scroll
UP

Dr Carlos Andres Peña Solorzano

Research Officer, Bioinformatics & cellular genomics Laboratory

Improving breast cancer diagnosis by applying artificial intelligence to mammograms

Current Research image

The Problem

Population-based breast screening programs have been shown to increase the likelihood of diagnosing cancer in its early stages. However, typical wait times for results are two weeks or more – the time it takes for two expert reads of a mammogram, and a third if interpretations differ. Despite this thorough approach, in 2018 more than 34,000 Australian women (in the age-group 50 to 74 years) were recalled for assessment, and later determined not to have breast cancer.

Artificial Intelligence (AI) has the potential to reduce the workload of radiologists by filtering out normal mammograms. But AI’s performance on cancer diagnosis has been limited, because the highly detailed information required to adequately ‘train’ the digital ‘brain’ has to date been lacking.
The project image

The Project

As part of the Breast Cancer AI project (BRAIx) and with the help of BreastScreen Victoria, Dr Carlos Peña Solorzano is working with radiologists to build a manually annotated database that can be used to ‘teach’ and verify the performance of AI-based interpretation of mammograms. This database will include highly specific visual details like cancer location and the contours of cancer tissue, which human experts use to visually identify cancers in mammograms.

Carlos' project will benchmark the use of this kind of database for the improvement of breast cancer diagnosis and localisation using AI on mammograms.

“Our results will inform the future development of free-to-use mammogram databases and provide a baseline for the wider medical research community developing automated tools for breast cancer diagnosis,” says Carlos.
Bio image

Dr Carlos Andres Peña Solorzano

Carlos is an expert in the use of AI for the processing of medical images, including CT (computed tomography), X-rays, and mammogram scans. His goal is to improve patient care through better use of AI in areas such as disease diagnosis, organ segmentation and automatic medical database labelling.

Carlos joined SVI’s Bioinformatics & Cellular Genomics Laboratory in 2020 as a postdoctoral researcher. He trained during his PhD to use AI for organ characterisation on post-mortem CT scan databases and has more than five years’ experience in the application of AI for image processing and process automation.