Radiomics to predict immunotherapy-induced pneumonitis: proof of concept

Abstract

We present the first reported work that explores the potential of radiomics to predict patients who are at risk for developing immunotherapy-induced pneumonitis. Despite promising results with immunotherapies, immune-related adverse events (irAEs) are challenging. Although less common, pneumonitis is a potentially fatal irAE. Thus, early detection is critical for improving treatment outcomes; an urgent need to identify biomarkers that predict patients at risk for pneumonitis exists. Radiomics, an emerging field, is the automated extraction of high fidelity, high-dimensional imaging features from standard medical images and allows for comprehensive visualization and characterization of the tissue of interest and corresponding microenvironment. In this pilot study, we sought to determine whether radiomics has the potential to predict development of pneumonitis. We performed radiomic analyses using baseline chest computed tomography images of patients who did (N = 2) and did not (N = 30) develop immunotherapy-induced pneumonitis. We extracted 1860 radiomic features in each patient. Maximum relevance and minimum redundancy feature selection method, anomaly detection algorithm, and leave-one-out cross-validation identified radiomic features that were significantly different and predicted subsequent immunotherapy-induced pneumonitis (accuracy, 100% [p = 0.0033]). This study suggests that radiomic features can classify and predict those patients at baseline who will subsequently develop immunotherapy-induced pneumonitis, further enabling risk-stratification that will ultimately lead to better treatment outcomes.

Immunotherapy and the role of imaging

“…In contrast to the standard qualitative imaging characteristics of neoplasms that are routinely evaluated on radiologic studies, advanced analytics like texture analysis have the potential to uncover tumor features that cannot be identified by visual assessment and can provide detailed characterization of the specific tissue or lesion of interest and the associated microenvironment. A recent proof‐of‐concept study reported that radiomics predicted pneumonitis in patients who received immunotherapy with an accuracy of 100% . Finally, radiogenomics, the linkage between imaging phenotypes and tumor genomics, may assist in the development of more robust stratification and endpoint imaging biomarkers for molecular‐targeted clinical trials.…”

“…A recent proof-of-concept study reported that radiomics predicted pneumonitis in patients who received immunotherapy with an accuracy of 100%. 50 Finally, radiogenomics, the linkage between imaging phenotypes and tumor genomics, may assist in the development of more robust stratification and endpoint imaging biomarkers for molecular-targeted clinical trials.…”

Imaging of Adverse Events Related to Checkpoint Inhibitor Therapy

“…To this point, initial research in the field of radiomics-based artificial intelligence are encouraging. In particular, Colen et al were able to identify radiomics features at baseline CT discriminating patients who subsequently developed ICIs-related pneumonitis (100% accuracy; p = 0.0033) [80]. Despite the small number of patients included, the study shows the potential of radiomics to identify reliable predictors of irAEs and to stratify patients according to risk before therapy is initiated.…”

Immune-Related Adverse Events: Pneumonitis

“…Recent advances in imaging techniques have allowed thoracic CT images to be analyzed at the voxel level to detect textural features which are associated with disease or health [116]. A similar approach led to the development of a radiomicbased algorithm, which predicted the onset of pneumonitis from pretreatment thoracic CT scans of patients who underwent ICI therapies [117]. These findings need to be externally validated but highlight the power of imaging as a biomarker of disease risk.…”