CUSAT software to help detect breast cancer

[1]

When breast cancer rules as the second leading cause of mortality among women, here’s some good news from researchers at Cochin University of Science and Technology (CUSAT).

Dr Tessamma Thomas from the department of electronics at CUSAT, in association with researcher Deepa Sankar, has developed a software to help radiologists in breast cancer detection.

The automated software will help to analyse mammograms and detect microcalcifications which are tiny deposits of calcium, with a size of, often, less than 0.5 mm in width, and are early indicators of breast cancer, says Dr Tessamma.

Currently, radiologists are relying on their eyes and brain for the detection process. The automatic mammogram classification and detection software package gives a classification accuracy of 98 per cent and detection accuracy of 91.25 per cent.

“These results were secured after analyzing the mammogram data available in the MIAS and DDSM online database,” explained Dr Tessamma.

“Currently, only the Regional Cancer Centre has the fully digital mammography unit which has computer-aided software detection for breast cancer and the equipment there costs `2.5 crore. The private hospitals don’t have the finances to buy this equipment.

They use the semi-digital machine and the development at CUSAT augurs well, but it has to undergo clinical trials,” said Dr Suma Mariam Jacob, radiologist at Lakeshore Hospital.

Researcher Dinesh Kumar, with the help of Dr Tessamma, has developed another software to aid orthopedic surgeons in scoliosis correction, which is claimed to be the first of its kind.
Scoliosis is an abnormal lateral curvature of the spine.

“For the rectification of this spinal curve, a measure of the spine curvature (cobb angle) is required. A new rule-based algorithm is developed forstrategic vertebrae selection and cobb angle and other scoliosis feature evaluation measurements,” says Dr Tessamma.

Using visual features, retrieval of cases of patients having similar diagnosis and also cases with visually similar but different diagnosis is possible, she said.