Project Ailanton

The main objective of the project is the development of an Artificial Intelligence (AI) solution applied to the E-Health sector, interconnected in the operational flow of Radiology so that the search, identification and classification of pulmonary nodules, on Computed Tomography (CT) examinations of the chest, are carried out automatically and their result is presented to radiologists without any slowdown in their normal operations. The purpose of the Project is to validate and test a system of multiple CNNs (convolutional neural networks), developed by Sferanet Srl, capable of localizing and classifying pulmonary nodules in CT images of the chest, having the ability to understand the rules that from time to time are followed by the Deep Learning system for the analysis and classification of lung lesions. The use of advanced technologies, based on the use of cloud-based systems and the transfer and analysis of medical images in DICOM format produced by CT equipment, are the baseline that make the model easily exportable.

Another goal is to automatically produce a graphical Structured Report summarizing the characteristics of all pulmonary nodules found and suggesting recommendations to follow, based on the Fleischner Society guidelines (2017). This may be possible by integrating the system with the RIS/PACS System present at the Fondazione Policlinico Universitario Campus Bio-Medico facility in Rome.

This is coupled with two other important components, which are those inherent to the design and implementation of a prototype, completely innovative to date, in support of percutaneous Cryoablation of already established pulmonary neoplastic lesions and that of the Medical Image Annotation System for training convolutional neural networks also in application areas other than that of the present project.

The Project, already underway, will be completed by May 2025.