Unbound MEDLINE

Differential diagnosis of follicular cancer and follicular adenoma using an expert system based on a set of qualitative signs of cell atypia.

Abstract

OBJECTIVE
To determine the efficiency of the developed expert system based on a set of qualitative signs of cell atypia and their weighting coefficients in the differential diagnosis of follicular cancer and follicular adenoma of the thyroid gland.
STUDY DESIGN
Sixty cytologic preparations of patients with histologic diagnosis of a follicular cancer and follicular adenoma were investigated.
RESULTS
Weighting coefficients for each sign of atypia for both forms of pathology have been calculated with the help of the deduced equation. This allowed creating an expert system by which the function of transforming qualitative signs of cell atypia to a quantitative form was realized. "Strength reserve" according to the diagnostic index value, coincidence of the verified diagnosis with the histologic conclusion, and its invariability for all 12 iterations testified to the reliability of an expert system. Preliminary trials showed the efficiency of an expert system for differentiating the nature of a thyroid follicular tumor to be 97.5%.
CONCLUSION
The developed expert system allows high efficiency in making a differential diagnosis of thyroid follicular cancer and follicular adenoma.

Authors

Kirillov V, Emeliyanova O

Institution

Laboratory for Electron Paramagnetic Resonance Dosimetry and Cytology, Belarusian State Medical University, Dzerzhinsky Avenue, 83, 220116 Minsk, Belarus. kirillov@bsmu.by

Source

Analytical and quantitative cytology and histology / the International Academy of Cytology [and] American Society of Cytology 33:5 2011 Oct pg 253-64

MeSH

Adenocarcinoma, Follicular
Adenoma
Adolescent
Adult
Aged
Biopsy
Child
Child, Preschool
Diagnosis, Differential
Expert Systems
Female
Humans
Image Interpretation, Computer-Assisted
Male
Middle Aged
Models, Biological
Software
Thyroid Gland
Thyroid Neoplasms
Young Adult

Pub Type(s)

Journal Article

Language

eng

PubMed ID

22611752