy were enrolled and received IBI363 at dose levels of 0.002/0.01/0.3/0.6 mg/kg every week, 0.3/0.6/1 mg/kg Q2W or 1.5/2/3/4 mg/kg Q3W. Endpoints included safety, ORR, DCR, DOR and PFS assessed by investigator per RECIST v1.1, and OS. Results: As of Nov 20, 2025, 136 pts with NSCLC were enrolled (prior treatment lines ≥2 72.1%; including 67 with squamous NSCLC [sqNSCLC] and 66 with adenocarcinoma). The median follow-up was 14.4 months (mos, range 0.8─30.6). Both the ORR and DCR in sqNSCLC or adenocarcinoma were consistent with those reported previously. In pts with sqNSCLC at 1/1.5 mg/kg (n=28) and 3 mg/kg Q3W (n=31) (58/59 with prior IO therapy), median PFS was 5.5 (95% CI 1.5, 8.3) and 10.1 mos (6.0, 14.0), median OS was 12.5 (7.6, 22.2; maturity 71.4%) and 18.2 mos (10.7, not calculable [NC]; maturity 48.4%), with 24-month OS rate of 29.9% (14.2, 47.3) and 47.8% (28.7, 64.7), respectively. Among those who had a confirmed complete response (n=1) or partial response (PR, n=17), the overall DOR was 10.3 mos (7.0, 13.4; maturity 72.2%). In pts with EGFR wt adenocarcinoma at 0.6/1/1.5 mg/kg (n=30) and 3 mg/kg Q3W (n=25) (all with prior IO therapy), median PFS was 2.7 (1.4, 5.1) and 4.2 mos (3.0, 7.0), median OS was 17.5 (7.1, NC; maturity 56.7%) and 15.2 mos (9.6, NC; maturity 56.0%), with 24-month OS rate of 40.2% (22.2, 57.7) and 42.7% (23.1, 61.0), respectively; the overall DOR reached 21.3 mos (4.0, 21.3; maturity 40.0%) in those achieving a confirmed objective response (n=10, all PRs). Smoking is a major factor affecting treatment efficacy in adenocarcinoma. In smokers (n=31) and non-smokers with adenocarcinoma (n=24), the median OS was 23.4 (11.3, NC; maturity 48.4%) and 11.5 mos (5.6, 19.6; maturity 66.7%), respectively. In the overall population (n=136), treatment-emergent adverse events (TEAEs) occurred in 135 pts (99.3% any grades; 48.5% ≥G3). TEAEs led to treatment discontinuation in 11 (8.1%) pts. TEAEs led to death in 5 pts (3.7%) with only 1 event (0.7%) considered treatment-related (unexplained death). The most common TEAEs were arthralgia (52.2%; 3.7% ≥G3), anemia (46.3%; 4.4% ≥G3), and rash (39.0%; 8.8% ≥G3). Conclusions: IBI363 was well tolerated with a manageable safety profile consistent with earlier reports, and continued to show strong and durable efficacy in pts with heavily pretreated NSCLC. Clinical trial information: NCT05460767 .

Education and decision-making autonomy as drivers of HPV screening outcomes in Nepal.

e22574 Background: Cervical cancer screening is uncommon in Nepal due to reasons including resource scarcity, lack of cancer screening knowledge, and sociocultural barriers to women’s healthcare decision-making. Scalable, low-resource HPV screening strategies may reduce disparities, but implementation determinants influencing screening knowledge, outcomes, and follow-up remain understudied. Methods: We conducted a pilot, community-based HPV screening study at two health fairs in Janakpur City, Nepal, in 2025. Clinician-collected cervical samples from 274 women were tested for high-risk HPV (HrHPV) using the AmpFire isothermal amplification assay. Participants completed a brief survey assessing age, education, prior cervical cancer screening awareness, source of screening information, and healthcare decision-making autonomy. Associations were evaluated using Pearson’s chi-square and Fisher’s exact tests, with stratified analyses adjusting for age and education. Results: Education level significantly influenced prior awareness of cervical cancer (χ2= 9.85, p = 0.02), though it was not significantly associated with procedural knowledge of vaginal sampling in univariate analysis (χ2= 4.31, p = 0.23). Women reporting their husband as the primary healthcare decision-maker had a higher HrHPV positivity rate compared with women reporting autonomous decision-making, but the difference was not significant (14.3% vs 9.6%; χ2= 0.93, p = 0.33). Among the 32 women testing positive for HrHPV, 62.5% completed follow-up Pap testing, with age group variation: under 30 (45.5%), 30-44 yo (70.6%) and 75.0% in those &gt; 45 yo. Only two women refused follow-up due to spousal decision-making, citing family opposition as the reason. Conclusions: Education significantly improved general awareness but not procedural understanding. While restricted autonomy showed a non-significant trend toward higher HrHPV positivity, spousal opposition remained a tangible barrier. Notably, the program achieved 62.5% follow-up retention, proving effective in navigating these social dynamics. Scalable strategies must therefore target procedural misconceptions and actively engage family decision-makers to support comprehensive care. Keywords: Cervical Cancer Screening, Human Papillomavirus (HPV), Nepal, Female Community Health Volunteers (FCHVs), Healthcare Decision-Making, LMICs.

The impact of reduction in intraoperative frozen sections on positive margin findings in the final pathology report.

e12621 Background: In October 2023, intraoperative frozen section (IOFS) was restricted to breast cancer patients after neoadjuvant chemotherapy (NACT) only due to shortage of pathologists at the largest tertiary center in Austria. Here, we investigate the impact of the omission of IOFS on positive margins in the final pathology report. Methods: In this retrospective analysis, breast cancer patients undergoing breast conserving surgery at the University Hospital of Vienna between April 2020 and October 2024 were included. Patients after NACT, those who underwent mastectomy, had skin and/or muscle infiltration or other locally advanced disease were excluded from this analysis. Results: A total of 697 patients, 382 with IOFS before October 2023 and 315 without IOFS thereafter, with a median age of 63 years have been included in this analysis. No differences between IOFS and no-IOFS group were seen regarding median BMI (25.3 vs 25.6 kg/m 2 ), median tumor size (12mm vs 13mm; 1 to 75mm), multifocal disease (39pts (10.2%) vs 34pts (10.8%)), receptor status and grading as well as use of sentinel node removal (340pts (89%) vs 271 (86%)). Intraoperative re-excisions due to frozen resection result or gut feeling of the surgeon differed significantly (145 (38%) in the IOFS vs 41 (13%) in the no-IOFS, p &lt; 0.001) with no difference regarding detection of residual cancer in those specimens (65 (44.8%) vs 16 (39%),p = 0.66). The final pathology result reported a significantly higher rate of positive margins in the no-IOFS group versus the IOFS group (94pts (29.8%) vs 82pts (21.5%), p = 0.011). This could be confirmed in the univariate logistic regression model (OR 1.56, CI1.11-2.2, p = 0.011) and was even more pronounced in the full multivariable logistic regression model (OR 2.06, CI1.36-3.14, p &lt; 0.001) as well as in the final model after backward stepwise model selection based on AIC (OR 2.05, CI1.38-3.09, p &lt; 0.001). Other factors that significantly impacted positive margins in the multivariate regression model were the presence of additional DCIS (OR 2.27, CI1.42-3.67, p &lt; 0.001), lobular cancer (OR 3.29, CI1.56-6.85, p &lt; 0.001) and multifocal disease (OR 3.22, CI1.86-5.58, p &lt; 0.001). Of 610 patients, who underwent sentinel node removal 55pts (16.2%) had a positive SLN in the IOFS and 36pts (13.3%) in the no-IOFS group, respectively (p = 0.2). Conclusions: The cut down on intraoperative frozen section in a tertiary center led to a significant short-term increase in positive margins resulting in 30% of patients in the final pathology report with a doubling in the likelihood when compared to the use of IOFS. This may result in major increase in direct as well as indirect re-operation costs and psychological distress of the patients. Further follow-up is needed to analyze long-term effects of omitting IOFS on oncological outcomes.