Multi-Objective-Based Radiomic Feature Selection for Lesion Malignancy Classification

Brain Tumor Recurrence vs. Radiation Necrosis Classification and Patient Survivability Prediction

GB (Glioblastoma WHO Grade 4) is the most aggressive type of brain tumor in adults that has a short survival rate even after aggressive treatment with surgery and radiation therapy. The changes in magnetic resonance imaging (MRI) for patients with GB after radiotherapy are indicative of either radiation-induced necrosis (RN) or recurrent brain tumor (rBT). Screening for rBT and RN at an early stage is crucial for facilitating faster treatment and better outcomes for the patients. Differentiating rBT from RN is challenging as both may present with similar radiological and clinical characteristics on MRI. Moreover, learning-based rBT versus RN classification using MRI may suffer from class imbalance due to a lack of patient data. While synthetic data generation using generative models has shown promise to address class imbalances, the underlying data representation may be different in synthetic or augmented data. This study proposes computational modeling with statistically rigorous repeated random sub-sampling to balance the subset sample size for rBT and RN classification. The proposed pipeline includes multiresolution radiomic feature (MRF) extraction followed by feature selection with statistical significance testing (p<0.05). The five-fold cross validation results show the proposed model with MRF features classifies rBT from RN with an area under the curve (AUC) of 0.892±0.055. Moreover, considering the dependence between survival time and censoring time (where patients are not followed up until death), the feasibility of using MRF radiomic features as a non-invasive biomarker to identify patients who are at higher risk of recurrence or radiation necrosis is demonstrated. The cross-validated results show that the MRF model provides the best overall survival prediction with an AUC of 0.77±0.032. Comparison with state-of-the-art methods suggest the proposed method is effective in RN versus rBT classification and patient survivability prediction.

Universal multi-factor feature selection method for radiomics-based brain tumor classification

Brain tumor mortality is high, and accurate classification before treatment can improve patient prognosis. Radiomics, which extracts numerous features from medical images, has been widely applied in brain tumor classification studies. Feature selection (FS) is a critical step in radiomics because it reduces redundant information and enhances classification performance. However, the lack of universal FS methods limits the development of radiomics-based brain tumor classification studies. To address this issue, we summarize the characteristics of the FS methods used in related studies and propose a universal method based on three selection factors called triple-factor cascaded selection (TFCS). Particularly, these factors correspond to the correlation between features and task labels, interdependence among features, and role of features in the model. The TFCS method divides FS into two steps. First, it utilizes mutual information to select features that are strongly correlated with the task and contain less redundant information. Recursive feature elimination is then employed to obtain the subset with the best classification performance. To validate the universality of the TFCS, we conducted experiments on seven datasets containing 13 brain tumor classification tasks and evaluated the overall performance using five types of indicators. Results: TFCS exhibited excellent overall performance for all tasks. Compared to the 13 related methods, it takes less time, has moderate parsimony, the best classification performance, adaptability, and stability, and shows better universality. Our study demonstrates that the reasonable utilization of multiple factors can enhance FS performance and provide new insights for future method design.

External validation of an MR-based radiomic model predictive of locoregional control in oropharyngeal cancer

OBJECTIVES

To externally validate a pre-treatment MR-based radiomics model predictive of locoregional control in oropharyngeal squamous cell carcinoma (OPSCC) and to assess the impact of differences between datasets on the predictive performance.

METHODS

Radiomic features, as defined in our previously published radiomics model, were extracted from the primary tumor volumes of 157 OPSCC patients in a different institute. The developed radiomics model was validated using this cohort. Additionally, parameters influencing performance, such as patient subgroups, MRI acquisition, and post-processing steps on prediction performance will be investigated. For this analysis, matched subgroups (based on human papillomavirus (HPV) status of the tumor, T-stage, and tumor subsite) and a subgroup with only patients with 4-mm slice thickness were studied. Also the influence of harmonization techniques (ComBat harmonization, quantile normalization) and the impact of feature stability across observers and centers were studied. Model performances were assessed by area under the curve (AUC), sensitivity, and specificity.

RESULTS

Performance of the published model (AUC/sensitivity/specificity: 0.74/0.75/0.60) drops when applied on the validation cohort (AUC/sensitivity/specificity: 0.64/0.68/0.60). The performance of the full validation cohort improves slightly when the model is validated using a patient group with comparable HPV status of the tumor (AUC/sensitivity/specificity: 0.68/0.74/0.60), using patients acquired with a slice thickness of 4 mm (AUC/sensitivity/specificity: 0.67/0.73/0.57), or when quantile harmonization was performed (AUC/sensitivity/specificity: 0.66/0.69/0.60).

CONCLUSION

The previously published model shows its generalizability and can be applied on data acquired from different vendors and protocols. Harmonization techniques and subgroup definition influence performance of predictive radiomics models.

KEY POINTS

• Radiomics, a noninvasive quantitative image analysis technique, can support the radiologist by enhancing diagnostic accuracy and/or treatment decision-making. • A previously published model shows its generalizability and could be applied on data acquired from different vendors and protocols.

Comparison between Deep Learning and Conventional Machine Learning in Classifying Iliofemoral Deep Venous Thrombosis upon CT Venography

In this study, we aimed to investigate quantitative differences in performance in terms of comparing the automated classification of deep vein thrombosis (DVT) using two categories of artificial intelligence algorithms: deep learning based on convolutional neural networks (CNNs) and conventional machine learning. We retrospectively enrolled 659 participants (DVT patients, 282; normal controls, 377) who were evaluated using contrast-enhanced lower extremity computed tomography (CT) venography. Conventional machine learning consists of logistic regression (LR), support vector machines (SVM), random forests (RF), and extreme gradient boosts (XGB). Deep learning based on CNN included the VGG16, VGG19, Resnet50, and Resnet152 models. According to the mean generated AUC values, we found that the CNN-based VGG16 model showed a 0.007 higher performance (0.982 ± 0.014) as compared with the XGB model (0.975 ± 0.010), which showed the highest performance among the conventional machine learning models. In the conventional machine learning-based classifications, we found that the radiomic features presenting a statistically significant effect were median values and skewness. We found that the VGG16 model within the deep learning algorithm distinguished deep vein thrombosis on CT images most accurately, with slightly higher AUC values as compared with the other AI algorithms used in this study. Our results guide research directions and medical practice.

IMAL-Net: Interpretable multi-task attention learning network for invasive lung adenocarcinoma screening in CT images

PURPOSE

Feature maps created from deep convolutional neural networks (DCNNs) have been widely used for visual explanation of DCNN-based classification tasks. However, many clinical applications such as benign-malignant classification of lung nodules normally require quantitative and objective interpretability, rather than just visualization. In this paper, we propose a novel interpretable multi-task attention learning network named IMAL-Net for early invasive adenocarcinoma screening in chest computed tomography images, which takes advantage of segmentation prior to assist interpretable classification.

METHODS

Two sub-ResNets are firstly integrated together via a prior-attention mechanism for simultaneous nodule segmentation and invasiveness classification. Then, numerous radiomic features from the segmentation results are concatenated with high-level semantic features from the classification subnetwork by FC layers to achieve superior performance. Meanwhile, an end-to-end feature selection mechanism (named FSM) is designed to quantify crucial radiomic features greatly affecting the prediction of each sample, and thus it can provide clinically applicable interpretability to the prediction result.

RESULTS

Nodule samples from a total of 1626 patients were collected from two grade-A hospitals for large-scale verification. Five-fold cross validation demonstrated that the proposed IMAL-Net can achieve an AUC score of 93.8% ± 1.1% and a recall score of 93.8% ± 2.8% for identification of invasive lung adenocarcinoma.

CONCLUSIONS

It can be concluded that fusing semantic features and radiomic features can achieve obvious improvements in the invasiveness classification task. Moreover, by learning more fine-grained semantic features and highlighting the most important radiomics features, the proposed attention and FSM mechanisms not only can further improve the performance but also can be used for both visual explanations and objective analysis of the classification results.

Hybrid Clinical-Radiomics Model for Precisely Predicting the Invasiveness of Lung Adenocarcinoma Manifesting as Pure Ground-Glass Nodule

RATIONALE AND OBJECTIVES

To identify whether the radiomics features of computed tomography (CT) allowed for the preoperative discrimination of the invasiveness of lung adenocarcinomas manifesting as pure ground-glass nodules (pGGNs) and further to develop and compare different predictive models.

MATERIALS AND METHODS

We retrospectively included 187 lung adenocarcinomas presenting as pGGNs (66 preinvasive lesions and 121 invasive lesions), which were randomly divided into the training and test sets (8:2). Radiomics features were extracted from non-enhanced CT images. Clinical features, including patient's demographic characteristics, smoking status, and conventional CT features that reflect tumor's morphology and surrounding information were also collected. Intraclass correlation coefficient and ℓ_2.1-norm minimization were used to identify influential feature subset which was then used to build three predictive models (clinical, radiomics, and clinical-radiomics models) with the gradient boosting regression tree classifier. The performances of the predictive models were evaluated using the area under the curve (AUC).

RESULTS

Of the 1409 radiomics features and 27 clinical feature subtypes, 102 features were selected to construct the hybrid clinical-radiomics model, which achieved the best discriminative power (AUC = 0.934 and 0.929 in training and test set). The radiomics model showed comparable predictive performance (AUC = 0.911 and 0.901 in training and test set) compared to the clinical model (AUC = 0.911 and 0.894 in training and test set).

CONCLUSION

The radiomics model showed good predictive performance in discriminating invasive lesions from preinvasive lesions for lung adenocarcinomas presenting as pGGNs. Its performance can be further improved by adding clinical features.

Radiomics Based Bayesian Inversion Method for Prediction of Cancer and Pathological Stage

OBJECTIVE

To develop a Bayesian inversion framework on longitudinal chest CT scans which can perform efficient multi-class classification of lung cancer.

METHODS

While the unavailability of large number of training medical images impedes the performance of lung cancer classifiers, the purpose built deep networks have not performed well in multi-class classification. The presented framework employs particle filtering approach to address the non-linear behaviour of radiomic features towards benign and cancerous (stages I, II, III, IV) nodules and performs efficient multi-class classification (benign, early stage cancer, advanced stage cancer) in terms of posterior probability function. A joint likelihood function incorporating diagnostic radiomic features is formulated which can compute likelihood of cancer and its pathological stage. The proposed research study also investigates and validates diagnostic features to discriminate accurately between early stage (I, II) and advanced stage (III, IV) cancer.

RESULTS

The proposed stochastic framework achieved 86% accuracy on the benchmark database which is better than the other prominent cancer detection methods.

CONCLUSION

The presented classification framework can aid radiologists in accurate interpretation of lung CT images at an early stage and can lead to timely medical treatment of cancer patients.

Deep Semantic Segmentation Feature-Based Radiomics for the Classification Tasks in Medical Image Analysis

Recently, an emerging trend in medical image classification is to combine radiomics framework with deep learning classification network in an integrated system. Although this combination is efficient in some tasks, the deep learning-based classification network is often difficult to capture an effective representation of lesion regions, and prone to face the challenge of overfitting, leading to unreliable features and inaccurate results, especially when the sizes of the lesions are small or the training dataset is small. In addition, these combinations mostly lack an effective feature selection mechanism, which makes it difficult to obtain the optimal feature selection. In this paper, we introduce a novel and effective deep semantic segmentation feature-based radiomics (DSFR) framework to overcome the above-mentioned challenges, which consists of two modules: the deep semantic feature extraction module and the feature selection module. Specifically, the extraction module is utilized to extract hierarchical semantic features of the lesions from a trained segmentation network. The feature selection module aims to select the most representative features by using a novel feature similarity adaptation algorithm. Experiments are extensively conducted to evaluate our method in two clinical tasks: the pathological grading prediction in pancreatic neuroendocrine neoplasms (pNENs), and the prediction of thrombolytic therapy efficacy in deep venous thrombosis (DVT). Experimental results on both tasks demonstrate that the proposed method consistently outperforms the state-of-the-art approaches by a large margin.

Intelligent immune clonal optimization algorithm for pulmonary nodule classification

Computer-aided diagnosis (CAD) of pulmonary nodules is an effective approach for early detection of lung cancers, and pulmonary nodule classification is one of the key issues in the CAD system. However, CAD has the problems of low accuracy and high false-positive rate (FPR) on pulmonary nodule classification. To solve these problems, a novel method using intelligent immune clonal selection and classification algorithm is proposed and developed in this work. First, according to the mechanism and characteristics of chaotic motion with a logistic mapping, the proposed method utilizes the characteristics of chaotic motion and selects the control factor of the optimal chaotic state, to generate an initial population with randomness and ergodicity. The singleness problem of the initial population of the immune algorithm was solved by the proposed method. Second, considering on the characteristics of Gaussian mutation operator (GMO) with a small scale, and Cauchy mutation operator (CMO) with a big scale, an intelligent mutation strategy is developed, and a novel control factor of the mutation is designed. Therefore, a Gauss-Cauchy hybrid mutation operator is designed. Ultimately, in this study, the intelligent immune clonal optimization algorithm is proposed and developed for pulmonary nodule classification. To verify its accuracy, the proposed method was used to analyze 90 CT scans with 652 nodules. The experimental results revealed that the proposed method had an accuracy of 97.87% and produced 1.52 false positives per scan (FPs/scan), indicating that the proposed method has high accuracy and low FPR.

Machine learning based congestive heart failure detection using feature importance ranking of multimodal features

In this study, we ranked the Multimodal Features extracted from Congestive Heart Failure (CHF) and Normal Sinus Rhythm (NSR) subjects. We categorized the ranked features into 1 to 5 categories based on Empirical Receiver Operating Characteristics (EROC) values. Instead of using all multimodal features, we use high ranking features for detection of CHF and normal subjects. We employed powerful machine learning techniques such as Decision Tree (DT), Naïve Bayes (NB), SVM Gaussian, SVM RBF and SVM Polynomial. The performance was measured in terms of Sensitivity, Specificity, Positive Predictive Value (PPV), Negative Predictive Value (NPV), Accuracy, False Positive Rate (FPR), and area under the Receiver Operating characteristic Curve (AUC). The highest detection performance in terms of accuracy and AUC was obtained with all multimodal features using SVM Gaussian with Sensitivity (93.06%), Specificity (81.82%), Accuracy (88.79%) and AUC (0.95). Using the top five ranked features, the highest performance was obtained with SVM Gaussian yields accuracy (84.48%), AUC (0.86); top nine ranked features using Decision Tree and Naïve Bayes got accuracy (84.48%), AUC (0.88); last thirteen ranked features using SVM polynomial obtained accuracy (80.17%), AUC (0.84). The findings indicate that proposed approach with feature ranking can be very useful for automatic detection of congestive heart failure patients and can be very helpful for further decision making by the clinicians and physicians in order to decrease the mortality rate.

Multifaceted radiomics for distant metastasis prediction in head & neck cancer

Accurately predicting distant metastasis in head & neck cancer has the potential to improve patient survival by allowing early treatment intensification with systemic therapy for high-risk patients. By extracting large amounts of quantitative features and mining them, radiomics has achieved success in predicting treatment outcomes for various diseases. However, there are several challenges associated with conventional radiomic approaches, including: (1) how to optimally combine information extracted from multiple modalities; (2) how to construct models emphasizing different objectives for different clinical applications; and (3) how to utilize and fuse output obtained by multiple classifiers. To overcome these challenges, we propose a unified model termed as multifaceted radiomics (M-radiomics). In M-radiomics, a deep learning with stacked sparse autoencoder is first utilized to fuse features extracted from different modalities into one representation feature set. A multi-objective optimization model is then introduced into M-radiomics where probability-based objective functions are designed to maximize the similarity between the probability output and the true label vector. Finally, M-radiomics employs multiple base classifiers to get a diverse Pareto-optimal model set and then fuses the output probabilities of all the Pareto-optimal models through an evidential reasoning rule fusion (ERRF) strategy in the testing stage to obtain the final output probability. Experimental results show that M-radiomics with the stacked autoencoder outperforms the model without the autoencoder. M-radiomics obtained more accurate results with a better balance between sensitivity and specificity than other single-objective or single-classifier-based models.

A CT-Based Radiomics Approach for the Differential Diagnosis of Sarcomatoid and Clear Cell Renal Cell Carcinoma

This study was aimed at building a computed tomography- (CT-) based radiomics approach for the differentiation of sarcomatoid renal cell carcinoma (SRCC) and clear cell renal cell carcinoma (CCRCC). It involved 29 SRCC and 99 CCRCC patient cases, and to each case, 1029 features were collected from each of the corticomedullary phase (CMP) and nephrographic phase (NP) image. Then, features were selected by using the least absolute shrinkage and selection operator regression method and the selected features of the two phases were explored to build three radiomics approaches for SRCC and CCRCC classification. Meanwhile, subjective CT findings were filtered by univariate analysis to construct a radiomics model and further selected by Akaike information criterion for integrating with the selected image features to build the fifth model. Finally, the radiomics models utilized the multivariate logistic regression method for classification and the performance was assessed with receiver operating characteristic curve (ROC) and DeLong test. The radiomics models based on the CMP, the NP, the CMP and NP, the subjective findings, and the combined features achieved the AUC (area under the curve) value of 0.772, 0.938, 0.966, 0.792, and 0.974, respectively. Significant difference was found in AUC values between each of the CMP radiomics model (0.0001 ≤ p ≤ 0.0051) and the subjective findings model (0.0006 ≤ p ≤ 0.0079) and each of the NP radiomics model, the CMP and NP radiomics model, and the combined model. Sarcomatoid change is a common pathway of dedifferentiation likely occurring in all subtypes of renal cell carcinoma, and the CT-based radiomics approaches in this study show the potential for SRCC from CCRCC differentiation.

A collection input based support tensor machine for lesion malignancy classification in digital breast tomosynthesis

Digital breast tomosynthesis (DBT) with improved lesion conspicuity and characterization has been adopted in screening practice. DBT-based diagnosis strongly depends on physicians' experience, so an automatic lesion malignancy classification model using DBT could improve the consistency of diagnosis among different physicians. Tensor-based approaches that use the original imaging data as input have shown promising results for many classification tasks. However, DBT data are pseudo-3D volumetric imaging as the slice spacing of DBT is much coarser than that of the in-plane resolution. Thus, directly constructing DBT as the third-order tensor in a conventional tensor-based classifier with introducing additional information to the original DBT data along the slice-spacing dimension will lead to inconsistency across all three dimensions. To avoid such inconsistency, we introduce a collection input based support tensor machine (CISTM)-based classifier that uses the tensor collection as input for classifying lesion malignancy in DBT. In CISTM, instead of introducing the third dimension directly into the geometry construction, the third-dimension structural relationship is related by weight parameters in the decision function, which is dynamically and automatically constructed during the classifier training process and is more consistent with the pseudo-3D nature of DBT. We tested our method on a DBT dataset of 926 images among which 262 were malignant and 664 were benign. We compared our method with the latest tensor-based method, KSTM (kernelled support tensor machine), which does not consider the unique non-uniform resolution property of DBT. Experimental results illustrate that the CISTM-based classifier is effective for classifying breast lesion malignancy in DBT and that it outperforms the KSTM-based classifier.

CT radiomics and PET radiomics: ready for clinical implementation?

INTRODUCTION

Today, rapid technical and clinical developments result in an increasing number of treatment options for oncological diseases. Thus, decision support systems are needed to offer the right treatment to the right patient. Imaging biomarkers hold great promise in patient-individual treatment guidance. Routinely performed for diagnosis and staging, imaging datasets are expected to hold more information than used in the clinical practice. Radiomics describes the extraction of a large number of meaningful quantitative features from medical images, such as computed tomography (CT) and positron emission tomography (PET). Due to the non-invasive nature and ability to capture 3D image-based heterogeneity, radiomic features are potential surrogate markers of the cancer phenotype. Several radiomic studies are published per day, owing to encouraging results of many radiomics-based patient outcome models. Despite this comparably large number of studies, radiomics is mainly studied in proof of principle concept. Hence, a translation of radiomics from a hot topic research field into an essential clinical decision-making tool is lacking, but of high clinical interest.

EVIDENCE ACQUISITION

Herein, we present a literature review addressing the clinical evidence of CT and PET radiomics. An extensive literature review was conducted in PubMed, including papers on robustness and clinical applications.

EVIDENCE SYNTHESIS

We summarize image-modality related influences on the robustness of radiomic features and provide an overview of clinical evidence reported in the literature. Today, more evidence has been provided for CT imaging, however, PET imaging offers the promise of direct imaging of biological processes and functions. We provide a summary of future research directions, which needs to be addressed in order to successfully introduce radiomics into clinical medicine. In comparison to CT, more focus should be directed towards harmonization of PET acquisition and reconstruction protocols, which is important for transferable modelling.

CONCLUSIONS

Both CT and PET radiomics are promising pre-treatment and intra-treatment biomarkers for outcome prediction. Most studies are performed in retrospective setting, however their validation in prospective data collections is ongoing.