Abstract
"Radiomics" extracts and mines a large number of medical imaging features in a non-invasive and cost-effective way. The underlying assumption of radiomics is that these imaging features quantify phenotypic characteristics of an entire tumor. In order to enhance applicability of radiomics in clinical oncology, highly accurate and reliable machine-learning approaches are required. In this radiomic study, 13 feature selection methods and 11 machine-learning classification methods were evaluated in terms of their performance and stability for predicting overall survival in head and neck cancer patients. Two independent head and neck cancer cohorts were investigated. Training cohort HN1 consisted of 101 head and neck cancer patients. Cohort HN2 (n = 95) was used for validation. A total of 440 radiomic features were extracted from the segmented tumor regions in CT images. Feature selection and classification methods were compared using an unbiased evaluation framework. We observed that the three feature selection methods minimum redundancy maximum relevance (AUC = 0.69, Stability = 0.66), mutual information feature selection (AUC = 0.66, Stability = 0.69), and conditional infomax feature extraction (AUC = 0.68, Stability = 0.7) had high prognostic performance and stability. The three classifiers BY (AUC = 0.67, RSD = 11.28), RF (AUC = 0.61, RSD = 7.36), and NN (AUC = 0.62, RSD = 10.52) also showed high prognostic performance and stability. Analysis investigating performance variability indicated that the choice of classification method is the major factor driving the performance variation (29.02% of total variance). Our study identified prognostic and reliable machine-learning methods for the prediction of overall survival of head and neck cancer patients. Identification of optimal machine-learning methods for radiomics-based prognostic analyses could broaden the scope of radiomics in precision oncology and cancer care.
Highlights
The emergence of “radiomics” [1] has expanded the scope of medical imaging in clinical oncology
As described earlier [30], we investigated 12 machine-learning classifiers belonging to the 12 classifier families: bagging (BAG), Bayesian (BY), boosting (BST), decision trees (DT), discriminant analysis (DA), generalized linear models (GLM), multiple adaptive regression splines (MARS), nearest neighbors (NN), neural networks (Nnet), partial least square and principle component regression (PLSR), random forests (RF), and support vector machines (SVM)
With increasing cohort sizes and expanding feature dimensions, radiomics targets a large pool of medical imaging data (“Big data”)
Summary
The emergence of “radiomics” [1] has expanded the scope of medical imaging in clinical oncology. It is hypothesized that these imaging features are enriched with crucial information regarding tumor phenotype [1, 2]. It has been stated that intra-tumor heterogeneity could have profound implications in clinical predictions (e.g., treatment response, survival outcomes, disease progression, etc.), and it is considered as a crucial factor for precision oncology and related research [3,4,5,6]. Some radio-genomic studies have reported associations between radiomic features and underlying gene expression patterns [2, 9, 11, 24, 25]. These reports indicate that radiomics could improve individualized treatment selection and monitoring. Radiomics is a novel and promising step forward toward the realization of precision oncology
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
