AIRA Matrix possesses expertise in AI-based quantification of histological parameters in disease models, with a library of existing models and the ability to rapidly develop custom algorithms to assess new models.
We were approached by a Israel-based pharmaceutical company to develop a custom solution for quantification of acute and chronic lung injury models.
The solution is aimed at ensuring a faster turnaround on results with enhanced accuracy and precision and reduced drug development times and costs.
This collaboration comes at an important juncture when the world unites in fighting the current, unprecedented pandemic healthcare crisis.
Spermatogenic Staging in Non Human Primates
We collaborated with a renowned pharmaceutical company based in Germany, for the development of an AI based solution to automate spermatogenic staging in Non Human Primates (NHP).
We leveraged our experience with a similar solution in rodents to develop a Deep Learning- based solution for accurate classification and quantification of spermatogenesis in sections of testis from NHP, first on Periodic Acid Schiff stained sections, with translation to Hematoxylin and Eosin stained sections.
The technical team at AIRA Matrix, worked closely with the domain experts on the client’s side to understand the requirement and collaboratively develop a solution to bring much-needed speed, objectivity, and reproducibility to the complex reporting workflow for spermatogenic staging in NHP.
Additionally the solution will help eliminate the need for an additional PAS stained section, thus further optimising the time and resources spent on reporting.
Hepatic Fibrosis Quantification
AIRA Matrix has extensive capabilities in the development of AI based quantification of histological parameters in in-vivo disease models, with the ability to work with different histochemical staining techniques.
We collaborated with a Belgium-based pharmaceutical company that has generated a set of agnostic monoclonal antibodies being tested in a variety of preclinical models of fibrotic, inflammatory autoimmune, and degenerative diseases.
The collaboration resulted in the expedited development of an AI-based solution for the quantification of collagen in Masson Trichrome stained rodent liver sections, thus improving the speed, accuracy, and precision of reporting hepatic fibrosis models.
Early Detection of Pancreatic Ductal Adenocarcinoma
Pancreatic Intraepithelial Neoplasia (PanIN) are microscopic epithelial precursor lesions of pancreatic ductal adenocarcinoma (PDAC). Existing imaging modalities cannot accurately identify PanIN lesions pre-operatively; they can be identified only on histopathological examination.
We collaborated with a leading European research institution to develop a Deep Learning-based method for detection and grading of PanIN lesions in histopathological sections of pancreas.
This algorithm provides quantification-based segmentation and classification of PanIN, facilitating early detection of precursors to PDAC, potentially accelerating pathological work up and patient outcomes.
The application can be further extended for differential diagnosis of precursor neoplastic lesions like intra-ductal papillary neoplasm and mucinous cystic neoplasm to develop a comprehensive early detection solution for PDAC precursors.
Severity Scoring in Rodent Cardiomyopathy
Progressive cardiomyopathy (PCM) is a common background change seen in rodents
The incidence and severity of PCM may increase as a result of cardiotoxicity, and toxicity-related morphologic changes that can overlap with those of PCM.
Consequently, there is a need for consistent and sensitive detection and quantification of PCM-related changes to help differentiate spontaneous findings from test article-related ones.
A leading North American pathology lab and the National Toxicology Program collaborated with us for development of a Deep Learning-based image analysis solution for accurate and precise identification and quantification of PCM histologic features in rats. The solution detects and quantifies major features of PCM: cardiomyocyte degeneration and necrosis, mononuclear cell infiltrate, and fibrosis.
This solution helps objective quantification of spontaneous background histologic changes, to aid pathologists discern cardiotoxicity-associated changes.