Back

Tumor Segmentation using temporal Independent Component Analysis for DCE-MRI

Lee, J.; Zhao, Q.; Kent, M.; Platt, S.

2022-12-19 bioinformatics
10.1101/2022.12.16.520830 bioRxiv
Show abstract

In this study, we developed the temporal independent component analysis (tICA) to solve the partial volume effect (PVE) in canine brain tumor segmentation. The performance of the tICA is compared to that of spatial ICA (sICA) and an expert manual delineation of tumors based on three criteria: percent volume overlap or Dice coefficient, and percent volume difference. Seven in vivo DCE-MRI datasets of spontaneously occurring canine brain tumors were acquired on a 3 Tesla MRI system. The mean value of percent volume overlap (76.00%) between sICA and manual segmentation is lower than that between tICA and manual segmentation, which is 81.11%. In conclusion, the performance of the two ICA methods for segmenting tumors is very close to that of the expert delineation method. However, the tICA has the advantage over the sICA method in its ability to separate independent tissue signals in a voxel containing different types of tissues.

Matching journals

The top 4 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.