A team from the University of Minnesota, inclusive of Masonic Cancer Center researcher and Division of Computational Health Sciences chieftain Rui Zhang, is utilizing artificial intelligence (AI) with a sizeable $1.2 million grant from the National Cancer Institute. Zhang's team, working closely with Ju Sun, an assistant professor of computer science, and Ying Cui from the University of California - Berkeley, is orchestrating a study that focuses on balancing imbalanced medical big data, which is frequently skewed by uneven patient health outcomes, as reported by the University's cancer research news portal.
The challenge ahead is stiff, not just for fleshing out the AI-powered tools that are often stumped by limited and variable patient data but also the fact that these tools are yet to be created, needing the team to rejig existing technologies that work best when they can draw on expansive data sets. Sun and Zhang's endeavor will "complement, inform and augment doctors," aiming to weave these AI models into a healthcare setting that can fine-tune itself to each unique patient case, according to CBS Minnesota's coverage of the ambitious project.
Zhang emphasized the urgent need to crack the conundrum of biased results due to unbalanced data pools, asserting that "The primary goal of the study is to solve an inevitable data problem with medical big data," during the process. This initiative could potentially carry implications that extend far beyond breast cancer, promising to enhance the predictive accuracy of AI across various diseases.
At the heart of this endeavor lies the objective of enabling physicians globally to foresee post-treatment cardiac problems in breast cancer survivors, touching upon the prospects of "using different types of medical records from patients across clinics and health systems," to "derive novel AI solutions" that will compensate for data imbalance and raise the predictive accuracy bar for medical practitioners," Ju Sun told CBS Minnesota. The researchers are currently arming themselves with electronic health records and developing novel optimization methods, poised to test their AI solutions on real-world data in the ensuing phase of the project.
