Preliminary results of the project A.I.D.A. (Auto Immunity: Diagnosis Assisted by computer)

摘要

In this paper, are presented the preliminary results of the A.I.D.A. (Auto Immunity: Diagnosis Assisted by computer) project which is developed in the frame of the cross-border cooperation Italy-Tunisia. According to the main objectives of this project, a database of interpreted Indirect ImmunoFluorescence (IIF) images on HEp 2 cells is being collected thanks to the contribution of Italian and Tunisian experts involved in routine diagnosis of autoimmune diseases. Through exchanging images and double reporting; a Gold Standard database, containing around 1000 double reported IIF images with different patterns including negative tests, has been settled. This Gold Standard database has been used for optimization of a computing solution (CAD-Computer Aided Detection) and for assessment of its added value in order to be used along with an immunologist as a second reader in detection of auto antibodies for autoimmune disease diagnosis. From the preliminary results obtained, the CAD appeared more powerful than junior immunologists used as second readers and may significantly improve their efficacy.