Given heterogeneous progression of Duchenne muscular dystrophy (DMD), accurately characterizing individual disease trajectories is critical for effective clinical trials and therapeutic development. In this study, we applied machine learning (ML) approaches leveraging disease progression models and baseline magnetic resonance (MR)‐derived radiomic features to quantify individual‐level DMD progression trajectories and support quantitative characterization of disease progression. Radiomic features were extracted from segmented soleus (SOL) and vastus lateralis (VL) fat‐fraction maps, generating 810 features per individual, with additional functional and demographic measures incorporated as covariates following feature selection. Multiple ML regressors were trained and validated to estimate MR‐spectroscopy fat‐fraction trajectory descriptive parameters: DPT 50 (age at half‐maximal increase), DP max (extent), γ (rate), S 0 (extrapolated measure at age 0). Key radiomic features were identified through feature importance analyses, followed by downstream analyses, including unsupervised clustering. The tree‐based XGBoost regression model demonstrated best future longitudinal fat‐fraction trajectories (SOL: mean absolute error (MAE) = 0.04, R 2 = 0.94; VL: MAE = 0.06, R 2 = 0.95). Subsequent downstream analyses highlighted potential key biomarkers clinically associated with DMD progression. Clustering analysis revealed clinically meaningful slow‐ and fast‐progression subgroups with interpretable decision criteria (SOL: DPT 50 = 12.21 vs. 14.83; γ = 8.07 vs. 7.54, VL: 10.37 vs. 14.21; γ = 7.72 vs. 7.09) that can inform clinical trial designs and therapeutic effects within homogenous groups. Individual‐level feature importance interpretation further shows the possibility of tailoring personalized intervention strategies. These findings highlight the potential of ML‐based trajectory modeling approaches leveraging baseline MRI radiomics for advancing AI‐assisted clinical trial development and precision medicine, charting a new direction for MRI‐guided disease modeling in neuromuscular disorders.
Population pharmacokinetic (PopPK) models, the gold standard for drug concentration prediction, are relatively complex and time‐consuming; machine learning (ML) excels with high‐dimensional data but remains exploratory. This study systematically evaluated seven ML algorithms—Extreme Gradient Boosting (XGBoost), Long Short‐Term Memory (LSTM), Bidirectional Long Short‐Term Memory (BiLSTM), Gated Recurrent Unit with Decay (GRUD), Phased Long Short‐Term Memory (PLSTM), Neural Controlled Differential Equations (NCDE), and Neural Ordinary Differential Equation (NODE)—and their heterogeneous XGBoost‐based ensembles against PopPK across three data settings: dense phase I trial data; simulated dense data: linear absorption and elimination (LL), Michaelis–Menten absorption with linear elimination (MMA), linear absorption with Michaelis–Menten elimination (MME); and sparse therapeutic drug‐monitoring (TDM) data. On two dense trial data, BiLSTM showed low systematic bias (F30 = 48.44%, MdPE = −6.86%; F30 = 50.62%, MdPE = 1.08%), whereas NCDE (F30 = 50.00%, MdPE = −19.39%) and NODE (F30 = 62.96%, MdPE = 11.77%) achieved the highest F30 values, respectively, compared with PopPK (F30 = 37.50%, MdPE = 24.73%; F30 = 61.73%, MdPE = 14.88%). BiLSTM showed robust extrapolation across LL, MMA, and MME datasets (F30 ≥ 75%, |MdPE| ≤ 10%). XGBoost lacked extrapolation ability, limiting ensemble gains. On sparse single‐point prior data, XGBoost (F30 = 23.61%, MdPE = −8.42%) was most robust, while PopPK failed to converge. On multi‐point sparse prior data, XGBoost (F30 = 26.84%, MdPE = −5.88%) outperformed PopPK (F30 = 23.63%, MdPE = −13.88%). In single‐point sparse posterior prediction, PopPK (F30 = 29.71%, MdPE = 2.32%) performed best, while several ensembles improved over their neural network components. Overall, BiLSTM showed strong performance in dense data, XGBoost performed best in the prior sparse scenarios, and PopPK retained the advantage after prior concentration information was incorporated. ML approaches achieved similar accuracy to PopPK with unique advantages, representing promising tools for drug concentration prediction.
Influenza viruses cause substantial global morbidity and mortality. Molnupiravir, an orally administered prodrug of β‐D‐N4‐hydroxycytidine, has broad activity against RNA viruses, including influenza. A phase 2a, randomized, two‐part human challenge trial in healthy adults (NCT05818124) was conducted to evaluate molnupiravir for influenza infection. In Part 1, 20 participants were inoculated with 10 6.26 TCID 50 of A/France/759/21 (H1N1) to establish infectivity, pathogenicity, and safety. After assessment of Part 1, a 10‐fold lower virus dose was selected for Part 2 to delay time to peak viral load (PVL). In Part 2, 141 participants were randomized 1:1:1:1 to receive molnupiravir (5 days, 800 mg twice daily) postexposure prophylaxis (PEP; initiated 12 hours after inoculation), molnupiravir treatment (initiated 48 hours after inoculation), oseltamivir treatment (5 days, 75 mg twice daily, initiated 48 hours after inoculation), or placebo. Primary endpoints, assessed by quantitative viral culture, were PVL in all participants for PEP and viral load area under the curve (VL‐AUC) in participants with confirmed infection for treatment. Lowering the viral inoculum delayed PVL by ~12 hours, although infection rates in the Part 2 placebo arm (40%) were lower than in Part 1 (75%). Molnupiravir PEP did not significantly reduce PVL compared with placebo but was associated with lower viral burden (difference in mean log 10 VL‐AUC –2.21 day*log 10 TCID 50 /mL; P = 0.03). No significant antiviral effect was observed with molnupiravir or oseltamivir treatment compared with placebo. Molnupiravir was well tolerated, with an adverse event profile similar to placebo. Although primary endpoints were not met, early administration of molnupiravir showed antiviral activity.
Green tea catechins modulate the activity of key drug pharmacokinetic determinants and consequently, concomitant consumption of green tea can significantly alter systemic drug exposure. This study aimed to investigate the influence of concomitant green tea catechin consumption on the pharmacokinetics of pazopanib, a tyrosine kinase inhibitor used in the treatment of advanced renal cell carcinoma and advanced soft tissue sarcoma. An open‐label, single‐dose, fixed‐sequence pharmacokinetic study in 15 healthy participants was conducted. Participants were administered a 200 mg pazopanib tablet on Day 1 (Control Period) and Day 15, and an 833 mg green tea extract tablet once daily on Days 10–18 (Green Tea Extract Period). Fourteen healthy participants completed the study. Concomitant green tea extract consumption significantly reduced pazopanib oral systemic exposure. The geometric mean ratios (Green Tea Extract/Control) for pazopanib maximum plasma concentration ( C max ), area under the plasma concentration–time curve (AUC) from time 0 to last quantifiable timepoint (AUC 0‐t ), AUC extrapolated to infinity (AUC 0‐∞ ) and half‐life ( t 1/2 ) were 0.44 (90% CI: 0.30–0.64, n = 14), 0.53 (90% CI: 0.41–0.68, n = 13), 0.57 (90% CI: 0.44–0.74, n = 13) and 1.12 (90% CI: 1.05–1.20, n = 13), respectively. Median pazopanib time of C max ( T max ) was slightly shorter in the Control Period (4.0 h) than the Green Tea Extract Period (4.5 h, P < 0.01, n = 14). Given the established concentration–response relationships for pazopanib anticancer effects, decreases in pazopanib systemic exposure with green tea catechin consumption may be clinically important.
Multiple drug hypersensitivity syndrome (MDHS) is defined as confirmed hypersensitivity to two or more chemically and pharmacologically unrelated drug entities. Few studies have characterized MDHS with systematic diagnostic confirmation. We aimed to determine the frequency, culprit drug patterns, and diagnostic findings of confirmed MDHS. We retrospectively reviewed patients evaluated for suspected drug hypersensitivity at a Korean tertiary allergy center from March 2017 to April 2026. MDHS required confirmed hypersensitivity to at least two unrelated drug entities by skin testing, serum‐specific IgE, or provocation testing; cross‐reactive reactions were not counted separately. Among 6684 evaluated patients, 3007 underwent diagnostic testing, and 1380 were tested for multiple suspected drugs. MDHS was confirmed in 94 patients (1.4%, 3.1%, and 6.8%, respectively). Anaphylaxis was the most common manifestation (38.3%). Most patients (84.0%) reacted to two entities, whereas 12.8% and 3.2% reacted to three and four entities, respectively. Leading culprits were β‐lactams ( n = 34), nonsteroidal anti‐inflammatory drugs (NSAIDs)/acetaminophen ( n = 34), antituberculosis agents ( n = 27), and histamine‐2 receptor antagonists ( n = 19). Recurrent combinations involved first‐line antituberculosis agents and NSAIDs plus ranitidine. Of 114 entities confirmed by drug provocation, 38 had negative preceding skin tests. Delayed‐reading intradermal testing was more frequently positive in drug reaction with eosinophilia and systemic symptoms (DRESS) than in maculopapular exanthema (MPE) (53.3% vs. 0.0%, P = 0.008). Confirmed MDHS was uncommon despite suspected hypersensitivity to multiple drugs. The clinical presentation resembled European cohorts, whereas the culprit spectrum reflected regional prescribing patterns. Systematically evaluating each suspected entity may preserve future treatment options.
