Page 51 - Status Report

The Context

Cervical cancer occurs when abnormal cells
on the cervix grow out of control. Most cervical
cancers are caused by a virus called the
Human Papillomavirus, or HPV. Cervical Cancer
is the largest killer cancer for women in India,
making India the Cervical Cancer capital of
the world. With an estimated 350million Indian
women aged between 21-65 years in the risk
category for Cervical Cancer; India is potentially
sitting on a ticking time-bomb. Like in the
case of most other cancers, early diagnosis
and treatment is critical for favourable
outcomes. For winning the war against cervical
cancer,diagnostic tests for cervical cancer have
to be easy, available at the point-of-care and
affordable.

The Initiative

The project proposes to build an affordable and
portable, ‘point-of-care’ Cervical Cancer Screening
tool to automate the analysis of the Pap smear
slides. The slides are stained,scanned, digitized and
then analyzed using computer algorithms to triage
them into normal, suspect and abnormal samples.
The images are then sent over a Tele-pathology
medium to pathologists for further confirmations
and recommendations. Multiple sites around rural
and urban Karnataka will be included in the initial
trial. Each site will receive a kit that includes Aindra’s
Auto Staining Device, Alexapath’s Slide Scanning
Equipment and a computer with a powerful graphics
processing unit to run the auto diagnostic. This kit
will pair semi-skilled workers capable of preparing
and scanning slides with robotics and a machine
learning/computer vision artificial intelligence. The
team aims to bring the cost of cytology down to
approximately INR 150 per diagnosis, a cost that
should be affordable for women across rural India.

The Outcome

The team has completed the development associated to hardware and have created a web
application to allow for control of the ADA stage. They have successfully developed a concept
prototype with established repeatability, positional accuracy, stabilized mechanical and electronic
components. The team has also developed the ability to accept tiled images from an Image
Acquisition Device and feed them into the program.

Status Report 51