Skip to Main Content

An official website of the United States government

Government Funding Lapse

Because of a lapse in government funding, the information on this website may not be up to date, transactions submitted via the website may not be processed, and the agency may not be able to respond to inquiries until appropriations are enacted. The NIH Clinical Center (the research hospital of NIH) is open. For more details about its operating status, please visit  cc.nih.gov. Updates regarding government operating status and resumption of normal operations can be found at OPM.gov.

Predicting patient outcomes with machine learning of cancer histopathology images

Principal Investigator
Name
Moritz Gerstung
Degrees
Ph.D
Institution
EMBL-EBI
Position Title
Group leader
Email
About this CDAS Project
Study
PLCO (Learn more about this study)
Project ID
PLCO-463
Initial CDAS Request Approval
Mar 21, 2019
Title
Predicting patient outcomes with machine learning of cancer histopathology images
Summary
Histopathology imaging, notably Hematoxylin and Eosin stained (H&E) images are widely used for cancer diagnosis, classification and prognosis. Unprecedented progress in computer vision has made the accurate quantification of digitalised histopathology images possible. The aim of this research proposal is to explore the utility of machine learning algorithms for disease diagnosis and prognosis. Deep convolutional neural network will be trained for classifying tumor and normal tissues of lung, ovarian and colorectal cancer and to extract histopathology patterns characteristic of different diseases and cell types. These features then feed into statistical models to predict tumor characterisations such as tumor lymphocyte infiltrating score and tumor proliferation rate. Finally, we will combine imaging features and clinical data to improve the survival predictions.
Aims

1. Extract imaging features from H&E images that can accurately distinguish tumor and normal tissues of lung, ovarian and colorectal cancer.
2. Predict tumor characteristics such as tumor infiltrating lymphocyte scores and tumor proliferation rate using imaging features.
3. Integrate imaging features and clinical follow up data to train prognostic models.

Collaborators

Yu Fu, yufu@ebi.ac.uk, EMBL-EBI
Alexander Jung, alexwjung@ebi.ac.uk, EMBL-EBI
Ramon Vinas, ramon.torne.17@ucl.ac.uk, UCL, EMBL-EBI