Lindau Alumna Works on Machine Learning Techniques for Better Cancer Care

Our newest blog post by Jude Dinely features the work of Lindau Alumna Harshita Sharma: 

Harshita Sharma of the University of Oxford, a young scientist who attended the 68th Lindau Nobel Laureate Meeting, develops such methods in her research. They include machine learning techniques, a hot topic in biomedical imaging. “We want to develop these applications for medical professionals so that it saves them time and effort,” says Sharma. The benefits could be particularly great in LMIC, where staff shortages and over-loaded clinicians are more common, adds Sharma.

Computational analyses can also save money. Focusing on digital pathology,  Sharma investigated during her PhD whether a more affordable, efficient staining technique for gastric carcinoma, H&E, could provide better analyses using machine learning. Currently, it is not the preferred choice of pathologists, as certain types of cancers are not easy to differentiate by eye using the technique. Initial accuracies of 75-80% are promising.

Read more about the issue of large inequalitiy in access to radiotherapy in cancer care between low-middle and high-income countries on our blog: Tackling the Silent Crisis.

An IAEA global directory of treatment machines per million people shows the radiotherapy ‘gap’. Credit: DIRAC Project, IAEA


Other news