ILDeasy

ILDeasy incorporates different AI models capable of automatically processing chest HRCT scans and analyzing the information contained in the images. The solution can detect pulmonary abnormalities and segment the main radiological findings, displaying the identified areas directly on the images and providing a quantitative assessment of their volume. Findings that can be analyzed include reticulation, ground-glass opacity, honeycombing, emphysema, bronchiectasis and pulmonary cysts.

The tool can also assess radiological patterns and characterize the presence of fibrotic or non-fibrotic interstitial lung disease, providing the professional with information on the probabilities associated with the different patterns and highlighting the variables that are most relevant to the result obtained.

One of ILDeasy’s distinguishing features is its ability to transform the information contained in medical images into quantitative data that can complement the specialist’s assessment. Automated segmentation makes it possible to visualize and quantify different radiological findings, facilitating their review and comparison.

In addition, ILDeasy is designed to work with clinical information associated with the imaging study. The system can integrate, among other data, results from pulmonary function tests such as FVC, TLC and DLCO, radiology and pulmonology reports, and other relevant clinical information.

Combining medical imaging and clinical data provides a more comprehensive view of the patient and makes it easier for professionals to review the relevant information for their assessment within a single environment.

ILDeasy is designed to support, not replace, the assessment of healthcare professionals. The results generated by the algorithms are presented visually and in association with the images, allowing specialists to review the findings, verify the results and, when necessary, modify the segmentations performed by the AI. Usability testing with medical professionals at Hospital Universitario La Paz highlighted the clarity of the information and the usefulness of viewing the results and images together.

The solution also incorporates features designed to facilitate workflow, including artifact identification and correction, anomaly detection, finding segmentation, radiological pattern assessment and report generation.

The development of ILDeasy is part of the SEPI-IA project, an R&D initiative focused on developing a program for the analysis and simulation of the progression of interstitial lung diseases using Artificial Intelligence.

The project has involved collaboration with Hospital Universitario La Paz, which participated in the project design, the generation of validation datasets and the clinical and usability assessment of the prototype. According to Emilio Cuesta, Head of the Cardiothoracic Radiology Section at Hospital Universitario La Paz and Associate Professor of Radiology and Physical Medicine at Universidad Autónoma de Madrid, “real-world anonymized clinical data and medical images have been used, including HRCT scans and different clinical variables.”

The clinical validation dataset used at La Paz “included patients with different diagnoses and a high representation of interstitial abnormalities, with segmentations reviewed by specialists.”

This evaluation process has made it possible to assess the performance of the different components of the solution and advance the robustness of the AI models under conditions close to the clinical environment.

According to Dr. Maria Molina, Head of the Interstitial Lung Disease Functional Unit at Bellvitge University Hospital and Associate Professor at the University of Barcelona, and Gold Medal recipient of the European Respiratory Society, ILDeasy represents “a step forward in the application of Artificial Intelligence to the analysis of interstitial lung diseases, providing healthcare professionals with objective, quantitative and visual information from medical images to facilitate their interpretation and support patient follow-up.”