Incorporating Radiomics into Machine Learning Models to Predict Outcomes of Neuroblastoma

Abstract

Neuroblastoma is one of the most common pediatric cancers. This study used machine learning (ML) to predict the mortality and a few other investigated intermediate outcomes of neuroblastoma patients non-invasively from CT images. Performances of multiple ML algorithms over retrospective CT images of 65 neuroblastoma patients are analyzed. An artificial neural network (ANN) is used on tumor radiomic features extracted from 3D CT images. A pre-trained 2D convolutional neural network (CNN) is used on slices of the same images. ML models are trained for various pathologically investigated outcomes of these patients. A subspecialty-trained pediatric radiologist independently reviewed the manually segmented primary tumors. Pyradiomics library is used to extract 105 radiomic features. Six ML algorithms are compared to predict the following outcomes: mortality, presence or absence of metastases, neuroblastoma differentiation, mitosis-karyorrhexis index (MKI), presence or absence of MYCN gene amplification, and presence of image-defined risk factors (IDRF). The prediction ranges over multiple experiments are measured using the area under the receiver operating characteristic (ROC-AUC) for comparison. Our results show that the radiomics-based ANN method slightly outperforms the other algorithms in predicting all outcomes except classification of the grade of neuroblastic differentiation, for which the elastic regression model performed the best. Contributions of the article are twofold: (1) noninvasive models for the prognosis from CT images of neuroblastoma, and (2) comparison of relevant ML models on this medical imaging problem.

Development and validation of a CT‐based radiomics signature for identifying high‐risk neuroblastomas under the revised Children's Oncology Group classification system

“…There have been some studies on radiomics in the pathological differentiation, molecular typing, and prediction of therapeutic effect of neuroblastomas. [9][10][11][12] These studies showed that radiomics has the potential for the accurate diagnosis of neuroblastomas. Therefore, radiomics is expected to provide a noninvasive evaluation method for risk stratification of neuroblastomas.…”

“…Thus, medical images are not limited to providing subjective visual information. There have been some studies on radiomics in the pathological differentiation, molecular typing, and prediction of therapeutic effect of neuroblastomas 9–12 . These studies showed that radiomics has the potential for the accurate diagnosis of neuroblastomas.…”

Computed tomography-based radiomics to assess risk stratification in pediatric malignant peripheral neuroblastic tumors

“…Our results proved valuable in stratifying and distinguishing high-risk groups with AUCs of 0.903 and 0.876, respectively, and seemed suitable for clinical application. In recent years, several radiomics studies of malignant PNTs have reported features associated with the prediction of tumor histology, [23] MYCN amplification status, [24][25][26] bone marrow involvement, [27] response to neoadjuvant chemotherapy, [28] recurrence, [29] prognosis, [30] and differentiation between highrisk and non-high-risk groups. [31] Nearly half of the radiomics studies were based on 18 F-fluorodeoxyglucose positron emission tomography/computed tomography ( 18 F-FDG PET/CT).…”

Artificial intelligence assists precision medicine in cancer treatment

“…Patients with poor surgical results may want to consider changing the surgical method or choosing non-surgical treatment. Patients with a higher risk of tumor recurrence and metastasis should continue to receive neoadjuvant radiotherapy and chemotherapy ( 21 , 26 , 178 – 180 ).…”