AI Helps Predict Treatment Response in Melanoma
A new study reveals that artificial intelligence can identify which melanoma patients will respond better to treatment. Research led by Xuefeng Wang, PhD, and Ahmad Tarhini, MD, PhD, at Moffitt Cancer Center shows that tertiary lymphoid structures, a cluster of immune cells that form in response to chronic inflammation such as cancer, can be used to predict patient outcomes.
“Tertiary lymphoid structures in tumors have emerged as novel predictive and prognostic biomarkers for immunotherapy,” Tarhini said. “Tertiary lymphoid structures in tumors suggest an immune system response that could correlate with improved patient outcomes.”
Moffitt’s work centers on using these tertiary lymphoid structures as biomarkers to enhance treatment guidance and outcomes for patients with cutaneous melanoma.
A study presented at the Society for Immunotherapy of Cancer annual meeting tackles one of the key challenges in tertiary lymphoid structure research: a lack of standardized methods for identifying them. By employing a state-of-the-art deep learning tool for automatic tertiary lymphoid structure detection, Wang’s team has found a way to analyze and quantify tertiary lymphoid structures in digital pathology slides with greater accuracy.
“Recent advancements in whole-slide image digitization and AI tools are making it feasible to predict the presence and characteristics of tertiary lymphoid structures directly from tissue slides,” Wang said.
For the phase 3 trial, the team used AI to analyze tertiary lymphoid structure density in samples from 442 patients with resected cutaneous melanoma who were treated with one of two immunotherapies: ipilimumab or interferon. The tool was able to characterize tertiary lymphoid structures and identify associations between their makeup and overall survival rates.
“We found a significant correlation between higher normalized tertiary lymphoid structure density, a measure of the number of tertiary lymphoid structures relative to the tissue area and improved survival,” Wang said.
“This suggests that tertiary lymphoid structures may indicate a stronger immune response against the tumor, which can translate into better outcomes for patients,” Tarhini added.
For clinicians, these results could mean the integration of tertiary lymphoid structure measurements into treatment planning, allowing them to offer more personalized treatment.
The team is now working on improving the accuracy of the tool and exploring its use in other cancer types.
During #SITC24, Drs. Xuefeng Wang, PhD (@xuefwang) and Ahmad Tarhini, MD, PhD (@ATarhiniMDPhD) joined me to discuss their research on using #AI to predict treatment response in #melanoma.
— W. Gregory Sawyer (@ProfGregSawyer) November 8, 2024
Read more: https://t.co/6stu2xInX0#MoffittSITC24 @sitcancer @MoffittNews pic.twitter.com/uE2qRLkIla
