Research Appraisals
Evidence-based critical appraisals of the latest medical research, systematically evaluated using Oxford CEBM methodology.
Showing 2 appraisals
mAbs
Optimization strategies for monoclonal antibody production: advances in simulation and artificial intelligence in bioprocessing
The rapid growth of monoclonal antibody (mAb) therapies has increased the need for efficient, scalable, and affordable manufacturing processes. However, mAb production remains complex because of nonlinear upstream cell culture behavior, expensive downstream purification, especially Protein-A chromatography, and plant-level bottlenecks that can increase cost, cycle time, and manuacturing uncertainty. This review examines recent developments in mAb manufacturing with focus on process simulation, mathematical optimization, and artificial intelligence/machine learning (AI/ML) across upstream processing (USP), downstream processing (DSP), and integrated plant-level operation. In USP, media optimization, dynamic feeding, high-density cultures, and continuous perfusion bioreactors are discussed in relation to productivity and critical quality attributes (CQAs). In DSP, alternative and intensified purification strategies are reviewed with a focus on recovery, impurity clearance, scalability, cost, and technology maturity. AI/ML applications are also discussed from early-stage development and cell-line screening to upstream control, CQA prediction, chromatography optimization, and downstream decision support. Despite these advancements, challenges such as data heterogeneity, limited standardized datasets, model transferability, and regulatory constraints remain important barriers to implementation. Overall, this review uniquely connects simulation and AI/ML approaches to practical optimization across the full mAb manufacturing workflow, including design, scheduling, debottlenecking, purification, monitoring, and quality prediction. The combination of process simulation, continuous bioprocessing, and AI/ML-based decision support may enable more flexible, reliable, and cost-effective mAb manufacturing. However, these benefits depend on validation through robust models, process-specific case studies, and technoeconomic analysis.
25 July 2026
Read appraisal →Journal of medical economics
From options to decisions: an innovative model for treatment sequencing in relapsing-remitting multiple sclerosis
AIMS: With the expanding range of disease-modifying therapies (DMTs) for relapsing-remitting multiple sclerosis (RRMS), clinicians face increasing complexity in defining optimal treatment sequences and timing of therapy switches. In this study, we adapted an earlier computer-assisted model to optimize therapeutic decisions and identify preferred treatment pathways, based on expert opinion and observations from clinical practice in Spain. MATERIALS AND METHODS: The original model was updated to integrate magnetic resonance imaging (MRI) activity data and define reaching an Expanded Disability Status Scale (EDSS) score of 3 - indicating moderate disability without ambulation impairment - as a criterion for switching treatment. Matrices were designed to model options when switching DMTs, triggered by either a lack of effectiveness or safety concerns. Change in DMT was based on a set of composite criteria (encompassing contributions from relapses, disability worsening, MRI activity, costs, and quality of life) according to the disease activity level. A maximum of three DMTs could be administered within the 8-year time horizon evaluated. RESULTS: The revised model identified high efficacy treatment, in particular cladribine tablets as the preferred initial DMT for patients with RRMS with mild or moderate disease activity. Of these patients with mild and moderate disease activity, most were switched to ofatumumab (79.4%) and ocrelizumab (75.9%), respectively, when disease progression occurred. Patients with high disease activity mostly received natalizumab if they were John Cunningham virus (JCV)-negative, or ocrelizumab if they were JCV-positive. LIMITATIONS: The model is informed by a Spain-based healthcare expert panel and has been evaluated using a simulated cohort of 10,000 patients with RRMS. CONCLUSION: This computational model helps to inform clinicians towards making optimal treatment decisions for RRMS, and identified high-efficacy DMTs as the preferred model-based option for all levels of disease activity, supporting early and effective control of disease activity in real-world clinical practice.
25 June 2026
Read appraisal →