A Review: Recent Approaches on Improving the Accuracy in Determining the Tumor and Cancer Grading in Histopathological Images Based on Scoring Ki67 Expression

¹ Faculty of Electronic Engineering Technology,Universiti Malaysia Perlis (UniMAP),02600, Arau, Perlis, Malaysia fahmiakmaldzulkifli@gmail.com yusoff@unimap.edu.my

² Department of Pathology,Hospital Tuanku Fauziah,Jalan Tun Abdul Razak, Pusat Bandar Kangar 01000, Kangar, Perlis, Malaysia ksyakir@gmail.com

³ Department of Pathology, Hospital Sultanah Bahiyah,Jln Langgar, Bandar Alor Setar 05460 Alor Setar, Kedah, Malaysia haidimohamad@gmail.com

^*Corresponding Author

Fahmi Akmal Dzulkifli

Faculty of Electronic Engineering Technology,

Universiti Malaysia Perlis (UniMAP)

02600, Arau, Perlis

Malaysia

Email: fahmiakmaldzulkifli@gmail.com

Abstract

Tumor markers are substances formed by cancer cells or other cells in the body, which aims to indicate the presence of cancer and provide information related to cancer growth. The expression of Ki67 is usually associated with tumor cell proliferation and growth. Basically, the pathologists will count the number of Ki67 expressions to determine the growth fraction of tumor cell proliferation. Conventional manual techniques have shown their drawbacks in counting Ki67 expression, such as time-consuming, low accuracy and reliability, poor reproducibility, and highly subjective. Therefore, this paper will focus on the recent studies in improving the quantification of Ki67 expression for determining the tumor or cancer grading. This study divides the approaches into three, which are double or multiple staining methods, image processing or image analysis software, and artificial intelligence.

Keywords:Artificial intelligence, Histopathological Images, Image processing, Ki67 Expression, Tumor Markers

Download the full article (PDF)