Experts are of the view that deep learning based solutions can significantly change outcomes of cancer therapy through tailor made personalised treatment regimens.
A case in point being prostate cancer, which is today the second most frequent cancer and the fifth leading cause of cancer death among men. However, it exhibits a broad disease spectrum ranging from indolent and slow growing which requires only watchful waiting to aggressive and lethal disease requiring immediate surgical intervention.
“Hence for effective treatment of the disease, it is not only essential to diagnose the disease early, but also to stratify and prognosticate it at the outset. At AIRA Matrix, in addition to screening and diagnosis, we have also developed solutions for accurate quantification of prognostic factors like Gleason scoring and tumour volume estimation. We are also working on developing deep learning networks to predict progression of cancer and its response to therapy,” explains Chaith Kondragunta, CEO, AIRA Matrix Pvt. Ltd.
AIRA Matrix provides artificial intelligence (AI) based solutions for image and data analysis in the life sciences industry. Its products and services are applicable in diverse areas like drug discovery, preclinical drug safety assessment, cancer diagnostics, ophthalmology and environmental monitoring.
The company has secured multiple patents in deep learning, image processing and digital pathology. Research and innovation are key to development of first in class products and solutions. The company firmly believes that it can be a global leader only if it invests in R&D across all the domains it operates.
“We regularly participate in international competitions to test ourselves. We finished 2nd in the global Gleason 2019 Challenge at MICCAI 2019. The challenge involved automating the Gleason score, the most commonly used method for grading prostate cancer. Similarly, as part of an NHSx UK challenge, we successfully developed an AI--based image analysis solution for Rapid On Site Evaluation of EBUS guided Transbronchial Needle Aspirate samples; a procedure applied for diagnosis and staging of lung cancers,” Kondragunta said.
In the discovery vertical, the company’s hybrid Deep Learning-based models predict potential toxic effects of a new drug molecule under investigation. These models analyse data from multiple modalities to help with the crucial go/no-go decisions in the selection of the safest drug molecule. The risk road map and failure point predictions provide significant reduction in the time and resources needed in this early phase of drug development. In addition, they reduce the need for animal sacrifice, providing a boost to humane animal research in line with the 3R principles of animal testing - Reduce, Refine, and Replace.
The company also provides solutions that improve the evaluation of parameters in in-vivo disease models used for efficacy studies. For example, in acute and chronic lung disease models, its solution quantifies multiple histopathological parameters for the accurate comparison of features between test and control groups. This helps provide objective, quantifiable and reproducible inferences on the efficacy of the drug molecule.
The company also helps improve the efficiency of pre-clinical toxicologic pathology workflows. Its tissue abnormality detection solution employs networks to pre-screen tissue images and help pathologists expedite the analyses of thousands of tissues included in these studies. This can lead to savings of several person-months that are otherwise spent on reporting toxicology studies.
It provides solutions that assist environmental monitoring in drug manufacturing areas that need to abide by stringent clean room regulations. The solutions aid by automated classification and quantification of contaminant micro-organisms. They also provide a traceable workflow with auditable assessment records – imperative for regulated settings.
“On the health care side, we develop AI-driven solutions that aid disease diagnosis and personalised healthcare. In the oncology domain, our predictive analytics solutions aim at precision medicine. In the ophthalmology domain, we develop solutions for automated screening, accurate grading and stratification of blindness causing diseases using deep learning networks to analyse retinal images,” Kondragunta concludes.
|