Regional registration of whole slide image stacks containing major histological artifacts

• Published in: BMC bioinformatics Vol. 21; no. 1; pp. 558 - 20
• Main Authors: Paknezhad, Mahsa, Loh, Sheng Yang Michael, Choudhury, Yukti, Koh, Valerie Koh Cui, Yong, Timothy Tay Kwang, Tan, Hui Shan, Kanesvaran, Ravindran, Tan, Puay Hoon, Peng, John Yuen Shyi, Yu, Weimiao, Tan, Yongcheng Benjamin, Loy, Yong Zhen, Tan, Min-Han, Lee, Hwee Kuan
• Format: Journal Article
• Language: English
• Published: London BioMed Central 04.12.2020; BioMed Central Ltd; Springer Nature B.V; BMC
• Subjects: Accuracy; Algorithms; Analysis; Bioinformatics; Biomedical and Life Sciences; Blood vessel 3D reconstruction; Blood Vessels - diagnostic imaging; Blood Vessels - pathology; Carcinoma, Renal Cell - blood supply; Computational Biology/Bioinformatics; Computer Appl. in Life Sciences; Data entry; Deformation; Evaluation; Humans; Image Processing, Computer-Assisted - methods; Image reconstruction; Image resolution; Imaging and image analysis; Imaging, Three-Dimensional - methods; Immunohistochemistry - methods; Immunohistochemistry images; Kidney cancer; Life Sciences; Literature reviews; Medical imaging; Medical imaging equipment; Methodology; Methodology Article; Methods; Microarrays; Microscopy; Multi-scale attention; Pathology; Regions; Registration; Rigid registration; Robustness (mathematics); Tissues; Whole slide images; Singapore; Rigid registration; Whole slide images; Multi-scale attention; Immunohistochemistry images; Blood vessel 3D reconstruction
• ISSN: 1471-2105; 1471-2105
• DOI: 10.1186/s12859-020-03907-6

Background High resolution 2D whole slide imaging provides rich information about the tissue structure. This information can be a lot richer if these 2D images can be stacked into a 3D tissue volume. A 3D analysis, however, requires accurate reconstruction of the tissue volume from the 2D image stack. This task is not trivial due to the distortions such as tissue tearing, folding and missing at each slide. Performing registration for the whole tissue slices may be adversely affected by distorted tissue regions. Consequently, regional registration is found to be more effective. In this paper, we propose a new approach to an accurate and robust registration of regions of interest for whole slide images. We introduce the idea of multi-scale attention for registration. Results Using mean similarity index as the metric, the proposed algorithm (mean ± SD 0.84 ± 0.11 ) followed by a fine registration algorithm ( 0.86 ± 0.08 ) outperformed the state-of-the-art linear whole tissue registration algorithm ( 0.74 ± 0.19 ) and the regional version of this algorithm ( 0.81 ± 0.15 ). The proposed algorithm also outperforms the state-of-the-art nonlinear registration algorithm (original: 0.82 ± 0.12 , regional: 0.77 ± 0.22 ) for whole slide images and a recently proposed patch-based registration algorithm (patch size 256: 0.79 ± 0.16 , patch size 512: 0.77 ± 0.16 ) for medical images. Conclusion Using multi-scale attention mechanism leads to a more robust and accurate solution to the problem of regional registration of whole slide images corrupted in some parts by major histological artifacts in the imaged tissue.