top of page
Writer's pictureSycai Technologies

Artificial Intelligence and Medical Imaging

Artificial intelligence (#AI), through Deep Learning (#DL), allows to analyse a great number of images and extract specific features. Convolutional neural networks are one of the most used systems for this purpose. These networks have a great capacity to recognize medical images. For this reason, AI have created great attention in its recent introduction in the medical field. These technologies are expected to solve many of the current problems in diagnosis and reduce miss rate of neoplasias due to its high accuracy in lesion recognition.

There are many fields in which AI techniques can be used to solve problems. For instance, AI has been successfully applied to gastrointestinal (GI) diseases. Recent research about AI software applied to endoscopy have showed that these techniques can improve accuracy in diagnosis of upper GI cancers such as Barrett´s adenocarcinoma or gastric cancer. AI also might have an especially important role in detecting and characterizing neoplastic colonic polyps, whose detection is crucial to reduce incidence of colorectal cancer.

A special mention must be done to the utility of these DL techniques to #radiological imaging. The technological improvements in this field,that have taken place in recent years, have led to the detection of many small lesions that were previously overlooked. However, this has also led to a larger amount of work for radiologists and other health professionals, who work in an already saturated system.

These radiological findings are especially important for pancreatic lesions, which have a malignant potential and must be properly identified and characterized. The incidence of pancreatic cystic lesions that are detected incidentally, is increasing dramatically. Most of those are small and probably benign, but the fact that some may develop or already harbour malignancy makes the correct diagnosis of this lesions crucial. In this fields, AI techniques, with its great capacity for image recognition, can play a particularly important role. This software may improve the results and shorten the amount of time devoted to the hard but crucial task of differentiating between the different kinds of cyst and its malignant potential.

Moreover, DL algorithms can also be seen as a tool to supervise some tasks and techniques to ensure that quality parameters are met, helping to “double check” the information provided for the clinician in a first approach, making it faster and reassuring the diagnosis.

AI can also be used for developing prognostic tools. It can use large amount of clinical data and mix it with information from imaging tests. This extremely large amount of information is impossible to handle for clinicians, but with the help of this tools, the analysis takes seconds and the conclusions can save time, tests and some time make the difference between life and death

So, AI can provide lots of opportunities to improve medical diagnosis, never replacing but complementing doctor's labour to make their jobs easier. And more important: Better diagnosis will improve patient’s life.




6 views0 comments

Comments


bottom of page