Multi-view radiomics and dosiomics analysis with machine learning for predicting acute-phase weight loss in lung cancer patients treated with radiotherapy

Abstract

We propose a multi-view data analysis approach using radiomics and dosiomics (R&D) texture features for predicting acute-phase weight loss (WL) in lung cancer radiotherapy. Baseline weight of 388 patients who underwent intensity modulated radiation therapy (IMRT) was measured between one month prior to and one week after the start of IMRT. Weight change between one week and two months after the commencement of IMRT was analyzed, and dichotomized at 5% WL. Each patient had a planning CT and contours of gross tumor volume (GTV) and esophagus (ESO). A total of 355 features including clinical parameter (CP), GTV and ESO (GTV&ESO) dose-volume histogram (DVH), GTV radiomics, and GTV&ESO dosiomics features were extracted. R&D features were categorized as first- (L1), second- (L2), higher-order (L3) statistics, and three combined groups, L1 + L2, L2 + L3 and L1 + L2 + L3. Multi-view texture analysis was performed to identify optimal R&D input features. In the training set (194 earlier patients), feature selection was performed using Boruta algorithm followed by collinearity removal based on variance inflation factor. Machine-learning models were developed using Laplacian kernel support vector machine (lpSVM), deep neural network (DNN) and their averaged ensemble classifiers. Prediction performance was tested on an independent test set (194 more recent patients), and compared among seven different input conditions: CP-only, DVH-only, R&D-only, DVH + CP, R&D + CP, R&D + DVH and R&D + DVH + CP. Combined GTV L1 + L2 + L3 radiomics and GTV&ESO L3 dosiomics were identified as optimal input features, which achieved the best performance with an ensemble classifier (AUC = 0.710), having statistically significantly higher predictability compared with DVH and/or CP features (p < 0.05). When this performance was compared to that with full R&D-only features which reflect traditional single-view data, there was a statistically significant difference (p < 0.05). Using optimized multi-view R&D input features is beneficial for predicting early WL in lung cancer radiotherapy, leading to improved performance compared to using conventional DVH and/or CP features.

A Multicentre Evaluation of Dosiomics Features Reproducibility, Stability and Sensitivity

“…Dosiomic is increasingly used in clinical studies aiming to improve the prediction of clinical outcomes, e.g., locoregional recurrence after IMRT for head and neck cancer [23] or local control after carbon-ion radiotherapy in skull-base chordoma [21]. Dosiomic features were analysed by machine learning for the prediction of acute-phase weight loss in lung cancer patients treated with radiotherapy [30]. Among preliminary multicentre experiences, Adachi et al [24] aimed at predicting radiation pneumonitis after lung stereotactic body radiation therapy using dosiomics.…”

What is plan quality in radiotherapy? The importance of evaluating dose metrics, complexity, and robustness of treatment plans

“…Texture analysis was recently proposed to characterise the spatial dose distribution. This is known as dosiomics [37][38][39][40][41][42] and is a powerful technique for characterising spatial and statistical distributions of pixel/voxel intensities in an image through the identification of patterns and voxel correlations. Dosiomics is a promising method for parameterising regions of interest and for producing intensity, textural and shape-based dose features that might be able to describe the dose distribution better than DVH-based metrics, as well as to potentially improve the predictive performance of TCP and NTCP models.…”

Multi-Organ Omics-Based Prediction for Adaptive Radiation Therapy Eligibility in Nasopharyngeal Carcinoma Patients Undergoing Concurrent Chemoradiotherapy

“…In this regard, several research groups have also suggested to incorporate dosimetric deviations in targets and/or OARs (such as parotid glands) as part of the ART regimen (12,(18)(19)(20). Of note, although Dosiomics has recently been studied for prediction of toxicity (32,34,(41)(42)(43) and prognosis (33,44) in cancer patients, its potential in treatment response prediction, in particular on the basis of the RECIST criteria, has not been reported. Future studies in this aspect are recommended to confirm its capability in this regard.…”

“…By contrast, dosiomics is capable of characterizing spatial pattern of local radiation dose distributions within the 8 studied VOIs. It has been extensively studied in various predictive modelling for cancer prognosis and treatment responses (32,33). In this study, dosiomic features of DVH curve points for the 8 VOIs were calculated based on the method adopted by Gabrys ́et al (34), examples include but not limited to maximum dose, minimum dose, mean dose, volume of the VOI receiving at least certain dose levels, and minimum dose received by certain volume of the VOI.…”