From ambiguities to insights in cancer diagnosis via query-based comparisons

Jeanne Kowalski, Conover Talbot, Hua L. Tsai, Nijaguna Prasad, Christopher Umbricht, Martha A. Zeiger

Research output: Contribution to journalArticle

Abstract

Because clinicians cannot determine suspicious thyroid neoplasms' malignancy pre- or intra-operatively, they perform 55,000 US thyroid operations annually, cannot manage patients optimally and, thus, need better diagnostic markers. Though diagnostic ambiguity spurs research, the quest for molecular markers that distinguish benign from malignant thyroid tumor classes remains unfulfilled. That eight subtypes (four per class) define thyroid tumors introduces a degree of heterogeneity that presents the major analytic impediment. A novel query-based comparison approach can, however, reliably distinguish benign from malignant lesions by examining subtype relations in class-specific gene expression. We introduce this approach through microarray analysis of thyroid tumor subtypes.

Original languageEnglish (US)
Pages (from-to)575-580
Number of pages6
JournalPattern Recognition
Volume42
Issue number4
DOIs
StatePublished - Apr 1 2009

    Fingerprint

Keywords

  • Gene expression
  • Query
  • Thyroid cancer

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Cite this