Predicting recurrence risks in lung cancer patients using multimodal radiomics and random survival forests

Abstract

.PurposeWe developed a model integrating multimodal quantitative imaging features from tumor and nontumor regions, qualitative features, and clinical data to improve the risk stratification of patients with resectable non-small cell lung cancer (NSCLC).ApproachWe retrospectively analyzed 135 patients [mean age, 69 years (43 to 87, range); 100 male patients and 35 female patients] with NSCLC who underwent upfront surgical resection between 2008 and 2012. The tumor and peritumoral regions on both preoperative CT and FDG PET-CT and the vertebral bodies L3 to L5 on FDG PET were segmented to assess the tumor and bone marrow uptake, respectively. Radiomic features were extracted and combined with clinical and CT qualitative features. A random survival forest model was developed using the top-performing features to predict the time to recurrence/progression in the training cohort (n = 101), validated in the testing cohort (n = 34) using the concordance, and compared with a stage-only model. Patients were stratified into high- and low-risks of recurrence/progression using Kaplan–Meier analysis.ResultsThe model, consisting of stage, three wavelet texture features, and three wavelet first-order features, achieved a concordance of 0.78 and 0.76 in the training and testing cohorts, respectively, significantly outperforming the baseline stage-only model results of 0.67 (p < 0.005) and 0.60 (p = 0.008), respectively. Patients at high- and low-risks of recurrence/progression were significantly stratified in both the training (p < 0.005) and the testing (p = 0.03) cohorts.ConclusionsOur radiomic model, consisting of stage and tumor, peritumoral, and bone marrow features from CT and FDG PET-CT significantly stratified patients into low- and high-risk of recurrence/progression.

Evaluating clinical and radiomic features for predicting lung cancer recurrence pre- and post-tumor resection

“…By incorporating radiomic features with clinically relevant covariates, we developed a post-tumor resection recurrence prediction model that has exhibited improved performance compared to traditional clinical prognostic markers such as TNM staging (p<0.01). With an AUC of 0.844, the PO model exhibits similar or improved performance to radiomics-based/deep-learning approaches [20][21][22]. However, we acknowledge that our results may be slightly optimistic as we did not employ feature selection separately on each fold.…”

Does FDG PET-Based Radiomics Have an Added Value for Prediction of Overall Survival in Non-Small Cell Lung Cancer?

“…Machine learning has been employed in the context of medical imaging for segmentation and malignancy characterization [31], in histopathology [32], and in the study of biomarkers [33]. Recent studies have begun to explore the potential role of machine learning networks in prognosis, applying machine learning models to predict OS in NSCLC, mainly integrating tumor stage with clinical factors showing promising results [34][35][36][37][38][39][40][41][42][43][44]. However, a possible research gap lies in the fact that although studies indicate the potential of radiomics to predict survival in oncological patients, the clinical settings and the most important radiomic variables for such prediction need to be better established.…”

“…In another group of studies, radiomic features were integrated with other clinical or imaging variables in fully integrated models, but their added value compared with a reference model was not assessed [38,39].…”

Current status and prospect of PET-related imaging radiomics in lung cancer

“…Christie et al. ( 50 ) developed a predictive model based on postoperative CT and PET/CT images of 135 NSCLC patients, stratifying patients into low and high recurrence/progression risk groups. Preoperative 18 F-FDG PET/CT radiomics, as discovered by Onozato et al.…”