Back

Deep learning-based adaptive detection of fetal nucleated red blood cells

Sun, C.; Wang, R.; Zhao, L.; Han, L.; Ma, S.; Liang, D.; Wang, L.; Tuo, X.; Zhong, D.; Li, Q.

2020-03-08 pathology
10.1101/2020.03.06.980227 bioRxiv
Show abstract

Aimthis study, we established an artificial intelligence system for rapid identification of fetal nucleated red blood cells (fNRBCs). MethodDensity gradient centrifugation and magnetic-activated cell sorting were used for the separation of fNRBCs from umbilical cord blood. The cell block technique was used for fixation. We proposed a novel preprocessing method based on imaging characteristics of fNRBCs for region of interest (ROI) extraction, which automatically segmented individual cells in peripheral blood cell smears. The discriminant information from ROIs was encoded into a feature vector and pathological diagnosis were provided by the prediction network. ResultsFour umbilical cord blood samples were collected and validated based on a large dataset containing 260 samples. Finally, the dataset was classified into 3,720 and 1,040 slides for training and testing, respectively. In the test set, classifier obtained 98.5% accuracy and 96.5% sensitivity. ConclusionTherefore, this study offers an effective and accurate method for fNRBCs preservation and identification.

Matching journals

The top 7 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.