Extraction Of Quantitative Features Research Articles

Feature-based PET/MRI radiomics in patients with brain tumors.

Radiomics allows the extraction of quantitative features from medical images such as CT, MRI, or PET, thereby providing additional, potentially relevant diagnostic information for clinical decision-making. Because the computation of these features is performed highly automated on medical images acquired during routine follow-up, radiomics offers this information at low cost. Further, the radiomics features can be used alone or combined with other clinical or histomolecular parameters to generate predictive or prognostic mathematical models. These models can then be applied for various important diagnostic indications in neuro-oncology, for example, to noninvasively predict relevant biomarkers in glioma patients, to differentiate between treatment-related changes and local brain tumor relapse, or to predict treatment response. In recent years, amino acid PET has become an important diagnostic tool in patients with brain tumors. Therefore, the number of studies in patients with brain tumors investigating the potential of PET radiomics or combined PET/MRI radiomics is steadily increasing. This review summarizes current research regarding feature-based PET as well as combined PET/MRI radiomics in neuro-oncology.

How does Radiomics actually work? - Review.

Personalized precision medicine requires highly accurate diagnostics. While radiological research has focused on scanner and sequence technologies in recent decades, applications of artificial intelligence are increasingly attracting scientific interest as they could substantially expand the possibility of objective quantification and diagnostic or prognostic use of image information.In this context, the term "radiomics" describes the extraction of quantitative features from imaging data such as those obtained from computed tomography or magnetic resonance imaging examinations. These features are associated with predictive goals such as diagnosis or prognosis using machine learning models. It is believed that the integrative assessment of the feature patterns thus obtained, in combination with clinical, molecular and genetic data, can enable a more accurate characterization of the pathophysiology of diseases and more precise prediction of therapy response and outcome.This review describes the classical radiomics approach and discusses the existing very large variability of approaches. Finally, it outlines the research directions in which the interdisciplinary field of radiology and computer science is moving, characterized by increasingly close collaborations and the need for new educational concepts. The aim is to provide a basis for responsible and comprehensible handling of the data and analytical methods used. KEY POINTS:: · Radiomics is playing an increasingly important role in imaging research.. · Radiomics has great potential to meet the requirements ofprecision medicine.. · Radiomics analysis is still subject to great variability.. · There is a need for quality-assured application of radiomics in medicine.. CITATION FORMAT: · Attenberger UI, Langs G, . How does Radiomics actually work? - Review. Fortschr Röntgenstr 2021; 193: 652 - 657.

Cardiac computed tomography radiomics: an emerging tool for the non-invasive assessment of coronary atherosclerosis.

In the last decades, significant advances have been made in the preventive approaches to cardiovascular disease. Even so, coronary artery disease remains one of the main causes of morbidity and mortality worldwide. Invasive imaging modalities, such as intravascular ultrasound or optical coherence tomography, have played a key role in the comprehension of the pathological processes underlying myocardial infarction and cerebrovascular disease. These imaging techniques have contributed greatly to the identification and phenotyping of the culprit lesion, the so-called vulnerable plaque. Coronary computed tomographic angiography (CCTA) has emerged in more recent years as the non-invasive modality of choice in the study of coronary atherosclerosis, showing in many studies a diagnostic yield comparable to invasive approaches. Moreover, being able to describe extra-luminal characteristics of the affected vessel, CCTA has greatly contributed towards shifting the attention of researchers from the mere quantification of luminal stenosis to the identification of adverse plaque features, which appear to have a stronger prognostic value. However, the identification of some of the hallmarks of vulnerable plaques is qualitative in nature and, therefore, subject to some degree of inter-reader variability. Moreover, CCTA is still unable to identify some fine markers of plaque vulnerability which can be detected by invasive techniques, such as neovascularization and plaque erosion, among others. Nonetheless, radiological images can be viewed as vast 3-D datasets which, via the use of recent technology, allow for the extraction of numerous quantitative features that may be used to accurately phenotype a given lesion. Radiomics is the process of extrapolating innumerable parameters from a given region of interest, with the goal of establishing correlations between quantitative variables and clinical data. These datasets can then be manipulated to create predictive models via the use of automated algorithms in a process called machine learning. As a result of these approaches, radiological images may offer information regarding the characterization of a plaque which can go much beyond the boundaries of what can be qualitatively asserted by the human eye, contributing to expanding the knowledge of the disease and ultimately assist clinical decisions. Thus far, radiomics has found its more consistent area of application in the field of oncology; to present date, the amount of clinical data regarding coronary artery disease is still relatively small, partly due to the technical difficulties associated with the implementation of such techniques to the study of a small and geometrically complex lesion such as the coronary plaque. The present review, after a summary of the imaging modalities most commonly used nowadays in the study of coronary plaques, will provide a perspective on the application of radiomic analysis to coronary artery disease.

Using deep learning to extract novel and quantitative imaging features from perfluoropropane contrast, sulphur hexafluoride contrast and non-contrast stress echocardiography images

Abstract Background Stress echocardiography has become established as the most widely applied non-invasive imaging test for diagnosis of coronary artery disease within the UK. However, stress echocardiography has been substantially qualitative, rather than quantitative, based on visual wall motion assessment. For the first time, we have identified and validated quantitative descriptors of cardiac geometry and motion, extracted from ultrasound images acquired using contrast agents in an automated way. Purpose To establish whether these novel imaging features can be generated in an automated, quantifiable and reproducible way from images acquired with perfluoropropane contrast, as well as investigating how these extracted measures compare to those extracted from sulphur hexafluoride contrast and non-contrast studies. Methods 100 patients who received perfluoropropane contrast during their stress echocardiogram were recruited. Their stress echocardiography images were processed through a deep learning algorithm. Novel feature values were recorded and a subset of 10 studies were repeated. The automated measures of global longitudinal strain (GLS) and ejection fraction (EF) extracted from these images were compared to values previously extracted from sulphur hexafluoride contrast and non-contrast images using the same software. Results A full set of 31 novel imaging features were successfully extracted from 79 studies acquired using the perfluoropropane contrast agent with a dropout rate of 14% (n=92, 8 incomplete image sets). Repeated analysis in a subset of 10 perfluoropropane cases demonstrated excellent reproducibility of the extracted feature values (R2=1). Automated values of GLS and EF, at both rest (GLS = −16.4±4.8%, EF = 63±13%) and stress stages (GLS = −17.7±5.8%, EF = 68±11%), were extracted from 83 perfluoropropane studies, with a dropout rate of 16% (n=99, fewer incomplete sets as short axis view not required). The ranges of GLS and EF measures extracted from the perfluoropropane images were comparable to the other contrast studies (n=222) (Rest GLS = −16.8±5.8%, Rest EF = 63±10%; Stress GLS = −19.1±6.7%, Stress EF = 71±9%) and non-contrast studies (n=86) (Rest GLS = −15.7±5.3%, Rest EF = 57±10%; Stress GLS = −17.3±6.4%, Stress EF = 61±14%). Conclusions Novel features and clinically relevant measures were extracted from images acquired using perfluoropropane contrast for the first time in a fully automated and reproducible way using a deep learning algorithm. The analysis failure rate and generated measures are comparable to those extracted from images using other commonly used sulphur hexafluoride contrast agents and non-contrast stress echocardiography studies. These findings demonstrate that deep learning algorithms can be used for automated quantitative analysis of stress echocardiograms acquired using various contrast agents and in non-contrast studies to improve stress echocardiography practice. Funding Acknowledgement Type of funding source: Private company. Main funding source(s): Lantheus Medical Imaging, Inc.

On the application of multi-parametric optical phenotyping of bacterial colonies for multipurpose microbiological diagnostics

The development of new diagnostics techniques and modalities is critical for early detection of microbial contamination. In this study, the novel integrated system for multi-parametric optical phenotyping and characterization of bacterial colonies, is presented. The system combines Mach-Zehnder interferometer with a spectral imaging system for capturing multispectral diffraction patterns and multispectral two-dimensional transmission maps of bacterial colonies, along with the simultaneous interferometric profilometry. The herein presented investigation was carried out on five representative bacteria species and nearly 3000 registered multispectral optical signatures. The interferograms were analyzed by four-step phase shift algorithm to reconstruct the colony profile to enable the obtaining of the comparable optical signatures. The dedicated image processing algorithms were used for extraction of quantitative features of these signatures. The random forest algorithm was applied for selection of the most predictive set of features, which were used in classification model based on Support-Vector Machine. Obtained results have shown that the use of multiple multispectral optical signatures provide a multi-parametric bacteria identification at an exceptionally high accuracy (99.4-100%), significantly better than in case of classification based on each of these signatures (multispectral diffraction patterns, two-dimensional transmission coefficient maps), separately. Obtained results revealed that analysis of multispectral signatures can also be applied for characterisation of physical, physicochemical and chemical properties of the bacterial colonies in the presence of the antimicrobial factors. Therefore, the proposed label-free, non-destructive optical technique has perspectives to be exploited in the multipurpose diagnostics and it can be used as a pre-screening tool in microbiological laboratories.

Diagnosis of Diabetic Retinopathy Utilizing Computer-Aided Diagnosis System

Purpose: Diabetes is considered one of most diseases spread among people; blindness is considered the most resulted effect. Diabetes can damage the retinal blood vessels and cause severe problems to the eyes, which may end with sight loss. Such medical condition is known as "diabetic retinopathy" (DR). In such a diagnosis, the retinal microvascular go through several stages of change threat. In the early stages of the DR, detecting the formation that happened to the retinal blood vessels helps prevent the disease's dangerous effects. Therefore, producing a method to diagnose the disease in the early stages is helpful. So, this work aimed to develop a system of detecting and classifying the retina formation, trying to avoid relevant effects.Methodology: The current method depends on vascular edges map extracted from images of retina captured by a fundus camera. Such a map been utilized to extract quantitative texture features. The system tested two independent groups of the region of interest in normal and abnormal images. The two sets were extracted from ground truth images of the 89 fundus images. Fundus images were annotated images from the Standard Diabetic Retinopathy Database (DIARETDB1).Findings: The system provided an accuracy of 71% with a sensitivity of 75%.Recommendation: The current work may open an opportunity for improvement for future work. Other methods may be reached to raise the accuracy and the sensitivity of the system. Besides, the current system may be tested on a larger sample size to study such effects. Finally, the ease of the current method makes it faster in adoption in the appropriate diagnosis especially in the early stages.

1251P Exploring quantitative biomarkers from different tumour volumes for radiomics in lung cancer

Recent advances in medical imaging technology allow the use of advanced image analysis methods by the extraction of quantitative features from medical images to characterize tumor biology in order to provide informations that may be useful to guide therapies and predict survival. In our previous experience, a radiomics signature mixing semantic and image-based features was able to predict the reduction of the target volume during chemoradiation in LA-NSCLC. The aim of this study is to investigate which 3D ROI contain more quantitative information that can be exploit for a radiomic approach in lung cancer.

MRI-based radiomics in breast cancer: feature robustness with respect to inter-observer segmentation variability

Radiomics is an emerging field using the extraction of quantitative features from medical images for tissue characterization. While MRI-based radiomics is still at an early stage, it showed some promising results in studies focusing on breast cancer patients in improving diagnoses and therapy response assessment. Nevertheless, the use of radiomics raises a number of issues regarding feature quantification and robustness. Therefore, our study aim was to determine the robustness of radiomics features extracted by two commonly used radiomics software with respect to variability in manual breast tumor segmentation on MRI. A total of 129 histologically confirmed breast tumors were segmented manually in three dimensions on the first post-contrast T1-weighted MR exam by four observers: a dedicated breast radiologist, a resident, a Ph.D. candidate, and a medical student. Robust features were assessed using the intraclass correlation coefficient (ICC > 0.9). The inter-observer variability was evaluated by the volumetric Dice Similarity Coefficient (DSC). The mean DSC for all tumors was 0.81 (range 0.19–0.96), indicating a good spatial overlap of the segmentations based on observers of varying expertise. In total, 41.6% (552/1328) and 32.8% (273/833) of all RadiomiX and Pyradiomics features, respectively, were identified as robust and were independent of inter-observer manual segmentation variability.

Radiomics analysis using stability selection supervised component analysis for right-censored survival data

Radiomics is a newly emerging field that involves the extraction of massive quantitative features from biomedical images by using data-characterization algorithms. Distinctive imaging features identified from biomedical images can be used for prognosis and therapeutic response prediction, and they can provide a noninvasive approach for personalized therapy. So far, many of the published radiomics studies utilize existing out of the box algorithms to identify the prognostic markers from biomedical images that are not specific to radiomics data. To better utilize biomedical images, we propose a novel machine learning approach, stability selection supervised principal component analysis (SSSuperPCA) that identifies stable features from radiomics big data coupled with dimension reduction for right-censored survival outcomes.The proposed approach allows us to identify a set of stable features that are highly associated with the survival outcomes in a simple yet meaningful manner, while controlling the per-family error rate. We evaluate the performance of SSSuperPCA using simulations and real data sets for non-small cell lung cancer and head and neck cancer, and compare it with other machine learning algorithms.The results demonstrate that our method has a competitive edge over other existing methods in identifying the prognostic markers from biomedical imaging data for the prediction of right-censored survival outcomes.

Radiomics: A primer for the radiation oncologist

PurposeRadiomics are a set of methods used to leverage medical imaging and extract quantitative features that can characterize a patient's phenotype. All modalities can be used with several different software packages. Specific informatics methods can then be used to create meaningful predictive models. In this review, we will explain the major steps of a radiomics analysis pipeline and then present the studies published in the context of radiation therapy. MethodsA literature review was performed on Medline using the search engine PubMed. The search strategy included the search terms “radiotherapy”, “radiation oncology” and “radiomics”. The search was conducted in July 2019 and reference lists of selected articles were hand searched for relevance to this review. ResultsA typical radiomics workflow always includes five steps: imaging and segmenting, data curation and preparation, feature extraction, exploration and selection and finally modeling. In radiation oncology, radiomics studies have been published to explore different clinical outcome in lung (n=5), head and neck (n=5), esophageal (n=3), rectal (n=3), pancreatic (n=2) cancer and brain metastases (n=2). The quality of these retrospective studies is heterogeneous and their results have not been translated to the clinic. ConclusionRadiomics has a great potential to predict clinical outcome and better personalize treatment. But the field is still young and constantly evolving. Improvement in bias reduction techniques and multicenter studies will hopefully allow more robust and generalizable models.

Will traditional biopsy be substituted by radiomics and liquid biopsy for breast cancer diagnosis and characterisation?

The diagnosis of breast cancer currently relies on radiological and clinical evaluation, confirmed by histopathological examination. However, such approach has some limitations as the suboptimal sensitivity, the long turnaround time for recall tests, the invasiveness of the procedure and the risk that some features of target lesions may remain undetected, making re-biopsy a necessity. Recent technological advances in the field of artificial intelligence hold promise in addressing such medical challenges not only in cancer diagnosis, but also in treatment assessment, and monitoring of disease progression. In the perspective of a truly personalised medicine, based on the early diagnosis and individually tailored treatments, two new technologies, namely radiomics and liquid biopsy, are rising as means to obtain information from diagnosis to molecular profiling and response assessment, without the need of a biopsied tissue sample. Radiomics works through the extraction of quantitative peculiar features of cancer from radiological data, while liquid biopsy gets the whole of the malignancy's biology from something as easy as a blood sample. Both techniques hopefully will identify diagnostic and prognostic information of breast cancer potentially reducing the need for invasive (and often difficult to perform) biopsies and favouring an approach that is as personalised as possible for each patient. Nevertheless, such techniques will not substitute tissue biopsy in the near future, and even in further times they will require the aid of other parameters to be correctly interpreted and acted upon.

Abstract WMP101: Prediction of Clinical Outcome in Supratentorial Intracerebral Hemorrhage: Application of Baseline Ct Scan Radiomics Feature Extraction and Machine Learning Classifiers

Aim: Radiomics refers to automatic extraction of numerous quantitative features from medical images to supplement visual assessment. Machine-learning algorithms provide a suitable statistical methodology for devising predictive classifiers based on large radiomics datasets. We aimed to predict intracerebral hemorrhage (ICH) outcome by applying machine-learning classifiers to both clinical data and hematoma radiomics features. Methods: Patients enrolled in the Yale Longitudinal Study of ICH were included if they had (1) spontaneous supratentorial ICH, (2) baseline CT scan, (3) known admission Glasgow Coma Scale (GCS), and (4) 3-month modified Rankin Scale (mRS). A total of 1134 radiomics features related to the intensity, shape, texture, and waveform were extracted from manually segmented ICH lesions on baseline CT. Clinical variables were patients’ age, gender, GCS, presence of intraventricular hemorrhage, and thalamic ICH. We calculated the averaged receiver operating characteristics (ROC) area under curve (AUC) in outcome prediction among 100 repeats of 5-fold cross-validation (x500 iterations) for different combinations of feature selection and machine-learning algorithms. Results: A total of 119 ICH patients were included, of whom 60 had poor outcome (mRS ≥4). Among different combinations, lasso regression feature selection and partial least square (PLS) classification model yielded the highest accuracy in outcome prediction (Figure), with an averaged (95% confidence interval) ROC AUC of 0.86 (0.83 - 0.89) using clinical variables “only”, versus 0.92 (0.89 - 0.95) using combination of clinical variables and 54 radiomics features selected by lasso regression. Among radiomics features selected by lasso regression, ICH lesion flatness had the highest variable importance and was the only shape feature selected. Conclusion: Addition of ICH lesion radiomics to clinical variables using machine-learning models can improve outcome prediction.

An automatic computer vision pipeline for the in-line monitoring of freeze-drying processes

Infrared imaging sensors were proposed in the past years for the real-time monitoring and control of the Vacuum Freeze Drying (VFD) process. In order to extract reliable, real-time, and quantitative features from images collected by these sensors, a robust and automated image processing pipeline is required. Two major issues that have to be addressed concern the object detection and segmentation step, which locates and segments the objects whose temperature needs to be measured, and the tracking step, which is about following the movement of objects of interest to correlate information across subsequent images. Traditional intensity-based image analysis techniques are not reliable in this specific application, since the temperature, which is the intensity of a thermal picture, varies with time, whereas deep learning-based techniques are more robust to such changes.In this work, an object detector network based on a Faster Region Convolutional Neural Network (Faster R-CNN), and a Kernelized Correlation Filter (KCF) tracker were combined to monitor the VFD process of products contained in glass vials, which is a common scenario in the pharmaceutical domain. The object detector was trained on nine experimental acquisitions (7981 images) and tested on nine more (7201 images). Data augmentation methods consisting in the generation of synthetic sequences were used to effectively increase the performance while reducing the cost of data acquisition and annotation. The proposed technique achieved a recall and precision equal to 99.6 %, with execution time compatible with a real-time application. The localization of the vials was precise enough to measure the mean temperature with an acceptable error.

Radiomics in Breast Imaging from Techniques to Clinical Applications: A Review.

Recent advances in computer technology have generated a new area of research known as radiomics. Radiomics is defined as the high throughput extraction and analysis of quantitative features from imaging data. Radiomic features provide information on the gray-scale patterns, inter-pixel relationships, as well as shape and spectral properties of radiological images. Moreover, these features can be used to develop computational models that may serve as a tool for personalized diagnosis and treatment guidance. Although radiomics is becoming popular and widely used in oncology, many problems such as overfitting and reproducibility issues remain unresolved. In this review, we will outline the steps of radiomics used for oncology, specifically addressing applications for breast cancer patients and focusing on technical issues.

Multi-Type Interdependent Feature Analysis Based on Hybrid Neural Networks for Computer-Aided Diagnosis of Epidermal Growth Factor Receptor Mutations

The mutation status of the epidermal growth factor receptor (EGFR) is an important clinical reference indicator for lung cancer diagnosis and treatment. However, the extraction of effective discriminative features for non-invasive computer-aided EGFR mutation prediction still poses a big challenge. In this paper, multiple types of features are designed and analyzed to address this problem. These features include clinical features based on prior medical knowledge and quantitative image features extracted by convolutional neural networks (CNN). A long short-term memory (LSTM) network is also introduced to exploit the dependency between these feature types and then fuse them. In particular, a CNN is constructed to extract quantitative features of computed-tomography (CT) images. Furthermore, a LSTM is utilized to analyze the dependency between these clinical and CT image features and generate a new feature representation for computer-aided diagnosis. For samples from the same category, the proposed method deal with feature representation variabilities arising from interdependencies in multi-type features and patient specificity. The multiple feature types of the collected clinical data are used to assess the proposed approach and other relevant algorithms. Our results demonstrate that the multi-type dependency-based feature representation shows superior performance (Accuracy = 75%, AUC = 0.78) compared to single-type feature representations. The proposed method is reliable to apply for diagnosing of the EGFR mutation status.

Multi-planar 3D breast segmentation in MRI via deep convolutional neural networks.

Nowadays, Dynamic Contrast Enhanced-Magnetic Resonance Imaging (DCE-MRI) has demonstrated to be a valid complementary diagnostic tool for early detection and diagnosis of breast cancer. However, without a CAD (Computer Aided Detection) system, manual DCE-MRI examination can be difficult and error-prone. The early stage of breast tissue segmentation, in a typical CAD, is crucial to increase reliability and reduce the computational effort by reducing the number of voxels to analyze and removing foreign tissues and air. In recent years, the deep convolutional neural networks (CNNs) enabled a sensible improvement in many visual tasks automation, such as image classification and object recognition. These advances also involved radiomics, enabling high-throughput extraction of quantitative features, resulting in a strong improvement in automatic diagnosis through medical imaging. However, machine learning and, in particular, deep learning approaches are gaining popularity in the radiomics field for tissue segmentation. This work aims to accurately segment breast parenchyma from the air and other tissues (such as chest-wall) by applying an ensemble of deep CNNs on 3D MR data. The novelty, besides applying cutting-edge techniques in the radiomics field, is a multi-planar combination of U-Net CNNs by a suitable projection-fusing approach, enabling multi-protocol applications. The proposed approach has been validated over two different datasets for a total of 109 DCE-MRI studies with histopathologically proven lesions and two different acquisition protocols. The median dice similarity index for both the datasets is 96.60 % (±0.30 %) and 95.78 % (±0.51 %) respectively with p < 0.05, and 100% of neoplastic lesion coverage.

Study Progress of Radiomics With Machine Learning for Precision Medicine in Bladder Cancer Management.

Bladder cancer is a fatal cancer that happens in the genitourinary tract with quite high morbidity and mortality annually. The high level of recurrence rate ranging from 50 to 80% makes bladder cancer one of the most challenging and costly diseases to manage. Faced with various problems in existing methods, a recently emerging concept for the measurement of imaging biomarkers and extraction of quantitative features called “radiomics” shows great potential in the application of detection, grading, and follow-up management of bladder cancer. Furthermore, machine-learning (ML) algorithms on the basis of “big data” are fueling the powers of radiomics for bladder cancer monitoring in the era of precision medicine. Currently, the usefulness of the novel combination of radiomics and ML has been demonstrated by a large number of successful cases. It possesses outstanding strengths including non-invasiveness, low cost, and high efficiency, which may serve as a revolution to tumor assessment and emancipate workforce. However, for the extensive clinical application in the future, more efforts should be made to break down the limitations caused by technology deficiencies, inherent problems during the process of radiomic analysis, as well as the quality of present studies.

MRI-derived radiomics: methodology and clinical applications in the field of pelvic oncology.

Personalized medicine aims at offering optimized treatment options and improved survival for cancer patients based on individual variability. The success of precision medicine depends on robust biomarkers. Recently, the requirement for improved non-biologic biomarkers that reflect tumor biology has emerged and there has been a growing interest in the automatic extraction of quantitative features from medical images, denoted as radiomics. Radiomics as a methodological approach can be applied to any image and most studies have focused on PET, CT, ultrasound, and MRI. Here, we aim to present an overview of the radiomics workflow as well as the major challenges with special emphasis on the use of multiparametric MRI datasets. We then reviewed recent studies on radiomics in the field of pelvic oncology including prostate, cervical, and colorectal cancer.

The current state of radiomics in pancreatic disease

Radiomics is the extraction of quantitative features from cross-sectional imaging and analysis of these data for decision support. As the role of medical imaging in patient care has increased exponentially over the past decade, so too have the data associated with these studies, allowing for more robust and meaningful exploration. Diseases of the pancreas requiring complex decision making including pancreatic cancer, pancreatic cystic neoplasms, and pancreatic neuroendocrine neoplasms, are the active subjects of radiomics research, and patients and clinicians stand to benefit from development of this technology in the near term. Radiomics has already revealed novel insights into tumor phenotype, disease biology, and patient outcomes, and will continue to become a more powerful tool in the care of patients with pancreatic disease. Key words: Pancreatic diseases; Radiology; Radiomics; Tumor biology

Deep learning in head & neck cancer outcome prediction

Traditional radiomics involves the extraction of quantitative texture features from medical images in an attempt to determine correlations with clinical endpoints. We hypothesize that convolutional neural networks (CNNs) could enhance the performance of traditional radiomics, by detecting image patterns that may not be covered by a traditional radiomic framework. We test this hypothesis by training a CNN to predict treatment outcomes of patients with head and neck squamous cell carcinoma, based solely on their pre-treatment computed tomography image. The training (194 patients) and validation sets (106 patients), which are mutually independent and include 4 institutions, come from The Cancer Imaging Archive. When compared to a traditional radiomic framework applied to the same patient cohort, our method results in a AUC of 0.88 in predicting distant metastasis. When combining our model with the previous model, the AUC improves to 0.92. Our framework yields models that are shown to explicitly recognize traditional radiomic features, be directly visualized and perform accurate outcome prediction.