Key Takeaways & Executive Findings
- •• • The hepatic spheroid model achieved an AUC of 0.92 (95% CI: 0.88-0.96) for DILI prediction, significantly outperforming 2D cultures (AUC=0.78) and animal models (AUC=0.71), enabling more accurate early-stage hepatotoxicity screening. • • The model demonstrated a positive predictive value of 89% for severe DILI compounds, with enhanced sensitivity for cholestatic and steatotic hepatotoxicants, reducing the risk of overlooking clinically relevant injuries. • • Transcriptomic analysis identified key predictive biomarkers, including genes involved in oxidative stress, mitochondrial dysfunction, and bile acid transport, providing mechanistic insights into DILI pathways. • • Integration of high-content imaging and biochemical assays with machine learning classifiers (e.g., random forest) yielded a robust predictive framework, achieving a sensitivity of 85% and specificity of 88% on the validation set, facilitating high-throughput screening in drug development.
Abstract
Drug-induced liver injury (DILI) remains a leading cause of drug attrition and post-marketing withdrawals. Traditional preclinical models often fail to predict human hepatotoxicity accurately. Here, we developed a novel hepatic spheroid model using primary human hepatocytes and evaluated its transcriptomic and functional responses to a panel of known hepatotoxicants. We integrated high-content imaging, biochemical assays, and RNA-seq data to train machine learning classifiers for DILI prediction. Our model achieved an area under the receiver operating characteristic curve (AUC) of 0.92 (95% CI: 0.88-0.96) on a validation set of 150 compounds, outperforming conventional 2D cultures (AUC=0.78) and animal models (AUC=0.71). Key predictive biomarkers included genes involved in oxidative stress, mitochondrial dysfunction, and bile acid transport. The spheroid model exhibited enhanced sensitivity for detecting cholestatic and steatotic hepatotoxicants, with a positive predictive value of 89% for compounds causing severe DILI. Our findings demonstrate that combining physiologically relevant 3D liver models with machine learning offers a robust platform for early DILI risk assessment, potentially reducing late-stage drug failures and improving patient safety.
1. Introduction
Drug-induced liver injury (DILI) remains a major challenge in pharmaceutical development, accounting for approximately 30% of all drug safety-related withdrawals. Conventional preclinical models, including 2D hepatocyte cultures and animal models, often fail to recapitulate human-specific metabolic and toxicological responses, leading to late-stage clinical failures and post-marketing safety issues. The lack of predictive in vitro systems that accurately reflect human liver physiology has created a critical bottleneck in early drug safety assessment.
To address this, we developed a physiologically relevant hepatic spheroid model using primary human hepatocytes, which better mimics the in vivo microenvironment, including cell-cell interactions and metabolic zonation. By integrating high-content imaging, biochemical assays, and transcriptomic profiling, we generated a comprehensive dataset of hepatotoxic responses. Machine learning models trained on this data demonstrated superior predictive performance compared to traditional models. This approach offers a scalable and cost-effective solution for early DILI risk assessment, potentially reducing drug attrition and improving patient safety.
Loading authentic research manuscript (Pages 1–5)...
ZHANG Wei, LI Ming, WANG Fang, et al. (2025). Predictive Modeling of Drug-Induced Liver Injury Using a Novel Hepatic Spheroid Model and Machine Learning Algorithms. Chinese Journal of New Drugs. https://doi.org/pub_80__articleID_232
Research & Educational Purpose Only:The translations, structured abstracts, analytical annotations, and data reports provided by SinoBioData are intended exclusively for academic research, internal corporate R&D, and educational benchmarking. They do not constitute formal engineering, chemical safety, legal, or professional advice.
Copyright & Intellectual Property Notice: Original copyright of the underlying source articles and experimental data remains with the respective authors, institutions, and original publishing journals. SinoBioData claims intellectual property only over its proprietary translations, analytical syntheses, and AEO structured enhancements in accordance with international fair use and academic citation principles.
Frequently Asked Questions
What are the key failure mechanisms of the spheroid model under chronic exposure conditions, and how does it compare to in vivo hepatotoxicity?
Under chronic exposure (up to 14 days), the spheroid model exhibited progressive steatosis and apoptosis, with a 2.5-fold increase in intracellular triglyceride accumulation and a 3-fold increase in caspase-3/7 activity compared to acute exposure. These changes correlated with transcriptomic signatures of lipotoxicity and mitochondrial dysfunction, consistent with in vivo findings. The model's predictive accuracy for chronic DILI was 0.89 (AUC), suggesting it can capture cumulative effects not observed in acute assays.
How does the cost and throughput of this spheroid-based assay compare to traditional 2D cultures for industrial screening?
The spheroid assay requires approximately 1.5 times the cost of 2D cultures due to specialized plates and longer culture times (7 days vs. 24 hours). However, it can be automated on high-content imaging platforms, achieving a throughput of 384-well format, which is comparable to 2D. The improved predictive performance reduces downstream costs by decreasing false positives and late-stage failures, offering a net cost benefit in drug development.
What are the scalability bottlenecks for using this model in large-scale drug screening, and how can they be overcome?
Scalability is limited by the availability of primary human hepatocytes and the need for consistent spheroid formation. We addressed this by using cryopreserved hepatocytes from multiple donors, achieving a batch-to-batch reproducibility of 92% (coefficient of variation <10% for ATP content). Additionally, we developed a semi-automated seeding protocol that reduces manual handling time by 40%, enabling screening of up to 10,000 compounds per month.
Can the machine learning model be generalized to predict DILI for novel chemical entities outside the training set?
The model was validated on an external test set of 50 compounds not used in training, achieving an AUC of 0.85. However, its performance decreases for compounds with novel mechanisms of toxicity not represented in the training data. To improve generalizability, we recommend incorporating additional mechanistic endpoints, such as mitochondrial membrane potential and reactive oxygen species, and using domain adaptation techniques to handle chemical diversity.
What are the regulatory implications of using this spheroid model for DILI risk assessment in IND submissions?
While the model is not yet accepted as a standalone regulatory tool, it can serve as a complementary assay to standard animal studies. Our data suggest that it could reduce the reliance on non-human primate studies by providing human-relevant toxicity data. We are currently engaging with regulatory agencies to qualify the model as a part of a weight-of-evidence approach, which may accelerate its adoption in drug development.
Related Technical Papers & Translations
Adverse Events Reporting System for Vaccine Safety Surveillance: A Comprehensive Analysis
Background: Adverse events following immunization (AEFI) are critical to monitor for vaccine safety. This study evaluates the performance of an adverse events reporting system (AERS) integrated with a vaccine adverse event reporting system (VAERS) to enhance surveillance. Methods: We analyzed data from multiple sources including the Vaccine Adverse Event Reporting System (VAERS), the Vaccine Safety Datalink (VSD), and the Clinical Immunization Safety Assessment (CISA) network. A novel framework was developed to integrate these systems, incorporating natural language processing for signal detection. Results: The integrated system improved detection of rare adverse events by 25% compared to traditional methods. The system identified new safety signals for influenza and COVID-19 vaccines. Conclusions: The proposed AERS framework enhances vaccine safety surveillance, enabling timely identification of potential risks. Integration of diverse data sources and advanced analytics is essential for robust pharmacovigilance.
Efficacy and Safety of Ferric Carboxymaltose in Treating Iron Deficiency Anemia: A Meta-Analysis of Randomized Controlled Trials
Background: Iron deficiency anemia (IDA) is a global health concern, and intravenous ferric carboxymaltose (FCM) has emerged as a promising treatment. This meta-analysis aimed to evaluate the efficacy and safety of FCM compared to other iron therapies or placebo in adults with IDA. Methods: We systematically searched PubMed, Embase, and Cochrane Library up to December 2024. Randomized controlled trials (RCTs) comparing FCM with active comparators or placebo in adults with IDA were included. The primary outcomes were change in hemoglobin (Hb) from baseline, and safety outcomes included adverse events (AEs) and serious adverse events (SAEs). Pooled estimates were calculated using random-effects models. Results: A total of 15 RCTs involving 4,856 patients were included. FCM significantly increased Hb levels compared to placebo (mean difference [MD] 1.2 g/dL, 95% CI 0.9-1.5) and was non-inferior to other intravenous iron preparations. The risk of AEs was similar between FCM and comparators (risk ratio [RR] 1.05, 95% CI 0.95-1.16), but FCM was associated with a lower risk of gastrointestinal AEs compared to oral iron. Serious adverse events were rare and comparable across groups. Conclusion: Ferric carboxymaltose is effective and safe for treating IDA, offering a convenient single-dose option with a favorable safety profile. These findings support its use in clinical practice.
Adverse Drug Reactions Associated with COVID-19 Vaccination: A Systematic Review and Meta-Analysis
Background: The rapid development and deployment of COVID-19 vaccines have been crucial in controlling the pandemic. However, adverse drug reactions (ADRs) associated with these vaccines have raised concerns. This systematic review and meta-analysis aimed to comprehensively evaluate the incidence and types of ADRs following COVID-19 vaccination. Methods: We systematically searched PubMed, Embase, and Cochrane Library from inception to December 2024. Randomized controlled trials and observational studies reporting ADRs after COVID-19 vaccination were included. A random-effects model was used to pool incidence rates, and subgroup analyses were performed by vaccine type and dose. Results: A total of 45 studies with 1,234,567 participants were included. The overall incidence of any ADR was 62.3% (95% CI: 58.1-66.4%). Common local reactions included injection site pain (48.2%), swelling (22.5%), and redness (18.7%). Systemic reactions included fatigue (34.6%), headache (28.9%), and myalgia (22.3%). Serious ADRs were rare (0.02%). Subgroup analysis showed higher incidence with mRNA vaccines compared to viral vector vaccines. Conclusion: COVID-19 vaccines are associated with a high incidence of mild-to-moderate ADRs, but serious ADRs are extremely rare. These findings support the overall safety of COVID-19 vaccination programs.