Browse Articles

First PROTAC gains FDA approval, bolstering targeted protein degradation and induced proximity ambitions

Perioperative approaches for patients with biliary tract cancer

Assembling a True “Olympic Gel” From over 16 000 Combinatorial DNA Rings

ABSTRACT Olympic gels are an elusive form of soft matter, comprising a 3D network of mechanically interlocked cyclic molecules. In the absence of defined network junctions, the high conformational freedom of the molecules was previously theorized to confer unique mechanical properties to Olympic gels, such as non‐linear elasticity and unconventional swelling characteristics. However, the synthesis of an Olympic gel exhibiting these intriguing features is challenging, since unintended crosslinking and polymerization processes are often favored over cyclization. Here, we report the successful assembly of a true Olympic gel from a library of DNA rings comprising more than 16 000 distinct molecules. Each of these rings contains a unique sequence domain that can be enzymatically activated to produce reactive termini that favor intramolecular cyclization. We characterized the genetic, mechanical, and structural characteristics of the material by next‐generation sequencing, oscillatory rheology, large‐scale computational simulations, atomic force microscopy, and cryogenic electron microscopy. Our results confirm the formation of a stable Olympic gel, which exhibits unique swelling behavior and an elastic response that is exclusively determined by entanglements, yet persists on long time scales. By combining concepts from polymer physics, synthetic biology, and DNA nanotechnology, this new material class provides a flexible experimental platform for future studies into the effects of network topology on macroscopic material properties and its function as a carrier of genetic information in biological and biomimetic systems. This work moreover demonstrates that exotic material properties can emerge in systems with a high compositional complexity that is more reminiscent of biological than synthetic matter.

Green synthesized selenium nanoparticles-based amelioration of Cd toxicity in radish using seed priming

Geranylgeraniol promotes osteoblast differentiation and inhibits osteoclastogenesis through MAPK and nuclear receptor signaling

Real-world outcomes of Black versus non-Hispanic White women with stage IV HER2-low breast cancer treated with trastuzumab deruxtecan at an urban academic center.

e13102 Background: Black women (BW) experience higher rates of advanced breast cancer and breast cancer mortality compared to Non-Hispanic White women (NHW). Trastuzumab deruxtecan (T-DXd) is an antibody-drug conjugate (ADC) approved for HER2-positive, low, and ultra-low metastatic breast cancers. Black patients comprised < 5% of participants in the trials that led to T-DXd approval, leaving gaps in our understanding of response/toxicity in this population. To address this, we aimed to compare baseline characteristics, treatment outcomes, and toxicity profiles between BW and NHW with HER2-low metastatic breast cancer (mBC) treated with T-DXd at a racially diverse, tertiary care institution. Methods: BW and NHW patients with HER2-low mBC treated with T-DXd at Emory University between 2019 and 2024 were retrospectively evaluated. Descriptive statistics were generated for all patient characteristics. Univariate analyses and logistic regression models were used to assess clinical response, duration on therapy, dose modifications, and toxicity by race. Results: Eighty-two women with HER2-low mBC treated with T-DXd were included (BW n = 35 [43%], NHW n = 47 [57%]). Mean age at diagnosis and prior lines of therapy were similar between groups. BW had a higher prevalence of ER/PR-negative tumors compared with NHW (p < 0.006). Median duration of T-DXd therapy was comparable (147 days NHW vs 126 days BW; p = 0.73), as were treatment discontinuation rates (82% vs 89%; p = 0.74). Discontinuation due to progressive disease was numerically higher in BW (80% vs 63%, p = 0.19). Best response outcomes did not differ significantly: stable disease was most common (51.0% NHW vs 51.4% BW), while progressive disease occurred numerically more frequently in BW (40.0% vs 22.5%; p = 0.18). Interstitial lung disease, growth factor use, and non-hematologic toxicities including fatigue, nausea, and constipation were similar by race. There were low and comparable rates of grade 3–4 hematologic toxicities. BW had a trend toward worse overall survival compared with NHW (HR 1.89, 95% CI 0.93–3.83, p = 0.08). Conclusions: Our institutional data demonstrate no statistically significant differences in treatment response, discontinuation rates, or toxicities between NHW and BW with HER2-low mBC treated with T-DXd, although BW showed numerically higher rates of treatment discontinuation due to progressive disease and worse overall survival compared with NHW. Although these associations did not meet statistical significance, the direction suggest a potential survival disparity that may be underpowered in this cohort. The inclusion of diverse patients in clinical trials is essential to identify and address potential disparities in treatment responses and toxicities for all patients, particularly as we study novel agents in the frontline metastatic and early-stage settings.

Estimated prevalence of polycythemia vera in the United States (2025-2030): SEER analysis with modeled reporting delay.

e18589 Background: Polycythemia vera (PV) is a myeloproliferative neoplasm characterized by overproduction of red blood cells, increased risk of cardiovascular events, and progression to myelofibrosis and acute myeloid leukemia. Determining PV prevalence is essential for characterizing disease burden; however, it is methodologically challenging and inconsistently reported. This study aims to generate contemporary representative estimates of PV prevalence in the United States (US). Methods: A modified life table approach was used to estimate PV prevalence based on delay-adjusted, age-specific incidence and age-of-onset-specific relative survival. Age-specific incidence rates (IR) were derived from the US Surveillance, Epidemiology, and End Results (SEER) database between 2001-2022 using ICD-O-3 histology code for PV of 9950/3 (SEER-21 November 2024 submission). Despite year-to-year variation, joinpoint analyses identified no significant trend in IR over the period 2001-2022. To minimize the impact of IR fluctuations on prevalence estimation, the average IR for 2018-2022 was applied to all years for prevalent estimation. An age-specific delay-adjustment correction factor was applied to IR to account for reporting lags in SEER. Relative survival was derived from 20-year observed PV survival in SEER-21, long-term (21-40 year) observed survival in published data, and expected survival in the general population. Age-specific US population estimates from the United Nations were used to estimate 2025-2030 prevalence. Results: The average age-adjusted IR in 2018-2022 was 1.4 per 100,000 person-years, and with delayed adjustment, 1.7 per 100,000 person-years. Observed age-specific IRs increased with age (Table 1), and the modeled delay-adjustment rate was 18-20% across age groups. The estimated PV prevalence rate in 2025 was 25.6 per 100,000, corresponding to 88,943 patients living with PV in the US. The estimated prevalence rate increased from 25.9 to 27.3 per 100,000 over 2026-2030, with the estimated prevalent cases increasing from 90,522 to 97,011. Conclusions: Using a large, nationally representative data source, we generated contemporary US PV prevalence estimates from 2025-2030, leveraging age-specific incidence, delayed reporting adjustment, and life-table survival modeling. These estimates will support more accurate characterization of disease burden and inform healthcare resource allocation for individuals living with PV in the US. Observed and delay-adjusted average IR by age groups (per 100,000) from 2018-2022 (SEER-21 Nov 2024 submission). Age, years Observed Cases Population Observed IR Delay-adjusted IR 0-29 242 297,902,560 0.1 0.1 30-39 529 106,479,102 0.5 0.6 40-49 1,114 95,845,657 1.2 1.4 50-59 2,067 96,400,850 2.1 2.6 60-69 3,345 83,866,510 4.0 4.8 70-79 2,976 50,604,812 5.9 7.0 80+ 1,872 25,861,928 7.2 8.6

Performance of ChatGPT in generating patient-facing cancer survivorship care plans.

1673 Background: Survivorship care plans (SCPs) summarize cancer treatments and provide evidence-based recommendations for surveillance, screening, and management of treatment-related complications. Despite the availability of guidelines and templates, SCPs remain underutilized in routine practice due to the time-intensive and error-prone nature of manual creation. Empowering cancer survivors to use publicly available large language models (LLM)–based chatbots, such as ChatGPT, may offer a feasible solution to this problem. Methods: Standardized patient profiles were developed for six common cancers: breast, colorectal, non–small cell lung, small cell lung, prostate cancer, and diffuse large B-cell lymphoma. Profiles were reviewed by a primary care physician and an oncologist to ensure clinical accuracy and representativeness of common survivorship scenarios. Using lay language to simulate real-world patient input, ChatGPT (version 5.2, publicly available at the time of the study) was prompted with each profile using the question, “What follow-up care do I need?” Common survivorship symptoms for each cancer were also tested. Responses were evaluated across five domains: medical accuracy (concordance with current ASCO or NCCN guidelines), thoroughness, patient safety and risk framing, clarity and patient comprehension, and actionability. Five-point Likert scales were used (1 = not at all; 5 = very much). Scores were averaged across two independent physician raters. Readability was assessed using the Flesch Reading Ease Score. Results: ChatGPT-generated SCPs demonstrated high medical accuracy (mean score 4.25/5), with no factual errors or hallucinations identified on manual review. Patient safety and risk framing scored 3.75/5; symptom red flags and care escalation guidance were appropriately stated in most scenarios. Thoroughness was moderate (3.50/5), with key survivorship elements, such as smoking cessation counseling or genetic risk considerations; sometimes omitted unless explicitly prompted. Readability was limited, with a mean Flesch Reading Ease Score of 40.5, corresponding to a college reading level and exceeding recommended readability standards for patient education materials. Clarity (3.38/5) and actionability (3.38/5) were similarly constrained due to dense medical language. Conclusions: ChatGPT demonstrated high medical accuracy in generating survivorship care plans across six common cancers, supporting the feasibility of using publicly available LLMs to guide survivorship care. However, improvements in comprehensiveness, readability, and patient-centered actionability are necessary before such tools can be safely integrated into clinical survivorship workflows.

Large language model–assisted decision-making in multidisciplinary tumor boards for colorectal cancer.

3521 Background: Evolving colorectal cancer management and rising incidence have increased the workload of multidisciplinary tumor boards (MTBs), placing growing time demands on healthcare professionals. This study evaluates large language models (LLMs) as decision-support tools by comparing AI-generated recommendations with MTB decisions. Methods: This retrospective study used two large language models, Gemini 2.5 (general-purpose) and MedGemma 27B (medical-domain specific), to generate treatment recommendations for patients discussed at MTB. MedGemma 27B was evaluated at two temperature settings (T = 0.0 and T = 1.0). MTB decisions served as the reference standard. Agreement was quantified using Cohen’s kappa, and classification performance was assessed with accuracy, F1 score, and recall. Recommendations were independently reviewed by two MTB members for clinical concordance and safety. Results: A total of 300 colorectal cancer cases were included. Gemini 2.5 demonstrated high agreement with MTB decisions (Cohen’s kappa = 0.792, p < 0.001), with 81.7% full concordance, significantly outperforming MedGemma 27B at both temperature settings (57.0% at T = 0.0 and 62.3% at T = 1.0, p < 0.001 for both). MedGemma 27B showed moderate agreement (k = 0.566 at T = 0.0, k = 0.610 at T = 1.0, p < 0.001), with a modest improvement in concordance at T = 1.0 (p = 0.038). In safety, Gemini 2.5 had the highest rate of clinically safe recommendations (83.7%), outperforming MedGemma at both temperature settings (p < 0.001), with no significant safety difference between MedGemma's temperature settings (p = 0.567). Subgroup analyses showed that Gemini 2.5 had higher discordance in stage IV (p = 0.033), recurrent cases (p < 0.001), and surgical/interventional or active surveillance decisions (p < 0.001). In contrast, MedGemma demonstrated increased discordance in stage II (p < 0.001), ECOG PS ≥2 (p = 0.023–0.036), MSI-unstable tumors (p < 0.001), recurrent disease (p = 0.008–0.010), and surgical/interventional or active surveillance decisions (p < 0.001). Conclusions: Large language models demonstrated meaningful alignment with multidisciplinary tumor board decisions, supporting their potential role in clinical decision support for colorectal cancer, while emphasizing the need for further refinement to ensure consistent and safe clinical integration. Comparison of large language models’ agreement and performance metrics with multidisciplinary tumor board decisions in colorectal cancer. LLM Fully concordant Partially concordant Discordant Accuracy F1 score Recall Cohen's kappa (κ) p value Gemini 2.5 245 (81.7 %) 23 (7.7 %) 32 (10.7 %) 85 % 0.79 0.82 0.792 <0.001** MedGemma 27B temperature 0.0 171 (57.0 %) 64 (21.3 %) 65 (21.7 %) 70.3% 0.64 0.61 0.566 <0.001** MedGemma 27B temperature 1.0 187 (62.3 %) 54 (18.0 %) 59 (19.7 %) 73.5% 0.71 0.66 0.610 <0.001** LLM, large language model.

Safety, pharmacokinetics, and immunogenicity of HLX6018, a monoclonal antibody targeting the GARP/TGF-β1 complex, in healthy subjects: A randomized, double-blind, placebo-controlled, phase I clinical study.

2526 Background: HLX6018 is a novel anti-GARP/TGF-β1 monoclonal antibody that inhibits TGF-β1 release and suppresses the activation, proliferation, and extracellular matrix secretion of fibroblasts. Preclinical studies have shown its efficacy in improving pulmonary fibrosis with a manageable safety profile. A phase 1 first-in-human study was conducted to evaluate the safety and tolerability of single-dose HLX6018 in healthy Chinese subjects. Methods: In this dose escalation phase 1 study, healthy subjects of age 18 to 55 were randomized to receive intravenous, single dose of either HLX6018 or placebo at 0.25 mg/kg, 1.0 mg/kg, 4.0 mg/kg, 12 mg/kg, 25 mg/kg, 50 mg/kg, and 70 mg/kg. A sentinel dosing approach was adopted: each dose group initially enrolled 2 subjects, with 1 receiving HLX6018 and the other receiving placebo in a blinded manner. These 2 subjects then entered a safety observation period after the infusion before the enrolment of remaining subjects for that dose group. The primary endpoint was safety. Secondary endpoints included pharmacokinetics (PK) and immunogenicity. Results: A total of 180 subjects were screened and 66 were randomized. 52 subjects received HLX6018 (0.25 mg/kg, 6; 1.0 mg/kg, 6; 4.0 mg/kg, 8; 12 mg/kg, 8; 25 mg/kg, 8; 50 mg/kg, 8; 70 mg/kg, 8) while 14 received placebo (2 subjects in each dose group). The median age was 42.0; 93.9% of the subjects were of Han ethnicity, 50.0% were male. Overall, 46 subjects (69.7%) experienced treatment-emergent adverse events (TEAEs), with 1 subject (1.5%) receiving HLX6018 at 25 mg/kg reporting a serious TEAE of osteonecrosis that was unrelated to HLX6018. 39 subjects (59.1%) experienced treatment-related adverse events (TRAEs). Most common TRAEs (≥ 10% in any dose group) included neutrophil count decreased (HLX6018 vs placebo group: 11.5% vs. 21.4%), injection site pain (11.5% vs. 21.4%), blood corticotrophin decreased (9.6% vs. 14.3%), blood triglycerides increased (9.6% vs. 14.3%) and blood follicle stimulating hormone increased (5.8% vs. 14.3%). There were no TEAEs leading to death, TRAEs leading to drug discontinuation, TRAEs of grade 3 or more in severity, or serious TRAEs. The incidence of TEAEs and TRAEs across the HLX6018 dose groups showed no clear dose-related pattern and was comparable to the placebo group. HLX6018 exhibited approximately linear PK characteristics after single intravenous infusion with the dose range of 0.25–70 mg/kg. Anti-drug antibody was detected in 3 subjects (5.8%) who received HLX6018; no neutralizing antibody was detected. Conclusions: HLX6018 is safe and well tolerated across the investigated doses. Further clinical investigation of its efficacy is warranted. Clinical trial information: NCT06310746 .

Kidney function trajectories in pediatric acute lymphoblastic leukemia.

e18513 Background: Acute kidney injury (AKI) is common in pediatric acute lymphoblastic leukemia (ALL) and can increase the risk of developing chronic kidney disease (CKD). However, AKI phenotypes vary widely in severity, duration, and number of episodes that may be characterized as patterns in estimated glomerular filtration rate (eGFR) trajectories. Distinct eGFR trajectories during therapy for ALL may indicate long-term CKD risk in pediatric ALL survivors. To our knowledge, eGFR trajectories have not been described in the pediatric ALL population. Objective: To identify groups based on eGFR trajectory during ALL therapy and to evaluate characteristics associated with declining eGFR groups. Methods: Electronic health record data were retrospectively collected from pediatric patients treated for de novo B- and T-ALL at Children’s Hospital of Philadelphia (2011-2015), Children’s Healthcare of Atlanta (2010-2018), and Texas Children’s Hospital (2011-2014), three large cancer centers in the Leukemia Electronic Abstraction of Records Network (LEARN). Patients with infant ALL were excluded. Patients were followed from the time of first chemotherapy course in LEARN through upfront chemotherapy completion, death, relapse, conditioning for stem cell transplant, or transfer of care, whichever came first. Latent class mixed models were used to determine course-level eGFR trajectory groups. Bivariate tests and multinomial logistic regressions were used to explore the association of covariates with specific trajectory groups. Results: Cohort characteristics (n=884) included: median (interquartile range) age 5.7 years (3.4-11.4), 56% male, median baseline eGFR 107.6 mL/min/1.73m 2 (89.8-126.8), 2% with Trisomy 21 (T21), and 54% with National Cancer Institute standard risk leukemia. Between 3-4 eGFR trajectory groups were identified in each course. All courses had stable eGFR trajectory groups distinguished by their course-specific baseline eGFR. Induction and Interim Maintenance-1 had additional groups with initial eGFR decline and recovery. Consolidation and Delayed Intensification courses had groups with eGFR decline. Bivariate analyses indicated that declining eGFR trajectory groups were significantly associated with lower course-specific baseline eGFR, AKI during the course, and T21; median number of eGFR values were not consistently associated with specific trajectory groups across courses. Multinomial logistic regression showed that those with higher baseline eGFR had increased odds of being in stable eGFR trajectory groups across all courses. Prior AKI was associated with increased odds of being in a lower eGFR trajectory group in Interim Maintenance-2. Conclusions: Pediatric patients with ALL have distinct eGFR trajectories during therapy. Patients with lower baseline eGFR, AKI, and a diagnosis of T21 may be at highest risk of eGFR decline during therapy and may benefit from increased screening for CKD.

First-line (1L) olomorasib + pembrolizumab in patients with <i>KRAS</i> G12C–mutant advanced NSCLC, and PD-L1 expression 0-49%, from the dose optimization cohorts of LOXO-RAS-20001 and SUNRAY-01.

8628 Background: For patients (pts) with 1L KRAS G12C-mutant advanced NSCLC and PD-L1 expression &lt;50%, the current SOC is pembrolizumab + platinum-based chemotherapy. Combining olomorasib, a KRAS G12C inhibitor, with pembrolizumab may overcome the limited efficacy of single-agent pembrolizumab in this population, while offering a chemotherapy-sparing targeted approach. Here we report an integrated analysis of 1L pts with PD-L1 expression 0-49% (&lt;1% and 1-49%) who received olomorasib + pembrolizumab from the dose optimization cohorts of LOXO-RAS-20001 and SUNRAY-01. Methods: Pts with advanced KRAS G12C-mutant NSCLC, known PD-L1 expression and ECOG PS 0-1 were randomized to olomorasib (50 or 100 mg, orally BID) with pembrolizumab (200 mg Q3W). One cycle of SOC pembrolizumab prior to enrollment was permitted. Objective response rate (ORR) was assessed in the efficacy evaluable population, with PD-L1 0-49% Other efficacy endpoints included best overall response (BOR) and disease control rate (DCR). Safety was assessed across all treated pts (PD-L1 0-100%). Results: As of 6 June 2025, 85 pts received olomorasib + pembrolizumab, of which 31 pts had tumors with PD-L1 expression 0-49% (&lt;1%: n=13; 1-49%: n=18). In pts with PD-L1 expression &lt;1%, no pts received a prior cycle of SOC, 23% of pts had baseline brain metastases, and 38% had an ECOG score of 1. For the pts with PD-L1 expression of 1-49%, 17% received 1 prior cycle of SOC, 17% had baseline brain metastases, and 56% had an ECOG score of 1. The ORR was 58% in pts with PD-L1 &lt;1%, 67% in pts with PD-L1 1-49% (Table). Across all treated pts the most common any grade treatment-related adverse events (TRAEs) were diarrhea (31%) and ALT/AST increased (26/24%). TRAEs led to permanent discontinuation of study treatment in 12% of pts (n=10). Conclusions: Olomorasib + pembrolizumab demonstrated promising efficacy in pts with 1L KRAS G12C-mutant advanced NSCLC, and PD-L1 expression &lt;1% and 1-49% with ORRs that compare favorably with historical outcomes for pembrolizumab and chemo-immunotherapy in unselected advanced NSCLC. Olomorasib + pembrolizumab is under evaluation in pts with 1L KRAS G12C-mutant metastatic NSCLC (SUNRAY-01, NCT06119581) and early-stage NSCLC (SUNRAY-02, NCT06890598). Clinical trial information: NCT04956640 , NCT06119581 . Response and time-to-event endpoints in the Efficacy evaluable population with PD-L1 0-49% (N=30*). Endpoint Pts with PD-L1 &lt;1%n=12 Pts with PD-L1 1-49%n=18 ORR † , % (n/N) 58.3 (7/12) 66.7 (12/18) BOR, n (%) CR † 1 (8.3) 1 (5.6) PR † 6 (50.0) 11 (61.1) SD 4 (33.3) 4 (22.2) PD 1 (8.3) 1 (5.6) NE 0 (0.0) 1 (5.6) DCR, % (n/N) 91.7 (11/12) 88.9 (16/18) † Includes responses confirmed and pending. *Data for 1 patient are not shown in the table due to incomplete target lesion assessment. CR, confirmed response; NE, not evaluable, PD, progressive disease; PR, partial response; SD, stable disease.

Insights into the impact of base-excision–repair genes on immunotherapy response in solid tumors.

e14548 Background: Base-excision repair (BER) is the primary pathway that removes ubiquitous single-base lesions from oxidative and alkylation stress, preserving genome integrity and preventing error-prone repair. When BER is disrupted, accumulation of DNA damage can heighten tumor immunogenicity—through increased mutational consequences and activation of cytosolic DNA–sensing pathways—creating a biologic context favorable to immune checkpoint blockade. We therefore asked whether BER gene alterations correlate with improved clinical outcomes to ICI across solid tumors. Methods: We pooled five published ICI cohorts with genomic/clinical annotation—DFCI Metastatic Melanoma (Science 2015); DFCI Clear-Cell Renal Cell Carcinoma (Science 2018/2019); MSK Melanoma (NEJM 2014); UCLA Metastatic Melanoma (Cell 2016); and MSK Immunotherapy/TMB (Nature Genetics 2019)—yielding 1,908 patients treated with anti-PD-1/PD-L1, anti-CTLA-4, or combination therapy across renal cell carcinoma, melanoma, non-small cell lung cancer, colorectal cancer, cancer of unknown primary, bladder cancer, breast cancer, glioma, esophagogastric cancer, non-melanoma skin cancer, and head &amp; neck cancer. Data were retrieved via cBioPortal. Patients were classified as BER-altered if they harbored ≥1 qualifying event in MUTYH, PARP1, or NTHL1 (loss-of-function or pathogenic/likely-pathogenic missense; deep deletions; amplifications excluded); comparator: triple-wildtype. OS from ICI initiation was compared by Kaplan–Meier/log-rank; Cox models adjusted for age, tumor type, ICI class, and sex. Subg roup analyses focused on melanoma, NSCLC, and CRC. Results: 78 patients were BER-altered, OF WHICH melanoma (53%), NSCLC (10%), and CRC (10%. Overall, BER-altered patients had longer OS than unaltered patients (mOS 32.4 vs 18.0 months; log-rank p = 0.0139). At the gene level, PARP1 alterations were associated with superior OS (median 32.4 vs 18.0 months; p = 0.0278). MUTYH alterations showed a favourable trend but did not reach statistical significance (mOS 49.3 vs 18.0 months; p = 0.299), and NTHL1 alterations were uncommon and showed similar survival to unaltered cases (mOS 20.1 vs 18.0 months; p = 0.870) By tumor type, NSCLC showed a favorable association (p = 0.0389), whereas melanoma was neutral (p = 0.561). Tumor mutational burden (exploratory): the BER-altered group had higher TMB than the unaltered group (median 21.7 vs 5.9 nonsynonymous mutations/Mb; p &lt; 0.001). Co-mutations (BER-altered): the most common co-mutations were KMT2D, GRIN2A, PTPRT, BRAF, MUC16, and TERT. Conclusions: Alterations in MUTYH/PARP1/NTHL1—particularly PARP1—identify a subset of ICI-treated patients with improved OS, with the clearest effect in NSCLC and neutrality in melanoma.These findings support prospective validation and consideration of BER status, alongside genomic context such as TMB, in immunotherapy stratification.

When to begin? A chart-based review of advanced care planning in cancer patients one year prior to death.

e24043 Background: Timely goals of care (GOC) conversations are a critical component of high-quality oncologic care. Despite increasing awareness of their benefit and importance, numerous studies have identified significant barriers to their completion for both patients and physicians. In this study, we review the chart documentation of patients who died on our oncology service at a large urban hospital. By reviewing the timeline of events and chart-documented language in the year prior to death, we hope to provide a new frame of reference for the known barriers to advanced care planning in actual practice. Methods: We performed a retrospective chart review of the last 90 deaths on the oncology service at two hospitals within our large, urban, academic medical center. Patients were included if they were age ≥ 20 and had a primary oncologist within our health system that they saw outpatient at least once. Reviewers used a chart abstraction form to track dates of diagnosis and death, dates of all GOC discussions within one year of death, who was present, and what topics were covered in both inpatient and outpatient visits. These data were then reviewed for factors that contributed to the timeliness of advanced care planning. Results: 97% of our patients (n= 87) had at least one GOC discussion in the year before death. Of those, 70% (n= 61) occurred within the last month of life during the terminal hospitalization. Only 16% (n=14) had any documented GOC discussion in the outpatient setting. GOC with patients and families was frequently limited to the last days of life and often initiated by a consult to the inpatient palliative care service. A number of records did reveal attempts to begin these discussions earlier, however with mixed success. Several factors appeared to challenge these efforts. These included: (1) patient or family lack of readiness to engage in advance care planning, (2) provider uncertainty or discomfort, and (3) rapid demise of the patient. No clear reason was apparent on chart review in some cases. Conclusions: In this retrospective review of 90 inpatient oncologic deaths, we see that despite strides made by major oncology institutions to promote advanced care planning early on in the patient’s disease course, significant delays continue to exist. While our data is limited to what is documented in the chart, it sheds new light on known barriers to completing GOC conversations in the inpatient setting for cancer patients.

Individualized neoantigen therapy intismeran autogene (intismeran) plus pembrolizumab (pembro) in resected melanoma: 5-year update of the KEYNOTE-942 study.

9500 Background: Intismeran is an mRNA-based individualized neoantigen therapy designed to increase endogenous neoantigen-specific antitumor T-cell responses. The phase 2b KEYNOTE-942 study of high-risk resected melanoma showed clinically meaningful improvements in recurrence-free survival (RFS) and distant metastasis–free survival (DMFS) with intismeran + pembro vs pembro alone at primary analysis. Intismeran + pembro had a manageable safety profile without potentiation of immune-related AEs vs pembro alone. Clinical benefits of intismeran + pembro were sustained at 3 y, with a 49% risk reduction of RFS and 62% risk reduction of DMFS. Intismeran + pembro also induced greater novel T-cell clonal expansion vs pembro, which was positively associated with RFS for intismeran + pembro but not pembro alone. We report results from KEYNOTE-942 after 5 y of planned follow-up. Methods: Eligible participants (pts; aged ≥18 y) with resected stage IIIB–IV cutaneous melanoma were randomized 2:1 to receive 9 doses of intismeran 1 mg IM Q3W + 18 doses of pembro 200 mg IV Q3W or 18 doses of pembro 200 mg Q3W alone. Primary endpoint was RFS; secondary endpoints included DMFS and safety. Exploratory endpoints included OS. No alpha was assigned to this analysis. Results: From Jul 2019 to Sep 2021,157 pts were randomized to intismeran + pembro (n=107) or pembro (n=50). With an additional 2 y of follow-up (data cutoff, Dec 15, 2025; median planned follow-up, 60.3 [range, 50.5–76.4] mo) after the 3-y analysis, minimal new events occurred. RFS risk reduction for intismeran + pembro vs pembro alone was 49% (HR, 0.51; 95% CI, 0.29–0.89), with landmark 5-y RFS rates of 68.8% (95% CI, 56.3%–78.3%) for intismeran + pembro vs 49.1% (95% CI, 33.3%–63.0%) for pembro alone. DMFS risk reduction was 59% (HR, 0.41; 95% CI, 0.20–0.84). In the intismeran + pembro arm, 7 pts (6.5%) died (disease progression, n=4) vs 7 pts (14.0%) in the pembro arm (disease progression, n=6). There was a trend for improved OS (HR, 0.47; 95% CI, 0.17–1.35); 5-y rate was 92.2% (95% CI, 84.2%–96.3%) for intismeran + pembro vs 71.3% (95% CI, 35.4%–89.6%) for pembro alone. The safety profile of intismeran was consistent with prior analyses. Conclusions: After a median 5 y of follow-up, intismeran + pembro continued to prolong RFS and DMFS, along with a trend for improved OS vs pembro alone in pts with high-risk resected melanoma. These long-term findings show that intismeran + pembro treatment benefits were sustained and durable over time, despite all pts having completed study treatment before primary analysis (2021). As reported previously, intismeran was well tolerated, with a manageable safety profile for intismeran + pembro. Intismeran + pembro is being evaluated in a phase 3 study of high-risk resected stage II–IV melanoma (INTerpath-001; NCT05933577) and in studies of pts with other malignancies. Clinical trial information: NCT03897881 .

Homologous recombination signature (HRDsig) in clinically advanced gallbladder adenocarcinoma (CAGAC): A genomic landscape study.

4138 Background: CAGAC is an aggressive malignancy with poor prognosis where chemotherapy and immunotherapy provide a limited benefit. There have been clinical trials investigating the benefit of PARP inhibitors in biliary tract cancers. HRDsig positivity in other types of cancer have shown to predict sensitivity to PARP inhibitors. We explored the genomic characteristics of CAGAC patients according to the HRDsig status using a large real-world genomic dataset. Methods: Comprehensive genomic profiling was performed on 2,423 cases of CAGAC to assess all types of genomic alterations (GA), including base substitutions, short insertions and deletions, copy number changes, rearrangements, and fusions. Microsatellite instability (MSI) status, tumor mutational burden (TMB), genomic ancestry, and trinucleotide signatures were determined from the sequencing data. PD-L1 was determined with the DAKO 22C3 IHC assay. HRDsig status was determined by measuring large scale copy number changes and DNA repair errors. Patients were grouped into HRDsig positive (HRDsig+) and negative (HRDsig-) based on results. Genes with GA ≥5% in either population were included in the analysis. Results: Among 2,423 total cases, 152 (6.3%) were HRDsig+. Between MRDsig+ and MRDsig-, there was no significant difference in gender (female 76.3% vs 69.1%, p = 0.061), age (median 66.5 vs 68.0 yrs; p = 0.065), genomic ancestry (EUR 67.6% vs 63.7%, p = 1.0), MSI-H status (0.0% vs 1.3%, p = 0.468), PD-L1 expression (Low 34.0% vs 23.4%, p = 0.529. High 1.9% vs 5.4%, p = 0.946). Median GA per tumor was the same between groups (median 5 vs 5, p = 0.723), but the proportion of TMB≥10 mut/Mbp were higher in HRDsig+ (12.5% vs 4.7%, p = 0.001). HRDsig+ had more frequent alterations in BRCA1 (9.2% vs 1.4%, p &lt; 0.001), BRCA2 (21.7% vs 1.9%; p &lt; 0.001), PALB2 (11.2% vs 0.4%; p &lt; 0.001), NF1 (9.9% vs 4.4%; p = 0.018), PIK3R1 (5.9% vs 2.5%; p = 0.05), and PTEN (11.8% vs 5.3%; p = 0.015). HRDsig- had more frequent alterations in ERBB2 (5.9% vs 19.0%, p &lt; 0.001), KRAS (5.9% vs 14.9%, p = 0.007), and MDM2 (3.3% vs 9.4%, p = 0.024). Conclusions: HRDsig positivity comprised approximately 6% of all CAGAC cases. HRDsig+ were associated with a higher proportion of mutations related to homologous recombination repair (HRR) genes and mTOR pathway, whereas HRDsig- demonstrated a higher proportion of common driver mutations such as ERBB2 and KRAS. Notably, there were cases with discordant HRDsig and HRR mutation status, which may be attributed to factors such as epigenetic alterations or mono-allelic alterations. These cases may represent the role of HRDsig complementing HRR mutation status and potentially have clinical implications in predicting response to PARP inhibitors.

Influence of oncogenic protein sequestration by platelets on early and stage-specific breast cancer detection.

e12601 Background: Mammography remains the standard modality for breast cancer (BC) screening but has several limitations. There is a critical need for minimally invasive markers that can aid in distinguishing benign, non-invasive lesions from those likely to progress to invasive BC. Platelets are nature’s transport mechanism for delivery of growth factors and inhibitors. Assaying the platelet proteome, Hessian Labs developed HeLP (Hessian Labs Personalized) test designed to identify early malignancy, as well as its grade, and invasiveness. The test is based on a discovery that platelets sequester oncogenic regulators against a concentration gradient in plasma early in tumor development. Hence, it has the potential to differentiate benign from invasive BC at various stages, stratifying the need for intervention. Methods: Platelet samples from 120 women with BC were analyzed using a discovery data-independent acquisition mass spectrometry (DIA-MS). The identified differentially expressed platelet-associated proteins, were prioritized using ClariaIQ , a proprietary computational platform that ranks proteins based on their relevance to stage-specific oncogenic pathways and established Cancer Hallmarks, to identify stage- and grade-associated proteomic signatures and create the HeLP panel. Results: DIA-MS identified multiple platelet sequestered proteins that were differentially expressed between indolent disease and invasive BC. ClariaIQ prioritization resulted in the selection of 13 proteins for inclusion in the HeLP breast cancer assay. We present two examples: SAA1 and SERPIN proteins demonstrated differential expression between healthy, benign breast lesions and malignant invasive disease (Table 1). SAA1 peaks in early BC and SERPIN in advanced stages. With SAA1, in HER negative BC the levels rise to an average of 285,000 RE units when compared to HER2 positive BC. Conclusions: Platelet sequestered protein signatures provide clinically relevant information regarding breast cancer status, grade, and invasiveness. The findings support the robustness of the signal, and a more patient-specific signature than traditional plasma or serum biomarkers. It suggests a role for HeLP as a minimally invasive adjunct to mammography to improve diagnostic stratification and therapeutic decision-making. Disease Group SAA1 Mean (RE) SAA1 SD (RE) SERPIN A3 Mean (RE) SERPIN A3 SD (RE) Healthy Controls 119,592 206,349 490,231 279,017 Benign Breast Lesions 622,187 50,855 341,102 153,144 Stage I Breast Cancer 1,223,865 1,878,654 742,409 466,468 Stage II Breast Cancer 663,369 1,970,265 922,381 844,512 Stage III Breast Cancer 437,853 662,211 705,994 597,399 Stage IV Breast Cancer 225,265 384,826 1,068,874 1,415,934 RE=Relative Expression units. N=20 subjects included in each group.

National mortality trends in colorectal cancer patients with psychiatric and respiratory comorbidities in the United States,1999-2023: A CDC WONDER analysis.

e15660 Background: Colorectal Cancer (CRC) is a leading cause of cancer-related deaths in the United States. Comorbidities such as tobacco use disorder, depression, anxiety and chronic pulmonary disease (COPD) may worsen outcomes via delayed diagnosis, reduced treatment tolerance and fragmented care. National long-term mortality trends and disparities among CRC patients with these comorbidities remain poorly defined. Methods: U.S. mortality data (1999-2023) from the CDC WONDER database was analyzed. CRC deaths (ICD-10 : C18-C20) with contributing causes of tobacco use disorder (F17), depression (F32), anxiety (F41) and COPD (J44) were identified. Age adjusted mortality rates were stratified by ten-year age group, sex, region, urbanization and place of death. Individuals aged 35-44 were excluded because of small case counts. Joinpoint regression assessed temporal trends via annual percentage change (APC) and average annual percent change (AAPC). Place of death was analyzed separately. Results: Stratified analyses demonstrated significant increase in colorectal cancer mortality across multiple subgroups.The steepest increases occurred in adults aged 45-54 (AAPC10.1%, p &lt; 0.001), and 55-64 (AAPC 6.0%, p &lt; 0.001) while trends among adults aged 75 and above were modest. Mortality increased in both males (AAPC 2.0%, p = 0.014) and females ( AAPC 2.5%, p = 0.002). Regionally the Midwest (AAPC 3.3%, p &lt; 0.001) and South (AAPC 3.4%, p &lt; 0.001) saw sustained increases, while Northeast remained stable.Nonmetropolitan areas showed greater mortality increase as compared to metropolitan areas ( AAPC 4.1% vs 2.3%, p &lt; 0.001). Most deaths occurred in inpatient hospital settings. Conclusions: Rising mortality among middle-aged CRC patients with psychiatric and respiratory comorbidities , particularly in rural and Southern regions necessitate urgent systemic reform. Integrated oncology models must prioritize medical, psychosocial and outpatient palliative care to ensure survival gains are equitably realized across all geographic and socioeconomic strata.

Functional domain disruption to define the molecular and clinical spectrum of germline <i>RAD51</i> family variants in hereditary breast cancer (BC).

572 Background: RAD51 paralogs (RAD51C, RAD51D, XRCC2) and RAD50 are crucial in DNA repair. Germline pathogenic variants (PVs) increase hereditary breast/ovarian cancer risk, but domain-specific impacts remain unclear. Methods: Retrospective cohort study (2019-2025) of breast cancer patients undergoing multigene testing. PVs were mapped to functional domains. Tumor features included histology, stage, grade, TILs, and molecular subtype. Associations were assessed using Fisher’s exact test. Results: Among 495 patients, 23 (4.6%) had RAD51 family PVs (RAD51C = 16, RAD51D = 4, RAD50 = 1, XRCC2 = 1). Median diagnosis age was 49. Synchronous malignancies were significantly more frequent in RAD51 carriers vs. BRCA1 carriers (34.8% vs.; OR 3.51, 95% CI 1.09–11.26). Most variants (22/23) were truncating (11 nonsense, 9 frameshift, 1 splice-site, 1 exon deletion). Domain mapping revealed disruption of critical regions: RAD51-like core domain truncations (RAD51D p.Arg186 ; RAD51C p.Arg193 ), RAD51C C-terminal truncations removing essential regions for complex integrity, and early RAD50 truncation predicted to abolish MRN complex function. Tumors were predominantly early-stage (65.2% stage I). Histology showed invasive carcinoma of no special type (69%) and invasive lobular carcinoma (25%), with strong enrichment of lobular histology in RAD51C carriers (p &lt; 0.001). RAD51C tumors were basal-like or luminal A, while the RAD50 tumor was luminal B HER2-positive. RAD51C tumors often had high nuclear grade (77.3%) but low TILs (71%). Conclusions: RAD51 family PVs are primarily truncating, disrupting key HRR domains and supporting loss-of-function as the main mechanism. Domain location correlates with distinct clinicopathological features, refining genotype–phenotype correlations for improved risk assessment.

Evaluating physician perspectives on neoadjuvant immunotherapy and chemotherapy in early-stage triple-negative breast cancer.

e12633 Background: The Keynote-522 trial established neoadjuvant chemo-immunotherapy as a new standard of care for patients with early-stage triple negative breast cancer (TNBC). However, questions remain regarding the optimal chemotherapy backbone, duration of immunotherapy and the potential for chronic toxicities in this curative population. This study aimed to identify Canadian practices for the management of early-stage TNBC, areas of clinical uncertainty and interest in future clinical trials. Methods: A voluntary electronic survey was distributed to Canadian medical oncologists via an email database compiled by the Ottawa Hospital Research Institute, REthinking Clinical Trials (REaCT) program. The survey focused on prescribing practices prior to and following the Keynote-522 trial and the incorporation of immunotherapy into the neoadjuvant treatment of early-stage TNBC. The analysis is reported descriptively. Results: Out of the 50 oncologists from across Canada who responded to the survey, 36 were eligible to complete the survey in its entirety. Among respondents, 83.3% adopted the Keynote-522 chemotherapy backbone as their first choice for early-stage TNBC. However, only 47% reported following the trial protocol exactly. Common modifications included utilizing dose-dense (every 2-week) administration of anthracycline and cyclophosphamide (AC) (50%) with 6-week pembrolizumab dosing intervals (50%). Concerns regarding toxicity (39%) and schedule impracticality (25%) were primary drivers for protocol divergence. While 86% of physicians prescribe immunotherapy for eligible stage II/III TNBC, 50% would also offer it for Stage I (T1N0) disease. Areas of clinical uncertainty were highlighted as the management of stage I disease and the tolerability of the Keynote-522 protocol in the elderly and comorbid patients. Optimal adjuvant treatment practices also remain uncertain, particularly in the setting of patients with pathological complete response (pCR), with 39% of respondents indicating they felt there was currently insufficient data to support the continuation of pembrolizumab in the adjuvant setting. Finally, 100% of respondents indicated a need for more response-adapted regimens to mitigate treatment-related toxicities. Conclusions: Canadian oncologists have rapidly adopted neoadjuvant immunotherapy for early-stage TNBC as standard of practice but frequently modify the chemotherapy backbone to improve tolerability and clinical efficiency. Significant uncertainty persists regarding the treatment of stage I disease and elderly or comorbid patients, in addition to optimal adjuvant treatment practices. While research is ongoing to address these issues, more is needed, particularly efforts to support response-adapted treatment approaches.