Short-term intra- and peri-tumoral spatiotemporal CT radiomics for predicting major pathological response to neoadjuvant chemoimmunotherapy in non-small cell lung cancer

Abstract

Purpose: This study aims to develop a short-term spatiotemporal CT radiomics model to predict the major pathological response (MPR) to neoadjuvant chemoimmunotherapy (NCI) in NSCLC by decoding the intra- and peri-tumoral imaging phenotypes.

Methods: A total of 352 patients undergoing curative surgery following NCI for NSCLC were enrolled from two centers, forming a training cohort (n = 186), an internal validation cohort (n = 80), and an external validation cohort (n = 86). Intra- and peri-tumoral CT radiomics features were computed to capture imaging phenotypes of the tumor microenvironment. Delta radiomics features were also calculated by quantifying changes in each radiomics feature. A support vector machine classifier was utilized to develop the short-term spatiotemporal model by analyzing changes in radiomics features.

Results: The multi-timepoint short-term spatiotemporal model, incorporating pre-treatment, post-treatment and delta radiomic features, achieved AUC values of 0.84, 0.77, and 0.75 in the training, internal validation, and external validation cohorts, respectively. These results significantly outperformed the RECIST model and pre-treatment model, with p-values < 0.05 indicating statistical significance.

Conclusion: This study demonstrates that short-term temporal analysis of intra- and peri-tumoral CT radiomics is a promising approach for predicting MPR to NCI in NSCLC. These findings underscore the potential of radiomics as a non-invasive tool for assessing treatment response and guiding personalized therapy in NSCLC patients.

Key points: Question Neoadjuvant chemoimmunotherapy has improved in major pathological response rate for non-small cell lung cancer (NSCLC), but it is unclear which patients will benefit most. Findings The multi-timepoint short-term spatiotemporal model based on CT pictures demonstrates high predictive performance for assessing major pathological response following neoadjuvant chemoimmunotherapy in NSCLC. Clinical relevance Short-term intra- and peri-tumoral CT radiomics is a promising approach for predicting major pathological response to neoadjuvant chemoimmunotherapy in NSCLC. These findings underscore the potential of radiomics as a non-invasive tool for assessing treatment response in NSCLC.

Keywords: Immune checkpoint inhibitors; Neoadjuvant therapy; Non-small cell lung carcinoma; Radiomics; Response prediction.