Lesion scannet: Artificial intelligence revolutionizes appendicitis diagnosis with 99% accuracy

Lesion scannet: Artificial intelligence revolutionizes appendicitis diagnosis with 99% accuracy

Research offers new knowledge to diagnose acute appendicitis, sudden inflammation of the worm process, which causes symptoms such as abdominal pain, vomiting and fever. Computer tomography (CT) is often used for diagnosis, but this can be difficult due to the location of the appendix and the complex anatomy of the colon. A new tool called "Lesionscannet" was developed to support radiologists in the precise detection of this disease.

Lesion scannet is a progressive model based on a deep neuronal network, more precisely a Convolutional Neural Network (CNN). This model analyzes CT images to identify signs of appendicitis with high accuracy. 2400 CT images were collected for the development of the model, which are a solid basis for training the network. An accuracy of 99% in the detection of appendicitis that reaches the model on test images is impressive.

The potential effects of this development could be significant. The use of such a powerful model could make the diagnosis of appendix inflammation more precise and faster, which in turn could lead to more efficient treatment. If this tool develops, it could possibly also be used in other medical areas, which makes it a versatile aid in the diagnostic image.

Basic terms and concepts

• acute appendicitis : quick inflammation of the worm process (appendix) with symptoms such as abdominal pain and fever.
• computer tomography (CT) : A medical imaging process that provides the body's cross -sectional images to diagnose diseases and injuries.
• Convolutional Neural Network (CNN) : A kind of artificial neural network that is particularly good when analyzing image data.
• Lesioncannet : A specific method based on CNN that was developed to recognize signs of diseases such as appendicitis on CT images.
• parameter : numbers used in a model to make predictions; The easier the model, the fewer parameters are needed.

abbreviations

• ct : computer tomography
• cnn : Convolutional Neural Network (folding neural network)

Lesion scannet: Precision in the diagnosis of acute appendicitis

The current research presents the Lesion Cannet model, a new Convolutional Neural Network (CNN) especially for computer-aided detection of acute appendicitis. This research work in particular emphasizes the precision and efficiency of this model in the processing of CT images, whereby it overcomes challenges such as anatomical variabilities of the location of the appendix.

methodology and model structure

• Data record: The study was based on an extensive data set, consisting of 2400 ct scan images, which was collected by the general surgery department at the Kanuni Sultan Süleyman training and research hospital in Istanbul, Turkey.
• model architecture: Lesion scannet is a lightweight model with only 765,000 parameters. It integrates several dual-core blocks, each of which includes folding, expansion, separable folding layers and skip connections.
• dualcernel blocks: These blocks process input images on two paths: one uses 3 × 3 filters, while the other 1 × 1 filter uses. This structure maximizes the identification efficiency by different filter sizes.

performance and generalization skills

• accuracy: The Lesion Cannet model achieved a considerable accuracy of 99% on the test data set. This result exceeds the performance of established benchmark-deep learning models for the detection of appendicitis.
• generalization ability: The effectiveness of lesion scannet was also supported by tests on a breast X-ray image for the detection of pneumonia and covid-19.

conclusions and future applications

The study shows that Lesion Cannet is superior in both its specificity and its efficiency in the detection of acute appendicitis. Despite the small number of parameters, the model offers a robust performance, which suggests its applicability to other medical fields, such as the detection of respiratory diseases.

The success of Lesioncannet can serve as the basis for future research and development work in medical image processing, with the potential to increase precision in the diagnosis of other diseases.

Source of research: https://pubmed.nlm.nih.gov/39654693