Dataset Information

Artificial intelligence defines protein-based classification of thyroid nodules

ABSTRACT: We applied PCT-DIA on a total of 1161 nodules from 1133 patients using either tissue cores (1 mm diameter; 0.5–1 mm depth) punched from regions of interest marked on retrospective FFPE tissue blocks or prospective cytology specimens from FNA aspirates. The samples comprise (i) a discovery set of FFPE samples from Singapore General Hospital (n = 579 nodules) where histopathological diagnoses were confirmed on central review by a board-certified pathologist; and independent test sets from twelve hospitals in China and Singapore consisting of (ii) retrospective test sets of FFPE samples (n = 288 nodules) with the same histopathological assessment and classification as the discovery sample set, and (iii) a prospective test set of FNA biopsies (n = 294 nodules) which were additionally scored by the Bethesda System for reporting thyroid cytopathology.

ORGANISM(S): Homo Sapiens

SUBMITTER: Tiannan Guo

PROVIDER: PXD036554 | iProX | Wed Sep 07 00:00:00 BST 2022

REPOSITORIES: iProX

ACCESS DATA

Publications

Artificial intelligence defines protein-based classification of thyroid nodules.

Sun Yaoting Y Selvarajan Sathiyamoorthy S Zang Zelin Z Liu Wei W Zhu Yi Y Zhang Hao H Chen Wanyuan W Chen Hao H Li Lu L Cai Xue X Gao Huanhuan H Wu Zhicheng Z Zhao Yongfu Y Chen Lirong L Teng Xiaodong X Mantoo Sangeeta S Lim Tony Kiat-Hon TK Hariraman Bhuvaneswari B Yeow Serene S Alkaff Syed Muhammad Fahmy SMF Lee Sze Sing SS Ruan Guan G Zhang Qiushi Q Zhu Tiansheng T Hu Yifan Y Dong Zhen Z Ge Weigang W Xiao Qi Q Wang Weibin W Wang Guangzhi G Xiao Junhong J He Yi Y Wang Zhihong Z Sun Wei W Qin Yuan Y Zhu Jiang J Zheng Xu X Wang Linyan L Zheng Xi X Xu Kailun K Shao Yingkuan Y Zheng Shu S Liu Kexin K Aebersold Ruedi R Guan Haixia H Wu Xiaohong X Luo Dingcun D Tian Wen W Li Stan Ziqing SZ Kon Oi Lian OL Iyer Narayanan Gopalakrishna NG Guo Tiannan T

Cell discovery 20220906 1

Determination of malignancy in thyroid nodules remains a major diagnostic challenge. Here we report the feasibility and clinical utility of developing an AI-defined protein-based biomarker panel for diagnostic classification of thyroid nodules: based initially on formalin-fixed paraffin-embedded (FFPE), and further refined for fine-needle aspiration (FNA) tissue specimens of minute amounts which pose technical challenges for other methods. We first developed a neural network model of 19 protein ...[more]

PMID: 36068205

Similar Datasets

Project description:This experiment contains the RNA-Seq samples only. Formalin-fixed, paraffin-embedded (FFPE) tissues are an invaluable resource for clinical research. However, nucleic acids extracted from FFPE tissues are fragmented and chemically modified making them challenging to use in molecular studies. We analysed 23 fresh-frozen (FF), 35 FFPE and 38 paired FF/FFPE specimens, representing six different human tissue types (bladder, prostate and colon carcinoma; liver and colon normal tissue; reactive tonsil) in order to examine the potential use of FFPE samples in next-generation sequencing (NGS) based retrospective and prospective clinical studies. Two methods for DNA and three methods for RNA extraction from FFPE tissues were compared and were found to affect nucleic acid quantity and quality. DNA and RNA from selected FFPE and paired FF/FFPE specimens were used for exome and transcriptome analysis. Preparations of DNA Exome-Seq libraries was more challenging (29.5 % success) than that of RNA-Seq libraries, presumably because of modifications to FFPE tissue-derived DNA. Libraries could still be prepared from RNA isolated from two-decade old FFPE tissues. Data were analysed using the CLC Bio Genomics Workbench and revealed systematic differences between FF and FFPE tissue-derived nucleic acid libraries. In spite of this, pairwise analysis of DNA Exome-Seq data showed concordance for 70-80 % of variants in FF and FFPE samples stored for fewer than three years. RNA-Seq data showed high correlation of expression profiles in FF/FFPE pairs (Pearson Correlations of 0.90 +/- 0.05), irrespective of storage time (up to 244 months) and tissue type. A common set of 1,494 genes was identified with expression profiles that were significantly different between paired FF and FFPE samples irrespective of tissue type. Our results are promising and suggest that NGS can be used to study FFPE specimens in both prospective and retrospective archive-based studies in which FF specimens are not available.

Dataset Information

Artificial intelligence defines protein-based classification of thyroid nodules

Publications

Artificial intelligence defines protein-based classification of thyroid nodules.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure