Searchable abstracts of presentations at key conferences in endocrinology
Endocrine Abstracts (2006) 11 P839

ECE2006 Poster Presentations Thyroid (174 abstracts)

Association between patient- and tumor-related factors and the gene expression profile of papillary thyroid cancer

S Szpak-Ulczok 1 , M Wiench 1 , M Oczko-Wojciechowska 1 , J Wloch 1 , K Fujarewicz 2 , A Swierniak 2 , D Lange 1 & B Jarzab 1


1Maria Sklodowska-Curie Memorial Cancer Center and Institute of Oncology, Gliwice, Poland; 2Silesian Technical University, Gliwice, Poland.


We performed the analysis of correlation between gene expression profile of the papillary thyroid cancer PTC tumors and known clinical factors influencing the outcome of the disease. Gene expression profile was assessed on Human Genome U133A array (Affymetrix). We examined tumors samples obtained from 49 patients diagnosed with PTC. For further evaluation we included following factors: sex, age, tumor size, capsule invasion, multifocality, vascular invasion, lymph node and distant metastases. We used corrected Welch t-test estimation to analyze how many genes are associated with each factor. We found out that poor differentiation/early recurrence (2 tumors) is connected with a prominent difference in gene expression profile (812 genes changed). For the further analysis we excluded these 2 samples (because of the large scale of difference between two poorly differentiated tumors and the rest of PTCs) and performed the analysis on the remaining 47 well-differentiated PTCs. The strongest factor in this group was sex with 10 sex-related genes. Large difference appeared for distant metastases: 1486 transcripts showing moderate level of statistical significance. To rank the remaining clinical factors, with low significance, we performed the analysis of uncorrected Welch t-test p-values. In such classification distant metastases were the strongest factor. We compared gene expression profile of 8 patients with metastases to 39 samples who did not present metastases within the short follow-up time. We selected 150 genes differentiating both groups of patients. By Support Vector Machine approach this dataset correctly predicted the status (metastases present or absent) in 45/49 patients. Our conclusions: 1) Some clinically relevant features in PTC (poor differentiation, sex) are strongly and significantly associated with gene expression profile. 2) We revealed that presence of distant metastases is related to gene expression and that it is possible to specify the metastatic signature in PTC.

Volume 11

8th European Congress of Endocrinology incorporating the British Endocrine Societies

European Society of Endocrinology 
British Endocrine Societies 

Browse other volumes

Article tools

My recent searches

No recent searches.