Automated Toxicology Screening System using Artificial Intelligence based

Automated Toxicology Screening System using Artificial Intelligence based Self Learning Imaging System with Pattern Identification Dr. Aparna Joshi Chief Medical Officer, Optra Systems (www. optrasystems. com)

WHY? Screening Strategies for Histopathological evaluation in Toxicology One of the major problems facing the research industry

Automated Toxicology Screening System Ø Design based on histological morphology of various tissues Ø In accordance with pathological lesions Ø To flag abnormal patterns in high throughput scenario Ø To “Learn on the Fly” without programmatic changes Ø To dynamically adapt to various boundary conditions Ø Personalizable and Customizable Ø Acceptability on variety of image acquisition devices Ø Can qualify images as Accept / Reject prior to screening “An artificial intelligence based self learning imaging system”

Optra. Assays™ (Image analysis libraries for histopathology in toxicological analysis) Ø Ø Ø Ø Ø Angiogenesis. dll Hemorrhage. dll Pan_Morphometry. dll Cellularity assessment. dll Necrosis. dll Infarction. dll Fibrosis. dll Mitoses. dll Micrometastasis. dll Adenomyosis. dll

Subcutaneous implantation of a cotton thread in the dorsum of rabbits (Oryctolagus cuniculus ) induced granulation tissue formation with angiogenesis. Automated detection of same. Pre Image Analysis Post Image Analysis Red blood cells Intravascular lumen Endothelial cells

Optra. Assays™ (Image analysis libraries for histopathology in toxicological analysis) Ø Ø Ø Ø Ø Angiogenesis. dll Hemorrhage. dll Pan_Morphometry. dll Cellularity assessment. dll Necrosis. dll Infarction. dll Fibrosis. dll Mitoses. dll Micrometastasis. dll Adenomyosis. dll

Automated detection of intratumoral hemorrhage in neuroblastoma Pre Image Analysis Post Image Analysis

Optra. Assays™ (Image analysis libraries for histopathology in toxicological analysis) Ø Ø Ø Ø Ø Angiogenesis. dll Hemorrhage. dll Pan_Morphometry. dll Cellularity assessment. dll Necrosis. dll Infarction. dll Fibrosis. dll Mitoses. dll Micrometastasis. dll Adenomyosis. dll

Automated Morphometry of Pancreatic Islet cells Pre Image Analysis Post Image Analysis

Automated Morphometry of Pancreatic Islet cells Pre Image Analysis Cellularity: Normocellular Area/Size : Normal Fibrosis: Absent Post Image Analysis

Automated Thyroid Histopathology Pre Image Analysis Lymphocytes vs Epithelial Cells Post Image Analysis Colloid Follicular Epithelial Cells

Optra. Assays™ (Image analysis libraries for histopathology in toxicological analysis) Ø Ø Ø Ø Ø Angiogenesis. dll Hemorrhage. dll Pan_Morphometry. dll Cellularity assessment. dll Necrosis. dll Infarction. dll Fibrosis. dll Mitoses. dll Micrometastasis. dll Adenomyosis. dll

Automated detection of Necrosis Pre Image Analysis Post Image Analysis

Optra. Assays™ (Image analysis libraries for histopathology in toxicological analysis) Ø Ø Ø Ø Ø Angiogenesis. dll Hemorrhage. dll Pan_Morphometry. dll Cellularity assessment. dll Necrosis. dll Adenomyosis. dll Infarction. dll Fibrosis. dll Mitoses. dll Micrometastasis. dll

Female rhesus monkeys (Macaca mulatta) were ovariectomized and treated with subcutaneous implants, each containing 200 mg estradiol. Adenomyosis at necropsy Automated detection of Adenomyosis Pre Image Analysis Post Image Analysis

Optra. Assays™ (Image analysis libraries for histopathology in toxicological analysis) Ø Ø Ø Ø Ø Angiogenesis. dll Hemorrhage. dll Pan_Morphometry. dll Cellularity assessment. dll Necrosis. dll Adenomyosis. dll Infarction. dll Fibrosis. dll Mitoses. dll Micrometastasis. dll

Summary Optra. Assays™ Ø Ø Ø Ø Ø Angiogenesis. dll Hemorrhage. dll Pan_Morphometry. dll Cellularity assessment. dll Necrosis. dll Infarction. dll Fibrosis. dll Mitoses. dll Micrometastasis. dll Adenomyosis. dll Automated Toxicology Screening System

Acknowledgement Development team at Optra Systems, India