Abstract
In this paper, a simple automatic Immunohistochemical (IHC) scoring method based on H-score has been described in which, tissue sections are segmented using Otsu method and each cell nuclei is classified into one of five different categories by multi-level thresholding of feature extracted using maximal separation (MS) techniques and finally, total score (TS) and diagnostic index (DI) are determined. The performance of the automatic scoring method is evaluated by comparing with manual scores using correlation coefficient (Cc), mean absolute difference (MAD) and 95% confidence interval (CI). According to the performance evaluation, DI score based on (blue component) had a strong agreement with both manual scores of tissue images. Therefore, automatic scoring based on technique has high potential in modern medical practice of diagnosis, prognosis and targeted therapy design for cancer patients.
Acknowledgements
The authors are expressing sincere thanks to Dr. Siow-Wee Chang, Faculty of Science, Institute of Biological Sciences, University of Malaya for providing IHC stained tissue images of OSCC used in this study. Special thanks to Dr. Anil, and Dr. Mathusudanan, Department of pathology, Meenakashi Mission Hospital & Research Center, Madurai, Tamilnadu, India., for their supports in understanding the concept and also for their valuable suggestion during this research work.
Conflict of Interest
No potential conflict of interest was reported by the authors.