PB0872 - Fully automated platelet Differential Interference Contrast image analysis via deep learning.

Monday, July 11, 2022

6:30 PM – 7:30 PM

Location: Exhibition

Presentation Format: In-Person

Presenting Author(s)

Carly Kempster

PhD student
University of Reading
Reading, England, United Kingdom

Co-Author(s)

GB

George Butler

Postdoctoral Research Fellow
Johns Hopkins University
Johns Hopkins School of Medicine, Maryland, United States
EK

Elina Kuznecova

BSc
University of Reading
Reading, England, United Kingdom
Kirk A. Taylor, PhD

Postdoctoral Research Associate
University of Reading
Reading, England, United Kingdom
NK

Neline Kriek

Postdoctoral Research Assistant
University of Reading
READING, England, United Kingdom
GL

Gemma Little

PhD Student
University of Reading
READING, England, United Kingdom
MS

Marcin A. Sowa

Postdoctoral researcher
NYU Grossman School of Medicine
New York, New York, United States
TS

Tanya Sage

Research Technician
University of Reading
READING, England, United Kingdom
LJ

Louise J. Johnson

Associate Professor, Ecology and Evolutionary Biology
University of Reading
READING, England, United Kingdom
JG

Jonathan M. Gibbins, PhD

Professor
Institute for Cardiovascular and Metabolic Research, School of Biological sciences, University of Reading, Reading, UK
Reading, England, United Kingdom
AP

Alice Y. Pollitt

Associate Professor
University of reading
Reading, England, United Kingdom

Background: Platelets mediate arterial thrombosis, a leading cause of myocardial infarction and stroke. During injury, platelets adhere and spread over exposed subendothelial matrix substrates of the damaged blood vessel wall. The mechanisms which govern platelet activation and their interaction with a range of substrates are therefore regularly assessed and investigated using platelet spreading assays. These assays often use differential interference contrast (DIC) microscopy to assess platelet morphology whereby analysis is performed by manually annotating the perimeter of each platelet.

Aims: To implement a convolutional neural network (CNN) to automate platelet spreading analyses and abrogate time consuming and biased manual analysis.

Methods: The CNN was trained using 120 generalised training images. Increasing the number of training images increased the mean average precision (mAP) of the CNN. Platelets were spread in the presence and absence of inhibitors (dasatinib, ibrutinib, and PRT-060318) or agonist (thrombin) known to impact platelet morphology, and over a range of different substrates (CRP-XL, vWF and fibrinogen). The CNN performance was compared to a trainer, and five manual annotators.

Results: A convolutional neural network (CNN) allowed fully automated analysis of platelet spreading assays captured by DIC microscopy when platelets were spread over a range of substrates, and when in the presence or absence of inhibitors or agonist. Significant variation was observed between individual annotators, highlighting biases associated with manual analyses. Results were consistent in quantifying the spread area of platelets, which were comparable to published literature.

Conclusion(s): The application of a CNN enables, for the first time, automated analysis of unlabelled platelet spreading assays captured by DIC microscopy which has the potential to standardise analyses across laboratories.