Digital Diagnostics

BACKGROUND: Optical coherence tomography is a modern high-tech, insightful approach to detecting pathologies of the retina and preretinal layers of the vitreous body. However, the description and interpretation of study findings require advanced qualifications and special training of ophthalmologists and are highly time-consuming for both the doctor and the patient. Moreover, mathematical models based on artificial neural networks now allow for the automation of many image processing tasks. Therefore, addressing the issues of automated classification of optical coherence tomography images using deep learning artificial neural network models is crucial.

AIM: To develop architectures of mathematical (computer) models based on deep learning of convolutional neural networks for the classification of retinal optical coherence tomography images; to compare the results of computational experiments conducted using Python tools in Google Colaboratory with single-model and multimodel approaches, and evaluate classification accuracy; and to determine the optimal architecture of models based on artificial neural networks, as well as the values of the hyperparameters used.

MATERIALS AND METHODS: The original dataset included >2,000 anonymized optical coherence tomography images of real patients, obtained directly from the device with a resolution of 1,920×969×24 BPP. The number of image classes was 12. To create the training and validation datasets, a subject area of 1,100×550×24 BPP was “cut out”. Various approaches were studied: the possibility of using pretrained convolutional neural networks with transfer learning, techniques for resizing and augmenting images, and various combinations of the hyperparameters of models based on artificial neural networks. When compiling a model, the following parameters were used: Adam optimizer, categorical_crossentropy loss function, and accuracy. All technological operations involving images and models based on artificial neural networks were performed using Python language tools in Google Colaboratory.

RESULTS: Single-model and multimodel approaches to the classification of retinal optical coherence tomography images were developed. Computational experiments on the automated classification of such images obtained from a DRI OCT Triton tomograph using various architectures of models based on artificial neural networks showed an accuracy of 98–100% during training and validation, and 85% during an additional test, which is a satisfactory result. The optimal architecture of the model based on an artificial neural network, a six-layer convolutional network, was selected, and the values of its hyperparameters were determined.

CONCLUSION: Deep training of convolutional neural network models with various architectures, as well as their validation and testing, resulted in satisfactory classification accuracy of retinal optical coherence tomography images. These findings can be used in decision support systems in ophthalmology.

BACKGROUND: In recent years, it has been critical to modify training methods and programs in numerous areas, including ultrasound diagnosis, with the use of various virtual and simulation devices. Because practical experience with employing such technologies in the teaching process is limited, there are few original studies on the subject in Russian and foreign literature.

AIM: To determine the possibilities and algorithms for using a virtual ultrasound simulator to train ultrasound specialists based on the results of related work, as well as to assess the benefits and drawbacks of simulators in comparison to conventional teaching methods.

MATERIALS AND METHODS: The results of using the Vimedix 3.2 virtual simulator in the teaching process were analyzed. Simulations of abdominal ultrasound, transthoracic echocardiography, and triplex scanning of major vessels were performed. The study included 26 residents specializing in ultrasound diagnosis and 37 physicians undergoing professional retraining courses.

RESULTS: Using a virtual simulator during the initial stage of training helps eliminate many of the challenges that residents and trainees encounter in clinical practice. The use of a simulator during testing appears to be less beneficial than during a practical examination employing ultrasound scanners and real patients.

CONCLUSIONS: The use of a simulator at the initial stage is advisable to get familiar with this research methodology. It is recommended to develop and use of additional teaching materials and programs in training. The advantages of the virtual simulator include ease of use during the initial stages of training, a steep learning curve, and the availability of an extensive database of pathological cases. The identified noncritical shortcomings require correction during further training in the clinic.

BACKGROUND: Aortic aneurysms are known as “silent killers” because this potentially fatal condition can be asymptomatic. The annual incidence of thoracic aortic aneurysms and ruptures is approximately 10 and 1.6 per 100,000 individuals, respectively. The mortality rate for ruptured aneurysms ranges from 94% to 100%. Early diagnosis and treatment can be life-saving. Artificial intelligence technologies can significantly improve diagnostic accuracy and save the lives of patients with thoracic aortic aneurysms.

AIM: This study aimed to assess the efficacy of artificial intelligence technologies for detecting thoracic aortic aneurysms on chest computed tomography scans, as well as the possibility of using artificial intelligence as a clinical decision support system for radiologists during the primary interpretation of radiological images.

MATERIALS AND METHODS: The results of using artificial intelligence technologies for detecting thoracic aortic aneurysms on non-contrast chest computed tomography scans were evaluated. A sample of 84,405 patients >18 years old was generated, with 86 cases of suspected thoracic aortic aneurysms based on artificial intelligence data selected and retrospectively assessed by radiologists and vascular surgeons. To assess the age distribution of the aortic diameter, an additional sample of 968 cases was randomly selected from the total number.

RESULTS: In 44 cases, aneurysms were initially identified by radiologists, whereas in 31 cases, aneurysms were not detected initially; however, artificial intelligence aided in their detection. Six studies were excluded, and five studies had false-positive results. Artificial intelligence aids in detecting and highlighting aortic pathological changes in medical images, increasing the detection rate of thoracic aortic aneurysms by 41% when interpreting chest computed tomography scans. The use of artificial intelligence technologies for primary interpretations of radiological studies and retrospective assessments is advisable to prevent underdiagnosis of clinically significant pathologies and improve the detection rate of pathological aortic enlargement. In the additional sample, the incidence of thoracic aortic dilation and thoracic aortic aneurysms in adults was 14.5% and 1.2%, respectively. The findings also revealed an age-dependent diameter of the thoracic aorta in both men and women.

CONCLUSION: The use of artificial intelligence technologies in the primary interpretation of chest computed tomography scans can improve the detection rate of clinically significant pathologies such as thoracic aortic aneurysms. Expanding retrospective screening based on chest computed tomography scans using artificial intelligence can improve the diagnosis of concomitant pathologies and prevent negative consequences.

BACKGROUND: The Order of the Ministry of Health of Russia “On Approval of the Procedure for the Provision of Medical Care to the Adult Population for Diseases of the Eye, Appendages, and Orbit” provides for equipping consultation and diagnostic departments of outpatient clinics with optical coherence tomographs. However, case follow-up in of patients with retinal pathology is most commonly performed in ophthalmology centers, limiting treatment accessibility for patients with primary (newly diagnosed) pathologies requiring immediate treatment initiation. The available approach requires modification and intensification, including the use of artificial intelligence technologies.

AIM: To develop methodological foundations for organizing follow-up care for patients with posterior segment eye diseases using an artificial intelligence-based clinical decision support system.

MATERIALS AND METHODS: The existing regulatory framework was analyzed based on the Constitution of the Russian Federation, federal laws, by-law framework, and judicial practice. A structured medical document describing an optical coherence tomography image was created using an expert method: a survey of 100 ophthalmologists with an appropriate education level, including additional professional training, engaged specialized medical care for patients with posterior segment eye diseases was performed.

RESULTS: Using an expert method, 123 binary features were selected to describe the structure of the macular area of the retina under normal and pathological conditions, with 26 features identified as predictors of a worsening clinical course of the disease.

CONCLUSION: The proposed classifier enabled the creation and training of a medical decision support system based on 60,000 medical images, which, as an information service, without making a diagnosis, can change the case follow-up process. Routing of patients is a primary service of the proposed system. If the clinical picture shows signs of deterioration, a referral to an ophthalmology center is considered to assess the course of the disease and provide specialized services, including high-tech medical care.

BACKGROUND: The incidence of circulatory system diseases in the Russian Federation has been steadily increasing during the last two decades, growing 2,047 times between 2000 and 2019. Vascular calcification involves the deposition of calcium salts in the artery wall, which leads to vascular wall remodeling. X-ray imaging is the gold standard for diagnosing of vascular calcification. However, because of the need to process an increasing amount of data in a shorter period of time, the number of diagnostic errors inevitably increases, and work efficiency inevitably decreases. The active development and introduction of artificial intelligence into clinical practice have created opportunities for specialists to address these issues.

AIM: To analyze the national and international literature on the use of artificial intelligence in the diagnosis of various vascular calcifications, summarize the prognostic value of vascular calcification, and evaluate aspects that prevent the diagnosis of vascular calcification without using artificial intelligence.

MATERIALS AND METHODS: A search was performed in PubMed, Web of Science, Google Scholar, and eLibrary. The search was conducted using the following keywords: artificial intelligence, machine learning, vascular calcification, and their analogues in Russian. The search covered the period from inception till July 2023.

RESULTS: The studies included in the review compared the diagnostic abilities of clinicians and artificial intelligence using the same images , with subsequent assessment of the accuracy, speed, and other parameters. The sites of vascular calcification varied, resulting in differences in their prognostic value.

CONCLUSION: Artificial intelligence has proven to be effective in the diagnosis of vascular calcification. In addition to improved accuracy and efficiency, the level of detail is superior to manual diagnosis methods. Artificial intelligence has advanced to the point that imaging specialists can automatically detect vascular calcification. Artificial intelligence can contribute to the successful development of X-ray imaging in the future.

The article presents a selective literature review on the use of computer vision algorithms for the diagnosis of liver and kidney neoplasms and urinary stones using computed tomography images of the abdomen and retroperitoneal space. The review included articles published between January 1, 2020, and April 24, 2023. Pixel-based algorithms showed the greatest diagnostic accuracy parameters for segmenting the liver and its neoplasms (accuracy, 99.6%; Dice similarity coefficient, 0.99). Voxel-based algorithms were superior at classifying liver neoplasms (accuracy, 82.5%). Pixel- and voxel-based algorithms fared equally well in segmenting kidneys and their neoplasms, as well as classifying kidney tumors (accuracy, 99.3%; Dice similarity coefficient, 0.97). Computer vision algorithms can detect urinary stones measuring 3 mm or larger with a high degree of accuracy of up to 93.0%. Thus, existing computer vision algorithms not only effectively detect liver and kidney neoplasms and urinary stones but also accurately determine their quantitative and qualitative characteristics. Evaluating voxel data improves the accuracy of neoplasm type determination since the algorithm analyzes the neoplasm in three dimensions rather than only the plane of one slice.