Texture analysis has arisen as a tool to explore the amount of data contained in images that cannot be explored by humans visually. Radiomics is a method that extracts a large number of features from ...radiographic medical images using data-characterisation algorithms. These features, termed radiomic features, have the potential to uncover disease characteristics. The goal of both radiomics and texture analysis is to go beyond size or human-eye based semantic descriptors, to enable the non-invasive extraction of quantitative radiological data to correlate them with clinical outcomes or pathological characteristics. In the latest years there has been a flourishing sub-field of radiology where texture analysis and radiomics have been used in many settings. It is difficult for the clinical radiologist to cope with such amount of data in all the different radiological sub-fields and to identify the most significant papers. The aim of this review is to provide a tool to better understand the basic principles underlining texture analysis and radiological data mining and a summary of the most significant papers of the latest years.
Abstract Purpose Fatty Liver Disease (FLD) is one of the most common diseases in liver. Early detection can improve the prognosis considerably. Using ultrasound for FLD detection is highly desirable ...due to its non-radiation nature, low cost and easy use. However, the results can be slow and ambiguous due to manual detection. The lack of computer trained systems leads to low image quality and inefficient disease classification. Thus, the current study proposes novel, accurate and reliable detection system for the FLD using computer-based training system. Materials and Methods One hundred twenty-four ultrasound sample images were selected retrospectively from a database of 62 patients consisting of normal and cancerous. The proposed training system was generated offline parameters using training liver image database. The classifier applied transformation parameters to an online system in order to facilitate real-time detection during the ultrasound scan. The system utilized six sets of features (a total of 128 features), namely Haralick, Basic geometric, Fourier transform, Discrete Cosine Transform, Gupta transform and Gabor transform. These features were extracted for both offline training and online testing. Levenberg-Marquardt Back Propagation Network (BPN) classifier was used to classify the liver disease into normal and abnormal categories. Results Random Partitioning approach was adapted to evaluate the classifier performance and compute its accuracy. Utilizing all the six sets of 128 features, the computer aided diagnosis (CAD) system achieved classification accuracy of 97.58%. Furthermore, the four performance metrics consisting of sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) realized 98.08%, 97.22%, 96.23%, and 98.59%, respectively. Conclusion The proposed system was successfully able to detect and classify the FLD. Furthermore, the proposed system was benchmarked against previous methods. The comparison established an advanced set of features in the Levenberg-Marquardt Back Propagation Network reports a significant improvement compared to the existing techniques.
A World Health Organization (WHO) Feb 2018 report has recently shown that mortality rate due to brain or central nervous system (CNS) cancer is the highest in the Asian continent. It is of critical ...importance that cancer be detected earlier so that many of these lives can be saved. Cancer grading is an important aspect for targeted therapy. As cancer diagnosis is highly invasive, time consuming and expensive, there is an immediate requirement to develop a non-invasive, cost-effective and efficient tools for brain cancer characterization and grade estimation. Brain scans using magnetic resonance imaging (MRI), computed tomography (CT), as well as other imaging modalities, are fast and safer methods for tumor detection. In this paper, we tried to summarize the pathophysiology of brain cancer, imaging modalities of brain cancer and automatic computer assisted methods for brain cancer characterization in a machine and deep learning paradigm. Another objective of this paper is to find the current issues in existing engineering methods and also project a future paradigm. Further, we have highlighted the relationship between brain cancer and other brain disorders like stroke, Alzheimer's, Parkinson's, and Wilson's disease, leukoriaosis, and other neurological disorders in the context of machine learning and the deep learning paradigm.
•Deep Learning (DL) technique is applied for detection of hypoechoic FLD and stratification of normal and abnormal US liver images under the class of Symtosis.•This paper provides comprehensive ...analysis and comparison of three ML-based classification methodologies: namely, support vector machines, extreme learning machines and deep learning.•A specialized deep learning operation called inception is comprehensively investigated.
Fatty Liver Disease (FLD) - a disease caused by deposition of fat in liver cells, is predecessor to terminal diseases such as liver cancer. The machine learning (ML) techniques applied for FLD detection and risk stratification using ultrasound (US) have limitations in computing tissue characterization features, thereby limiting the accuracy.
Under the class of Symtosis for FLD detection and risk stratification, this study presents a Deep Learning (DL)-based paradigm that computes nearly seven million weights per image when passed through a 22 layered neural network during the cross-validation (training and testing) paradigm. The DL architecture consists of cascaded layers of operations such as: convolution, pooling, rectified linear unit, dropout and a special block called inception model that provides speed and efficiency. All data analysis is performed in optimized tissue region, obtained by removing background information. We benchmark the DL system against the conventional ML protocols: support vector machine (SVM) and extreme learning machine (ELM).
The liver US data consists of 63 patients (27 normal/36 abnormal). Using the K10 cross-validation protocol (90% training and 10% testing), the detection and risk stratification accuracies are: 82%, 92% and 100% for SVM, ELM and DL systems, respectively. The corresponding area under the curve is: 0.79, 0.92 and 1.0, respectively. We further validate our DL system using two class biometric facial data that yields an accuracy of 99%.
DL system shows a superior performance for liver detection and risk stratification compared to conventional machine learning systems: SVM and ELM.
Cerebral small vessel disease (cSVD) has a crucial role in lacunar stroke and brain hemorrhages and is a leading cause of cognitive decline and functional loss in elderly patients. Based on ...underlying pathophysiology, cSVD can be subdivided into amyloidal and non-amyloidal subtypes. Genetic factors of cSVD play a pivotal role in terms of unraveling molecular mechanism. An important pathophysiological mechanism of cSVD is blood-brain barrier leakage and endothelium dysfunction which gives a clue in identification of the disease through circulating biological markers. Detection of cSVD is routinely carried out by key neuroimaging markers including white matter hyperintensities, lacunes, small subcortical infarcts, perivascular spaces, cerebral microbleeds, and brain atrophy. Application of neural networking, machine learning and deep learning in image processing have increased significantly for correct severity of cSVD. A linkage between cSVD and other neurological disorder, such as Alzheimer's and Parkinson's disease and non-cerebral disease, has also been investigated recently. This review draws a broad picture of cSVD, aiming to inculcate new insights into its pathogenesis and biomarkers. It also focuses on the role of deep machine strategies and other dimensions of cSVD by linking it with several cerebral and non-cerebral diseases as well as recent advances in the field to achieve sensitive detection, effective prevention and disease management.
•Holistic information in COVID-19 patients with imaging and non-imaging data can help predict patient outcome in terms of the need for ICU admission.•Validation of model over multiple sites is ...important to establish its generalizablity.•Both volume and radiomic features of pulmonary opacities are key to quantifying the extent of lung involvement.
Display omitted
While image analysis of chest computed tomography (CT) for COVID-19 diagnosis has been intensively studied, little work has been performed for image-based patient outcome prediction. Management of high-risk patients with early intervention is a key to lower the fatality rate of COVID-19 pneumonia, as a majority of patients recover naturally. Therefore, an accurate prediction of disease progression with baseline imaging at the time of the initial presentation can help in patient management. In lieu of only size and volume information of pulmonary abnormalities and features through deep learning based image segmentation, here we combine radiomics of lung opacities and non-imaging features from demographic data, vital signs, and laboratory findings to predict need for intensive care unit (ICU) admission. To our knowledge, this is the first study that uses holistic information of a patient including both imaging and non-imaging data for outcome prediction. The proposed methods were thoroughly evaluated on datasets separately collected from three hospitals, one in the United States, one in Iran, and another in Italy, with a total 295 patients with reverse transcription polymerase chain reaction (RT-PCR) assay positive COVID-19 pneumonia. Our experimental results demonstrate that adding non-imaging features can significantly improve the performance of prediction to achieve AUC up to 0.884 and sensitivity as high as 96.1%, which can be valuable to provide clinical decision support in managing COVID-19 patients. Our methods may also be applied to other lung diseases including but not limited to community acquired pneumonia. The source code of our work is available at https://github.com/DIAL-RPI/COVID19-ICUPrediction.
The introduction of machine learning in digital pathology has deeply impacted the field, especially with the advent of whole slide image (WSI) analysis. In this review, we tried to elucidate the role ...of machine learning algorithms in diagnostic precision, efficiency, and the reproducibility of the results. First, we discuss some of the most used tools, including QuPath, HistoQC, and HistomicsTK, and provide an updated overview of machine learning approaches and their application in pathology. Later, we report how these tools may simplify the automation of WSI analyses, also reducing manual workload and inter-observer variability. A novel aspect of this review is its focus on open-source tools, presented in a way that may help the adoption process for pathologists. Furthermore, we highlight the major benefits of these technologies, with the aim of making this review a practical guide for clinicians seeking to implement machine learning-based solutions in their specific workflows. Moreover, this review also emphasizes some crucial limitations related to data quality and the interpretability of the models, giving insight into future directions for research. Overall, this work tries to bridge the gap between the more recent technological progress in computer science and traditional clinical practice, supporting a broader, yet smooth, adoption of machine learning approaches in digital pathology.
IL-6 (interleukin-6) has important roles in atherosclerosis pathophysiology. To determine if anti-IL-6 therapy warrants evaluation as an adjuvant stroke prevention strategy in patients with carotid ...atherosclerosis, we tested whether circulating IL-6 levels predict carotid plaque severity, vulnerability, and progression in the prospective population-based CHS (Cardiovascular Health Study).
Duplex carotid ultrasound was performed at baseline and 5 years. Baseline plaque severity was scored 0 to 5 based on North American Symptomatic Carotid Endarterectomy Trial grade of stenosis. Plaque vulnerability at baseline was the presence of markedly irregular, ulcerated, or echolucent plaques. Plaque progression at 5 years was a ≥1 point increase in stenosis severity. The relationship of baseline plasma IL-6 levels with plaque characteristics was modeled using multivariable linear (severity) or logistic (vulnerability and progression) regression. Risk factors of atherosclerosis were included as independent variables. Stepwise backward elimination was used with
>0.05 for variable removal. To assess model stability, we computed the
-value or minimum strength of association (odds ratio scale) that unmeasured confounders must have with log IL-6 and the outcome to suppress the association. We performed internal validation with 100 bootstrap samples.
There were 4334 participants with complete data (58.9% women, mean age: 72.7±5.1 years), including 1267 (29.2%) with vulnerable plaque and 1474 (34.0%) with plaque progression. Log IL-6 predicted plaque severity (β=0.09,
=1.3×10
), vulnerability (OR, 1.21 95% CI, 1.05-1.40;
=7.4×10
,
-value=1.71), and progression (OR, 1.44 95% CI, 1.23-1.69,
=9.1×10
,
-value 2.24). In participants with >50% predicted probability of progression, mean log IL-6 was 0.54 corresponding to 2.0 pg/mL. Dichotomizing IL-6 levels did not affect the performance of prediction models.
Circulating IL-6 predicts carotid plaque severity, vulnerability, and progression. The 2.0 pg/mL cutoff could facilitate the selection of individuals that would benefit from anti-IL-6 drugs for stroke prevention.
Radiogenomics, a combination of “Radiomics” and “Genomics,” using Artificial Intelligence (AI) has recently emerged as the state-of-the-art science in precision medicine, especially in oncology care. ...Radiogenomics syndicates large-scale quantifiable data extracted from radiological medical images enveloped with personalized genomic phenotypes. It fabricates a prediction model through various AI methods to stratify the risk of patients, monitor therapeutic approaches, and assess clinical outcomes. It has recently shown tremendous achievements in prognosis, treatment planning, survival prediction, heterogeneity analysis, reoccurrence, and progression-free survival for human cancer study. Although AI has shown immense performance in oncology care in various clinical aspects, it has several challenges and limitations. The proposed review provides an overview of radiogenomics with the viewpoints on the role of AI in terms of its promises for computational as well as oncological aspects and offers achievements and opportunities in the era of precision medicine. The review also presents various recommendations to diminish these obstacles.
The pathogenesis of neuropsychiatric systemic lupus erythematosus (NPSLE) is widely unknown, and the role of autoantibodies is still undetermined.
To identify brain-reactive autoantibodies possibly ...related to NPSLE, immunofluorescence (IF) and transmission electron microscopy (TEM) on rat and human brains were performed. ELISA was used to reveal the presence of known circulating autoantibodies, while western blot (WB) was applied to characterize potential unknown autoantigen(s).
We enrolled 209 subjects, including patients affected by SLE (n=69), NPSLE (n=36), Multiple Sclerosis (MS, n=22), and 82 age- and gender-matched healthy donors (HD). Autoantibody reactivity by IF was observed in almost the entire rat brain (cortex, hippocampus, and cerebellum) using sera from NPSLE and SLE patients and was virtually negative in MS and HD. NPSLE showed higher prevalence (OR 2.4; p = 0.047), intensity, and titer of brain-reactive autoantibodies than SLE patients. Most of the patient sera with brain-reactive autoantibodies (75%) also stained human brains. Double staining experiments on rat brains mixing patients' sera with antibodies directed against neuronal (NeuN) or glial markers showed autoantibody reactivity restricted to NeuN-containing neurons. Using TEM, the targets of brain-reactive autoantibodies were located in the nuclei and, to a lesser extent, in the cytoplasm and mitochondria. Given the high degree of colocalization between NeuN and brain-reactive autoantibodies, we assumed NeuN was a possible autoantigen. However, WB analysis with HEK293T cell lysates expressing or not expressing the gene encoding for NeuN protein (RIBFOX3) showed that patients' sera carrying brain-reactive autoantibodies did not recognize the NeuN corresponding band size. Among the panel of NPSLE-associated autoantibodies (e.g., anti-NR2, anti-P-ribosomal protein, antiphospholipid) investigated by ELISA assay, only the anti-β2-glycoprotein-I (aβ2GPI) IgG was exclusively found in those sera containing brain-reactive autoantibodies.
In conclusion, SLE and NPSLE patients possess brain-reactive autoantibodies but with higher frequency and titers found in NPSLE patients. Although many target antigens of brain-reactive autoantibodies are still undetermined, they likely include β2GPI.