A new method of customised, single patient trials are proving to be more effective than traditional, templated clinical treatment with the help of Artificial Intelligence (AI). As part of the initial trials initiated by the Singapore's Institute for Digital Medicine at the NUS Yong Loo Lin School of Medicine (NUS Medicine) have been observed to improve patient outcomes through dynamic and economical solutions.

The researchers at the NUS are focusing on delivering AI technologies which would integrate innovative advances in medicine and digital technology to further advance the revolutionary trial design protocols and targeted healthcare solutions that deliver faster, and more effective clinical interventions. 

The Institute (also known as WisDM) has gained fame for discovering effective drug mixes against COVID-19, and pioneered digital therapies to treat cognitive decline for post-brain radiation therapy and other oncology patients as well as other ageing and illness-related challenges. 

The NUS created an interactive digital platform called IDentif.AI (Optimising Infectious Disease Combination Therapy with Artificial Intelligence), that uses AI to compute the most effectual combination of drugs and doses which has helped the NUS researchers discover the most optimal drug combination regimen against COVID-19 comprises remdesivir, lopinavir and ritonavir.

NUS Medicine has nine new Translational Research Programmes (TRPs) under which WisDM is one of the new programmes. The programme aims to design a strong and coherent scientific base for the delivery of impactful and meaningful research outcomes for the School and Singapore’s health system. 

The IDentif.AI platform went through a pool of 12 drugs that were identified based on their status of being under evaluation in multiple clinical trials. The AI algorithm, unlike other AI algorithms, did not rely on using pre-existing data to train algorithms and predict treatment regimens. It created its own experiments that executed various permutations of drugs and doses to crowdsource the live virus to determine the combinations that optimise antiviral activity. The IDentif.AI investigates a relationship between drugs and doses to efficacy and safety using a quadratic algebraic algorithm. 

This allows the algorithm to run more than 530,000 possible combinations and identify only a few hundred experiments within 2 weeks. Through the platform’s ability to leverage unforeseen drug interactions within each combination, optimised recommendations of the drugs and corresponding doses were then suggested. These results have been studied by a various team of international collaborators on another strain of SARS-CoV-2, and two study protocols have been cleared to enable clinicalstudies should they be needed.

NUS has now created another AI-enabled platform like the IDentif.AI platform called CURATE.AI. The platform provides actionable Nof-1 (i.e. single patient) combination therapy for a patient's care; the trials are tailored for the patient based on their profiles, to develop drug therapies and interventions that achieve better outcomes for patients. 

The AI-platform CURATE.AI goes on to adjust the needed drug doses for the optimisation of combination therapy as patient responses are recorded. Assistant Professor Raghav Sundar from the Department of Medicine and WisDM at NUS Medicine, and Consultant with the Department of Haematology-Oncology at the National University Cancer Institute, Singapore (NCIS) said, “In the current clinical context, the doses of chemotherapy drugs given in combination can be further optimised. Drug dosing in cancer treatments is typically based on the degree of side effects experienced by the patient. With CURATE.AI, each patient’s recommended dose is calibrated using clinicaldata generated from their individual response to treatment. This may redefine how we care for patients and leverage digital medicine to treat cancers.”

Another study based on CURATE.AI was done to address ageing and illness-related challenges in cognitive and physical performance, such as diabetes, cognitive decline and Alzheimer’s disease. Using the data provided by patients such as training intensity, current performance level, degree of improvement, etc a personalised three-dimension (3D) profile was constructed to identify the performance of different patients under different approaches and intensities. CURATE.AI created distinct profiles to train and improve individual performance at digital therapy.

As the next step, the team now has funding to run clinical studies to address the cognitive depreciation in patients who have undergone radiation therapy to the brain. This study design will pair CURATE.AI with dynamically changing intensities of the software to provide diagnostic information regarding each patient’s responses. These responses will, in turn, be used to personalise treatment. It is envisioned that this study will eventually lead to a therapy that can be remotely deployed in patient’s homes.

Beyond clinical diagnosis, the team envisions to use the AI and digital solutions to build sustainable and cost-effective methods and treatments that can be used to treat all communities globally. Challenging the status quo early, their work centres on precision testing enabled by N-of-1 trial designs that have proven more effective than standardised clinical trial methods.