Citations

Abstract Rationale Inflammatory bowel disease (IBD) is increasingly recognized as a systemic disorder, with extraintestinal manifestations (EIMs) involving multiple organ systems. Pulmonary involvement, particularly interstitial lung disease (ILD), remains an underrecognized manifestation of IBD. Its prevalence, clinical characteristics, and risk factors in IBD remain poorly defined. Methods We conducted a retrospective study of adults (≥18 years) with IBD who had at least one encounter in our health system between January 1, 2013 and December 31, 2023. IBD, and ILD cases were identified using ICD-9/ICD-10 codes, with ILD diagnoses confirmed by thoracic radiologists. Demographic, clinical, pulmonary, smoking, medication, and surgical data were abstracted. Associations between patient characteristics and ILD were evaluated using both a cohort design and a matched case-control design (1:3 ratio). Mortality was assessed using a Cox proportional hazards model with ILD as a time-dependent covariate. Results Among 875 patients with IBD, 109 (12%) had ILD. Radiographic patterns included nonspecific interstitial pneumonia (NSIP, 27%), indeterminate (26%), definite/probable UIP (24%), hypersensitivity pneumonitis (15%), and organizing pneumonia (8%). Pulmonary function testing demonstrated predominantly restrictive physiology and impaired diffusion capacity. In multivariable analysis, higher BMI (OR 1.04), bronchiectasis (OR 2.65), ex-smoking (OR 2.15), and use of adalimumab (OR 2.04) or infliximab (OR 2.14) were independently associated with ILD. The matched case-control design confirmed these findings and also identified older age (OR 1.04) as a significant factor. ILD was strongly associated with mortality (HR, 3.78; 95% CI, 2.50-5.72; p < 0.001). Conclusions ILD occurs in a meaningful subset of patients with IBD and is associated with older age, higher BMI, prior smoking, bronchiectasis, and TNF inhibitor use. ILD confers a markedly increased risk of death, highlighting the importance of early recognition and close monitoring in this population. This abstract is funded by: None

Abstract Scientific foundation models learn high-dimensional representations from diverse data modalities, yet what they encode and how to extract that knowledge remain open questions. Here we show that probing the internal representations of Evo 2, a 7-billion-parameter genomic foundation model, enables accurate and interpretable genetic variant effect prediction. We introduce a covariance-based probe that captures second-order structure from Evo 2 sequence embeddings to predict variant pathogenicity across variant types and functional consequences, matching or exceeding specialized predictors within their domains. To ground these predictions in known biological mechanisms, we train a complementary panel of probes on existing annotations to detect which genomic properties are disrupted by a variant. This categorized evidence is then integrated with each variant’s genomic context through a language model to generate variant-specific mechanistic hypotheses. Our pathogenicity predictions correlate with experimental measures of variant function, clinical penetrance, and biobank disease associations while the mechanistic hypotheses are consistent with expert reviews, known mechanism classes, and downstream molecular readouts. We release pathogenicity scores, disruption profiles, and contextualized interpretations for 4.2 million variants from the ClinVar database as an open resource through the Evo Variant Effect Explorer (EVEE). More broadly, this structured probing approach offers a general framework for interrogating foundation models across scientific disciplines and grounding their outputs in existing domain concepts.

Abstract Objectives Patients referred for specialized care often arrive with outside medical records (OMRs) compiled into multi-report PDFs that include imaging, pathology, and clinical notes in unstructured formats. Reviewing these records is time consuming and mentally taxing, increasing the risk of delayed care, clinician frustration, and missed information affecting quality of care. This study aimed to automate the segmentation, classification, and date extraction of scanned OMRs, with a focus on records relevant to breast cancer care. Materials and Methods We used optical character recognition (OCR) to extract machine-readable text from 1303 scanned PDF documents from 116 distinct external institutions. Gemini 1.5, a large language model (LLM), was then used to segment multi-report files into individual documents, classify them into clinically meaningful categories such as mammograms and pathology reports, and extract study dates to build diagnostic timelines. Document categories were informed by clinical workflows in a breast cancer center. Results The system achieved an F1 score of 0.95 for segmentation, 0.96 for classification, and 0.90 for date extraction. In a pilot of 45 records reviewed by clinicians, only 2 classification errors and 1 date error were reported. Clinicians estimated that the tool reduced OMR review time by 40%, improved workflow efficiency, and increased satisfaction. Discussion Our findings demonstrate that combining OCR with LLMs can significantly enhance the processing of unstructured medical records, reducing manual burden and supporting timely clinical decision-making. Conclusion This study demonstrates the successful application of OCR and LLMs for organizing scanned OMRs within a specialty clinic. By automating a previously manual process, the approach supports scalable review of incoming outside records and has potential for adaptation to other clinical workflows. Future work will focus on evaluating the system across additional specialties and institutions.

Spontaneous pneumothorax (SP) is defined by the presence of air in the pleural space arising from neither trauma nor iatrogenic causes. It can develop secondary to various etiologies. The familial clustering observed in some patients with SP supports the view that genetic factors play a role in the pathogenesis of SP. Several recognizable pneumothorax-associated genetic syndromes exist, including Birt-Hogg-Dubé syndrome (BHD), tuberous sclerosis complex-associated lymphangioleiomyomatosis (TSC-LAM), Marfan syndrome (MFS), cystic fibrosis (CF), alpha-1 antitrypsin deficiency (AATD), vascular Ehlers-Danlos syndrome (vEDS), Loeys-Dietz syndrome (LDS), and a few others. Recognition of these syndromes underlying SP facilitates optimal management and counseling regarding prognosis and potential comorbidities. In this review, we systematically summarize several genetic syndromes associated with pneumothorax, in which SP may manifest either as an initial presenting symptom or as a potential complication that adversely affects the prognosis.

Background Multimorbidity is common in patients with atrial fibrillation (AF); however, the impact of the number and type of comorbid conditions on outcomes remains uncertain. Methods This cohort study included patients with new‐onset AF from a Midwest region between 2013 and 2017. Eighteen chronic conditions at the time of AF were classified into groups: cardiometabolic, other somatic, and mental health. Cox regression determined associations between the number of each condition type with death, ischemic stroke/transient ischemic attack, and congestive heart failure, stratified by age. Results Among 16 509 patients with AF (mean age, 74 years; 43% women), the mean number of cardiometabolic, other somatic, and mental health conditions was 2.7, 1.4, and 0.5, respectively. The number and type of conditions had a varying impact on outcomes and differed by age. A higher number of cardiometabolic conditions were associated with increased risk of death within 90 days only in people aged ≥85 years (≥4 conditions versus 0: hazard ratio [HR], 1.74 [95% CI, 1.05–2.87]), whereas for death after 90 days, associations were strongest in the youngest age group (<65 years: HR, 1.83 [95% CI, 1.35–2.46]; 65–74 years: HR, 1.34 [95% CI, 1.04–1.73]; 75–84 years: HR, 1.32 [95% CI, 1.09–1.60]; ≥85 years: HR, 1.14 [95% CI, 0.96–1.35]). Associations with outcomes were generally strongest in the youngest age group and attenuated with older age for higher number of other somatic conditions, whereas the pattern was less consistent for mental health conditions. Conclusions Along with cardiometabolic‐related conditions, other somatic and mental health conditions are important predictors of outcomes in AF, with effects differing by age, and should be considered when caring for these patients.

Abstract Background: Young BRCA carriers may undergo breast conserving surgery (BCS) or mastectomy with or without contralateral risk reducing mastectomy at the time of an index breast cancer diagnosis. A variety of factors may influence surgical decision making. Methods: The BRCA BCY Collaboration (NCT03673306) is an international, hospital-based, retrospective cohort study conducted at 109 centres across 5 continents including women with germline pathogenic or likely pathogenic variants (PV) in BRCA1/2 diagnosed between 2000 and 2020 with invasive stage I-III breast cancer prior to or at age 40 years. The primary objective of the present analysis was to evaluate surgical management in these patients. Breast surgery was defined as BCS, unilateral mastectomy, or bilateral mastectomy administered as part of primary treatment for a first diagnosis of invasive breast cancer. Multivariable logistic regression was used to determine factors associated with receipt of bilateral mastectomy. Results: Of 5660 eligible BRCA1/2 carriers, 4715 had unilateral stage I-III breast cancer with surgical treatment details available. The median age of the cohort was 35 years (IQR, 31-38 years). Overall, 1805 (38.3%) patients underwent BCS, 1752 (37.2%) underwent unilateral mastectomy, and 1158 (24.6%) underwent bilateral mastectomy. The proportion of young BRCA1/2 carriers undergoing BCS decreased over the study period (55.7% in 2000 vs. 27.0% in 2020), whereas the use of bilateral mastectomy significantly increased (8.0% in 2000 vs. 44.4% in 2020). Bilateral mastectomy was more common in women who had genetic testing performed before (57.0%) or within 6 months of a breast cancer diagnosis (38.5%), compared to those tested after diagnosis (5.1%, p&lt;0.001). On multivariable analysis, compared to those tested after diagnosis, earlier genetic testing remained the strongest predictor of bilateral mastectomy receipt (testing prior to diagnosis: OR 20.7 [95% CI 15.3-28.1]; testing at diagnosis: OR 6.8 [95% CI 5.4-8.7]). Other predictors of bilateral mastectomy receipt included geographic region, later year of diagnosis, smaller tumor size, node negative disease, and neoadjuvant treatment. In subgroup analysis including 2327 patients who had genetic testing performed before or within 6 months of diagnosis, 42.0% underwent bilateral mastectomy, with the lowest rates reported in Asia/Africa (26.1%) and the highest rates reported in North America (66.4%, p&lt;0.001). In adjusted subgroup analyses, BRCA1 PV, later year of diagnosis, smaller tumor size, node negative disease, neoadjuvant treatment, and geographic region remained independently associated with bilateral mastectomy. Conclusions: In this large, international study of young affected BRCA carriers, geographic region and timing of genetic testing strongly impacted surgical management. These data support early genetic testing for young women with breast cancer to help optimize local therapy decisions. Citation Format: S. M. Wong, E. Blondeaux, V. Delucchi, F. Coussy, R. Bernstein Molho, S. Linn, A. Di Meglio, A. Kwong, K. Pogoda, E. Agostinetto, I. Nevelsteen, J. Balmana, H. C. Moore, A. Toss, A. Ferrari, C. Rousset-Jablonski, T. Renaud, K. Phillips, F. A. PeccatorI, M. Agustina Ipina, S. Paluch-Shimon, J. Ryu, W. Kui, M. Lee, R. Fruscio, M. Rozenblit, C. Vernieri, A. Matikas, M. Dieci, F. J. Couch, L. De Marchis, S. E. Hwang, D. Aguilar-y Mendez, D. Can Guven, M. Swain, C. Gianni, S. Spinaci, G. Montagna, M. Lambertini, H. Kim. Surgical management of young BRCA1/2 carriers with early onset breast cancer [abstract]. In: Proceedings of the San Antonio Breast Cancer Symposium 2025; 2025 Dec 9-12; San Antonio, TX. Philadelphia (PA): AACR; Clin Cancer Res 2026;32(4 Suppl):Abstract nr PD12-01.

Interstitial lung abnormalities (ILAs) are incidental nondependent radiologic findings that may portend early or future interstitial lung disease (ILD), but do not meet specific criteria at the time of presentation. They are subclinical by definition and found more commonly in older ever-smokers undergoing computed tomography (CT) imaging for other indications, including cardiac or lung cancer screening programs. As ILA prevalence increases, driven by an aging population and heightened awareness in younger patients, understanding risk factors for their development and progression has gained recent interest, particularly for optimizing subsequent ILD outcomes. This narrative review summarizes current ILA definitions, epidemiology, risk stratification, and management, while highlighting current challenges and knowledge gaps.

Abstract Background Prostate cancer is a heterogeneous disease comprising diverse cell types within the tumor microenvironment that support cancer cell survival. Using single cell and spatial transcriptomics, we previously identified subpopulations of treatment-persistent cancer cells exhibiting regenerative properties and a distinct molecular signature. Herein, we aimed to characterize the locked ‘Persist’ signature (PS) and its association with outcomes in a large cohort of non-metastatic prostate cancer tumors. Methods Patients with non-metastatic prostate cancer that underwent Decipher prostate genomic classifier and whole-transcriptome profiling on biopsy (BX, n=146,941) and radical prostatectomy (RP, n=55,329) specimens between October 2016 and February 2024 were included (Veracyte Inc., San Diego, CA, USA). Baseline clinical and pathological data from deidentified patients as well as signature scores were retrieved from the Decipher Genomic Resource for Intelligent Discovery (GRID) registry (Clinical Trials identifier: NCT02609269). The PS score based on the expression of 17 genes was calculated for all samples. Results PS was significantly associated with adverse clinical and pathological variables. In biopsy sample, PS was higher in cases with worse T3/4 clinical stage, Grade group (GG4-5), and high/very high NCCN risk (p&lt;0.001 for all). In RP sample, PS was significantly higher in patients with GG4-5, extraprostatic extension, seminal vesicle invasion and lymph node involvement (p&lt;0.001 for all). PS was also higher in patients with very high Decipher (&gt;0.85) in both BX and RP (p&lt;0.001 for both). When evaluating the prognosis of PS in multiple retrospective cohorts, PS in biopsy was prognostic of adverse pathology features at RP. PS was also prognostic of distance metastasis (DM) in multiple RP cohorts with long-term follow up (META855). In the ARNEO trial testing apalutamide in addition to ADT as neo-adjuvant treatment prior RP, PS was prognostic of early recurrence. In the Stampede trial, PS was prognostic of progression free survival, PCSM and OS in both non-metastatic and metastatic disease. When associating PS with other GRID signatures, PS was higher in samples with homologous recombination deficiency, TP53 mutations, Luminal B and luminal proliferating subtypes. On the other hand, PS was lower in samples with low androgen receptor-activity, RB-loss and neuroendocrine-like tumors. Finally, PS was positively correlated with radiation sensitivity signatures. Conclusions Persist is a novel molecular signature associated with more aggressive tumor characteristics. Future studies aim to further examine beyond this signature potential to identify patients at risk of aggressive prostate cancer, and understand what the optimal treatments for those with treatment-persistent prostate cancer cells are. Citation Format: Francesco Ballesio, Sinja Taavitsainen, Mohammed Alshalalfa, Antti Kiviaho, Mazlina Ismail, Reetta Nätkin, Alexander Giesen, Thomas Munro, James A. Proudfoot, Hyunnam Ryu, Elai Davicioni, Matthew C. Cooperberg, R. Jeffrey. Karnes, Ruochen Jiang, Steven Joniau, Gerhardt Attard, Frank Claessens, Matti Nykter, Alfonso Urbanucci. Persist: a molecular signature for prostate cancer patients with preexisting treatment-persistent cells associated with aggressive disease [abstract]. In: Proceedings of the AACR Special Conference in Cancer Research: Innovations in Prostate Cancer Research and Treatment; 2026 Jan 20-22; Philadelphia PA. Philadelphia (PA): AACR; Cancer Res 2026;86(2_Suppl):Abstract nr B079.