Author/Editor		Šiška, Arijan; Babić, Ankica; Prevodnik-Kloboves, Veronika; Stanta, Giorgio; Pavešić, Nikola
Title		Studies of responsiveness in the R-EGFR and Rcerb-B2 oncogenes spaces: an attempt to classify thyroid lesions
Type		članek
Source		In: Zajc B, editor. Zbornik 6. elektrotehniške in računalniške konference ERK'97. Zvezek B. Računalništvo in informatika, umetna inteligenca, robotika, razpoznavanje vzorcev - 3. letna konferenca SDRV, biomedicinska tehnika, močnostna elektrotehnika, didaktika, študentski članki; 1997 sep 25-27; Portorož. Ljubljana: Slovenska sekcija IEEE,
Publication year		1997
Volume		str. 387-90
Language		eng
Abstract		Diagnosis of thyroid lesions is very difficult process implying a need for a reliable and non-invasive. Among verity of diagnostic approaches suitable, for clinical analysis of thyroid lesions, the two oncogenes (R-EGFR and Rcerb-B2) are evaluated for their discriminative power. In our study we collected patient material with classified examples of normal tissue and four different lesion types (goiters, adenoma, follicular and papillary carcinoma) and respective oncogene measurements. We then attempted to evaluate discriminating performance of the oncogenes. Standard multivariate statistics have not performed satisfactory partly due to the data distribution and partly due to presence of the noise. Therefore we have developed a method for the classification purpose, which was based on principle of minimizing generalized classification error. Results of the separation between papillary carcinoma and normal tissue, as well as papillary and follicular carcinoma was about 70%, other classification atempts ended up in poor results. In general, misclassifications cold be explained with the data quality (noise) and, when it came to distinguishing benign lesions, with responsiveness of the oncogenes to tumor tissues.
Descriptors		THYROID NEOPLASMS PROTO-ONCOGENES DISCRIMINANT ANALYSIS