Intelligent diagnostics of intervertebral hernia lumbar with telemedicine

An important stage in predicting the resorption of intervertebral hernias is the identification of parameters describing the condition of the lumbar spine based on magnetic resonance imaging (MRI). The article presents a software solution for intelligent recognition of intervertebral discs, muscles, and hernias on MRI images of the lumbosacral spine, detection, and prediction with a high probability of hernia resorption using neural network methods. Purpose: the study aims at developing intelligent software for diagnosing lumbar intervertebral hernias based on MRI data with telemedicine access. Materials and methods. An artificial neural network has been implemented as a decision support tool for predicting the probability of hernia resorption to draw up a treatment plan or confirm the need for surgery based on MRI image recognition. Results. To train the neural network, pre-labeled MRI images with intervertebral hernias with and without resorption in patients of different ages were used, taking into account the morphometric parameters: muscle size, adipose tissue, the presence of a healthy yellow ligament. To increase the volume of the training data sample, the data augmentation method was used. The unique data set for the training sample was 2336 images. Findings. The developed software allows you to recognize the parameters of the lumbar spine according to MRI data using intelligent computer vision technologies. This result is useful for automating decision support for predicting hernia resorption and choosing treatment tactics. An important result of the development for practice is the implementation of telemedicine access to the software using a Telegram bot.