10173060A Full-Stack AI Platform for Predicting Efficacy, Toxicity, and Cost–Effectiveness of Complex Drug CocktailsClosedFast Start ResponseInnovate UKCombining multiple drugs into a single treatment-known as a drug cocktail-can be significantly more effective than administering single agents alone. This approach is widely used in cancer and antibiotic therapies, as attacking diseases from multiple angles helps prevent the development of drug resistance. However, identifying the optimal combination and dosage is highly complex. Experimentally testing every possible combination is unfeasible due to the sheer number of potential permutations. Prosemble, a spin-out from King's College London, has developed an innovative AI platform to address this challenge, complementing its award-winning protein nanoparticle technology designed to deliver drugs efficiently. The company has created, coded, and trained a proprietary machine learning system that understands drug interactions in three-dimensional space rather than as isolated data points. This enables Prosemble to predict both the efficacy and safety of drug combinations well before costly clinical trials commence. The AI platform has already successfully predicted clinical outcomes for 21 drugs developed by major pharmaceutical companies, including Eli Lilly and AstraZeneca. Additionally, Prosemble's system incorporates advanced toxicity assessment using AI models that integrate multiple data types, thereby helping to avoid harmful side effects while prioritising the most effective treatments. A distinctive feature of Prosemble's platform is its financial risk analysis, which evaluates the cost implications of prediction errors, such as: 1\. Correctly identifying toxic drugs (no cost) 2\. Advancing safe drugs (net gain) 3\. Incorrectly discarding safe drugs (opportunity cost) 4\. Advancing toxic drugs (high-risk, costly failures) This balanced approach ensures safety and efficacy while optimising development costs. Prosemble is currently collaborating with Johnson & Johnson to pilot this platform for lung cancer treatment. Lung cancer remains one of the deadliest cancers worldwide, causing approximately 1.8M deaths annually, including around 40K in the UK. The 10-year survival rate is just 10%. Treatment costs are substantial; chemotherapy in the UK averages £10K per cycle, with most patients requiring four to six cycles---total treatment costs of around £60K. Managing side effects, which affect up to 85% of patients, can add further £60K. Common side effects include fatigue, hair loss, mucositis, neuropathy, organ toxicity, and vomiting. This partnership aims to develop safer, more effective drug combinations more rapidly and affordably. Beyond oncology, Prosemble's adaptable platform has potential applications in neurodegenerative diseases such as Alzheimer's and consumer health. By combining advanced AI, innovative drug delivery, and cost-sensitive risk analysis, Prosemble offers a smarter, safer path to developing complex drug treatments of the future.