Informatics model predicts breast cancer recurrence

Thursday, December 3 | 11:00 a.m.-11:10 a.m. | SSQ01-04 | Room E450A
In this presentation, researchers from Johns Hopkins University will present their work on using informatics methods to predict the potential for breast cancer to recur.

In the new era of computer-learning techniques and personalized medicine, the researchers sought to determine if they could utilize both radiological variables from multiparametric MRI and clinicopathologic parameters to identify the potential for breast cancer recurrence in hormone-positive patients, according to presenter Michael Jacobs, PhD.

The team, which included medical physicists, computer scientists, radiologists, and an oncologist, collaborated to develop an unsupervised machine-learning informatics method that can be applied to both MRI and clinical variables, Jacobs said. The results were then compared with Oncotype DX, a 21-gene assay used by oncologists to help determine if chemotherapy may be useful.

While the results are preliminary, it appears that the machine-learning methods can be applied to patients to determine if they need treatment, Jacobs said.

"Moreover, we can relate the underlying biology to radiological and clinicopathologic variables," he said. "These findings will help us understand the complex interactions between each variable."

Dive deeper into this topic by attending this Thursday morning session.

Page 1 of 775
Next Page