Artificial Intelligence in Radiology is here to stay!

The changes of Artificial Intelligence in Radiology are here to stay!

As mudanças da Inteligência Artificial na Radiologia vieram para ficar!

“If Artificial Intelligence (AI) is the new electricity, today’s coal is the data.”

It is known that the influence of AI in Radiology and diagnostic imaging is growing and its impact is transformative, as was the case with digital images and PACS systems (Picture Archiving and Communication System), which brought a new proposal to carry out reports from imaging exams, allowing large amounts of data to be stored.

What previously prevailed the feeling of fear in the face of novelty, such as, for example, inquiries about the end of the radiologist professional; today AI comes to give strength to PACS. This intelligence becomes a powerful tool that provides greater agility and assertiveness in the report, collaborating, more and more, for precision and personalized medicine. 

And along with AI, new terms such as machine learning, deep learning, neural network, CAD (computer-aided diagnosis), CBIR (contentbased image retrival), are introduced in the day-to-day debate on how to enhance these technologies in Radiology . 

Claudia da Costa Leite, from the Radiology and Oncology department of the Faculty of Medicine of the University of São Paulo (FMUSP), points out that there are several possibilities for the application of AI in Image, such as the “use of algorithms for patient flow, definition of imaging protocols, synthetic images, quality control, radiation dose control, computer-assisted diagnosis, automatic lesion detection, automatic interpretation of findings, radiomic, radiogenomic, among others ”.

That is, if before the AI ​​the imaging exams had quantitative and diagnostic value, now, the specialist has qualitative data about the patient in his hands, obtaining information that goes beyond, for example, detecting whether a tumor is malignant or benign. With the support of AI, radiology provides data such as the presence of mutations in a tumor, the chance of response to treatment, recurrence and patient survival. 

Marcel Koenigkam Santos, from the Center for Image Sciences and Medical Physics (CCIFM) at the Ribeirão Preto School of Medicine at the University of São Paulo (FMRP-USP), argues that, far beyond “lowering the stack” of tests, intelligence artificial in Radiology comes to reduce the time of action in urgent cases, streamline interpretation and issue of reports, increase the degree of confidence in diagnoses, make the analysis of images more objective and reproducible, offer more reliable prognostic information, assist in teaching and learning about imaging, and finally, definitively inserting radiology into the concept of precision medicine and multidisciplinary patient assessment. ” 

More than changing the work of professionals in the field, AI and its tools propose to change the perspective of work, that is, tasks and simple exams can be interpreted by AI, but the doctor’s role in verification / validation / decision therapy should not be threatened.

While the AI ​​provides accurate information for the professional, he must know how to integrate it with the patient’s clinical data, contributing in a more holistic way to the diagnosis. 

Therefore, AI provides this professional with greater attention to the patient and a greater field of investigation, since he will have in his hands data from several sources besides the image.