Assessment of parathyroid gland cellularity by digital slide analysis

Rotem Sagiv, Bertha Delgado, Oleg Lavon, Vladislav Osipov, Re'em Sade, Sagi Shashar, Ksenia M. Yegodayev, Moshe Elkabets, Ben Zion Joshua

Research output: Contribution to journalArticlepeer-review

Abstract

Background: Computer-aided examination of digital tissue images has attracted attention in recent years. Application in the field of parathyroid pathology has not been studied previously. It holds a potential to assist in the examination of parathyroid gland adenoma or hyperplasia. Objectives: To explore parathyroid cell detection of slide images by digital tissue analysis and compare the results to standard human processing. Methods: 47 incisional biopsies of healthy appearing parathyroid glands were evaluated for their cellularity level. First, by the standard examination using microscopy by three independent pathologists. We compared the mean cellularity grading of the pathologists to the output of a computerized cell detection software. Results: A disagreement was found between the standard human cellularity grading and the digital analysis output. However, the digital analysis reaches a 94% specificity and 48% sensitivity to predict high cellularity (>60% parenchymal cells). Conclusions: Digital analysis of parathyroid tissue can be used as a tool for hypercellularity elimination, therefore assisting in the diagnosis of parathyroid cell hyperplasia. Additional studies using more advanced algorithms are necessary for further precision enhancement.

Original languageEnglish
Article number151907
JournalAnnals of Diagnostic Pathology
Volume58
DOIs
StatePublished - 1 Jun 2022

Keywords

  • Computer-assisted image analysis
  • Hyperparathyroidism
  • Light Microscopy
  • Parathyroid Adenoma
  • Parathyroidectomy
  • Surgical Pathology

ASJC Scopus subject areas

  • Pathology and Forensic Medicine

Fingerprint

Dive into the research topics of 'Assessment of parathyroid gland cellularity by digital slide analysis'. Together they form a unique fingerprint.

Cite this