Advanced Morphology Unit for Realtime Analysis (AMURA)

Problem Statement:

In the Indian scenario, most of the population lives in the rural than the urban areas, hence most people have access to Primary Healthcare Centres (PHCs) only. The PHCs with restricted resources (Microscopes and Scanners) and expertise (Trained pathologists), cannot perform screening for early detection of cancer. Although the patients get referred to a tertiary hospital for incision biopsy, patient trauma caused, and expenditure incurred is high. Cytology being non-invasive, affordable, and accessible will be thus more favourable as an oral cancer screening tool.

Concept:

Manual cytopathology diagnosis is less accurate and is not used for screening especially for early detection. Our aim with AMURA was to automate the analysis of the H&E stained cytology slides, to develop a more accurate screening method. The system we developed was able to achieve 90% overall accuracy. We achieved a substantial increase in the sensitivity of low stages of cancer (from 25% to 73%) when compared to manual cytology. To further increase the sensitivity in lower grades of cancer, we are currently working with our clinical collaborators to develop a system that enables molecular cytology at point of care. The system developed would be able to do fluorescence imaging by utilizing devices acquired with a fraction of the cost of normal automated fluorescence imaging devices.

Technology/Science involved:

Mechatronic systems, Cytology, Oncology, Point-of-Care tools, Machine learning, Image processing

Key Roles/ Responsibilities :

Team Lead, Prototype Development Engg

My Tasks:

• Designed & Rendered the entire system 

• Developed Engineering drawings based on ASME Y14.5  standards for manufacturing 

• Components orders and procurement 

• Complete system integration and subsystem level calibration 

• Worked on high precision slide scanning mechanism and developed hardware and software backlash compensation