Purpose: The aim of this study was to assess the potential of machine learning with multiparametric magnetic resonance imaging (mpMRI) for the early prediction of pathological complete response (pCR) to neoadjuvant chemotherapy (NAC) and of survival outcomes in breast cancer patients. Materials and Methods: This institutional review board–approved prospective study included 38 women (median age, 46.5 years; range, 25–70 years) with breast cancer who were scheduled for NAC and underwent mpMRI of the breast at 3 T with dynamic contrast-enhanced (DCE), diffusion-weighted imaging (DWI), and T2-weighted imaging before and after 2 cycles of NAC. For each lesion, 23 features were extracted: qualitative T2-weighted and DCE-MRI features according to BI-RADS (Breast Imaging Reporting and Data System), quantitative pharmacokinetic DCE features (mean plasma flow, volume distribution, mean transit time), and DWI apparent diffusion coefficient (ADC) values. To apply machine learning to mpMRI, 8 classifiers including linear support vector machine, linear discriminant analysis, logistic regression, random forests, stochastic gradient descent, decision tree, adaptive boosting, and extreme gradient boosting (XGBoost) were used to rank the features. Histopathologic residual cancer burden (RCB) class (with RCB 0 being a pCR), recurrence-free survival (RFS), and disease-specific survival (DSS) were used as the standards of reference. Classification accuracy with area under the receiving operating characteristic curve (AUC) was assessed using all the extracted qualitative and quantitative features for pCR as defined by RCB class, RFS, and DSS using recursive feature elimination. To overcome overfitting, 4-fold cross-validation was used. Results: Machine learning with mpMRI achieved stable performance as shown by mean classification accuracies for the prediction of RCB class (AUC, 0.86) and DSS (AUC, 0.92) based on XGBoost and the prediction of RFS (AUC, 0.83) with logistic regression. The XGBoost classifier achieved the most stable performance with high accuracies compared with other classifiers. The most relevant features for the prediction of RCB class were as follows: changes in lesion size, complete pattern of shrinkage, and mean transit time on DCE-MRI; minimum ADC on DWI; and peritumoral edema on T2-weighted imaging. The most relevant features for prediction of RFS were as follows: volume distribution, mean plasma flow, and mean transit time; DCE-MRI lesion size; minimum, maximum, and mean ADC with DWI. The most relevant features for prediction of DSS were as follows: lesion size, volume distribution, and mean plasma flow on DCE-MRI, and maximum ADC with DWI. Conclusions: Machine learning with mpMRI of the breast enables early prediction of pCR to NAC as well as survival outcomes in breast cancer patients with high accuracy and thus may provide valuable predictive information to guide treatment decisions.

In this study, a multi-stage optimization procedure is proposed to develop deep neural network models which results in a powerful deep learning pipeline called intelligent deep learning (iDeepLe). The proposed pipeline is then evaluated by a challenging real-world problem, the modeling of the spectral acceleration experienced by a particle during earthquakes. This approach has three main stages to optimize the deep model topology, the hyper-parameters, and its performance, respectively. This pipeline optimizes the deep model via adaptive learning rate optimization algorithms for both accuracy and complexity in multiple stages, while simultaneously solving the unknown parameters of the regression model. Among the seven adaptive learning rate optimization algorithms, Nadam optimization algorithm has shown the best performance results in the current study. The proposed approach is shown to be a suitable tool to generate solid models for this complex real-world system. The results also show that the parallel pipeline of iDeepLe has the capacity to handle big data problems as well.

Purpose: Engineering design and operational decisions depend largely on deep understanding of applications that requires assumptions for simplification of the problems in order to find proper solutions. Cutting-edge machine learning algorithms can be used as one of the emerging tools to simplify this process. In this paper, we propose a novel scalable and interpretable machine learning framework to automate this process and fill the current gap.Design/methodology/approach: The essential principles of the proposed pipeline are mainly (1) scalability, (2) interpretibility and (3) robust probabilistic performance across engineering problems. The lack of interpretibility of complex machine learning models prevents their use in various problems including engineering computation assessments. Many consumers of machine learning models would not trust the results if they cannot understand the method. Thus, the SHapley Additive exPlanations (SHAP) approach is employed to interpret the developed machine learning models.Findings: The proposed framework can be applied to a variety of engineering problems including seismic damage assessment of structures. The performance of the proposed framework is investigated using two case studies of failure identification in reinforcement concrete (RC) columns and shear walls. In addition, the reproducibility, reliability and generalizability of the results were validated and the results of the framework were compared to the benchmark studies. The results of the proposed framework outperformed the benchmark results with high statistical significance.Originality/value: Although, the current study reveals that the geometric input features and reinforcement indices are the most important variables in failure modes detection, better model can be achieved with employing more robust strategies to establish proper database to decrease the errors in some of the failure modes identification.