Accurate prediction of drug-drug interactions (DDIs) is crucial for ensuring drug safety and supporting pharmaceutical development. Existing methods still face challenges in jointly modeling molecular structural information and biomedical semantic contexts. To address this issue, we propose DSAT-DDI, a Dual-channel Structure-Aware Transformer framework for DDI prediction. The core design of DSAT-DDI lies in incorporating structure-aware attention into both molecular and knowledge graph representation learning, enabling the model to capture local chemical environments and semantic neighborhoods while preserving global dependency modeling. In addition, a bidirectional cross-attention fusion mechanism is introduced to dynamically integrate molecular and biomedical representations and generate pair-aware drug embeddings. Experimental results on the DrugBank and KEGG datasets show that DSAT-DDI consistently outperforms representative molecular-based, knowledge graph-based, and multimodal baselines across multiple evaluation metrics. These results demonstrate the effectiveness of structure-aware attention and bidirectional cross-attention fusion for robust DDI prediction.

With the rapid development of artificial intelligence in healthcare, medication recommendation systems have become important tools for achieving personalized treatment. However, existing methods still have limitations in modeling drug molecular functions and perceiving dynamic changes in patient conditions. To address these issues, this paper proposes MoTRANet, a safety-aware combinatorial medication recommendation model. The model designs a dynamic patient modeling module that integrates GRU, Transformer, and diagnosis-procedure cross attention to effectively capture the dynamic evolution of patient health status. On the drug side, an ECFP-enhanced graph attention network and LightGCN-based motif propagation are employed to construct drug representations with molecular functional information. Furthermore, a DDI-aware encoder and a curriculum learning strategy are introduced to dynamically balance recommendation accuracy and medication safety during training. Experimental results on two public datasets, MIMIC III and MIMIC-IV, demonstrate that MoTRANet outperforms existing mainstream baseline methods in terms of Jaccard, F1, and PRAUC, while achieving a relatively low DDI rate, validating the model's accuracy and safety. Future work will focus on multi center real-world data validation, dose sensitivity analysis, and model interpretability.

Isavuconazonium sulfate is the water-soluble prodrug of the active triazole antifungal agent isavuconazole. Isavuconazonium sulfate is approved for the treatment of invasive aspergillosis and invasive mucormycosis in pediatric and adult populations. In the present analysis, an isavuconazole physiologically-based pharmacokinetic (PBPK) model was built and verified in adult and pediatric populations using the Simcyp® simulator. The verified pediatric model was then used to predict the magnitude of DDI risk in pediatric populations vs. adult populations. The simulations conducted were designed to match the observed clinical trials conducted in adults but with pediatric equivalent doses based on clinical guidelines. Based on the results of the analysis, isavuconazole DDI risk was similar between pediatric and adult populations. When the isavuconazole perpetrator DDI simulations were compared to adults, there was a trend of larger geometric mean (GM) AUCinf and Cmax ratios in patients 1 to < 3 years old receiving the 15 mg/kg dose. The GM AUCinf ratio in this study population ranged from 7 to 36% larger than predicted in the adult population. Based on the results of this analysis and literature reports, it is advisable to show greater caution when considering potential DDI in patients < 3-year-old, especially when receiving larger isavuconazonium sulfate doses. In addition, caution should be used in patients < 3 years-old when considering coadministration of isavuconazole with narrow therapeutic index medications such as digoxin and warfarin or medications with large variability in oral bioavailability such as midazolam.

Suzetrigine (VX-548) is a first-in-class, oral, non-opioid, FDA-approved analgesic for the treatment of moderate to severe acute pain. Based on Phase 1 clinical drug-drug interaction (DDI) studies, suzetrigine is a weak-to-moderate cytochrome P450 3A (CYP3A) inducer and thus has the potential to decrease exposure and impact efficacy of CYP3A substrates, such as apixaban and rivaroxaban, which are commonly used as anticoagulants peri- and post-operatively. Given this, it is important to assess the DDIs of apixaban and rivaroxaban with suzetrigine. Physiologically based pharmacokinetic (PBPK) models are commonly used to assess DDIs mediated through modulation of CYP3A and were used here to quantify the potential effect of co-administration of suzetrigine on apixaban and rivaroxaban PK. The PBPK models were validated using observed data from suzetrigine, apixaban, and rivaroxaban clinical DDI studies. The validated models predict that suzetrigine decreases apixaban AUCinf by 6.1% (90% confidence interval [CI]: 5.7% to 6.4%) and Cmax by 1.8% (90% CI: 1.7% to 1.9%), and rivaroxaban AUCinf by 8.4% (90% CI: 7.2% to 9.6%) and Cmax by 3.5% (90% CI: 3.3% to 3.7%). The predicted small reductions in apixaban and rivaroxaban exposures (AUCinf <10%) are not expected to be clinically relevant since a 20% no-effect boundary is typically used to determine the clinical impact of a DDI. The PBPK model predictions are consistent with the apixaban and rivaroxaban labels, which do not warn against co-administration with weak or moderate CYP3A inducers, such as suzetrigine.

Addressing knowledge gaps at drug approval through drug-drug interaction (DDI)-related postmarketing requirements (PMRs) and commitments (PMCs) is essential for safe and effective use. This analysis reviewed CYP3A-related PMRs and PMCs, the most evaluated enzyme in postmarketing studies, to characterize their scope. CYP3A-related PMRs, PMCs, and labeling recommendations were identified for small-molecule new molecular entities approved by the U.S. Food and Drug Administration (2015-2024) using the Certara Drug Interaction Database (DIDB). The latest product labels from Drugs@FDA were reviewed for updates. Drugs were categorized as CYP3A substrates, inhibitors, or inducers based on PMR/PMC language. Of 332 approved drugs, 51 (15%) had CYP3A-related PMRs/PMCs, totaling 91 studies (53 PMRs, 37 PMCs, and 1 unspecified). About two-thirds were oncology drugs, including 25 kinase inhibitors. As substrates, 37 drugs had 65 requests: 26 with inhibitors (all PMRs) and 39 with inducers (30 PMCs, 8 PMRs, and 1 unspecified); 20 included physiologically based pharmacokinetic (PBPK) modeling. As precipitants, 25 drugs had 26 requests: 12 as inhibitors (11 PMRs, 1 PMC), 14 as inducers (8 PMRs, 6 PMCs), with only 1 PBPK option. Requests most often involved studies with a CYP3A substrate (N = 24), moderate inducer (N = 21), strong inducer (N = 15), strong inhibitor (N = 10), and moderate inhibitor (N = 9). Initial labeling recommendations were issued for 60 studies (65%; 35 drugs), and 53 studies (58%; 29 drugs) resulted in label updates. Overall, CYP3A-mediated DDIs remain a key focus of postmarketing evaluation, particularly for oncology drugs. PMRs and PMCs are critical for resolving uncertainties and informing label updates.

Endogenous biomarkers have gained considerable attention as tools for assessing transporter-mediated drug-drug interactions (tDDIs). While plasma coproporphyrin I is considered a validated biomarker for assessing hepatic organic anion transporting polypeptide (OATP) 1B-mediated DDI risk, data supporting biomarkers for other major drug transporters, particularly renal transporters, remains limited. The International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) included endogenous biomarker-based assessments in its recent guidance, ICH M12, as an alternative approach for evaluating tDDIs, which is anticipated to broaden the use of this strategy in early clinical development. However, methodological gaps remain in biomarker measurement, including study design, biological matrix selection, and data analysis. Importantly, consensus is still needed on the most appropriate biomarkers for renal organic cation transporter (OCT) 2 and multidrug and toxin extrusion (MATE) transporters 1 and 2-K, such as creatinine, N[1]-methylnicotinamide, and N[1]-methyladenosine, given the influence of various extrinsic and intrinsic factors on their performance. This review aims to summarize the current state of clinical evidence for endogenous biomarkers of major hepatic and renal transporters, including study design considerations relevant to meaningful data interpretation. It also briefly summarizes reported Physiologically based pharmacokinetic (PBPK) models of these biomarkers, highlights key gaps in current understanding, and identifies future work needed to expand the application of a biomarker-based tDDI strategy.

Drug-Drug Interaction (DDI) prediction is crucial for ensuring the safety and efficacy of combination drug therapies. However, most existing methods lack uncertainty estimation, limiting their reliability and practical adoption. While conformal prediction (CP) offers guaranteed coverage in classification tasks, its direct application to DDI prediction faces challenges in achieving valid marginal coverage and efficient prediction set sizes. To overcome these limitations, this work introduces WeightedCPDDI, the first conformal prediction framework tailored for DDI tasks. WeightedCPDDI incorporates a novel weighting mechanism to achieve user-specific coverage and enhance prediction efficiency. Evaluation across transductive and inductive settings demonstrates WeightedCPDDI's strong performance. In the transductive setting, it achieves target coverage with smaller or comparable prediction sets, indicating improved calibration and efficiency. In the inductive setting, WeightedCPDDI remains competitive with or better than state-of-the-art methods in both seen-unseen and unseen-unseen scenarios, with the largest efficiency gains in the challenging unseen-unseen drug-pair setting, where it reduces prediction set sizes relative to the strongest baselines while maintaining the target coverage rate. Its strong performance across various backbone models highlights its generalizability and practical reliability for DDI prediction under uncertainty.

Targeted drug therapies offer a promising approach for treating complex diseases, with combinational drug therapies often employed to enhance therapeutic efficacy. However, unintended drug-drug interactions (DDIs) may undermine treatment outcomes or cause adverse side effects. In this work, we propose a novel joint learning framework for the simultaneous prediction of effective drug combinations and DDIs, based on coupled tensor-tensor factorization. Specifically, we model drug combination therapies and DDI by representing drug-drug-disease associations and DDI profiles as coupled three-way tensors. To address the challenges of data incompleteness and sparsity, the proposed model integrates auxiliary drug similarity information, such as chemical structure similarities, drug-specific side effects, drug target profiles, and drug inhibition data on cancer cell lines, within a multiview learning framework. For optimization, we adopt a modified alternating direction method of multipliers (ADMM) algorithm with non-negativity constraints. In addition to standard tensor completion tasks, we further evaluate the proposed method under a more realistic new-drug prediction setting, where all interactions involving a previously unseen drug are withheld. This scenario closely aligns with real-world applications, in which reliable predictions for emerging or under-studied compounds are essential. We evaluate the proposed method (named SI-ADMM for Side Information-ADMM) on a comprehensive dataset compiled from multiple sources, including DrugBank, the Continuous Drug Combination Database (CDCDB), the Side Effect Resource (SIDER), and PubChem. Our experiments show that SI-ADMM maintains robust performance and achieves the best results comparing to other tensor factorization approaches, with or without auxiliary information, particularly in the new-drug prediction setting. The implementation of our method is publicly available at: https://github.com/Xiaoge-Zhang/SI-ADMM.

Drug-drug interactions (DDIs) can alter therapeutic efficacy or cause severe adverse reactions, posing major risks in polypharmacy. Accurate DDI prediction is therefore important for medication safety and early stage drug screening. Although molecular graph learning has been widely used for DDI prediction, many methods still rely on atom-centered propagation and global graph aggregation, which may obscure interaction-relevant local structural patterns. To address this limitation, we propose SSE-DDI, a molecular structure-driven framework centered on selective substructure encoding for relation-specific DDI prediction. SSE introduces a bond-level substructure-selective encoder in molecular line graphs, where chemical bonds serve as fundamental representation units and DDI-relevant local patterns are emphasized before drug-level representation learning. The encoder performs selective information routing between adjacent bond states and condenses interaction-relevant substructures into compact molecular representations. These representations are further refined by an edge-fusion graph transformer. To complement local structural modeling, SSE-DDI refines SMILES-derived Morgan-fingerprint similarity profiles to suppress noisy global structural signals without external biomedical annotations. Experiments on DrugBank and Twosides under transductive and inductive settings show that SSE-DDI outperforms representative baselines across multiple metrics. Ablation and visualization analyses verify the effectiveness of selective encoding and highlight DDI-relevant molecular substructures. The source code is available at https://github.com/qichi77/SSE-DDI.

The Office of Clinical Pharmacology (OCP) is a part of the Center for Drug Evaluation and Research (CDER) at the FDA. OCP probably comprises the largest group of scientists with expertise in, and involved with, the evaluation of all aspects of clinical pharmacology in new drug development, approval, and lifecycle management. Over a series of two parts, details are covered about the growth of this office over the last 40 years from one small Division of Biopharmaceutics (DB) in 1986, with staff of approximately 30 members carrying a large backlog of applications, to the current OCP of approximately 270 scientists with needed resources to carry out thousands of reviews and industry meetings in a timely manner. The growth of OCP is detailed with milestone developments in the areas of biopharmaceutics (BP) and clinical pharmacology (CP) science, which became a part of the review focus of new drug and biological applications over the last 40 years. Relevant regulations, guidances and some of the seminal articles/drug approvals highlighting critical contributions of the emerging components of OCP are also briefly covered. The parallel growth of CDER and enactment of new critical legislation requiring this growth are also summarized. In this first part, the organizational evolution of the OCP and the first segment of its scientific growth, i.e., bioanalysis, specific topics in the areas of bioavailability (BA) and bioequivalence (BE), biosimilars, and new drug applications (NDAs) based on Sect. 505(b)(2) of the Federal Food, Drug, and Cosmetic Act (FDC Act), are covered. Part II covers the areas of metabolism and drug‒drug interactions (DDIs), including DDI-induced cardiotoxicity of terfenadine, special populations, pediatric efficacy extrapolation, pharmacogenomics, quantitative clinical pharmacology (QCP), and conclusions.

The Office of Clinical Pharmacology (OCP) is a part of the Center for Drug Evaluation and Research (CDER) at the FDA. OCP is likely the largest group of scientists with expertise in the evaluation of all aspects of clinical pharmacology in new drug development, approval, and life cycle management. This article details the growth of this office over the last 40 years from one small Division of Biopharmaceutics (DB) in 1986, with staff of about 30 members carrying a large backlog of applications, to the current OCP of approximately 270 scientists with needed resources to carry out thousands of reviews and industry meetings in a timely manner. The growth of OCP is detailed with milestone developments in the areas of biopharmaceutics (BP) and CP science, which became a part of the review focus of new drug and biological applications over the last 40 years. Part I covered the organizational evolution of the OCP and the first segment of its scientific growth, i.e., bioanalysis, specific topics in the areas of bioavailability (BA) and bioequivalence (BE), biosimilars, and new drug applications (NDAs) on the basis of Sect. 505(b)(2) of the Federal Food, Drug, and Cosmetic Act (FDC Act). Part II covers the areas of metabolism and drug‒drug interactions (DDI), including DDI-induced fatal terfenadine cardiotoxicity, special populations, pediatric efficacy extrapolation, pharmacogenomics, quantitative clinical pharmacology (QCP), and conclusions.