Site Editor
Gregory J. Riely, MD, PhD
Gregory J. Riely, MD, PhD
Posted: Friday, November 8, 2024
Stephanie T. Schmidt, PhD, of The University of Texas MD Anderson Cancer Center, Houston, and colleagues developed an analysis and integration framework to identify whether neoadjuvant platinum-based chemotherapy, nivolumab, nivolumab plus chemotherapy, nivolumab plus ipilimumab, and/or a combination of all regimens impact the immunomicrobial landscape of non–small cell lung cancer (NSCLC) to determine potential biomarkers of response. These researchers are currently refining their approach to validate potential response biomarkers, and their data were presented during the 2024 Society for Immunotherapy of Cancer (SITC) Annual Meeting (Abstract 1242).
“We report a flexible, integrative approach that can handle data with missing values and maximize use of available multiplatform data to determine treatment-specific impacts and biomarkers,” mentioned the study authors. “We use it to perform the first integrated examination of the immunomicrobial effects of neoadjuvant chemotherapy/nivolumab/ipilimumab combinations in NSCLC.”
The investigators enrolled a total of 150 patients from the NEOSTAR trial—which included patients with resectable NSCLC who received nivolumab with or without ipilimumab, as well as those treated with neoadjuvant chemotherapy or chemoimmunotherapy before surgery; and the ICON trial, which recruited individuals with resectable NSCLC who received chemotherapy alone or no treatment before surgery. The pretreatment stool of these patients was analyzed via ribosomal RNA gene profiling, and surgically resected tumors/tissue underwent immune profiling.
More than 1,100 measurements were recorded, of which 65 were able to distinguish treatment groups spanning all flow-cytometry data of adjacent uninvolved lung, stool, and tumor microbial profiling. Measurements were prioritized by the SHAPley post hoc explainability metric and retained from classifier-identified readings.
To flag measurements for removal when patients were incorrectly classified, a leave-one-out approach was implemented. According to the data, early fusion appeared to better represent treatment-specific outcomes than late fusion, supporting its potential to handle interactions between data sets and thereby highlighting the prognostic value of integrated analysis in NSCLC.
Disclosure: Disclosure information was not provided.
JNCCN Spotlights
