Accurate Molecular Classification of Kidney Cancer Subtypes Using MicroRNA Signature

详细信息    查看全文

• 作者：Youssef M. Youssef^a ; ^b ; Nicole M.A. White^a ; ^b ; Jö ; rg Grigull^c ; Adriana Krizova^a ; ^b ; Christina Samy^a ; ^b ; Salvador Mejia-Guerrero^a ; ^b ; Andrew Evans^a ; ^d ; George M. Yousef^a ; ^b ; ^{yousefg@smh.ca}
• 关键词：Biomarker ; Cancer ; Clear cell renal cell carcinoma ; Chromophobe ; Diagnosis ; Kidney cancer ; MicroRNA ; miRNA ; Oncocytoma ; Papillary ; Pathology ; Profiling ; Statistical classifier ; Prognosis ; RCC ; Renal cancer ; Subtypes ; Tumour markers ; Unclassified
• 刊名：European Urology
• 出版年：2011
• 出版时间：May 2011
• 年：2011
• 卷：59
• 期：5
• 页码：721-730
• 全文大小：838 K

文摘

Background

Renal cell carcinoma (RCC) encompasses different histologic subtypes. Distinguishing between the subtypes is usually made by morphologic assessment, which is not always accurate.

Objective

Our aim was to identify microRNA (miRNA) signatures that can distinguish the different RCC subtypes accurately.

Design, setting, and participants

A total of 94 different subtype cases were analysed. miRNA microarray analysis was performed on fresh frozen tissues of three common RCC subtypes (clear cell, chromophobe, and papillary) and on oncocytoma. Results were validated on the original as well as on an independent set of tumours, using quantitative reverse transcription-polymerase chain reaction (qRT-PCR) analysis with miRNA-specific primers.

Measurements

Microarray data were analysed by standard approaches. Relative expression for qRT-PCR was determined using the ΔΔC_T method, and expression values were normalised to small nucleolar RNA, C/D box 44 (SNORD44, formerly RNU44). Experiments were done in triplicate, and an average was calculated. Fold change was expressed as a log₂ value. The top-scoring pairs classifier identified operational decision rules for distinguishing between different RCC subtypes and was robust under cross-validation.

Results and limitations

We developed a classification system that can distinguish the different RCC subtypes using unique miRNA signatures in a maximum of four steps. The system has a sensitivity of 97 % in distinguishing normal from RCC, 100 % for clear cell RCC (ccRCC) subtype, 97 % for papillary RCC (pRCC) subtype, and 100 % accuracy in distinguishing oncocytoma from chromophobe RCC (chRCC) subtype. This system was cross-validated and showed an accuracy of about 90 % . The oncogenesis of ccRCC is more closely related to pRCC, whereas chRCC is comparable with oncocytoma. We also developed a binary classification system that can distinguish between two individual subtypes.

Conclusions

MiRNA expression patterns can distinguish between RCC subtypes.