Genomic Data Research Articles

Functionality of highly diverged Imd like genes identified in stinkbugs and bedbugs.

The IMD pathway is an innate immune signalling pathway regulating antibacterial humoral defence responses, and is highly conserved among diverse insects and other organisms. However, genomic studies among hemipterans suggested that certain insects from this order may lack the Imd gene, a key component of the IMD immune pathway. Our previous work identified Imd in the brown-winged green stinkbug Plautia stali, which showed low sequence homology to other insect Imd genes but exhibited functionality in the IMD pathway. Using the P. stali Imd sequence as a query, we surveyed genomic data of the common bedbug Cimex lectularius and the brown marmorated stinkbug Halyomorpha halys, both of which were thought to lack an Imd gene, and identified Imd-like gene sequences in each. RNA interference experiments demonstrated that, in both species, these Imd-like genes function similarly to the canonical Imd gene. 3D protein structure predictions confirmed that, despite extensive sequence divergence, the Imd-like translation products were structurally similar to Imd proteins of other organisms. High levels of sequence diversity and positive selection were observed with some Imd-like genes of bedbugs and stinkbugs, suggesting an ongoing evolutionary arms race between insect hosts and their microbial symbionts.

Tertiary lymphoid structures and their association with immune checkpoint inhibitor response and survival outcomes in patients with non-small cell lung cancer.

2634 Background: Immune checkpoint inhibitor (ICI)-based therapy is currently the first-line treatment for patients with lung adenocarcinoma (LUAD) and lung squamous cell carcinoma (LUSC) without actionable mutations. However, the commonly utilized biomarkers, including PD-L1 protein expression and tumor mutation burden, are not sufficiently accurate to predict the treatment response from ICI in this patient population. As tumor microenvironment (TME) and tertiary lymphoid structure (TLS) play a significant role in antitumor immunity, we explore these immunophenotypic factors to determine the potential biomarkers in patients with LUAD or LUSC. Methods: We evaluated all patients with LUAD or LUSC from three publicly available data and two novel retrospective cohorts for transcriptomic-based immune TME subtype classification (immune-hot vs. immune-cold) and TLS signature, along with associated clinical and genomic data. Those with other histological subtypes of non-small cell lung cancer or those who harbored EGFR mutations or ALK rearrangements were excluded from our study. The cellular decomposition within tumor samples was calculated using the deconvolutional Kassandra algorithm. Survival analysis was evaluated using log-rank test and multivariate Cox regression adjusted by PD-L1 status, KEAP1/STK11/KRAS/TP53 mutational status, immune TME subtype, and TLS signature. All statistical analyses were performed using Python. Results: A total of 514 patients were included from five cohorts, with 272 and 505 having genomic and transcriptomic data, respectively. 59% of patients with LUAD or LUSC exhibited an immune-cold phenotype, which correlated with adverse overall survival (OS) and progression-free survival (PFS) than immune-hot phenotype in LUAD. However, the ICI response rates were similar in both groups. Superior PFS and ICI response rates were observed in patients with high TLS signatures (> 88th percentile) in LUAD, even after multivariate adjustments. Immune signatures that were positively associated with ICI response included the infiltration and trafficking of T and NK cells for LUAD and B-cell percentage for LUSC. In contrast, CD8 + T-cell abundance did not correlate with ICI response. The presence of KEAP1 or STK11 mutations also did not affect the response rates but were associated with shorter OS and PFS. Conclusions: Transcriptomic-based immune-hot TME and high TLS signature may serve as novel predictive and prognostic biomarkers in patients with LUAD, while the presence of KEAP1 or STK11 mutations only offered prognostic values. Further prospective studies are warranted to expand to other treatment combinations with PD-(L)1 inhibitors.

Treatment patterns, genomic characteristics, and outcomes among patients with metastatic lobular breast cancer.

1119 Background: Invasive lobular carcinoma (ILC) is characterized by loss of E-cadherin expression and accounts for 10-15% of breast cancer diagnoses. ILC differs from the more common invasive carcinoma of no special type in the pattern of metastatic spread and genomic characteristics; however, clinicopathologic characteristics among patients with metastatic ILC are not well described. Here, we present comprehensive treatment, genomic, and outcome data in a large single-center cohort of patients with metastatic ILC. Methods: Patients were identified for inclusion in this retrospective study if they had ILC histology on early-stage breast biopsy/surgical pathology and/or CDH1 mutation on metastatic site biopsy; all patients were required to have MSK-IMPACT somatic next generation sequencing (NGS) data available. Clinicopathologic characteristics were abstracted from the EMR. The Kaplan-Meier method was used to estimate overall survival (OS). The log-rank test was used to compare OS by ILC subtype and by genetic mutations. Wilcoxon rank sum test and Kruskal-Wallis test were used to compare number of treatment lines by receptor status. Results: 654 patients were included, of whom 99.8% were female, 89% were white, and 96% non-Hispanic. 438 (67%) had recurrent disease whereas 212 (33%) had de novo metastatic disease. Among 307 with ILC histologic subtype data available, 139 (45%) had classic type, 65 (21%) had pleomorphic, 45 (15%) had mixed, and 58 (19%) had other ILC subtypes. 454 (87%) had hormone receptor-positive (HR+) disease, 45 (9.1%) had HER2+ disease, and 50 (9.5%) had triple negative disease at metastatic diagnosis. In the total cohort, median number of treatment lines for metastatic disease was 4 (IQR 2-7) and median number of chemotherapies was 2 (IQR 1-3). In the HR+ cohort, median number of endocrine therapies was 2 (IQR 1-3). Among patients with genomic data from a biopsy obtained within 2 months of metastatic diagnosis, 79% had a CDH1 mutation, 48% had a PIK3CA mutation, 5.6% had an AKT1 mutation, 11% had a PTEN mutation, 9.3% had an ESR1 mutation, 17% had a HER2 mutation, and median tumor mutation burden (TMB) was 4 (IQR 3-7); 17% had TMB ³10. Median OS in the total cohort was 4.4 years (95%CI 4.1-4.8). OS did not differ significantly by ILC subtype ( p = 0.8). OS differed significantly by CDH1 mutation status (wt 5.3 years, 95%CI 4.1-6.6; mut 3.7 years, 95%CI 3.5-4.2, p = 0.01), PIK3CA status (wt 4.6 years, 95%CI 4.0-5.8; mut 3.4 years, 95%CI 3.1-3.9, p < 0.001), and PTEN status (wt 4.2 years, 95%CI 3.7-4.5; mut 3.4 years, 95%CI 2.2-4.3, p = 0.008). OS did not differ significantly by HER2 , AKT1 , or ESR1 mutation status. Conclusions: In a large single-center cohort of patients with metastatic ILC, OS did not vary by ILC subtype, but did differ significantly by CDH1, PIK3CA, and PTEN mutation status. This underscores the prognostic importance of NGS in metastatic ILC.

Development and real-world validation of a multi-modal machine learning model to predict response to immune checkpoint inhibitors in cancer patients.

e14532 Background: Immune checkpoint inhibitors (ICIs) have transformed cancer care, offering the promise of durable, life-changing responses for cancer patients. However, these benefits are realized in only about 20% of cases, with most patients failing to respond. This disparity underscores the critical need for more advanced predictive tools to refine patient selection and fully unlock the potential of ICIs. We present a machine learning (ML) model, called ioBERT, that predicts ICI response using multi-modal data, including clinical information, drug target and structure features, and genomic alteration data commonly captured by commercial NGS panels in real-world clinical settings. Methods: We curated a clinicogenomics dataset of ~2,900 cancer patients treated with FDA-approved anti-PD1 and anti-CTLA4 therapies. Inclusion criteria for ioBERT development included real-world ICI outcomes (rw-OS), clinical data (e.g., disease type, subtype, sex, tumor stage), and genomic alteration data for ~200 genes commonly assessed by commercial NGS panels. NGS profiles were filtered to biopsies collected within 2 years before ICI initiation. Samples with whole-exome data were subset to these ~200 genes. Multi-modal inputs were processed using custom and publicly available encoders, including BioBERT and ProteinBERT, to generate numerical embeddings, which were integrated into a neural network to model relationships with rw-OS outcomes. ioBERT predictions were validated on hold-out and independent rw-datasets. Results: ioBERT outperformed top-performing models from the Anti-PD1 Response Prediction DREAM Challenge in predicting ICI response, as measured by the concordance index (C-index). Its performance was independently validated in a real-world dataset, outperforming existing ICI biomarkers including tumor mutational burden, and PD1 expression where available. Direct comparisons in real-world datasets were limited by other models’ reliance on RNAseq, which is not routinely available in clinical settings. Conclusions: Our study demonstrates the value of integrating multi-modal data to achieve a comprehensive view of patient and tumor biology, enabling more accurate ICI response predictions with ML. Our approach has the potential to improve patient outcomes by identifying those most likely to benefit from ICI treatments and reducing unnecessary therapies. Furthermore, ioBERT can be integrated with off-the-shelf NGS assays, facilitating rapid clinical translation without the need for bespoke assays or costly sequencing methods.

Using Genomics to Develop Personalized Cardiovascular Treatments.

Advances in genomic technologies have significantly enhanced our understanding of both monogenic and polygenic etiologies of cardiovascular disease. In this review, we explore how the utilization of genomic information is bringing personalized medicine approaches to the forefront of cardiovascular disease management. We discuss how genomic data can resolve diagnostic uncertainty, support cascade screening, and inform treatment strategies. The role that genome-wide association studies have had in identifying thousands of risk variants for polygenic cardiovascular diseases, and how these insights, harnessed through the development of polygenic risk scores, could advance personalized risk prediction beyond traditional clinical algorithms. We detail how pharmacogenomics approaches leverage genotype information to guide drug selection and mitigate adverse events. Finally, we present the paradigm-shifting approach of gene therapy, which holds the promise of being a curative intervention for cardiovascular conditions.

Paclitaxel-related type I Kounis Syndrome in a very young patient with HER2-positive breast cancer and the role of genomics to disentangle a complex therapeutic scenario: a case report and narrative review.

Paclitaxel-related type I Kounis Syndrome in a very young patient with HER2-positive breast cancer and the role of genomics to disentangle a complex therapeutic scenario: a case report and narrative review.

The HealthTree Cure Hub registry: A patient-centered, multicenter approach to advancing multiple myeloma care.

e19569 Background: Multiple myeloma (MM) is the second most common hematologic malignancy, with an increasing incidence worldwide. Clinical registries provide valuable real-world insights into disease epidemiology, treatment patterns, and outcomes, serving as essential resources for research and clinical decision-making. The HealthTree CureHub Registry is a multicenter, patient-driven registry designed to collect and analyze longitudinal clinical and omic data to enhance treatment strategies and improve patient outcomes across multiple diseases. Methods: The HealthTree Cure Hub Registry is a digital health patient advocacy platform that harmonizes electronic health records and patient-reported outcomes for hematologic malignancies, currently for 7,917 patients, as of January 2025. Data collection includes, but is not limited to, demographics, diagnosis history, treatment regimens, and genetic markers. Patients voluntarily contribute their health information through the HealthTree platform through connections based on the FHIR standard, allowing continuous updates and integration with clinical and omic data across 380 unique facilities. Results: The distribution of the patients gathered in the registry includes 6456 (82%) diagnosed with MM, 949 (12%) with SMM, 438 (5.5%) with MGUS, and 73 (0.9%) with PCL. The median age at MM diagnosis is 60 years (IQR: 53–66), and the cohort comprises 58% female and 42% male patients. The registry includes comprehensive cytogenetic data, Among patients with available FISH data (n = 2542), the most common abnormalities included trisomies (44%), the most prevalent trisomies included trisomy 9 (25%), trisomy 11 (20%), trisomy 15 (18%), and trisomy 5 (11%). Deletion 13q showed in 40%, 1q21 additions (36%), which is frequently linked to poor prognosis and increased treatment resistance, and t(11;14) (2%), which is frequently linked to BCL-2 inhibition sensitivity. High-risk chromosomal translocations showed for t(4;14) in 12%, t(14;16) (6%) and t(14;20) (2%), additionally, del(17p) was found in 16% of patients. Conclusions: The HealthTree Cure Hub Registry serves as a comprehensive real-world data platform, supporting clinicians, researchers, and policymakers in understanding MM progression. By integrating patient-reported outcomes with clinical and genomic data, the registry enhances research capabilities and fosters personalized treatment strategies. Continued expansion and data curation will further strengthen its role in improving MM care worldwide.

Molecular testing practice patterns for lung cancer based on gender: Insights into disparities and opportunities for precision medicine.

e20523 Background: Targeted and biomarker-driven immunotherapies have revolutionized outcomes of lung cancer, yet gender disparities may impact care. This study explores gender-specific patterns in testing, biomarker distribution, and clinical decisions in lung cancer. Methods: Genomic data from 2,743 lung cancer patients across multiple oncology centers, encompassing 3,326 NGS-based multigene variant profiling tests performed at Datar Cancer Genetics was used for this analysis. These tests utilized either tumor tissue DNA (ttDNA, n=1,395) or circulating cell-free DNA (cfDNA, n=1,931) to identify actionable alterations. Results: Median age was 63 years for men and 60 years for women. Histopathological subtypes included adenocarcinoma (86%), squamous carcinoma (9%), neuroendocrine tumors (3%), and adenosquamous carcinoma (2%). Adenocarcinoma was more common in women (90.53%) than men (82.55%), while squamous cell carcinoma was higher in men (11.77%) than women (4.26%). Adenosquamous carcinoma was rare and similar across genders (around 2%). Small cell carcinoma was more prevalent in men (2.19%) than women (1.42%). In India, lung cancer incidence is higher in men (76%), largely due to greater tobacco consumption among men than women. However, in the testing cohort, women constituted 40% of the group, suggesting a higher adoption of molecular testing among this gender. This referral bias may arise from women's higher likelihood of having targetable driver mutations as non-smokers. The data reveals gender disparities in molecular testing, with women more likely to undergo testing and show higher rates of targetable alterations. While biological differences partly explain these trends, factors like physician bias, resource access, and patient engagement, may also contribute, ultimately affecting clinical outcomes. Conclusions: Addressing barriers to molecular testing, which may include clinician gender biases, is critical for advancing equity in precision oncology. Future prospective studies are warranted to validate these findings and explore underlying reasons for observed discrepancies. By ensuring equitable access to advanced molecular testing technologies across genders, precision oncology can fully realize its potential in improving lung cancer outcomes. Distribution of tissue-based key genetic alterations across genders in lung cancer. Gene Altered Adenocarcinoma Lung Squamous Carcinoma Lung Female(N=426) Male(N=605) p-value(Chi-square test) Female(N=22) Male(N=90) p-value(Chi-square test) EGFR 49.30% 31.07% <0.00001 31.82% 13.33% 0.038377 KRAS 12.44% 21.82% 0 .000112 4.55% 6.67% 0.712536 BRAF 2.82% 2.98% 0.881623 0% 0% - MET 3.99% 3.80% 0.877053 0% 5.56% 0.786588 ERBB2 2.11% 4.30% 0.056454 0% 0% - ALK 9.39% 5.95% 0.037436 4.55% 2.22% 0.545174 ROS1 3.76% 2.98% 0.489456 0.00% 0.00% - RET 0.94% 2.48% 0.070172 4.55% 1.11% 0.27555

Digital Health and Genomics.

Digital Health and Genomics.

Prognostic impact and racial disparities of chromosome 8q gains in prostate cancer: Insights from integrated molecular profiling.

e17121 Background: Chromosome arm 8q is frequently amplified in prostate cancer, harboring genes that drive tumor progression, therapy resistance, and poor outcomes. Key loci include MYC , which regulates cell proliferation; AR , central to prostate cancer biology and resistance; and RAD21 , essential for chromosomal stability. While the prognostic implications of 8q gains have been explored, their role in driving racial disparities in prostate cancer outcomes remains underexplored. This study investigates chromosome 8q amplification as a prognostic biomarker, examining its differential prevalence and impact across racial groups to elucidate its contribution to disparities in disease progression and patient survival. Methods: Using cBioPortal's clinical and genomic data from Memorial Sloan Kettering Cancer Center, this work examined two datasets from 2021 and 2024. Comprehensive genomic profiling included copy number alterations, mutational co-occurrence of important 8q genes, and mutational analysis. Multivariate Cox regression and Kaplan-Meier survival analysis were conducted to evaluate the impact of 8q amplifications on overall survival (OS) and disease outcomes, stratified by racial demographic groups and disease stages. Results: In this cohort of 1,577 prostate cancer patients, 652 (41%) exhibited chromosome 8q amplification, which correlated with aggressive tumor phenotypes. The most common amplifications were observed in PTEN (21%), FOXA1 (17%), AR (16%), and MYC (7%),, with notable co-occurrences observed between MYC-RAD21, MYC-NBN, and PRDM14-ELOC ( P = <0.001), linking these alterations to genomic instability. Kaplan-Meier analysis showed patients with 8q gain had significantly inferior overall survival (HR = 2.31, 95% CI: 1.99 - 2.69, P < 0.001, mOS gain vs neutral: 35mo vs 55mo). Black patients exhibited a higher frequency of 8q gain (10% vs. 6.1%, P = 0.003) and distinct oncogenic profiles, including MYC (12%), RAD21 (7%), and LYN (4%) alterations. While Black patients experienced inferior overall outcomes, chromosome 8q amplification was a consistent predictor of poor prognosis across all racial groups. Conclusions: This study establishes chromosome 8q amplification as a robust prognostic biomarker in prostate cancer. The integration of clinical and genomic data highlights the potential of 8q amplification to guide personalized therapeutic strategies. These findings underscore the value of incorporating 8q analysis into clinical risk stratification frameworks to enhance treatment decision-making and address racial disparities in prostate cancer outcomes.

Active cancer in the era of COVID-19: A multivariable meta-analysis vs non-cancer controls with variant imputation.

e23092 Background: Patients with active cancer are at increased risk from SARS-CoV-2 infection, yet the influence of recent cancer treatment, tumour subtype, and specific variants of concern (VOCs) remains unclear. We conducted a multivariable-adjusted meta-analysis to compare mortality and clinical severity between individuals with active cancer (diagnosed or treated within three years) and non-cancer controls, spanning pre- to post-Omicron periods. Methods: We searched Medline, Embase, Cochrane Central, and the WHO COVID-19 Research Database up to 25 November 2024 for eligible cohort studies enrolling ≥10 active cancer patients and reporting multivariable-adjusted SARS-CoV-2 outcomes against non-cancer comparators. We evaluated mortality (all-cause or COVID-19–specific) and severity (WHO ordinal scale) by tumor type, metastatic status, and dominant VOC, imputed via genomic data from NCBI GenBank and GISAID. Random-effects models generated pooled odds ratios (ORs) and 95% confidence intervals. Results: From 30 cohort studies (n = 281,270 with active cancer; n = 18.88 million controls), overall pooled analysis showed that active cancer significantly increased the odds of mortality (OR 1.70, 95% CI 1.36-2.12) and hospitalization (OR 1.58, 95% CI 1.22-2.06). Hematologic malignancies had higher ORs for mortality (2.10, 95% CI 1.43-3.07) than solid tumors (1.40, 95% CI 1.12-1.73), although the difference was not statistically significant ( p = 0.07). Thoracic and colorectal tumors, as well as metastatic disease, were associated with notably increased mortality risk. When stratified by VOC, Alpha (OR 4.59) and Omicron (OR 2.74) produced greater excess mortality in cancer patients compared to earlier SARS-CoV-2 lineages. Conclusions: Active malignancy elevates mortality and hospitalization relative to individuals without cancer, particularly among those with thoracic, colorectal, or metastatic disease. VOC-specific trends indicate greater vulnerability during Alpha and Omicron waves. These findings highlight the need for vigilant genomic surveillance and tailored clinical strategies to protect high-risk cancer populations against emerging future SARS-CoV-2 variants.

Clinical characteristics and recurrence patterns in Hispanic vs. non-Hispanic patients with leiomyosarcoma: Findings from a retrospective single-center cohort study.

e23544 Background: Leiomyosarcoma (LMS), a rare and aggressive malignancy, commonly arises from the uterus, retroperitoneum, and extremities, carrying a poor prognosis. This study investigates differences in clinical characteristics, treatment patterns, outcomes, and genomic data from next-generation sequencing (NGS) between Hispanic and non-Hispanic patients with non-metastatic and metastatic LMS. The goal is to identify disparities to inform personalized treatment strategies and improve outcomes for underrepresented populations. Methods: This single-center retrospective cohort study, conducted under an IRB-approved protocol, evaluated 190 patients with biopsy-proven LMS confirmed by sarcoma pathologists at the University of Miami Sylvester Cancer Center. NGS data were collected for patients tested between January 2009 and December 2024. Demographics, like sex, ethnicity, and age, along with clinical variables such as tumor location, size, grade, and recurrence, were gathered from electronic medical records and stored in REDCap. Kaplan-Meier and logistic regression models were used for survival analysis, adjusting for confounding variables. Patients with incomplete clinical histories were excluded. Results: The dataset included 190 patients (154 females, 33 males) with a median age of 55.8 years. Of these, 55 were metastatic, and 132 were nonmetastatic. Among nonmetastatic patients, 81 were high grade, including 55 non-Hispanic and 26 Hispanic patients. In the nonmetastatic, high-grade LMS cohort, disparities were observed: 5-year overall survival (OS) was 87.7% for non-Hispanics compared to 76.2% for Hispanics (p=0.045; hazard ratio: 1.96). Median disease-free survival (DFS) was shorter for Hispanics (14.3 months) than for non-Hispanics (50.7 months), indicating earlier recurrence among Hispanics (p=0.32). At 24 months, DFS was 39.4% for Hispanics versus 72.2% for non-Hispanics. Initial therapies, including radiation, surgery, and systemic treatments, showed no significant differences in recurrence rates, though non-Hispanic patients were more likely to receive radiation. Conclusions: These findings highlight survival disparities between Hispanics and non-Hispanics, potentially driven by earlier recurrence. NGS data from 38 patients' diagnostic tissue will be analyzed to identify potential drivers of these disparities and inform equitable treatment strategies. Clinical characteristics in patients with high grade, nonmetastatic leiomyosarcoma. Not Hispanic or Latino (N=55) Hispanic or Latino (N=26) Overall (N=81) Retroperitoneum 12.7% 11.5% 12.3% Trunk/Extremity 23.6% 34.6% 27.2% Uterine 63.6% 53.8% 60.5% First Therapy - Radiation 3.6% 0% 2.5% First Therapy - Surgery 76.4% 73.1% 75.3% First Therapy - Systemic 18.2% 19.2% 18.5% Had Recurrence 61.8% 65.4% 63.0% Had NGS 76.4% 65.4% 72.8%

AI-based predictive tool for identifying druggable mutations in lung cancer using nationwide comprehensive genomic profiling data.

e20642 Background: Comprehensive genomic profiling (CGP) plays a pivotal role in precision medicine. However, the low probability of discovering mutation-based treatments, despite the financial and time burden, may prevent eligible patients from undergoing CGP. To enhance the efficiency and efficacy of cancer precision medicine, it is critical to identify patients who are likely to benefit from CGP. This study aims to identify patient characteristics associated with the discovery of mutation-based treatments through CGP and to develop an intuitive AI tool to predict the probability of identifying druggable mutations. Methods: We retrospectively analyzed data from 3,470 lung cancer patients (Cohort 1) who underwent CGP between Jun 2019 and Nov 2023 and were registered in the Center for Cancer Genomics and Advanced Therapeutics (C-CAT) database, which covers 99.7% of CGP performed in Japan. Using clinical information available prior to CGP, we developed an eXtreme Gradient Boosting (XGBoost) model to predict the detection of druggable mutations. SHapley Additive exPlanations (SHAP) was employed to extract features that contribute to the model prediction. Using the identified clinical factors as input, another AI model was built and deployed as a smartphone application to predict the probability of identifying druggable mutations. The app’s performance was tested on clinical data from 1,307 lung cancer patients (Cohort 2) who underwent CGP between Dec 2023 and Nov 2024. Results: The predictive AI model trained on Cohort 1 achieved an AUROC of 0.851 (sensitivity: 0.825, specificity: 0.733). A separate model, excluding patients with at least one druggable mutation detected by small companion diagnostic tests before CGP, showed an AUROC of 0.791. Positive SHAP values were associated with adenocarcinoma and the number of metastatic sites, while negative SHAP values were with male sex and smoking history. These relationships were consistent across tissue and liquid CGP cases. Notably, among tissue CGP cases, patients with lung or bone metastases exhibited a significantly higher rate of druggable mutation detection compared to those without (lung: p < 0.001, bone: p < 0.05). A streamlined AI model was retrained using the most influential clinical factors and deployed as a smartphone application. When tested on Cohort 2, the app demonstrated predictive accuracy with an AUROC of 0.766 (sensitivity: 0.716, specificity: 0.695) and a Brier score of 0.188. Conclusions: This study identified key clinical factors predictive of druggable mutation detection in lung cancer through explainable AI analysis of nationwide CGP data. Based on these findings, a smartphone application was developed to predict the probability of identifying druggable mutations. This app could expand access to targeted therapies by facilitating broader utilization of CGP tests.

Analysis of BAI3 as a potential biomarker for lung cancer immunotherapy.

e20556 Background: Brain angiogenesis inhibitor protein 3 (BAI3/ADGRB) belongs to the family of adhesion G protein-coupled receptors (GPCRs). Recently, emerging evidence has shown that BAI3 plays a role in the sensitivity of biliary duct cancer to immunotherapy. Here we explored the relationship between BAI3 mutation and potential effect of immunotherapy for lung cancer based on multidimensional data. Methods: The whole exome sequencing (WES) data of immune checkpoint inhibitor (ICI) treated lung cancer patients (n=137) were derived from cBioPortal website ( https://www.cbioportal.org ) to analyze the association between BAI3 mutation and efficacy of ICIs therapy. The TCGA genomic and RNA data for 963 lung cancers were used to evaluate tumor mutational burden (TMB) differences and potential mechanism between BAI3 mutation group and wildtype group. Results: In immunotherapy cohort, 10.9% (15/137) patients harbored BAI3 mutation. BAI3 mutant lung cancers had higher TMB level than wildtype patients (P<0.001). Patients with BAI3 mutation were identified to be associated with prolonged progression-free survival (PFS) than wildtype group (HR=0.25, 95%CI: 0.09-0.69, P=0.0036). A multivariable analysis using Cox proportional-hazards regression demonstrated that BAI3 mutation was significantly associated with better PFS (HR=0.22; 95%CI: 0.078-0.62; P = 0.004), adjusting for histology type, PD-L1 expression, gender, smoke status and treatment regimen. In TCGA cohort, 13.1% (126/963) lung cancers had BAI3 mutation. BAI3 mutant patients also had higher TMB than wildtype group (P<0.001). RNA data revealed that enhanced anti-tumor immunity, characterized by a higher abundance of CD8 T cells, was observed in BAI3 mutant tumors (P=0.028). Conclusions: BAI3-mutated lung cancer patients have a higher TMB, longer PFS with immunotherapy and show higher abundance of CD8 T cells. These results indicates that BAI3 mutation may serve as a potential biomarker of ICI benefit in lung cancers. Moreover, further clinical insights and prospective validation studies are warranted.

Genomic landscape of clinically acquired resistance alterations in patients treated with KRASG12C inhibitors.

Genomic landscape of clinically acquired resistance alterations in patients treated with KRASG12C inhibitors.

Advancements and implications of artificial intelligence for early detection, diagnosis and tailored treatment of cancer.

Advancements and implications of artificial intelligence for early detection, diagnosis and tailored treatment of cancer.

Clinical characteristics and treatment outcomes of hepatosplenic T-cell lymphoma: Mayo Clinic experience.

7075 Background: Hepatosplenic T-cell lymphoma (HSTCL) is a rare, aggressive peripheral T-cell lymphoma arising primarily from γδ T-cells. It carries a poor prognosis and resists conventional chemotherapy. Most reports on HSTCL are case-based. This study comprehensively analyzes a large cohort, evaluating treatment strategies and survival outcomes. Methods: This retrospective study included patients (pts) with pathologically confirmed HSTCL diagnosed between 2000-2024, consecutively seen at Mayo Clinic MN. Clinical, pathological, genomic, and treatment-related data were extracted when available. Descriptive statistics were used to summarize baseline characteristics. Time-to-event analyses, including Kaplan-Meier estimates, median overall survival (OS), and survival time estimates were conducted from the date of diagnosis. Results: A total of 20 patients with newly diagnosed HSTCL were included, with a median age of 57 years (range: 35-71). The cohort was predominantly male (70%) and non-Hispanic (93%). Molecular data was available for five patients, revealing abnormalities in STAT5B, MLL3 deletion, TP53, EZH2, TERT, and NF1 E291D . The median follow-up was 27.6 months (m) with a median OS of 17.6 m (95% CI: 11.2 - NA). First-line treatment was anthracycline-based in 68% of pts and non-anthracycline-based in 32%, with higher response rates in the latter group (50% vs.83%). Although non-anthracycline regimens showed a trend toward improved 3-year OS (100% vs. 29%) the difference was not statistically significant (p = 0.16). Achieving a complete response to first-line therapy was also associated with a trend towards a better 3-year OS compared to refractory disease (80% vs. 50%, p = 0.17). Most pts (79%) underwent hematopoietic stem cell transplant (HSCT), primarily allogeneic, with only one receiving autologous HSCT. First-line therapy before HSCT was evenly distributed between anthracycline (55%) and non-anthracycline (45%) regimens. HSCT recipients had significantly higher 3-year OS than non-recipients (83% vs. 33%, p = 0.017). Notably, the patient with a TP53 mutation has remained in remission for over a year post-allogeneic HSCT. Conclusions: HSTCL predominantly affects younger pts, with nearly half dying within a year. Allogeneic HSCT, rarely used in other NHL subtypes, improved survival. Non-anthracycline regimens and achieving CR trended toward better outcomes. Our study, leveraging a sizable cohort, highlights the need for targeted research and novel therapies to improve HSTCL management. Summary of survival outcomes in HSTCL. Characteristics Median OS (y) 3-Year OS (95% CI) 1.48 [0.93- NA] 46% [0.26 - 0.81] Transplant Transplant NA [NA - NA] 83% [0.58 - 1.00] No Transplant 1.02 [0.93 - NA] 33% [0.07 - 1.00] Treatment Anthracycline Based 1.02 [0.36 - NA] 29% [0.11 - 0.73] Non-Anthracycline Based 4.61 [NA- NA] 100% [1.00 - 1.00]

Genomic landscape and biomarker analyses utilizing circulating-tumor DNA in advanced esophageal squamous cell carcinoma: Sub-analysis of SCRUM-MONSTAR GOZILA.

4024 Background: Advanced esophageal squamous cell carcinoma (ESCC) is a cancer type with a poor prognosis, with limited survival benefits from current multimodal approaches. The incomplete understanding of molecular mechanisms in advanced ESCC has hindered the development of effective targeted therapies, emphasizing the critical need for identifying predictive biomarkers and novel therapeutic targets. Methods: SCRUM-MONSTAR GOZILA is a nationwide plasma-based genomic profiling study utilizing Guardant360 in Japan, which aimed to analyze circulating tumor DNA (ctDNA) genomic alterations in patients with advanced solid tumors, including ESCC. We evaluated the genomic landscape with advanced ESCC patients and investigated associations between genomic alterations and overall survival (OS) using the log-rank test. The correlation between progression-free survival (PFS) and blood tumor mutation burden (bTMB) in immune checkpoint inhibitor (ICI) monotherapy was also assessed using multiple cut-off values (2, 4, 6, 8, and 10 mutation/Mb). Results: The present study included 313 patients with available genomic and clinical data. The gene alteration spectrum comprised mutations (single nucleotide variants, 71.6%; and insertions/deletions, 10.7%), copy number alterations (CNAs, 17.3%), and fusions (0.48%). TP53 was the most frequently altered gene (88.5%), followed by PIK3CA (36.4%), NFE2L2 (24.3%), CCND1 (22.4%), EGFR (20.1%), ATM (16.3%), FGFR1 (10.2%), BRCA2 (10.2%), MET (9.6%) and ARID1A (9.6%). Regarding the survival outcomes, PIK3CA CNA was significantly associated with worse OS compared to those with PIK3CA wild type [hazard ratio (HR), 1.84; 95% confidence interval (CI), 1.24–2.74; p-value, 0.0002], and PIK3CA mutation showed a trend toward shorter OS (HR, 1.43; 95%CI, 0.94–2.17; p-value, 0.06). Patients with both PIK3CA mutation and CNA exhibited significantly worse OS compared to those with PIK3CA wild type (HR, 1.94; 95%CI, 0.85–4.45; p-value, 0.03). Both FGFR1 CNA and mutation were associated with poorer OS (HR, 1.98; 95%CI, 1.03–3.79; p-value, 0.005; and HR, 2.84; 95%CI, 0.89–9.07; p-value, 0.002, respectively). CNA in CCND1 and EGFR , and mutation in NFE2L2 and RB1 also significantly correlated with worse OS (any p-value≤0.01). Among 142 patients treated with ICI monotherapy, no statistically significant differences in PFS were observed at any cut-off value of bTMB (any p-value > 0.1). Conclusions: This comprehensive analysis of ctDNA profiles revealed distinct genomic alterations with prognostic significance in advanced ESCC. Multiple alterations demonstrated significant associations with poor OS, meanwhile bTMB was not validated as an effective predictive biomarker for ICI efficacy. These findings provide insights into potential therapeutic targets and prognostic biomarkers in advanced ESCC. Clinical trial information: 2021-GB-009 .

Effect of extrachromosomal DNA (ecDNA) on MYCN amplified neuroblastoma and patient outcomes.

3090 Background: Recurrent cytogenetic abnormalities represent candidate therapeutic targets for children with neuroblastoma (NB). MYCN oncogene amplification is associated with significantly worse survival rates for children with NB and remains one of the primary predictors of patient prognosis. MYCN amplifications in NB can be found both within the linear genome and on circular extrachromosomal DNA (ecDNA), and therapeutic targeting of the mechanisms underlying MYCN amplification represents a novel and promising strategy in NB. However, the molecular features and clinical and biological significance of these amplifications in NB tumors are not sufficiently understood. Methods: Whole genome and RNA sequencing data were analyzed for NB cell lines and NCI TARGET NB samples using AmpliconSuite software for ecDNA identification and characterization. GISTIC was used for identification of recurrently amplified regions. Gene expression levels were determined using StringTie, and gene clustering heatmaps were generated using FeatureCounts software. For differential gene expression analyses, samples were divided into ecDNA + and ecDNA - , and genes contained on ecDNA were compared to the same regions on linear DNA across samples using DESeq2. Associations between ecDNA quantity, content, and patient survival were performed using multivariate Cox regression survival analysis. Associations of gene expression with patient survival were performed using the R2 Platform. The efficacy of targeting ecDNA-associated gene products was assessed using live cell imaging and cell viability assays. Results: WGS analysis confirmed 7/20 NB patient tumors from the TARGET database to be ecDNA amplified with 1-5 independent ecDNA elements and MYCN gene expression correlated with the ecDNA copy number. ecDNAs in MYCN -amplified neuroblastoma cell lines contained distinct gene combinations and possessed unique structures. MYCN overexpression in NB cells has been shown to be associated with replication stress (RS), and tumor cells containing ecDNA are hyper-reliant on the DNA damage response (DDR) kinase CHK1 to manage heightened replication stress. Expression of the CHK1 gene was associated with neuroblastoma patient outcomes and neuroblastoma was most significantly associated with CHK1 RNA dependency. We further validated CHK1i as a promising therapeutic strategy in MYCN amplified NB, as CHK1 inhibition with the novel inhibitor BBI-2779 was most effective against ecDNA+, MYCN -amplified neuroblastoma cell lines. Conclusions: Our results emphasize the critical role of ecDNA in NB. We identify a synthetic lethality axis shaped by ecDNA MYCN amplification and CHK1 dependence. We further demonstrate the feasibility of targeting this vulnerability through CHK1 inhibition, thus offering new avenues for treatment in MYCN amplified tumors.

Differential predictive impact of PD-L1 expression on immunotherapy outcomes and immunophenotype in squamous versus non-squamous NSCLC.

8562 Background: PD-L1 tumor proportion score (TPS) is a key biomarker for immune checkpoint inhibitors (ICIs) efficacy in non-small cell lung cancer (NSCLC), but its predictive value in patients (pts) with squamous (SQ) histology remains uncertain, highlighting the need for histology-specific studies. Methods: Clinicopathologic, genomic, and outcomes data were collected and analyzed from advanced NSCLC pts treated with ICIs ± chemotherapy (CT) at Dana-Farber Cancer Institute, Memorial Sloan Kettering Cancer Center and MD Anderson Cancer Center. Cox regression tested the association between PD-L1 levels and survival to ICIs by histology, adjusting for potential confounders such as treatment regimen and line. Multiplexed immunofluorescence (mIF) on baseline tissue samples quantified CD8+, PD1+, CD8+/PD1+, and FOXP3+ densities, stratifying by PD-L1 TPS and histology. Results: Among 4967 NSCLC pts treated with ICIs ± CT, 727 (14.6%) had SQ histology. Among pts with available PD-L1 TPS, 1359 (37.9%) had TPS <1%, 1061 (29.6%) 1–49%, and 1167 (32.5%) ≥50%. Increasing PD-L1 TPS of <1%, 1–49% and ≥50% correlated with significant stepwise improvements in progression-free (PFS) and overall survival (OS) in pts with NonSQ NSCLC but not in those with SQ (Table 1). In SQ NSCLCs, there was no difference in PFS and OS between pts with PD-L1 TPS of 1–49% vs ≥50%, while only a dichotomized PD-L1 TPS (<1% vs ≥1%) was predictive of longer survival in this histology (PFS adjusted hazard ratio [aHR]: 0.72, p<0.01; OS aHR: 0.76, p=0.02). Comparing histologies, PFS and OS to ICIs ± CT were similar between SQ and NonSQ NSCLCs in PD-L1 TPS subgroups of <1% and 1–49%. However, among pts with a PD-L1 TPS ≥50%, those with NonSQ NSCLC had longer survival compared to SQ (PFS aHR: 1.30, p=0.01; OS aHR: 1.43, p<0.01), indicating stronger predictive value of increasing PD-L1 TPS levels only in NonSQ. mIF analysis (229 samples: 22 SQ, 207 NonSQ) showed lower intratumoral CD8+, PD1+, CD8+/PD1+, and FOXP3+ densities in SQ vs NonSQ. Increasing PD-L1 TPS significantly correlated with higher CD8+ cells in NonSQ NSCLCs (R = 0.25, p<0.01) but not in SQ (R = -0.034, p = 0.89). A similar association was observed for PD1+, CD8+/PD1+, and FOXP3+ cells. Conclusions: Increasing PD-L1 levels show stepwise PFS and OS improvements in NonSQ but not in SQ NSCLCs, where TPS acts as a dichotomous (<1% vs ≥1%) rather than continuous predictor. These findings have implications for treatment decision making as well as ICIs trial design and interpretation. PD-L1 TPS<1% vs 1-49% PD-L1 TPS<1% vs ≥50% PD-L1 TPS1-49% vs ≥50% SQ PFS mo aHR, p 4.0 vs 6.6 0.72, <0.01 4.0 vs 6.2 0.71, 0.01 6.6 vs 6.2 0.99, 0.95 NonSQ PFS mo aHR, p 4.6 vs 5.8 0.79, <0.01 4.6 vs 8.2 0.56, <0.01 5.8 vs 8.2 0.70, <0.01 SQ OS mo aHR, p 13.0 vs 17.0 0.77, 0.04 13.0 vs 17.5 0.76, 0.06 17.0 vs 17.5 0.98, 0.92 NonSQ OS mo aHR, p 14.7 vs 18.3 0.81, <0.01 14.7 vs 27.7 0.59, <0.01 18.3 vs 27.7 0.72, <0.01