New candidates to discriminate among thyroid carcinomas, benign lesions and normal tissues based on transcriptome and genome-wide methylation analysis

International Conference on Cytopathology

August 31-September 02, 2015 Toronto, Canada

Silvia Regina Rogatto

Sao Paulo State University, Brazil

Posters-Accepted Abstracts: J Cytol Histol

Abstract :

The majority of thyroid lesions diagnosed is proven to be benign after surgery. High accuracy in pre-operative diagnosing of thyroid tumors particularly follicular subtypes is still a challenge. Expression profiling (8x60K, Agilent Technologies) was carried out in 61 papillary carcinomas (PTC) and 13 adjacent non-neoplastic tissues (NT). A panel of 28 transcripts was further evaluated by RT-qPCR in benign thyroid lesions (BTL) and a diagnostic algorithm was trained (86 PTC, 23 NT and 8 BTL) and validated (120 PTC, 10 NT and 140 BTL). Methylation profiling (450 K, Illumina) was performed in 50 NT, 17 BTL, 60 PTC, 10 follicular, 1 poorly differentiated and 3 anaplastic carcinomas. Combined expression of CLDN10, HMGA2 and LAMB3 achieved 94% of sensitivity and 96% of specificity in the validation set of samples. Cases with higher algorithm scores were associated with lymph nodes involvement. Aberrant methylation in nine loci showed sensitivity of 92% and specificity of 77% to discriminate between benign lesions from thyroid carcinomas. Three of 9 loci were able to stratify patients with PTC according to the risk of developing lymph node metastasis. We also developed a prognostic classifier based on five hypomethylated probes which was able to predict poor outcome in patients with malignant tumors (Sensitivity: 77%; Specificity: 90%). In overall, transcritpome and methylome analysis revealed genes with potential to differentiate PTC and BTL/NL tissues as well as cases with lymph nodes involvement. In addition, methylome analysis allowed the identification of a classifier useful to predict outcome in patients with thyroid carcinoma.

Biography :

Email: silvia.rogatto2@gmail.com