Quantitative image analysis of nuclear chromatin distribution for cytological diagnosis.

Ryotaro Jingu, Masafumi Ohki, Sumiko Watanabe, Sadafumi Tamiya, Setsuo Sugishima, Tsunehisa Kaku
Author Information
  1. Ryotaro Jingu: Department of Health Sciences, Graduate School of Medical Sciences, Fukuoka, Japan.

Abstract

OBJECTIVE: We previously reported on the classification of the nuclear chromatin distribution into 3 types, that is peripheral (P), mixed (M) and central (C), which were related to the malignancy of cervical glandular lesions. However, the classification was subjective as it was performed by visual assessment. In the present study, quantitative assessment of nuclear chromatin distribution using image analysis by computer was applied for objective classification of cervical squamous epithelial lesions, which are the most common cervical lesions.
STUDY DESIGN: A total of 130 cells (44 cells of type P, 78 of type M and 8 of type C) from cytological specimens of cervical squamous epithelial lesions (dysplasia, squamous cell carcinoma) were analyzed. An image analysis program was developed as a plug-in macro program of an existing image processing software. The radial distribution (RD) value, which represents the gradient of the staining intensity from the center to the edge of a nucleus, was defined as an index of the chromatin distribution.
RESULTS: The RD values calculated in type P, type M and type C cells showed significant statistical differences as assessed by the t test (p < 0.001).
CONCLUSIONS: Quantification of the nuclear chromatin distribution by image analysis is fast and highly objective. The RD value could be useful as an index for malignancy.

MeSH Term

Carcinoma, Squamous Cell
Cell Nucleus
Chromatin
Cytodiagnosis
Female
Humans
Image Processing, Computer-Assisted
Uterine Cervical Neoplasms

Chemicals

Chromatin

Word Cloud

Created with Highcharts 10.0.0distributiontypechromatinimagenuclearcervicallesionsanalysisclassificationPMCsquamouscellsRDmalignancyassessmentobjectiveepithelialcytologicalprogramvalueindexOBJECTIVE:previouslyreported3typesperipheralmixedcentralrelatedglandularHoweversubjectiveperformedvisualpresentstudyquantitativeusingcomputerappliedcommonSTUDYDESIGN:total13044788specimensdysplasiacellcarcinomaanalyzeddevelopedplug-inmacroexistingprocessingsoftwareradialrepresentsgradientstainingintensitycenteredgenucleusdefinedRESULTS:valuescalculatedshowedsignificantstatisticaldifferencesassessedttestp<0001CONCLUSIONS:QuantificationfasthighlyusefulQuantitativediagnosis

Similar Articles

Cited By