We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

eess.IV

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Electrical Engineering and Systems Science > Image and Video Processing

Title: Ki-67 Index Measurement in Breast Cancer Using Digital Image Analysis

Abstract: Ki-67 is a nuclear protein that can be produced during cell proliferation. The Ki67 index is a valuable prognostic variable in several kinds of cancer. In breast cancer, the index is even routinely checked in many patients. Currently, pathologists use the immunohistochemistry method to calculate the percentage of Ki-67 positive malignant cells as Ki-67 index. The higher score usually means more aggressive tumor behavior. In clinical practice, the measurement of Ki-67 index relies on visual identifying method and manual counting. However, visual and manual assessment method is timeconsuming and leads to poor reproducibility because of different scoring standards or limited tumor area under assessment. Here, we use digital image processing technics including image binarization and image morphological operations to create a digital image analysis method to interpretate Ki-67 index. Then, 10 breast cancer specimens are used as validation with high accuracy (correlation efficiency r = 0.95127). With the assistance of digital image analysis, pathologists can interpretate the Ki67 index more efficiently, precisely with excellent reproducibility.
Subjects: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV)
Cite as: arXiv:2209.13155 [eess.IV]
  (or arXiv:2209.13155v1 [eess.IV] for this version)

Submission history

From: Hsiang-Wei Huang [view email]
[v1] Tue, 27 Sep 2022 04:48:57 GMT (925kb)

Link back to: arXiv, form interface, contact.