Citations

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.

Background Acute kidney injury (AKI) is associated with high morbidity and mortality in critically ill patients. Multicenter analyses describing the epidemiology of AKI in the cardiac intensive care unit (CICU) are scarce. Here, we describe the prevalence, predictors, and outcomes associated with AKI in a contemporary multicenter CICU population. Methods The Critical Care Cardiology Trials Network (CCCTN) is a collaboration of CICUs in North America coordinated by the TIMI Study Group (Boston, MA). We evaluated patient and hospital-level outcomes as a function of creatinine-only Kidney Disease: Improving Global Outcomes (KDIGO) AKI stage within the CCCTN registry. Logistic regression was used to assess associations with in-hospital mortality. Results Among 21 603 admissions, the overall prevalence of AKI was 23.7%, which was composed of KDIGO stage 1: 12.7%, stage 2: 1.7%, and stage 3: 9.2%. A higher prevalence and severity of AKI were seen in patients with baseline kidney disease, hypertension, or diabetes and in admissions with heart failure, cardiogenic shock, and cardiac arrest ( P < .001 for all). There was a stepwise increase in hospital mortality across AKI stages (stage 1: 23%, stage 2: 39%, stage 3: 42%; P < .0001). The increased risk of death associated with progressive AKI stage persisted after multivariate adjustment (adjusted OR: stage 1: 2.1, stage 2: 3.7, stage 3: 3.4). Conclusions AKI occurs in 1 in 4 admissions to the CICU, and a higher AKI stage is incrementally associated with in-hospital mortality, highlighting a need to develop strategies to mitigate AKI and its complications.

Background Cardiovascular biomarkers such as NT‐proBNP (N‐terminal pro‐B‐type natriuretic peptide) are used for heart failure risk stratification in patients with acquired heart disease, but there are limited data about its role in patients with congenitally corrected transposition of the great arteries. The current study aims to assess the role of serial NT‐proBNP measurements for risk stratification in adults with congenitally corrected transposition of the great arteries. Methods We studied adults with congenitally corrected transposition of the great arteries with ≥2 NT‐proBNP measurements (2003–2024). Temporal change in NT‐proBNP was calculated as relative change from baseline levels. Results Of 152 patients (age 47±13 years, 45% men), baseline NT‐proBNP was 539 (interquartile range, 193–1549) pg/mL, and relative change from baseline levels of NT‐proBNP was 34% (interquartile range, −11 to 67). Of 152 patients, 45 (30%) had cardiovascular events (heart failure hospitalization [n=29], ventricular assist device [n=5], heart transplantation [n=11], and death [n=10]), yielding a 10‐year incidence of 26%. Higher baseline NT‐proBNP and relative change from baseline levels were associated with cardiovascular events. Serial NT‐proBNP measurements had incremental prognostic value compared with baseline measurement. The correlates of higher NT‐proBNP levels were atrial fibrillation, hypertension, right ventricular dilation and dysfunction, and atrioventricular valve regurgitation. Renin–angiotensin–aldosterone system antagonist was associated with lower NT‐proBNP levels and lower risk of cardiovascular events. Conclusions NT‐proBNP testing is inexpensive, widely available, and noninvasive, making it well suited for longitudinal surveillance in adults with congenital heart disease. Most of the correlates for elevated NT‐proBNP levels were modifiable, providing target for therapeutic interventions.

Abstract Purpose: In RET-rearranged non–small cell lung cancer (NSCLC), treatment options after first-generation selective RET inhibitors (SRI) are limited, and the value of SRI retreatment remains unclear. This study evaluates outcomes of SRI retreatment. Experimental Design: This multicenter retrospective study included patients with advanced RET-rearranged NSCLC who received ≥2 SRI-based therapy lines. Of 411 SRI-treated patients, 41 (10%) underwent SRI retreatment. Outcomes included objective response rate (ORR), progression-free survival (PFS), 6-month PFS, time to treatment failure, adverse events (AE). Results: Among 41 patients, 14 (34%) discontinued the initial SRI because of toxicity and 27 (66%) because of progression. After discontinuation for toxicity, all switched to an alternate SRI, achieving an ORR of 67%, median PFS of 9.9 months, and 6-month PFS of 80.3%. AEs reoccurred in nine patients (64%), with grade ≥3 AEs in 3 (21%) who switched at full dose. Among patients who discontinued because of progression, SRI monotherapy (n = 13) achieved an ORR of 23%, median PFS of 7 months, and 6-month PFS of 61.5%. Benefit was observed in patients with brain-only or oligoprogressive disease or dose reduction on first SRI. Combination therapy (n = 14; targeted agents, n = 11; chemotherapy, n = 3) achieved an ORR of 39%, median PFS of 4 months, and 6-month PFS of 27.3%. Conclusions: Same-class SRI switch after toxicity is feasible and clinically active but warrants cautious, dose-adjusted switching to mitigate toxicities. The efficacy of SRI rechallenge after progression seems limited overall. However, selected patients, such as those with brain-only or oligoprogressive disease, may derive benefit.

Abstract Specialized immune niches located in the stroma-tumor junction promote interactions between tumor-resident antigen-presenting cells and effector lymphocytes that are required for immunotherapy efficacy. We have hypothesized that this stromal-based immune crosstalk may be regulated by provisional extracellular matrix signals, analogous to the early-stages of wound healing where provisional matrix hosts robust immune activity. A cardinal provisional matrix remodeling event, the regulated proteolysis of the large proteoglycan versican (VCAN) generates an immunomodulatory fragment (matrikine), versikine, that engages adaptive immunity. We have previously shown that VCAN proteolysis correlates with T-cell infiltration across multiple solid and liquid cancers with the most immune-evasive cancers showing low to negligible rates of VCAN proteolysis (e.g., 0% in pancreatic adenocarcinoma). A recent clinical trial generated informative prospective data linking VCAN proteolysis with immunotherapy outcomes: patients whose metastatic colorectal tumors categorized as VCAN proteolysis-weak (VPW, 60% of cases) demonstrated inferior responses to checkpoint inhibition (CPI, pembrolizumab) and worse outcomes compared to VCAN proteolysis-predominant (VPP, 40%) cases. Leveraging these prospective trial data, we hypothesized that versikine therapy may rationally reverse the immunosuppression associated with the prevalent VPW phenotype. To tailor versikine therapy to specific human biomarker-stratified tumor phenotypes, we studied experimental models replicating partially T-permissive VPW-VCANlo human cancers and poor-prognosis, immune-exclusionary VPW-VCANhi cancers. Across VPW-like tumors, versikine promoted the stromal accumulation of CD40lghiIcoshiFoxp3neg Tfh-like CD4+ T cells that engaged dendritic cells (DC) and CD8+ T cells in “immune triads” even at baseline (prior to CPI), concurrently with the intratumoral expansion of activated CD8+ T cells expressing cytolytic molecules (granzymes, perforin). In VPW-VCANlo tumors, versikine delivered as LNP-mRNA or recombinant protein demonstrated monotherapy activity, dependent on Batf3-lineage DC and CD8+ T cells. When combined with single-agent CPI, versikine cured most VPW-VCANlo tumors with memory to tumor re-challenge. In VPW-VCANhi tumors, versikine synergized with combination immunotherapy modulating “mature DCs enriched in immunoregulatory molecules” (mregDC). Our data provide mechanistic rationale for the biomarker-stratified, therapeutic restoration of a provisional matrix-regulated immune signaling network to reverse resistance to CPI. Moreover, the data highlight the need to reevaluate prior approaches to globally disrupt tumor stroma in favor of biomarker-driven, pathway-focused harnessing of stromal signals that modulate anti-tumor immunity. Citation Format: Daniel J. Lagal, Duncan Hong, Athanasios Papadas, Emma Geatches, George Yacu, Surabhi Naik, Alicia Gibbons, Alexander Cicala, Kersi Pestonjamasp, Peter Toth, Kristina Matkowskyj, Dustin A. Deming, Fotis Asimakopoulos. Biomarker-driven restoration of tumor provisional matrix signaling network reverses resistance to checkpoint inhibition immunotherapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 6746.

Abstract Background: Prostate cancer (CaP) is among the most common cancers in U.S. men, with Black men experiencing a 1.7-fold higher incidence and nearly twice the mortality rate compared with non-Hispanic White men. While early detection has improved survival, differences remain in recovery and quality of life (QoL). Sexual dysfunction after treatment affects identity, relationships, and well-being, yet stigma, cultural norms, and limited access to patient-centered care prevent open discussion. Few initiatives address the intersection of identity, stigma, and survivorship in this population. Community town halls provide a trusted, culturally responsive space for shared learning, peer connection, and open dialogue on survivorship and sexual health. Methods: Two community-engaged CaP town halls were convened by the Multidisciplinary Health Outcomes Research and Economics (MORE) Lab at the University of Oklahoma Health Campus (OUHC) to explore survivorship issues among Black CaP survivors. The first, held virtually in June 2024, included watch parties in Georgia and Texas. The second, in November 2024, was hybrid, hosted at OUHC and via Teams with a watch party at the Willa D. Johnson Recreation Center, a community center in Oklahoma City. Panelists included oncologists, psychologists, survivors, partners, and Community Advisory Board members. Both town halls featured survivor storytelling, provider Q&A, and interactive discussion. Post-event surveys captured demographics, satisfaction, and impact. Open-discussion transcripts were analyzed using Braun and Clarke’s reflexive thematic analysis, guided by a constructivist framework. Results: A total of 124 participants attended both town halls. Survey respondents (n=17; 13.7%) provided demographic and satisfaction data; remaining insights came from all participants who engaged in open dialogue. Respondents ranged in age from 19-76 years (mean=48.7 ± 21.8). Of the 17 respondents, 93% rated the events as very good or excellent; 23.5% were CaP survivors, all identifying as Black. Five themes emerged: (1) Masculinity, stigma, and silence limit help-seeking; (2) Sexual dysfunction and identity loss cause major emotional burdens; (3) Navigation challenges highlight systemic barriers; (4) Spousal and psychological support aid coping; and (5) Survivorship empowerment. Participants valued hearing “real stories from real men,” validating their experiences. Conclusions: Community town halls are a feasible and culturally relevant strategy to engage Black CaP survivors in survivorship and sexual health discussions. By centering lived experiences and facilitating two-way communication between survivors, clinicians, and researchers, these events build trust and reduce stigma. Future work will assess changes in knowledge, engagement, and confidence in care navigation and explore scaling this community-academic model to improve survivorship QoL. Citation Format: Parisa Ghasemi, Gaurav Kumar, Ernest T. Kaninjing, Adanma Ayanambakkam, Yaw A. Nyame, Amy Siston, Jordan M. Neil, Adam C. Alexander, Andrew G. McIntosh, Yan Zhao, Zsolt Nagykaldi, Kathleen A. Dwyer, Mary Ellen Young, Sabrina Dickey, Daniel Morton, Opeyemi Bolajoko, Darla E. Kendzor, Motolani E. Ogunsanya. Survivorship and sexual health: What we learned from Black men in community town halls [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 2483.

Abstract Background: Multiple myeloma is the second most common adult blood cancer and remains incurable, despite recent advances. The IMWG/IMS recently published updated risk criteria, Consensus Genomic Staging (CGS), combining genomic events with clinical measures to identify high-risk patients (Avet-Loiseau, et al. JCO 2025). We applied CGS risk criteria to baseline samples from the Multiple Myeloma Research Foundation CoMMpass Study (NCT01454297), a prospective, longitudinal observational study that enrolled more than1100 participants between 2011-2017. Methods: We assessed patient risk following CGS criteria. Criteria 1: del(17p), with a cutoff of &gt;20% clonal fraction, and/or TP53 mutation. Criteria 2: an IgH translocation including t(4;14), t(14;16), or t(14;20) along with 1q+ and/or del(1p32). Criteria 3: monoallelic del(1p32) along with 1q+ or biallelic del(1p32). Criteria 4: beta-2 microglobulin &gt;= 5.5 mg/L with normal creatinine. Patients meeting one or more criteria were classified as High Risk. To test if risk showed association with genetic ancestry, we grouped patients into those with high African genetic ancestry and those with low African genetic ancestry. Genetic ancestry was measured using ADMIXTURE in comparison to 1000 Genomes. Results: We evaluated the CGS risk status of 877 CoMMpass patients, finding 69.2% (573/828) to be Standard Risk (SR) and 30.8% (255/828) to be High Risk (HR). HR patients showed worse progression-free survival (median PFS 26.3 months vs. 42.5 months; log rank test, p &lt; 0.0001) and overall survival (median OS 63.9 months vs. median not reached; p &lt; 0.0001). 10.6% (91/858) of patients met Criteria 1, 10.7% (93/873) met Criteria 2, 5.6% (49/877) met Criteria 3, and 8.7% (73/835) met Criteria 4. We found 14.2% (123/867) of patients to have high African-like genetic ancestry (AFR-high) and 85.8% (744/867) to have low African-like genetic ancestry (AFR-low). We observed significantly fewer HR patients within the AFR-high group (20%, 23/115) than in the AFR-low group (32.4%, 231/712) (Fisher’s exact test, p = 0.0065). The AFR-high group met HR criteria at a lower rate across all criteria, but only with Criteria 3 was the difference significant (p-value = 0.03). Higher levels of creatinine in AFR-high patients may have impacted meeting the normal creatinine requirement of Criteria 4. Conclusions: HR patients comprise 30.8% of CoMMpass subjects and show significantly worse PFS and OS than SR patients. AFR-high patients met HR criteria at a lower rate than AFR-low patients according to the CGS criteria. Clinical trials that utilize CGS criteria to determine patient eligibility may be challenged to enroll a representative patient population, including AFR-high patients. Screening and eligibility criteria should be carefully considered to mitigate potential enrollment disparities. Citation Format: Steven M. Foltz, Chaitanya R. Acharya, Alexander Gout, Yuxin Jin, David E. Avigan, Ravi Vij, Shaji Kunnathu Kumar, Sagar Lonial, Hearn Cho, Katie Wozniak, Jonathan J. Keats, Craig Cole, John D. Carpten, George J. Mulligan. Multiple myeloma risk classification of CoMMpass using Consensus Genomic Staging identifies differences between patients with high and low African genetic ancestry [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2026; Part 1 (Regular Abstracts); 2026 Apr 17-22; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2026;86(7 Suppl):Abstract nr 3887.