AI Provides Personalized Treatment Strategies for Esophageal Cancer

A research team from the University of Texas at Arlington (UTA) has developed an algorithm to provide optimal personalized treatment options for esophageal cancer patients, according to a January study published in the Journal of Mathematical Biology.

Esophageal cancer can be challenging to detect and treat effectively, as early signs of the disease – like heartburn and indigestion – are common, and there is no standard screening test for it. As a result, esophageal cancer often goes undetected until it is in its later stages.

At this point, there are fewer effective treatment options available to patients, and this lack of options contributes to the condition’s 21.7 percent five-year survival rate. Risk factors for esophageal cancer also vary, but include alcohol and tobacco use, gastric reflux, and older age.

These variations make treating patients difficult.

“Esophageal cancer is an unusually diverse disease with many heterogeneities and cancerous origins, which have been shown to be leading factors for drug resistance in the patients,” said Souvik Roy, PhD, an assistant professor of mathematics at UTA who led the research, in a news release. “That’s why timely and effective treatment strategies are key to combating it.”

To improve patient outcomes, the research team designed a modeling framework that is designed to account for esophageal cancer’s heterogeneities in order to help understand the disease’s progression and guide treatment.

“Imagine a software application where you can input real-time esophageal cancer data along with available drug information and possible patient interactions, then receive a recommended optimal course of treatments,” Roy stated. “It would make the clinician’s job much easier and, hopefully, improve survival rates for this dreaded disease.”

The UTA framework is made up of three steps: using a pharmacokinetic, mathematics-based model to determine the ideal pharmaceutical dose for each patient; conducting sensitivity analysis to identify factors that facilitate the cancer’s progression; and leveraging an “optimal control model” to select drug combinations and dosage profiles.

The framework was tested using synthetic data, and the model successfully recommended optimal treatment strategies with high accuracy.

“This work presents a new mathematical framework that depicts a new way of modeling cancer heterogeneities by using different interaction laws and assessing combination treatment strategies, all targeting a class of signaling pathways that are over-expressive in esophageal cancer patients,” said Roy. “We’re hopeful this will add a new tool to the armamentarium clinicians can use to fight this disease.”

Other research teams at UTA are also utilizing advanced computational models to track disease progression and enhance treatment.

Last week, researchers shared that they have developed a tool to predict how an individual’s Alzheimer's disease will progress over time.

Tracking the progression of Alzheimer's is key to helping patients and their caregivers prepare for the increasing support needs that come along with the disease over time. To help anticipate these care needs, the research team built a model to pinpoint where a patient is within the Alzheimer's disease-development spectrum.

The model can successfully project a patient’s clinical status and disease trajectory by coding them into five stages of disease development – normal cognition, significant memory concern (SMC), early mild cognitive impairment (EMCI), late mild cognitive impairment (LMCI), and Alzheimer’s disease.