AI Platform Recommends Lower, More Personalized Chemotherapy Doses
Researchers have developed an artificial intelligence solution than can recommend personalized chemotherapy doses for cancer patients, lowering prescribed dosages by 20 percent on average.
Source: Getty Images
By - Researchers from the National University of Singapore (NUS) have shown that that an artificial intelligence (AI) platform designed to recommend optimal chemotherapy doses from cancer patients achieved promising results and lowered prescribed dosages by an average of 20 percent, according to findings presented at ASCO 2022.
Chemotherapy is a common part of many cancer patients’ treatment plans, but concerns about side effects and toxicity is a concern for many patients and their families. Chemotherapy dosages differ by patient and cancer type, among other factors, but higher doses of chemotherapy can result in more severe side effects, such as low blood cell counts, hair loss, liver problems, and vomiting. In cases where higher radiation doses are needed, other measures must be taken to protect healthy cells and manage side effects.
For this reason, standard or low-dose chemotherapy is usually prescribed, if possible. These still have the potential for side effects, but they are often less severe. Some research has shown that low-dose chemotherapy achieves similar results when compared to standard-dose, while also resulting in less toxicity and side effects. More research in this area is needed, but the potential for lower side effects without sacrificing treatment effectiveness has created interest in optimal or personalized chemotherapy dosing.
In the NUS study, researchers developed and tested CURATE.AI, a solution designed to take clinical data, including medications, dosages, and the presence of cancer biomarkers, to create an individualized digital profile that it can use to make personalized dosage recommendations over the course of chemotherapy treatment.
“Chemotherapy treatment is often given at fixed doses, based on certain patient parameters. However, these toxicity-guided doses may not result in optimal response to treatment,” said Professor Dean Ho, head of the Department of Biomedical Engineering at NUS and co-author of the study, in the press release. “Using CURATE.AI, which is efficacy-driven, we hope to help doctors to quickly identify the optimal doses that are customized for each patient at different stages of the treatment cycle. The ultimate goal is to improve patient and treatment outcomes.”
In a pilot clinical trial, CURATE.AI was used as a clinical decision support tool during chemotherapy treatment for 10 patients with advanced solid tumors and predominantly metastatic colorectal cancers. Over the course of treatments, participating clinicians accepted roughly 96.7 percent of the dosages recommended by CURATE.AI. Some of the recommended dosages patients received were approximately 20 percent lower than standard dosages on average.
Following this study, the researchers plan to undertake a larger, randomized clinical trial to validate their findings. They will also expand their research to examine CURATE.AI’s performance when evaluating patients with other types of cancer and additional comorbidities. Currently, the team is working to launch a similar clinical trial to optimize personalized immunotherapy dosing for cancer patients.
