Predicting overall survival from tumor dynamics metrics using parametric statistical and machine learning models: application to patients with RET-altered solid tumors.

In oncology drug development, tumor dynamics modeling is widely applied to predict patients' overall survival (OS) via parametric models. However, the current modeling paradigm, which assumes a disease-specific link between tumor dynamics and survival, has its limitations. This is particularly evident in drug development scenarios where the clinical trial under consideration contains patients with tumor types for which there is little to no prior institutional data. In this work, we propose the use of a pan-indication solid tumor machine learning (ML) approach whereby all three tumor metrics (tumor shrinkage rate, tumor regrowth rate and time to tumor growth) are simultaneously used to predict patients' OS in a tumor type independent manner. We demonstrate the utility of this approach in a clinical trial of cancer patients treated with the tyrosine kinase inhibitor, pralsetinib. We compared the parametric and ML models and the results showed that the proposed ML approach is able to adequately predict patient OS across RET-altered solid tumors, including non-small cell lung cancer, medullary thyroid cancer as well as other solid tumors. While the findings of this study are promising, further research is needed for evaluating the generalizability of the ML model to other solid tumor types.

Population pharmacokinetics and CD20 binding dynamics for mosunetuzumab in relapsed/refractory B-cell non-Hodgkin lymphoma.

Mosunetuzumab (Mosun) is a CD20xCD3 T-cell engaging bispecific antibody that redirects T cells to eliminate malignant B cells. The approved step-up dose regimen of 1/2/60/30 mg IV is designed to mitigate cytokine release syndrome (CRS) and maximize efficacy in early cycles. A population pharmacokinetic (popPK) model was developed from 439 patients with relapsed/refractory B-Cell Non-Hodgkin lymphoma receiving Mosun IV monotherapy, including fixed dosing (0.05-2.8 mg IV every 3 weeks (q3w)) and Cycle 1 step-up dosing groups (0.4/1/2.8-1/2/60/30 mg IV q3w). Prior to Mosun treatment, ~50% of patients had residual levels of anti-CD20 drugs (e.g., rituximab or obinutuzumab) from prior treatment. CD20 receptor binding dynamics and rituximab/obinutuzumab PK were incorporated into the model to calculate the Mosun CD20 receptor occupancy percentage (RO%) over time. A two-compartment model with time-dependent clearance (CL) best described the data. The typical patient had an initial CL of 1.08 L/day, transitioning to a steady-state CL of 0.584 L/day. Statistically relevant covariates on PK parameters included body weight, albumin, sex, tumor burden, and baseline anti-CD20 drug concentration; no covariate was found to have a clinically relevant impact on exposure at the approved dose. Mosun CD20 RO% was highly variable, attributed to the large variability in residual baseline anti-CD20 drug concentration (median = 10 µg/mL). The 60 mg loading doses increased Mosun CD20 RO% in Cycle 1, providing efficacious exposures in the presence of the competing anti-CD20 drugs. PopPK model simulations, investigating Mosun dose delays, informed treatment resumption protocols to ensure CRS mitigation.

Deep-NCA: A deep learning methodology for performing noncompartmental analysis of pharmacokinetic data.

Noncompartmental analysis (NCA) is a model-independent approach for assessing pharmacokinetics (PKs). Although the existing NCA algorithms are very well-established and widely utilized, they suffer from low accuracies in the setting of sparse PK samples. In response, we developed Deep-NCA, a deep learning (DL) model to improve the prediction of key noncompartmental PK parameters. Our methodology utilizes synthetic PK data for model training and uses an innovative patient-specific normalization method for data preprocessing. Deep-NCA demonstrated adequate performance across six previously unseen simulated drugs under multiple dosing, showcasing effective generalization. Compared to traditional NCA, Deep-NCA exhibited superior performance for sparse PK data. This study advances the application of DL to PK studies and introduces an effective method for handling sparse PK data. With further validation and refinement, Deep-NCA could significantly enhance the efficiency of drug development by providing more accurate NCA estimates while requiring fewer PK samples.

SARS-CoV-2 Antiviral Prescribing Gaps Among Nonhospitalized High-Risk Adults.

Within a multistate clinical cohort, SARS-CoV-2 antiviral prescribing patterns were evaluated from April 2022-June 2023 among nonhospitalized patients with SARS-CoV-2 with risk factors for severe COVID-19. Among 3247 adults, only 31.9% were prescribed an antiviral agent (87.6% nirmatrelvir/ritonavir, 11.9% molnupiravir, 0.5% remdesivir), highlighting the need to identify and address treatment barriers.

Artificial Intelligence for Quantitative Modeling in Drug Discovery and Development: An Innovation and Quality Consortium Perspective on Use Cases and Best Practices.

Recent breakthroughs in artificial intelligence (AI) and machine learning (ML) have ushered in a new era of possibilities across various scientific domains. One area where these advancements hold significant promise is model-informed drug discovery and development (MID3). To foster a wider adoption and acceptance of these advanced algorithms, the Innovation and Quality (IQ) Consortium initiated the AI/ML working group in 2021 with the aim of promoting their acceptance among the broader scientific community as well as by regulatory agencies. By drawing insights from workshops organized by the working group and attended by key stakeholders across the biopharma industry, academia, and regulatory agencies, this white paper provides a perspective from the IQ Consortium. The range of applications covered in this white paper encompass the following thematic topics: (i) AI/ML-enabled Analytics for Pharmacometrics and Quantitative Systems Pharmacology (QSP) Workflows; (ii) Explainable Artificial Intelligence and its Applications in Disease Progression Modeling; (iii) Natural Language Processing (NLP) in Quantitative Pharmacology Modeling; and (iv) AI/ML Utilization in Drug Discovery. Additionally, the paper offers a set of best practices to ensure an effective and responsible use of AI, including considering the context of use, explainability and generalizability of models, and having human-in-the-loop. We believe that embracing the transformative power of AI in quantitative modeling while adopting a set of good practices can unlock new opportunities for innovation, increase efficiency, and ultimately bring benefits to patients.

Machine Learning-Based Quantification of Patient Factors Impacting Remission in Patients With Ulcerative Colitis: Insights from Etrolizumab Phase III Clinical Trials.

Etrolizumab, an investigational anti-ß7 integrin monoclonal antibody, has undergone evaluation for safety and efficacy in phase III clinical trials on patients with moderate to severe ulcerative colitis (UC). Etrolizumab was terminated because mixed efficacy results were shown in the induction and maintenance phase in patients with UC. In this post hoc analysis, we characterized the impact of explanatory variables on the probability of remission using XGBoost machine learning (ML) models alongside with the SHapley Additive exPlanations framework for explainability. We used patient-level data encompassing demographics, physiology, disease history, clinical questionnaires, histology, serum biomarkers, and etrolizumab drug exposure to develop ML models aimed at predicting remission. Baseline covariates and early etrolizumab exposure at week 4 in the induction phase were utilized to develop an induction ML model, whereas covariates from the end of the induction phase and early etrolizumab exposure at week 4 in the maintenance phase were used to develop a maintenance ML model. Both the induction and maintenance ML models exhibited good predictive performance, achieving an area under the receiver operating characteristic curve (AUROC) of 0.74 ± 0.03 and 0.75 ± 0.06 (mean ± SD), respectively. Compared with placebo, the highest tertile of etrolizumab exposure contributed to 15.0% (95% confidence interval (CI): 9.7-19.9) and 17.0% (95% CI: 8.1-26.4) increases in remission probability in the induction and maintenance phases, respectively. Additionally, the key covariates that predicted remission were CRP, MAdCAM-1, and stool frequency for the induction phase and white blood cells, fecal calprotectin and age for the maintenance phase. These findings hold significant implications for establishing stratification factors in the design of future clinical trials.

Current Status and Future Directions: The Application of Artificial Intelligence/Machine Learning for Precision Medicine.

Technological innovations, such as artificial intelligence (AI) and machine learning (ML), have the potential to expedite the goal of precision medicine, especially when combined with increased capacity for voluminous data from multiple sources and expanded therapeutic modalities; however, they also present several challenges. In this communication, we first discuss the goals of precision medicine, and contextualize the use of AI in precision medicine by showcasing innovative applications (e.g., prediction of tumor growth and overall survival, biomarker identification using biomedical images, and identification of patient population for clinical practice) which were presented during the February 2023 virtual public workshop entitled "Application of Artificial Intelligence and Machine Learning for Precision Medicine," hosted by the US Food and Drug Administration (FDA) and University of Maryland Center of Excellence in Regulatory Science and Innovation (M-CERSI). Next, we put forward challenges brought about by the multidisciplinary nature of AI, particularly highlighting the need for AI to be trustworthy. To address such challenges, we subsequently note practical approaches, viz., differential privacy, synthetic data generation, and federated learning. The proposed strategies - some of which are highlighted presentations from the workshop - are for the protection of personal information and intellectual property. In addition, methods such as the risk-based management approach and the need for an agile regulatory ecosystem are discussed. Finally, we lay out a call for action that includes sharing of data and algorithms, development of regulatory guidance documents, and pooling of expertise from a broad-spectrum of stakeholders to enhance the application of AI in precision medicine.

Explainable deep learning for tumor dynamic modeling and overall survival prediction using Neural-ODE.

While tumor dynamic modeling has been widely applied to support the development of oncology drugs, there remains a need to increase predictivity, enable personalized therapy, and improve decision-making. We propose the use of Tumor Dynamic Neural-ODE (TDNODE) as a pharmacology-informed neural network to enable model discovery from longitudinal tumor size data. We show that TDNODE overcomes a key limitation of existing models in its ability to make unbiased predictions from truncated data. The encoder-decoder architecture is designed to express an underlying dynamical law that possesses the fundamental property of generalized homogeneity with respect to time. Thus, the modeling formalism enables the encoder output to be interpreted as kinetic rate metrics, with inverse time as the physical unit. We show that the generated metrics can be used to predict patients' overall survival (OS) with high accuracy. The proposed modeling formalism provides a principled way to integrate multimodal dynamical datasets in oncology disease modeling.

REFINED-CNN framework for survival prediction with high-dimensional features.

Robust and accurate survival prediction of clinical trials using high-throughput genomics data is a fundamental challenge in pharmacogenomics. Current machine learning tools often provide limited predictive performance and model interpretation in these settings. In the present study, we extend the application of REFINED-CNN from regression tasks to making survival predictions, by mapping high-dimensional RNA sequencing data into REFINED images which are conducive to CNN modeling. We show that the REFINED-CNN survival model can be easily adapted to new tasks of a similar nature (e.g., predicting on new cancer types) using transfer learning with a low number of patients. Furthermore, the model can also be interpreted both locally and globally through risk score back propagation that quantifies each feature (e.g., gene) importance in survival prediction task for the patient or cancer type of interest.

Machine Learning for Exposure-Response Analysis: Methodological Considerations and Confirmation of Their Importance via Computational Experimentations.

Exposure-response (E-R) is a key aspect of pharmacometrics analysis that supports drug dose selection. Currently, there is a lack of understanding of the technical considerations necessary for drawing unbiased estimates from data. Due to recent advances in machine learning (ML) explainability methods, ML has garnered significant interest for causal inference. To this end, we used simulated datasets with known E-R "ground truth" to generate a set of good practices for the development of ML models required to avoid introducing biases when performing causal inference. These practices include the use of causal diagrams to enable the careful consideration of model variables by which to obtain desired E-R relationship insights, keeping a strict separation of data for model-training and for inference generation to avoid biases, hyperparameter tuning to improve the reliability of models, and estimating proper confidence intervals around inferences using a bootstrap sampling with replacement strategy. We computationally confirm the benefits of the proposed ML workflow by using a simulated dataset with nonlinear and non-monotonic exposure-response relationships.

Risk Factors of Hyperglycemia After Treatment With the AKT Inhibitor Ipatasertib in the Prostate Cancer Setting: A Machine Learning-Based Investigation.

PURPOSE: Hyperglycemia is a major adverse event of phosphatidylinositol 3-kinase/AKT inhibitor class of cancer therapeutics. Machine learning (ML) methodologies can identify and highlight how explanatory variables affect hyperglycemia risk. METHODS: Using data from clinical trials of the AKT inhibitor ipatasertib (IPAT) in the metastatic castrate-resistant prostate cancer setting, we trained an XGBoost ML model to predict the incidence of grade ≥2 hyperglycemia (HGLY ≥ 2). Of the 1,364 patients included in our analysis, 19.4% (n = 265) of patients had HGLY ≥2 events with a median time of first onset of 28 days (range, 0-753 days), and 30.0% (n = 221) of patients on an IPAT regimen had at least one HGLY ≥2 event compared with 7.0% (n = 44) of patients on placebo. RESULTS: An 11-variable XGBoost model predicted HGLY ≥2 events well with an AUROC of 0.83 ± 0.02 (mean ± standard deviation). Using SHapley Additive exPlanations analysis, we found IPAT exposure and baseline HbA1c levels to be the strongest predictors of HGLY ≥2, with additional predictivity of baseline measurements of fasting glucose, magnesium, and high-density lipoproteins. CONCLUSION: The findings support using patients' prediabetic status as a key factor for hyperglycemia monitoring and/or trial exclusion criteria. Additionally, the model and relationships between explanatory variables and HGLY ≥2 described herein can help identify patients at high risk for hyperglycemia and develop rational risk mitigation strategies.

Evidence for SARS-CoV-2 Delta and Omicron co-infections and recombination.

BACKGROUND: Between November 2021 and February 2022, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) Delta and Omicron variants co-circulated in the United States, allowing for co-infections and possible recombination events. METHODS: We sequenced 29,719 positive samples during this period and analyzed the presence and fraction of reads supporting mutations specific to either the Delta or Omicron variant. FINDINGS: We identified 18 co-infections, one of which displayed evidence of a low Delta-Omicron recombinant viral population. We also identified two independent cases of infection by a Delta-Omicron recombinant virus, where 100% of the viral RNA came from one clonal recombinant. In the three cases, the 5' end of the viral genome was from the Delta genome and the 3' end from Omicron, including the majority of the spike protein gene, though the breakpoints were different. CONCLUSIONS: Delta-Omicron recombinant viruses were rare, and there is currently no evidence that Delta-Omicron recombinant viruses are more transmissible between hosts compared with the circulating Omicron lineages. FUNDING: This research was supported by the NIH RADx initiative and by the Centers for Disease Control Contract 75D30121C12730 (Helix).

Applying interpretable machine learning workflow to evaluate exposure-response relationships for large-molecule oncology drugs.

The application of logistic regression (LR) and Cox Proportional Hazard (CoxPH) models are well-established for evaluating exposure-response (E-R) relationship in large molecule oncology drugs. However, applying machine learning (ML) models on evaluating E-R relationships has not been widely explored. We developed a workflow to train regularized LR/CoxPH and tree-based XGboost (XGB) models, and derive the odds ratios for best overall response and hazard ratios for overall survival, across exposure quantiles to evaluate the E-R relationship using clinical trial datasets. The E-R conclusions between LR/CoxPH and XGB models are overall consistent, and largely aligned with historical pharmacometric analyses findings. Overall, applying this interpretable ML workflow provides a promising alternative method to assess E-R relationships for impacting key dosing decisions in drug development.

Incomplete Penetrance of Population-Based Genetic Screening Results in Electronic Health Record.

The clinical value of population-based genetic screening projects depends on the actions taken on the findings. The Healthy Nevada Project (HNP) is an all-comer genetic screening and research project based in northern Nevada. HNP participants with CDC Tier 1 findings of hereditary breast and ovarian cancer syndrome (HBOC), Lynch syndrome (LS), or familial hypercholesterolemia (FH) are notified and provided with genetic counseling. However, the HNP subsequently takes a "hands-off" approach: it is the responsibility of notified participants to share their findings with their healthcare providers, and providers are expected to implement the recommended action plans. Thus, the HNP presents an opportunity to evaluate the efficiency of participant and provider responses to notification of important genetic findings, using electronic health records (EHRs) at Renown Health (a large regional hospital in northern Nevada). Out of 520 HNP participants with findings, we identified 250 participants who were notified of their findings and who had an EHR. 107 of these participants responded to a survey, with 76 (71%) indicating that they had shared their findings with their healthcare providers. However, a sufficiently specific genetic diagnosis appeared in the EHRs and problem lists of only 22 and 10%, respectively, of participants without prior knowledge. Furthermore, review of participant EHRs provided evidence of possible relevant changes in clinical care for only a handful of participants. Up to 19% of participants would have benefited from earlier screening due to prior presentation of their condition. These results suggest that continuous support for both participants and their providers is necessary to maximize the benefit of population-based genetic screening. We recommend that genetic screening projects require participants' consent to directly document their genetic findings in their EHRs. Additionally, we recommend that they provide healthcare providers with ongoing training regarding documentation of findings and with clinical decision support regarding subsequent care.

SARS-CoV-2 variant Delta rapidly displaced variant Alpha in the United States and led to higher viral loads.

We report on the sequencing of 74,348 SARS-CoV-2 positive samples collected across the United States and show that the Delta variant, first detected in the United States in March 2021, made up the majority of SARS-CoV-2 infections by July 1, 2021 and accounted for >99.9% of the infections by September 2021. Not only did Delta displace variant Alpha, which was the dominant variant at the time, it also displaced the Gamma, Iota, and Mu variants. Through an analysis of quantification cycle (Cq) values, we demonstrate that Delta infections tend to have a 1.7× higher viral load compared to Alpha infections (a decrease of 0.8 Cq) on average. Our results are consistent with the hypothesis that the increased transmissibility of the Delta variant could be due to the ability of the Delta variant to establish a higher viral load earlier in the infection as compared to the Alpha variant.

Using symptom-based case predictions to identify host genetic factors that contribute to COVID-19 susceptibility.

Epidemiological and genetic studies on COVID-19 are currently hindered by inconsistent and limited testing policies to confirm SARS-CoV-2 infection. Recently, it was shown that it is possible to predict COVID-19 cases using cross-sectional self-reported disease-related symptoms. Here, we demonstrate that this COVID-19 prediction model has reasonable and consistent performance across multiple independent cohorts and that our attempt to improve upon this model did not result in improved predictions. Using the existing COVID-19 prediction model, we then conducted a GWAS on the predicted phenotype using a total of 1,865 predicted cases and 29,174 controls. While we did not find any common, large-effect variants that reached genome-wide significance, we do observe suggestive genetic associations at two SNPs (rs11844522, p = 1.9x10-7; rs5798227, p = 2.2x10-7). Explorative analyses furthermore suggest that genetic variants associated with other viral infectious diseases do not overlap with COVID-19 susceptibility and that severity of COVID-19 may have a different genetic architecture compared to COVID-19 susceptibility. This study represents a first effort that uses a symptom-based predicted phenotype as a proxy for COVID-19 in our pursuit of understanding the genetic susceptibility of the disease. We conclude that the inclusion of symptom-based predicted cases could be a useful strategy in a scenario of limited testing, either during the current COVID-19 pandemic or any future viral outbreak.

Positive predictive value highlights four novel candidates for actionable genetic screening from analysis of 220,000 clinicogenomic records.

PURPOSE: To identify conditions that are candidates for population genetic screening based on population prevalence, penetrance of rare variants, and actionability. METHODS: We analyzed exome and medical record data from >220,000 participants across two large population health cohorts with different demographics. We performed a gene-based collapsing analysis of rare variants to identify genes significantly associated with disease status. RESULTS: We identify 74 statistically significant gene-disease associations across 27 genes. Seven of these conditions have a positive predictive value (PPV) of at least 30% in both cohorts. Three are already used in population screening programs (BRCA1, BRCA2, LDLR), and we also identify four new candidates for population screening: GCK with diabetes mellitus, HBB with ß-thalassemia minor and intermedia, PKD1 with cystic kidney disease, and MIP with cataracts. Importantly, the associations are actionable in that early genetic screening of each of these conditions is expected to improve outcomes. CONCLUSION: We identify seven genetic conditions where rare variation appears appropriate to assess in population screening, four of which are not yet used in screening programs. The addition of GCK, HBB, PKD1, and MIP rare variants into genetic screening programs would reach an additional 0.21% of participants with actionable disease risk, depending on the population.

Drugs affecting renin-angiotensin-aldosterone system and the cancer risk: A meta-analysis of nested case-control studies.

Multiple studies have discussed the associations between drugs affecting the renin-angiotensin-aldosterone system and the cancer risk, but their consequence s were conflicting. A meta-analysis of nested case-control studies published regarding this subject was conducted in our study, aims to estimate the association between ACEI/ARB and the cancer risk. Pubmed database was searched up to February, 1 2016 to identify eligible nested case-control studies, and we used Newcastle-Ottawa Scale (NOS) to assess quality of the studies. Pooled odds ratio (OR) and 95% confidence intervals (CIs) were calculated (with fixed effect model: Mantel-Haenszel). Publication bias and heterogeneity were evaluated before the calculation. Subgroup analysis and sensitivity analysis were also performed. Seven studies contributed to the analysis. Overall, ACEI/ARB use was not associated with the risk of cancer (OR=0.99, 95% CI 0.97-1.01), nor in long-term use patients (OR=0.97, 95% CI 0.92-1.01). ACEI may decrease cancer risk (OR=0.90, 95% CI 0.82-0.99). We observed no significant publication bias. In conclusion, ACEI/ARB use was not associated with cancer risk, nor in long-term use patients, but ACEI use may decrease cancer risk. More researches are needed to confirm these findings.