Rare diseases individually impact fewer than 1 in 2,000 people. Nonetheless, there are over 7,000 known types, which together make a significant global impact. In the Asia-Pacific region, around 258 million people are affected by rare diseases, the highest in the world, with over 45 million cases in Southeast Asia alone. This substantial figure underscores major challenges in treatment due to the diverse needs of this patient group, leading to notable healthcare inequalities and difficulties in recruiting for clinical trials. Furthermore, within the limited patient population, each individual’s condition varies over time, emphasizing the urgent need for accessible and tailored treatments for these patients, while also pointing out the difficulties in developing therapies for rare diseases.
To meet the demand for effective treatment for rare diseases without relying on extensive population data, researchers from the Institute for Digital Medicine (WisDM) at the Yong Loo Lin School of Medicine, National University of Singapore (NUS Medicine) utilized a small data set from a single patient afflicted with a rare disease to determine his treatment, yielding promising outcomes. Under the leadership of Professor Dean Ho, Director of WisDM at NUS Medicine, the team embarked on a clinical trial for a patient diagnosed with Waldenström macroglobulinemia, a rare blood condition that affects approximately three individuals per million each year. They employed an AI-driven platform known as CURATE.AI. Unlike traditional AI systems that require large data sets, CURATE.AI adapts treatment dosages dynamically based on the individual patient’s responses. Since the trial commenced in October 2021, there has been a notable improvement in the patient’s red blood cell counts, allowing him to avoid blood transfusions. Importantly, the patient experienced minimal serious side effects from the treatment, and hospital visits were significantly reduced.
During the trial, the research team collaborated with doctors from the National University Cancer Institute, Singapore (NCIS) to establish drug dosages for the patient, guided by the CURATE.AI platform. These dosages were determined based on the patient’s own responses, making this treatment strategy groundbreaking. In comparison to the standard care regimen, the doses recommended during the trial were lower and well-tolerated by the patient, achieving sustainable disease control. Consequently, the patient managed to save about USD 8,000 (approximately SGD 10,500) on medication costs over the first two years of treatment.
The ongoing trial utilizing CURATE.AI’s recommendations is now recruiting new eligible patients. The findings from the initial two years of the trial have been published in the journal NPJ Digital Medicine, part of the Nature Portfolio.
Prof Ho commented, “Every patient is unique, and even the same patient may change over time. It’s crucial that treatments adapt to the patient’s needs. Our research demonstrates the potential of using small data for treating very rare diseases, filling in the gaps left by traditional big data approaches, particularly when large-scale trials are impractical due to limited patient numbers. CURATE.AI’s method, which customizes treatment based on small datasets, presents a viable solution for the urgent and complex demand for personalized approaches to rare diseases.” Prof Ho also serves as the Head of the Department of Biomedical Engineering at the NUS College of Design and Engineering and is the Director of the NUS N.1 Institute for Health.
Dr Sanjay de Mel, Senior Consultant in the Division of Hematology at NCIS and the trial’s clinical lead, remarked, “Achieving an effective treatment response while minimizing side effects is critical for patients with Waldenström macroglobulinemia. Because patients can vary widely in their reactions to treatments and the side effects they may face, a personalized approach to dosing is essential to accommodate this individual variability.”
About CURATE.AI
The CURATE.AI platform employs an AI-based technology that offers actionable, individualized therapy throughout the patient’s care. It customizes trials based on each person’s unique data to create drug therapies and interventions that yield better results for patients. By constantly adjusting drug doses based on patient responses, CURATE.AI enhances treatment optimization.
Traditionally, AI models analyze large datasets from numerous patients to train and test an algorithm, which is then validated with data from different patient groups. This approach can lead to advances in AI but often fails to translate into real-world applications, resulting in predictions that lack practical implementation.
In contrast, CURATE.AI’s methodology is pioneering, as it utilizes small data to inform real-time treatment decisions. Within WisDM’s clinical trials, clinicians frequently approve and adjust medication dosages based on insights from this tailored data platform. Because CURATE.AI relies on single patient data, it circumvents the requirement for large patient samples, which often lead to non-personalized trial strategies that are impractical for rare diseases.
In a previous pilot study conducted alongside a hospital in the US, a patient suffering from advanced prostate cancer received a recommendation for a 50% dose reduction of an investigational drug, enhancing effectiveness. As a result, the patient was able to return to an active lifestyle since the lower dosage was better tolerated. In another case in Singapore, a patient with advanced cancer who had a reduced dose of nab-paclitaxel experienced a decrease in lung tumor size and a halt in cancer progression, allowing for prolonged treatment compared to the typical duration for patients receiving this drug.
