Cancer researchers at the University of Tennessee Health Science Centre have published three groundbreaking studies demonstrating how advanced artificial intelligence and population health analytics can identify and address barriers to quality cancer care.

Funded by a Practice Transformation and Enhancement grant from the Tennessee Department of Health, the work documents key technical steps leading to the opening of a currently enrolling clinical trial in UT Health Science Centre’s Department of Radiation Oncology.

The trial is testing the real-world impact of the ENRICH (Equitable Needs-based Radiotherapy Interruption Care for Health) platform.

ENRICH is the first personalised social support platform specifically designed to improve cancer radiotherapy completion rates among rural and underserved populations anywhere in the United States.

ENRICH integrates advanced AI-driven analytics with community-based human support to reduce treatment interruptions, a major driver of cancer mortality.

David L. Schwartz, MD, professor and chair of the Department of Radiation Oncology, serves as co-principal investigator alongside Arash Shaban-Nejad, PhD, MPH, associate professor in the Department of Paediatrics, Centre for Biomedical Informatics, and director of Population and Precision Health, and Altha Stewart, MD, professor of psychiatry and the senior associate dean for Community Health Engagement in the College of Medicine.

The collaborative team brings together expertise in clinical oncology, biomedical informatics, artificial intelligence, population health, cancer navigation, and community-engaged research to address a critical challenge: patient access to cancer radiotherapy (RT).

The team has published three manuscripts demonstrating the ability to leverage advanced analytics for deploying ENRICH.

The three papers address:

• An AI-enhanced EHR (electronic health record) and population-health analytics framework for identifying patients at risk of radiation therapy interruptions;
• A comparative evaluation of LLMs (large language models) and BioBERT (a biomedical language representation model) for cancer-diagnosis classification;
• And an explainable machine-learning model predicting county-level lung cancer mortality and its geographic and socioeconomic determinants.

Taken together, these papers confirm that automated population health data collection, AI-enhanced data preparation, explainable machine learning, and geospatial analysis can help providers individually identify and support at-risk Tennessee cancer patients through their treatment.

Cancer mortality in Tennessee is currently the fifth worst in the U.S. and is particularly high in the state’s rural areas.

Most cancer patients require radiotherapy to cure or control disease.

Cure typically requires one to two months of daily treatment at specialised facilities and can produce severe toxicity and patient hardship.

Unplanned radiotherapy interruptions as brief as two days are closely associated with shorter survival.

Dr. Schwartz’s group characterised geographic and demographic trends in unplanned radiotherapy interruption rates in both western and eastern Tennessee.

Nearly one-quarter of Tennessee patients undergoing radiotherapy miss two or more scheduled treatment days, with approximately 10% missing a week or more.

These interruptions significantly reduce treatment effectiveness and survival rates.

Social support interventions specific to the needs of radiotherapy patients to prevent such lapses remain completely undeveloped.

That’s where ENRICH steps in.

ENRICH is an individualised, community-anchored support intervention designed to enhance radiotherapy access.

ENRICH leverages formally trained Community Health Support Specialists to directly address modifiable social barriers to radiotherapy, including social isolation, health information challenges, and/or lack of personal resources.

This is complemented by automated data analytics to leverage clinical, geospatial, and population data to identify patients at highest risk for interruption.

The programme is currently funded by the Tennessee Department of Health Practice Transformation and Enhancement Health Resiliency Grant and is recruiting for a 200-patient clinical trial in Memphis.

“Our research demonstrates that we can accurately identify which patients are at highest risk for treatment interruption before they even begin therapy,” Dr. Schwartz said.

“This allows us to proactively deploy support resources where they’re most needed, potentially improving both treatment completion and cancer outcomes while simultaneously addressing health equity.”

Dr. Shaban-Nejad, who leads the AI and data analytics strategy for ENRICH, added, “The integration of advanced AI with community-based patient support represents a new paradigm in precision medicine, one that recognises that optimal cancer care requires addressing not just the biology of the disease, but also the social and logistical challenges patients face.”

With support from the College of Medicine and the Tennessee Department of Health, the research team is actively working to scale the ENRICH programme beyond Memphis and Knoxville.

The goal is to implement this evidence-based intervention with community partners throughout Tennessee.

“The Tennessee Department of Health’s investment in this work reflects its forward-thinking leadership to propel health care transformation,” Dr. Schwartz said.

“By funding both technical innovation and downstream clinical implementation, they’re enabling us to develop solutions that can serve all folks throughout our state, from Memphis to Knoxville and all rural communities in between.”

The potential impact of this research extends well beyond radiation oncology.

The methodologies developed by the team establish a practical framework for addressing stubborn health care quality and access challenges facing any patient living in any community across the state.

Source: University of Tennessee Health Science Center