Computer Aided Diagnosis of Systemic Lupus Erythematosus Through Antinuclear Antibody Image Pattern Recognition

This study focuses on diagnosis of systemic lupus erythematosus through pattern recognition of antinuclear Antibodies. SLE is an autoimmune disease. In the human body the immune system attacks its own healthy cells, this results in widespread inflammation and tissue damage and affects the organs like skin, brain, lungs, kidneys, and blood vessels. ANA test (screening test) detects antinuclear antibodies in your blood. After this ANA screening test, Investigation called the anti-dsDNA (anti-double stranded DNA) test is conducted for confirmation of SLE. This study is to replace this anti-dsDNA confirmatory test with pattern recognition of anti-nuclear antibodies which is more economic and time savings. Majorly in three patterns SLE probability is high. They are: homogenous pattern, cytoplasmic pattern, speckled pattern. From the image of the ANA screening test, the single cell pattern image is cropped and a watershed algorithm is used to get segmentation of the nucleus, followed by extracting of shape features from the segmented image. With these shape features we can classify the patterns of ANA cells. This pattern recognition is used for pre-screening. Further, along with appropriate patterns of SLE, clinical biomarkers, blood biomarkers, and organ specific biomarkers are also considered and get accurate diagnosis of SLE