Diversity and Inclusion in Drug Development: Rethinking Intrinsic and Extrinsic Factors with Patient Centricity

Explainable machine learning prediction of edema adverse events in patients treated with tepotinib

Tepotinib is approved for the treatment of patients with non-small-cell lung cancer harboring MET exon 14 skipping alterations. While edema is the most prevalent adverse event (AE) and a known class effect of MET inhibitors including tepotinib, there is still limited understanding about the factors contributing to its occurrence. Herein, we apply machine learning (ML)-based approaches to predict the likelihood of occurrence of edema in patients undergoing tepotinib treatment, and to identify factors influencing its development over time. Data from 612 patients receiving tepotinib in five Phase I/II studies were modeled with two ML algorithms, Random Forest, and Gradient Boosting Trees, to predict edema AE incidence and severity. Probability calibration was applied to give a realistic estimation of the likelihood of edema AE. Best model was tested on follow-up data and on data from clinical studies unused while training. Results showed high performances across all the tested settings, with F1 scores up to 0.961 when retraining the model with the most relevant covariates. The use of ML explainability methods identified serum albumin as the most informative longitudinal covariate, and higher age as associated with higher probabilities of more severe edema. The developed methodological framework enables the use of ML algorithms for analyzing clinical safety data and exploiting longitudinal information through various covariate engineering approaches. Probability calibration ensures the accurate estimation of the likelihood of the AE occurrence, while explainability tools can identify factors contributing to model predictions, hence supporting population and individual patient-level interpretation.

Asia-Inclusive Global Development of Enpatoran: Results of an Ethno-Bridging Study, Intrinsic/Extrinsic Factor Assessments and Disease Trajectory Modeling to Inform Design of a Phase II Multiregional Clinical Trial

Enpatoran is a novel, highly selective, and potent dual toll-like receptor (TLR)7 and TLR8 inhibitor currently under development for the treatment of autoimmune disorders including systemic lupus erythematosus (SLE), cutaneous lupus erythematosus (CLE), and myositis. The ongoing phase II study (WILLOW; NCT05162586) is evaluating enpatoran for 24 weeks in patients with active SLE or CLE and is currently recruiting. To support development of WILLOW as an Asia-inclusive multiregional clinical trial (MRCT) according to International Conference on Harmonisation E5 and E17 principles, we have evaluated ethnic sensitivity to enpatoran based on clinical pharmacokinetic (PK), pharmacodynamic (PD), and safety data from an ethno-bridging study (NCT04880213), supplemented by relevant quantitative PK, PD, and disease trajectory modeling (DTM) results, and drug metabolism/disease knowledge. A single-center, open-label, sequential dose group study in White and Japanese subjects matched by body weight, height, and sex demonstrated comparable PK and PD properties for enpatoran in Asian vs. non-Asian (White and other) subjects across single 100, 200, and 300 mg orally administered doses. DTM suggested no significant differences in SLE disease trajectory for Asian vs. non-Asian individuals. Aldehyde oxidase (AOX) is considered to be a key contributor to enpatoran metabolism, and a literature review indicated no relevant ethnic differences in AOX function based on in vitro and clinical PK data from marketed drugs metabolized by AOX, supporting the conclusion of low ethnic sensitivity for enpatoran. Taken together, the inclusion of Asian patients in MRCTs including WILLOW was informed based on a Totality of Evidence approach.

Machine Learning in Modeling Disease Trajectory and Treatment Outcomes: An Emerging Enabler for Model-Informed Precision Medicine

The increasing breadth and depth of resolution in biological and clinical data, including -omics and real-world data, requires advanced analytical techniques like artificial intelligence (AI) and machine learning (ML) to fully appreciate the impact of multi-dimensional population variability in intrinsic and extrinsic factors on disease progression and treatment outcomes. Integration of advanced data analytics in Quantitative Pharmacology is crucial for drug-disease knowledge management, enabling precise, efficient and inclusive drug development and utilization - an application we refer to as model-informed precision medicine. AI/ML enables characterization of the molecular and clinical sources of heterogeneity in disease trajectory, advancing end point qualification and biomarker discovery, and informing patient enrichment for proof-of-concept studies as well as trial designs for efficient evidence generation incorporating digital twins and virtual control arms. Explainable ML methods are valuable in elucidating predictors of efficacy and safety of pharmacological treatments, thereby informing response monitoring and risk mitigation strategies. In oncology, emerging opportunities exist for development of the next generation of disease models via ML-assisted joint longitudinal modeling of high-dimensional biomarker data such as circulating tumor DNA and radiomics profiles as predictors of survival outcomes. Finally, mining real-world data leveraging ML algorithms enables understanding of the impact of exclusion criteria on clinical outcomes, thereby informing rational design of appropriately inclusive clinical trials through data-driven broadening of eligibility criteria. Herein, we provide an overview of the aforementioned contexts of use of ML in drug-disease modeling based on examples across multiple therapeutic areas including neurology, rare diseases, autoimmune diseases, oncology and immuno-oncology.

Tumor growth inhibition modeling in patients with second line biliary tract cancer and first line non-small cell lung cancer based on bintrafusp alfa trials

This analysis aimed to quantify tumor dynamics in patients receiving either bintrafusp alfa (BA) or pembrolizumab, by population pharmacokinetic (PK)-pharmacodynamic modeling, and investigate clinical and molecular covariates describing the variability in tumor dynamics by pharmacometric and machine-learning (ML) approaches. Data originated from two clinical trials in patients with biliary tract cancer (BTC; NCT03833661) receiving BA and non-small cell lung cancer (NSCLC; NCT03631706) receiving BA or pembrolizumab. Individual drug exposure was estimated from previously developed population PK models. Population tumor dynamics models were developed for each drug-indication combination, and covariate evaluations performed using nonlinear mixed-effects modeling (NLME) and ML (elastic net and random forest models) approaches. The three tumor dynamics' model structures all included linear tumor growth components and exponential tumor shrinkage. The final BTC model included the effect of drug exposure (area under the curve) and several covariates (demographics, disease-related, and genetic mutations). Drug exposure was not significant in either of the NSCLC models, which included two, disease-related, covariates in the BA arm, and none in the pembrolizumab arm. The covariates identified by univariable NLME and ML highly overlapped in BTC but showed less agreement in NSCLC analyses. Hyperprogression could be identified by higher tumor growth and lower tumor kill rates and could not be related to BA exposure. Tumor size over time was quantitatively characterized in two tumor types and under two treatments. Factors potentially related to tumor dynamics were assessed using NLME and ML approaches; however, their net impact on tumor size was considered as not clinically relevant.

Clinical Pharmacology of Brigatinib: A Next-Generation Anaplastic Lymphoma Kinase Inhibitor

Brigatinib, a next-generation anaplastic lymphoma kinase (ALK) inhibitor designed to overcome mechanisms of resistance associated with crizotinib, is approved for the treatment of ALK-positive advanced or metastatic non-small cell lung cancer. After oral administration of single doses of brigatinib 30-240 mg, the median time to reach maximum plasma concentration ranged from 1 to 4 h. In patients with advanced malignancies, brigatinib showed dose linearity over the dose range of 60-240 mg once daily. A high-fat meal had no clinically meaningful effect on systemic exposures of brigatinib (area under the plasma concentration-time curve); thus, brigatinib can be administered with or without food. In a population pharmacokinetic analysis, a three-compartment pharmacokinetic model with transit absorption compartments was found to adequately describe brigatinib pharmacokinetics. In addition, the population pharmacokinetic analyses showed that no dose adjustment is required based on body weight, age, race, sex, total bilirubin (< 1.5× upper limit of normal), and mild-to-moderate renal impairment. Data from dedicated phase I trials have indicated that no dose adjustment is required for patients with mild or moderate hepatic impairment, while a dose reduction of approximately 40% (e.g., from 180 to 120 mg) is recommended for patients with severe hepatic impairment, and a reduction of approximately 50% (e.g., from 180 to 90 mg) is recommended when administering brigatinib to patients with severe renal impairment. Brigatinib is primarily metabolized by cytochrome P450 (CYP) 3A, and results of clinical drug-drug interaction studies and physiologically based pharmacokinetic analyses have demonstrated that coadministration of strong or moderate CYP3A inhibitors or inducers with brigatinib should be avoided. If coadministration with a strong or moderate CYP3A inhibitor cannot be avoided, the dose of brigatinib should be reduced by approximately 50% (strong CYP3A inhibitor) or approximately 40% (moderate CYP3A inhibitor), respectively. Brigatinib is a weak inducer of CYP3A in vivo; data from a phase I drug-drug interaction study showed that coadministration of brigatinib 180 mg once daily reduced the oral midazolam area under the plasma concentration-time curve from time zero to infinity by approximately 26%. Brigatinib did not inhibit CYP1A2, CYP2B6, CYP2C8, CYP2C9, CYP2C19, or CYP2D6 at clinically relevant concentrations in vitro. Exposure-response analyses based on data from the ALTA (ALK in Lung Cancer Trial of AP26113) and ALTA-1L pivotal trials of brigatinib confirm the favorable benefit versus risk profile of the approved titration dosing regimen of 180 mg once daily (after a 7-day lead-in at 90 mg once daily).

Disease trajectory of SLE clinical endpoints and covariates affecting disease severity and probability of response: Analysis of pooled patient-level placebo (Standard-of-Care) data to enable model-informed drug development

Systemic lupus erythematosus (SLE) is an autoimmune disease affecting multiple organ systems. Many investigational agents have failed or shown only modest effects when added to standard of care (SoC) therapy in placebo-controlled trials, and only two therapies have been approved for SLE in the last 60 years. Clinical trial outcomes have shown discordance in drug effects between clinical endpoints. Herein, we characterized longitudinal disease activity in the SLE population and the sources of variability by developing a latent disease trajectory model for SLE component endpoints (Systemic Lupus Erythematosus Disease Activity Index [SLEDAI], Physician's Global Assessment [PGA], British Isles Lupus Assessment Group Index [BILAG]) and composite endpoints (Systemic Lupus Erythematosus Responder Index [SRI], BILAG-based Composite Lupus Assessment [BICLA], and Lupus Low Disease Activity State [LLDAS]) using patient-level historical SoC data from nine phase II and III studies. Across all endpoints, in predictions up to 52 weeks from the final disease trajectory model, the following baseline covariates were associated with a greater decrease in SLE disease activity and higher response to placebo + SoC: Hispanic ethnicity from Central/South America, absence of hypocomplementemia, recent SLE diagnosis, and high baseline disease activity score using SLEDAI and BILAG separately. No discernible differences were observed in the trajectory of response to placebo + SoC across different SoC medications (antimalarial and immunosuppressant such as mycophenolate, methotrexate, and azathioprine). Across all endpoints, disease trajectory showed no difference in Asian versus non-Asian patients, supporting Asia-inclusive global SLE drug development. These results describe the first population approach to support a model-informed drug development framework in SLE.

Asia-Inclusive Clinical Research and Development Enabled by Translational Science and Quantitative Clinical Pharmacology: Toward a Culture That Challenges the Status Quo

Access lag to innovative therapies in Asian populations continues to present a challenge to global health. Recent progressive changes in the global regulatory landscape, including newer guidelines, are enabling simultaneous global drug development and near-simultaneous global drug registration. The International Conference on Harmonization (ICH) E17 guideline outlines general principles for the design and analysis of multiregional clinical trials (MRCTs). We posit that translational research and quantitative clinical pharmacology tools are core enablers for Asia-inclusive global drug development aligned with ICH E17 principles. Assessment of ethnic sensitivity should be initiated early in the development lifecycle to inform the need for, and extent of, Asian phase I ethno-bridging data. Relevant ethno-bridging data may be generated as standalone Asian phase I trials, as part of Western First-In-Human trials, or under accelerated development settings as a lead-in phase in an MRCT. Quantitative understanding of human clearance mechanisms and pharmacogenetic factors is vital to forecasting ethnic sensitivity in drug exposure using physiologically-based pharmacokinetic models. Stratification factors to control heterogeneity in MRCTs can be identified by reverse translational research incorporating pharmacometric disease models and model-based meta-analyses. Because epidemiological variations can extend to the molecular level, quantitative systems pharmacology models may be useful in forecasting how molecular variation in therapeutic targets or pathway proteins across populations might impact treatment outcomes. Through prospective evaluation of conservation in drug- and disease-related intrinsic and extrinsic factors, a pooled East Asian region can be implemented in Asia-inclusive MRCTs to maximize efficiency in substantiating evidence of benefit-risk for the region at-large with a Totality of Evidence approach.

Pharmacogenomics and health disparities, are we helping?

Pharmacogenomics has been at the forefront of precision medicine during the last few decades. Precision medicine carries the potential of improving health outcomes at both the individual as well as population levels. To harness the benefits of its initiatives, careful dissection of existing health disparities as they relate to precision medicine is of paramount importance. Attempting to address the existing disparities at the early stages of design and implementation of these efforts is the only guarantee of a successful just outcome. In this review, we glance at a few determinants of existing health disparities as they intersect with pharmacogenomics research and implementation. In our opinion, highlighting these disparities is imperative for the purpose of researching meaningful solutions. Failing to identify, and hence address, these disparities in the context of the current and future precision medicine initiatives would leave an already strained health system, even more inundated with inequality.

How resilient were we in 2021? Results of a LinkedIn Survey including biomedical and pharmaceutical professionals using the Benatti Resiliency Model

Enhancing resiliency should elevate innovation and efficiency in biomedical research and development (R&D); however, compared with other professions, data on practice of resilience is lacking. Using the Benatti Resiliency Model (5 anchors: Well-Being, Self-Awareness, Brand, Connection, and Innovation), we surveyed professionals, including those in biomedical and pharmaceutical R&D. A structured LinkedIn questionnaire (March 16-May 23, 2021), surveyed each model anchor using five categories. One hundred fifty-eight participants (~6% student/trainee, 18%, 27%, and 49% in 1-5, 5-15 or >15 years post-terminal degree) took the survey (90 in biomedical and pharmaceutical R&D). Over 50% chose "always"/"often" across questions, except external influence or engagement. The question with one of the lowest "always" scores (~15%) was "I get feedback on my influence and impact in my career" in Brand, highlighting areas for leadership development and coaching. In the anchor of Well-being, nutrition and stress management also received some lowest "always" scores (~15% for both). Connection and Innovation scores trended slightly higher in biomedical and pharmaceutical R&D. No students/trainees chose "always" in Brand, indicating evolution of brand maturity over time. Self- and survey-assessed resiliency scores were associated (r_s  = 0.37, p < 0.0001). Our survey yielded actionable insights on Resilience, including "best practices" through an open-ended question for one thing most useful to boost resilience in the survey and is the first application of the Benatti Model for crowdsourced research.