Multiomic analysis identifies a high-risk signature that predicts early clinical failure in DLBCL.

Recent genetic and molecular classification of DLBCL has advanced our knowledge of disease biology, yet were not designed to predict early events and guide anticipatory selection of novel therapies. To address this unmet need, we used an integrative multiomic approach to identify a signature at diagnosis that will identify DLBCL at high risk of early clinical failure. Tumor biopsies from 444 newly diagnosed DLBCL were analyzed by WES and RNAseq. A combination of weighted gene correlation network analysis and differential gene expression analysis was used to identify a signature associated with high risk of early clinical failure independent of IPI and COO. Further analysis revealed the signature was associated with metabolic reprogramming and identified cases with a depleted immune microenvironment. Finally, WES data was integrated into the signature and we found that inclusion of ARID1A mutations resulted in identification of 45% of cases with an early clinical failure which was validated in external DLBCL cohorts. This novel and integrative approach is the first to identify a signature at diagnosis, in a real-world cohort of DLBCL, that identifies patients at high risk for early clinical failure and may have significant implications for design of therapeutic options.

Assessing and attenuating the impact of selection bias on spatial cluster detection studies.

Spatial cluster analyses are commonly used in epidemiologic studies of case-control data to detect whether certain areas in a study region have an excess of disease risk. Case-control studies are susceptible to potential biases including selection bias, which can result from non-participation of eligible subjects in the study. However, there has been no systematic evaluation of the effects of non-participation on the findings of spatial cluster analyses. In this paper, we perform a simulation study assessing the effect of non-participation on spatial cluster analysis using the local spatial scan statistic under a variety of scenarios that vary the location and rates of study non-participation and the presence and intensity of a zone of elevated risk for disease for simulated case-control studies. We find that geographic areas of lower participation among controls than cases can greatly inflate false-positive rates for identification of artificial spatial clusters. Additionally, we find that even modest non-participation outside of a true zone of elevated risk can decrease spatial power to identify the true zone. We propose a spatial algorithm to correct for potentially spatially structured non-participation that compares the spatial distributions of the observed sample and underlying population. We demonstrate its ability to markedly decrease false positive rates in the absence of elevated risk and resist decreasing spatial sensitivity to detect true zones of elevated risk. We apply our method to a case-control study of non-Hodgkin lymphoma. Our findings suggest that greater attention should be paid to the potential effects of non-participation in spatial cluster studies.

Synoptic Reporting by Summarizing Cancer Pathology Reports using Large Language Models.

Background: Synoptic reporting, the documenting of clinical information in a structured manner, is known to improve patient care by reducing errors, increasing readability, interoperability, and report completeness. Despite its advantages, manually synthesizing synoptic reports from narrative reports is expensive and error prone when the number of structured fields are many. While the recent revolutionary developments in Large Language Models (LLMs) have significantly advanced natural language processing, their potential for innovations in medicine is yet to be fully evaluated. Objectives: In this study, we explore the strengths and challenges of utilizing the state-of-the-art language models in the automatic synthesis of synoptic reports. Materials and Methods: We use a corpus of 7,774 cancer related, narrative pathology reports, which have annotated reference synoptic reports from Mayo Clinic EHR. Using these annotations as a reference, we reconfigure the state-of-the-art large language models, such as LLAMA-2, to generate the synoptic reports. Our annotated reference synoptic reports contain 22 unique data elements. To evaluate the accuracy of the reports generated by the LLMs, we use several metrics including the BERT F1 Score and verify our results by manual validation. Results: We show that using fine-tuned LLAMA-2 models, we can obtain BERT Score F1 of 0.86 or higher across all data elements and BERT F1 scores of 0.94 or higher on over 50% (11 of 22) of the questions. The BERT F1 scores translate to average accuracies of 76% and as high as 81% for short clinical reports. Conclusions: We demonstrate successful automatic synoptic report generation by fine-tuning large language models.

Marginal zone lymphoma international prognostic index: a unifying prognostic index for marginal zone lymphomas requiring systemic treatment.

Background: Marginal zone lymphomas (MZL), comprised of three unique but related subtypes, lack a unifying prognostic score applicable to all the patients in need for systemic chemotherapy and/or immunotherapy. Methods: Patients from the prospective NF10 study (NCT02904577) with newly diagnosed MZL and receiving frontline systemic therapy at diagnosis or after observation were used to train a prognostic model. The primary endpoint was progression-free survival (PFS) from start of treatment. The model was externally validated in a pooled analysis of two independent cohorts from the University of Iowa and Mayo Clinic Molecular Epidemiology Resource and the University of Miami. Findings: We identified 501 eligible patients. After multivariable modeling, lactate dehydrogenase (LDH) above upper normal limit, hemoglobin <12 g/dL, absolute lymphocyte count <1 × 109/L, platelets <100 × 109/L, and MZL subtype (nodal or disseminated) were independently associated with inferior PFS. The proposed MZL International Prognostic index (MZL-IPI) combined these 5 factors, and we defined low (LRG, 0 factors, 27%), intermediate (IRG, 1-2 factors, 57%) and high (HRG, 3+ factors, 16%) risk groups with 5-y PFS of 85%, 66%, and 37%, respectively (c-Harrell = 0.64). Compared to the LRG, the IRG (Hazard Ratio [HR] = 2.30, 95% CI 1.39-3.80) and HRG (HR = 5.41, 95% CI 3.12-9.38) had inferior PFS. Applying the MZL-IPI to the pooled US cohort (N = 353), 94 (27%), 192 (54%), and 67 (19%) patients were classified as LRG, IRG, and HRG, respectively, and the model was validated for PFS (log-rank test p = 0.0018; c-Harrell = 0.578, 95% CI 0.54-0.62). The MZL-IPI was also prognostic for OS in both the training and the external validation sets. Interpretation: MZL-IPI is a new prognostic score for use in all patients with MZL considered for systemic treatment. Funding: The MER was supported by P50 CA97274 and U01 CA195568.

Defining Primary Refractory Large B-cell Lymphoma.

Patients with large B-cell lymphoma (LBCL) that fail to achieve a complete response (CR) or relapse early after anthracycline-containing immunochemotherapy (IC) have a poor prognosis and are commonly considered "primary refractory disease". However, different definitions of primary refractory disease are used in the literature and clinical practice. In this study, we ex-amined variation in the time to relapse used to define refractory status and association with sur-vival outcomes in patients with primary refractory LBCL in a single-center prospective cohort with a validation in an independent multi-center cohort. Newly diagnosed LBCL patients were enrolled in the Molecular Epidemiological Resource cohort (MER; N=949) or the Lymphoma Epidemiology of Outcomes cohort (LEO; N=2,755) from 9/2002 to 5/2021. Primary refractory LBCL was defined as no response (SD) or progressive disease (PD) during or by the end of frontline (1L) IC (primary PD; PPD), partial response at end of treatment (EOT PR), or relapse within 3-12 months after achieving CR at EOT to 1L IC (early relapse). In the MER cohort, pa-tients with PPD had inferior OS (2-year OS rate 15% MER, 31% LEO) when compared to other subgroups considered in defining primary refractory disease, EOT PR (2-year OS rate 38% MER, 50% LEO) and early relapse (2-year OS rate 44% MER, 58% LEO). Among patients re-ceiving frontline IC with curative intent, we identified that patients with PPD are the key sub-group with poor outcomes. We propose a definition of primary refractory LBCL as SD or PD during or by the end of 1L treatment.

Follicular lymphoma B cells exhibit heterogeneous transcriptional states with associated somatic alterations and tumor microenvironments.

Follicular lymphoma (FL) is an indolent non-Hodgkin lymphoma of germinal center origin, which presents with significant biologic and clinical heterogeneity. Using RNA-seq on B cells sorted from 87 FL biopsies, combined with machine-learning approaches, we identify 3 transcriptional states that divide the biological ontology of FL B cells into inflamed, proliferative, and chromatin-modifying states, with relationship to prior GC B cell phenotypes. When integrated with whole-exome sequencing and immune profiling, we find that each state was associated with a combination of mutations in chromatin modifiers, copy-number alterations to TNFAIP3, and T follicular helper cells (Tfh) cell interactions, or primarily by a microenvironment rich in activated T cells. Altogether, these data define FL B cell transcriptional states across a large cohort of patients, contribute to our understanding of FL heterogeneity at the tumor cell level, and provide a foundation for guiding therapeutic intervention.

Low-frequency and rare genetic variants associated with rheumatoid arthritis risk.

Rheumatoid arthritis (RA) has an estimated heritability of nearly 50%, which is particularly high in seropositive RA. HLA alleles account for a large proportion of this heritability, in addition to many common single-nucleotide polymorphisms with smaller individual effects. Low-frequency and rare variants, such as those captured by next-generation sequencing, can also have a large role in heritability in some individuals. Rare variant discovery has informed the development of drugs such as inhibitors of PCSK9 and Janus kinases. Some 34 low-frequency and rare variants are currently associated with RA risk. One variant (19:10352442G>C in TYK2) was identified in five separate studies, and might therefore represent a promising therapeutic target. Following a set of best practices in future studies, including studying diverse populations, using large sample sizes, validating RA and serostatus, replicating findings, adjusting for other variants and performing functional assessment, could help to ensure the relevance of identified variants. Exciting opportunities are now on the horizon for genetics in RA, including larger datasets and consortia, whole-genome sequencing and direct applications of findings in the management, and especially treatment, of RA.

Modeling variation in mixture effects over space with a Bayesian spatially varying mixture model.

Mixture analysis is an emerging statistical tool in epidemiological research that seeks to estimate the health effects associated with mixtures of several exposures. This approach acknowledges that individuals experience many simultaneous exposures and it can estimate the relative importance of components in the mixture. Health effects due to mixtures may vary over space driven by to political, demographic, environmental, or other differences. In such cases, estimating a global mixture effect without accounting for spatial variation would induce bias in effect estimates and potentially lower statistical power. To date, no methods have been developed to estimate spatially varying chemical mixture effects. We developed a Bayesian spatially varying mixture model that estimates spatially varying mixture effects and the importance weights of components in the mixture, while adjusting for covariates. We demonstrate the efficacy of the model through a simulation study that varies the number of mixtures (one and two) and spatial pattern (global, one-dimensional, radial) and magnitude of mixture effects, showing that the model is able to accurately reproduce the spatial pattern of mixture effects across a diverse set of scenarios. Finally, we apply our model to a multi-center case-control study of non-Hodgkin lymphoma (NHL) in Detroit, Iowa, Los Angeles, and Seattle. We identify significant spatially varying positive and inverse associations with NHL for two mixtures of pesticides in Iowa and do not find strong spatial effects at the other three centers. In conclusion, the Bayesian spatially varying mixture model represents a novel method for modeling spatial variation in mixture effects.

Tumor mutational load is prognostic for progression to therapy among high-count monoclonal B-cell lymphocytosis.

ABSTRACT: High-count monoclonal B-cell lymphocytosis (HCMBL) is a precursor condition to chronic lymphocytic leukemia (CLL). We have shown that among individuals with HCMBL, the CLL-International Prognostic Index (CLL-IPI) is prognostic for time-to-first therapy (TTFT). Little is known about the prognostic impact of somatically mutated genes among individuals with HCMBL. We sequenced DNA from 371 individuals with HCMBL using a targeted sequencing panel of 59 recurrently mutated genes in CLL to identify high-impact mutations. We compared the sequencing results with that of our treatment-naïve CLL cohort (N = 855) and used Cox regression to estimate hazard ratios and 95% confidence intervals (CIs) for associations with TTFT. The frequencies of any mutated genes were lower in HCMBL (52%) than CLL (70%). At 10 years, 37% of individuals with HCMBL with any mutated gene had progressed requiring treatment compared with 10% among individuals with HCMBL with no mutations; this led to 5.4-fold shorter TTFT (95% CI, 2.6-11.0) among HCMBL with any mutated gene vs none, independent of CLL-IPI. When considering individuals with low risk of progression according to CLL-IPI, those with HCMBL with any mutations had 4.3-fold shorter TTFT (95% CI, 1.6-11.8) vs those with none. Finally, when considering both CLL-IPI and any mutated gene status, we observed individuals with HCMBL who were high risk for both prognostic factors had worse prognosis than patients with low-risk CLL (ie, 5-year progression rate of 32% vs 21%, respectively). Among HCMBL, the frequency of somatically mutated genes at diagnosis is lower than that of CLL. Accounting for both the number of mutated genes and CLL-IPI can identify individuals with HCMBL with more aggressive clinical course.

The Lymphoma Epidemiology of Outcomes cohort study: Design, baseline characteristics, and early outcomes.

To address the current and long-term unmet health needs of the growing population of non-Hodgkin lymphoma (NHL) patients, we established the Lymphoma Epidemiology of Outcomes (LEO) cohort study (NCT02736357; https://leocohort.org/). A total of 7735 newly diagnosed patients aged 18 years and older with NHL were prospectively enrolled from 7/1/2015 to 5/31/2020 at 8 academic centers in the United States. The median age at diagnosis was 62 years (range, 18-99). Participants came from 49 US states and included 538 Black/African-Americans (AA), 822 Hispanics (regardless of race), 3386 women, 716 age <40 years, and 1513 rural residents. At study baseline, we abstracted clinical, pathology, and treatment data; banked serum/plasma (N = 5883, 76.0%) and germline DNA (N = 5465, 70.7%); constructed tissue microarrays for four major NHL subtypes (N = 1189); and collected quality of life (N = 5281, 68.3%) and epidemiologic risk factor (N = 4489, 58.0%) data. Through August 2022, there were 1492 deaths. Compared to population-based SEER data (2015-2019), LEO participants had a similar distribution of gender, AA race, Hispanic ethnicity, and NHL subtype, while LEO was underrepresented for patients who were Asian and aged 80 years and above. Observed overall survival rates for LEO at 1 and 2 years were similar to population-based SEER rates for indolent B-cell (follicular and marginal zone) and T-cell lymphomas, but were 10%-15% higher than SEER rates for aggressive B-cell subtypes (diffuse large B-cell and mantle cell). The LEO cohort is a robust and comprehensive national resource to address the role of clinical, tumor, host genetic, epidemiologic, and other biologic factors in NHL prognosis and survivorship.

Prediction of outcomes for high-count monoclonal B lymphocytosis using an epigenetic and immunogenetic signature.

ABSTRACT: Monoclonal B-cell lymphocytosis (MBL) progresses to chronic lymphocytic leukemia (CLL) requiring therapy at 1% to 5% per year. Improved prediction of progression would greatly benefit individuals with MBL. Patients with CLL separate into 3 distinct epigenetic subtypes (epitypes) with high prognostic significance, and recently the intermediate epitype has been shown to be enriched for high-risk immunoglobulin lambda variable (IGLV) 3-21 rearrangements, impacting outcomes for these patients. Here, we employed this combined strategy to generate the epigenetic and light chain immunoglobulin (ELCLV3-21) signature to classify 219 individuals with MBL. The ELCLV3-21 high-risk signature distinguished MBL individuals with a high probability of progression (39.9% and 71.1% at 5 and 10 years, respectively). ELCLV3-21 improved the accuracy of predicting time to therapy for individuals with MBL compared with other established prognostic indicators, including the CLL international prognostic index (c-statistic, 0.767 vs 0.668, respectively). Comparing ELCLV3-21 risk groups in MBL vs a cohort of 226 patients with CLL revealed ELCLV3-21 high-risk individuals with MBL had significantly shorter time to therapy (P = .003) and reduced overall survival (P = .03) compared with ELCLV3-21 low-risk individuals with CLL. These results highlight the power of the ELCLV3-21 approach to identify individuals with a higher likelihood of adverse clinical outcome and may provide a more accurate approach to classify individuals with small B-cell clones.

Matched control analysis suggests that R-CHOP followed by (R)-ICE may improve outcome in non-GCB DLBCL compared with R-CHOP.

ABSTRACT: Rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP) is considered the standard-of-care for patients with advanced-stage diffuse large B-cell lymphoma (DLBCL), despite findings that patients with nongerminal center B-cell like (non-GCB) have significantly worse outcome with this regimen. We evaluated the prognostic significance of baseline risk factors, including cell of origin (COO) classified by the Hans algorithm, within an alternative chemoimmunotherapy program. At Memorial Sloan Kettering Cancer Center (MSK), 151 patients with DLBCL received sequential R-CHOP induction and (R)-ICE (rituximab, ifosfamide, carboplatin, and etoposide) consolidation. Outcome analysis based on COO was validated with a propensity score-matched cohort treated with R-CHOP from the Mayo Clinic component of the Molecular Epidemiology Resource (MER). Among the patients with GCB (n = 69) and non-GCB (n = 69) at MSK, event-free survival (EFS) of non-GCB was superior to that of GCB (hazard ratio [HR], 0.53; 95% confidence interval [CI], 0.29-0.98). Overall survival (OS) demonstrated an association in the same direction but was not statistically significant (HR, 0.68; 95% CI, 0.33-1.42). Propensity score-matched patients from MSK (n = 108) demonstrated a small attenuation in the HRs for EFS (HR, 0.57; 95% CI, 0.27-1.18) and OS (HR, 0.76; 95% CI, 0.33-1.79) and were no longer statistically significant. In contrast, the matched MER cohort (n = 108) demonstrated an EFS association (HR, 1.17; 95% CI, 0.70-1.95) and OS association (HR, 1.13; 95% CI, 0.64-2.00) in the opposite direction, but were also not statistically significant. R-CHOP induction and (R)-ICE consolidation may overcome the negative prognostic impact of the non-GCB phenotype, per the Hans algorithm, and can be preferentially selected for this population. This trial was registered at www.ClinicalTrials.gov as #NCT00039195 and #NCT00712582.

Early complete response as a validated surrogate marker in extranodal marginal zone lymphoma systemic therapy.

ABSTRACT: Extranodal marginal zone lymphoma (EMZL) has a very indolent course, and the validation of surrogate markers could accelerate novel therapies. Although prognostic markers do exist, no surrogate markers have been validated in EMZL. We hypothesized that time to complete response within 24 months (TTCR24) and complete response (CR) at 24 months (CR24) could be valid surrogate markers of progression-free survival (PFS). The International Extranodal Lymphoma Study Group 19 phase 3 trial showed the advantage of double therapy (rituximab + chlorambucil) over single therapy (rituximab or chlorambucil) on PFS. We used 2 recently published single-trial approaches to assess whether TTCR24 and CR24 were good surrogate markers of 8-year PFS (8y-PFS). Among the 401 patients, 264 (66%) reached a CR in the first 24 months, of which 222 (84%) remained in CR at month 24. The cumulative incidence of CR over time was significantly higher in patients under double therapy (hazard ratio, 1.75; P < .001). The double therapy arm was associated with a higher CR24 rate, a shorter TTCR24, and a longer 8y-PFS. The estimated proportion of treatment effect on 8y-PFS explained by TTCR24 was 95% (95% confidence interval [CI], 0.27-1.87). CR24 was also a strong surrogate marker because it mediated 90% (95% CI, 0.51-2.22) of the treatment effect on PFS and its natural indirect effect was significant throughout the follow-up. We found that TTCR24 predicted 95% and that CR24 mediated 90% of the treatment effect on long-term PFS. Therefore, TTCR24 and CR24 could be used in clinical trials as informative and valid early indicators of treatment effect on PFS. This trial was registered at www.clinicaltrials.gov as #NCT00210353.

Matching-adjusted indirect comparison from the Lymphoma Epidemiology of Outcomes Consortium for Real World Evidence (LEO CReWE) study to a clinical trial of mosunetuzumab in relapsed or refractory follicular lymphoma.

Mosunetuzumab is a novel bispecific antibody targeting epitopes on CD3 on T cells and CD20 on B cells with the goal of inducing T-cell mediated elimination of malignant B cells. A recent pivotal phase I/II clinical trial (GO29781) demonstrated that mosunetuzumab induced an overall response rate of 80%, complete response rate of 60%, and a median progression-free survival of 17.9 months in patients with relapsed/refractory (r/r) follicular lymphoma (FL) following at least two prior lines of systemic therapy, including alkylator and anti-CD20 antibody-based therapy. Historical data from cohorts receiving therapy for r/r FL can provide some context for interpretation of single-arm trials. We compared the results from the mosunetuzumab trial to outcomes from a cohort of patients with r/r FL from the LEO Consortium for Real World Evidence (LEO CReWE). We applied clinical trial eligibility criteria to the LEO CReWE cohort and utilized matching-adjusted indirect comparison weighting to balance the clinical characteristics of the LEO CReWE cohort with those from the mosunetuzumab trial. Overall response rates (73%, 95% CI:65-80%) and complete response rates (53%, 95% CI:45-61%) observed in the weighted LEO CReWE cohort were lower than those reported on the mosunetuzumab trial (ORR=80%, 95% CI:70-88%; CR=60%, 95% CI:49-70% respectively). Progression-free survival at 12 months was similar in the weighted LEO CReWE (60%, 95% CI:51-69%) and the mosunetuzumab trial (PFS 58%, 95% CI:47-68%). Sensitivity analyses examining the impact of matching variables, selection of line of therapy, and application of eligibility criteria, provide context for best practices in this setting.

Evolving treatment patterns and improved outcomes in relapsed/refractory mantle cell lymphoma: a prospective cohort study.

Over the last two decades, the frontline therapy for mantle cell lymphoma (MCL) has evolved. However, the impact of subsequent lines of therapy on survival outcomes has not been well characterized. In this study, we investigated the treatment patterns and survival outcomes in patients with relapsed/refractory (R/R) MCL treated with second-line (2 L) therapy. Adult patients with newly diagnosed MCL from 2002 to 2015 were enrolled in a prospective cohort study. Clinical characteristics, 2 L treatment details, and outcomes were compared between patients who received 2 L treatment between 2003-2009 (Era 1), 2010-2014 (Era 2), and 2015-2021 (Era 3). 2 L treatment was heterogenous in all eras, and there was a substantial shift in the pattern of 2 L therapy over time. The estimated 2-year EFS rate was 21% (95% CI, 13-35), 40% (95% CI, 30-53), and 51% (95% CI, 37-68) in Era 1-3 respectively, and the 5-year OS rate was 31% (95% CI, 21-45), 37% (95% CI, 27-50), and 67% (95% CI, 54-83) in Era 1-3, respectively. These results provide real-world evidence on evolving treatment patterns of 2 L therapy based on the era of relapse. The changes in 2 L treatment correlated with improved EFS and OS, suggesting that treatment advances are associated with improved outcomes in patients with R/R MCL.

Gene-respiratory disease interactions for rheumatoid arthritis risk.

OBJECTIVE: We aimed to identify gene by respiratory tract disease interactions that increase RA risk. METHODS: In this case-control study using the Mass General Brigham Biobank, we matched incident RA cases, confirmed by ACR/EULAR criteria, to four controls on age, sex, and electronic health record history. Genetic exposures included a validated overall genetic risk score (GRS) for RA, a Human Leukocyte Antigen (HLA) GRS for RA, and the MUC5B promoter variant, an established risk factor for RA-associated interstitial lung disease (ILD). Preceding respiratory tract diseases came from diagnosis codes (positive predictive value 86%). We estimated attributable proportions (AP) and multiplicative odds ratios (OR) with 95% confidence intervals (CI) for RA for each genetic and respiratory exposure using conditional logistic regression models, adjusting for potential confounders. RESULTS: We identified 653 incident RA cases and 2,607 matched controls (mean 54 years, 76% female). The highest tertile of the overall GRS and the HLA GRS were both associated with increased RA risk (OR 2.28, 95% CI 1.89,2.74; OR 2.02, 95% CI 1.67-2.45). ILD and the HLA GRS exhibited a synergistic relationship for RA risk (OR for both exposures 4.30, 95% CI 1.28,14.38; AP 0.51, 95% CI-0.16,1.18). Asthma and the MUC5B promoter variant also exhibited a synergistic interaction for seropositive RA (OR for both exposures 2.58, 95% CI 1.10,6.07; AP 0.62, 95% CI 0.24,1.00). CONCLUSION: ILD-HLA GRS and asthma-MUC5B promoter variant showed synergistic interactions for RA risk. Such interactions may prove useful for RA prevention and screening.

Artificial Intelligence-based Efficacy Prediction of Phase 3 Clinical Trial for Repurposing Heart Failure Therapies.

Introduction: Drug repurposing involves finding new therapeutic uses for already approved drugs, which can save costs as their pharmacokinetics and pharmacodynamics are already known. Predicting efficacy based on clinical endpoints is valuable for designing phase 3 trials and making Go/No-Go decisions, given the potential for confounding effects in phase 2. Objectives: This study aims to predict the efficacy of the repurposed Heart Failure (HF) drugs for the Phase 3 Clinical Trial. Methods: Our study presents a comprehensive framework for predicting drug efficacy in phase 3 trials, which combines drug-target prediction using biomedical knowledgebases with statistical analysis of real-world data. We developed a novel drug-target prediction model that uses low-dimensional representations of drug chemical structures and gene sequences, and biomedical knowledgebase. Furthermore, we conducted statistical analyses of electronic health records to assess the effectiveness of repurposed drugs in relation to clinical measurements (e.g., NT-proBNP). Results: We identified 24 repurposed drugs (9 with a positive effect and 15 with a non-positive) for heart failure from 266 phase 3 clinical trials. We used 25 genes related to heart failure for drug-target prediction, as well as electronic health records (EHR) from the Mayo Clinic for screening, which contained over 58,000 heart failure patients treated with various drugs and categorized by heart failure subtypes. Our proposed drug-target predictive model performed exceptionally well in all seven tests in the BETA benchmark compared to the six cutting-edge baseline methods (i.e., best performed in 266 out of 404 tasks). For the overall prediction of the 24 drugs, our model achieved an AUCROC of 82.59% and PRAUC (average precision) of 73.39%. Conclusion: The study demonstrated exceptional results in predicting the efficacy of repurposed drugs for phase 3 clinical trials, highlighting the potential of this method to facilitate computational drug repurposing.

Leveraging multi-source to resolve inconsistency across pharmacogenomic datasets in drug sensitivity prediction.

Pharmacogenomics datasets have been generated for various purposes, such as investigating different biomarkers. However, when studying the same cell line with the same drugs, differences in drug responses exist between studies. These variations arise from factors such as inter-tumoral heterogeneity, experimental standardization, and the complexity of cell subtypes. Consequently, drug response prediction suffers from limited generalizability. To address these challenges, we propose a computational model based on Federated Learning (FL) for drug response prediction. By leveraging three pharmacogenomics datasets (CCLE, GDSC2, and gCSI), we evaluate the performance of our model across diverse cell line-based databases. Our results demonstrate superior predictive performance compared to baseline methods and traditional FL approaches through various experimental tests. This study underscores the potential of employing FL to leverage multiple data sources, enabling the development of generalized models that account for inconsistencies among pharmacogenomics datasets. By addressing the limitations of low generalizability, our approach contributes to advancing drug response prediction in precision oncology.

Multiomic Analysis Identifies a High-Risk Metabolic and TME Depleted Signature that Predicts Early Clinical Failure in DLBCL.

PURPOSE: 60-70% of newly diagnosed diffuse large B-cell lymphoma (DLBCL) patients avoid events within 24 months of diagnosis (EFS24) and the remainder have poor outcomes. Recent genetic and molecular classification of DLBCL has advanced our knowledge of disease biology, yet were not designed to predict early events and guide anticipatory selection of novel therapies. To address this unmet need, we used an integrative multiomic approach to identify a signature at diagnosis that will identify DLBCL at high risk of early clinical failure. PATIENTS AND METHODS: Tumor biopsies from 444 newly diagnosed DLBCL were analyzed by WES and RNAseq. A combination of weighted gene correlation network analysis and differential gene expression analysis followed by integration with clinical and genomic data was used to identify a multiomic signature associated with high risk of early clinical failure. RESULTS: Current DLBCL classifiers are unable to discriminate cases who fail EFS24. We identified a high risk RNA signature that had a hazard ratio (HR, 18.46 [95% CI 6.51-52.31] P < .001) in a univariate model, which did not attenuate after adjustment for age, IPI and COO (HR, 20.8 [95% CI, 7.14-61.09] P < .001). Further analysis revealed the signature was associated with metabolic reprogramming and a depleted immune microenvironment. Finally, WES data was integrated into the signature and we found that inclusion of ARID1A mutations resulted in identification of 45% of cases with an early clinical failure which was validated in external DLBCL cohorts. CONCLUSION: This novel and integrative approach is the first to identify a signature at diagnosis that will identify DLBCL at high risk for early clinical failure and may have significant implications for design of therapeutic options.

Modeling historic neighborhood deprivation and non-Hodgkin lymphoma risk.

Many studies have identified associations between neighborhood deprivation and disease, emphasizing the importance of social determinants of health. However, when studying diseases with long latency periods such as cancers, considering the timing of exposures for deprivation becomes more important. In this study, we estimated the associations between neighborhood deprivation indices at several time points and risk of non-Hodgkin lymphoma (NHL) in a population-based case-control study at four study centers - Detroit, Iowa, Los Angeles County, and Seattle (1998-2000). We used the Bayesian index regression model and residential histories to estimate neighborhood deprivation index effects in crude models and adjusted for four chemical mixtures measured in house dust and individual-level covariates. We found that neighborhood deprivation in 1980, approximately twenty years before study entry, provided better model fit than did neighborhood deprivation at 1990 and 2000. We identified several statistically significant associations between neighborhood deprivation in 1980 and NHL risk in Iowa and among long-term (20+ years) residents of Detroit. The most important variables in these indices were median gross rent as a percentage of household income in Iowa and percent of single-parent households with at least one child and median household income in Detroit. Associations remained statistically significant after adjustment for individual-level covariates and chemical mixtures, providing evidence for historic neighborhood deprivation as a risk factor for NHL and motivating future research to uncover the specific carcinogens driving these associations in deprived areas.