Molecular classification of melanomas and nevi using gene expression microarray signatures and formalin-fixed and paraffin-embedded tissue

Stephen S. Koh, Michael L. Opel, Jia Perng J. Wei, Kenneth Yau, Rashmi Shah, Mercedes E. Gorre, Eric Whitman, Paul K. Shitabata, Yong Tao, Alistair J. Cochran, Payam Abrishami, Scott W. Binder

Research output: Contribution to journalArticle

35 Citations (Scopus)

Abstract

Melanoma may be difficult to identify histologically and relatively high rates of misdiagnosis leads to many malpractice claims. Currently separation of melanomas from nevi is based primarily on light microscopic interpretation of hematoxylin and eosin stained sections with limited assistance from immunohistology. To increase the accuracy of discrimination of benign and malignant melanocytic lesions we identified DNA microarray-derived gene expression profiles of different melanocytic lesions and evaluated the performance of these gene signatures as molecular diagnostic tools in the molecular classification and separation of melanomas and nevi. Melanocyte-derived cells were isolated by laser capture microdissection from 165 formalin-fixed and paraffin-embedded melanocytic nevi and melanoma tissue sections. RNA was isolated, amplified, labeled, and hybridized to a custom DNA microarray. In all 120 samples were used to identify differentially expressed genes and generate a gene expression classifier capable of distinguishing between melanomas and nevi. These classifiers were tested by the leave-one-out method and in a blinded study. RT-PCR verified the results. Unsupervised hierarchical clustering identified two distinct lesional groups that closely correlated with the histopathologically identified melanomas and nevi. Analysis of gene expression levels identified 36 significant differentially expressed genes. In comparison with nevi, melanomas expressed higher levels of genes promoting signal transduction, transcription, and cell growth. In contrast, expression of L1CAM (homolog) was reduced in melanomas relative to nevi. Genes differentially expressed in melanomas and nevi, on the basis of molecular signal, sub classified a group of unknown melanocytic lesions as melanomas or nevi and had high concordance rates with histopathology. Gene signatures established using DNA microarray gene expression profiling can distinguish melanomas from nevi, indicating the feasibility of using molecular classification as a supplement to standard histology. Our successful use of a standard formalin-fixed and paraffin-embedded tissue further supports the practicability of combining molecular diagnostic testing with histopathology in evaluation of difficult melanocytic lesions.

Original languageEnglish (US)
Pages (from-to)538-546
Number of pages9
JournalModern Pathology
Volume22
Issue number4
DOIs
StatePublished - Apr 1 2009

Fingerprint

Nevi and Melanomas
Transcriptome
Paraffin
Formaldehyde
Oligonucleotide Array Sequence Analysis
Genes
Melanoma
Molecular Diagnostic Techniques
Neural Cell Adhesion Molecule L1
Laser Capture Microdissection
Gene Expression
Pigmented Nevus
Molecular Pathology
Nevus
Malpractice
Melanocytes
Gene Expression Profiling
Hematoxylin
Eosine Yellowish-(YS)
Diagnostic Errors

All Science Journal Classification (ASJC) codes

  • Pathology and Forensic Medicine

Cite this

Koh, Stephen S. ; Opel, Michael L. ; Wei, Jia Perng J. ; Yau, Kenneth ; Shah, Rashmi ; Gorre, Mercedes E. ; Whitman, Eric ; Shitabata, Paul K. ; Tao, Yong ; Cochran, Alistair J. ; Abrishami, Payam ; Binder, Scott W. / Molecular classification of melanomas and nevi using gene expression microarray signatures and formalin-fixed and paraffin-embedded tissue. In: Modern Pathology. 2009 ; Vol. 22, No. 4. pp. 538-546.
@article{260df06760ed44dfa8e7b4ed9438cf83,
title = "Molecular classification of melanomas and nevi using gene expression microarray signatures and formalin-fixed and paraffin-embedded tissue",
abstract = "Melanoma may be difficult to identify histologically and relatively high rates of misdiagnosis leads to many malpractice claims. Currently separation of melanomas from nevi is based primarily on light microscopic interpretation of hematoxylin and eosin stained sections with limited assistance from immunohistology. To increase the accuracy of discrimination of benign and malignant melanocytic lesions we identified DNA microarray-derived gene expression profiles of different melanocytic lesions and evaluated the performance of these gene signatures as molecular diagnostic tools in the molecular classification and separation of melanomas and nevi. Melanocyte-derived cells were isolated by laser capture microdissection from 165 formalin-fixed and paraffin-embedded melanocytic nevi and melanoma tissue sections. RNA was isolated, amplified, labeled, and hybridized to a custom DNA microarray. In all 120 samples were used to identify differentially expressed genes and generate a gene expression classifier capable of distinguishing between melanomas and nevi. These classifiers were tested by the leave-one-out method and in a blinded study. RT-PCR verified the results. Unsupervised hierarchical clustering identified two distinct lesional groups that closely correlated with the histopathologically identified melanomas and nevi. Analysis of gene expression levels identified 36 significant differentially expressed genes. In comparison with nevi, melanomas expressed higher levels of genes promoting signal transduction, transcription, and cell growth. In contrast, expression of L1CAM (homolog) was reduced in melanomas relative to nevi. Genes differentially expressed in melanomas and nevi, on the basis of molecular signal, sub classified a group of unknown melanocytic lesions as melanomas or nevi and had high concordance rates with histopathology. Gene signatures established using DNA microarray gene expression profiling can distinguish melanomas from nevi, indicating the feasibility of using molecular classification as a supplement to standard histology. Our successful use of a standard formalin-fixed and paraffin-embedded tissue further supports the practicability of combining molecular diagnostic testing with histopathology in evaluation of difficult melanocytic lesions.",
author = "Koh, {Stephen S.} and Opel, {Michael L.} and Wei, {Jia Perng J.} and Kenneth Yau and Rashmi Shah and Gorre, {Mercedes E.} and Eric Whitman and Shitabata, {Paul K.} and Yong Tao and Cochran, {Alistair J.} and Payam Abrishami and Binder, {Scott W.}",
year = "2009",
month = "4",
day = "1",
doi = "10.1038/modpathol.2009.8",
language = "English (US)",
volume = "22",
pages = "538--546",
journal = "Modern Pathology",
issn = "0893-3952",
publisher = "Nature Publishing Group",
number = "4",

}

Koh, SS, Opel, ML, Wei, JPJ, Yau, K, Shah, R, Gorre, ME, Whitman, E, Shitabata, PK, Tao, Y, Cochran, AJ, Abrishami, P & Binder, SW 2009, 'Molecular classification of melanomas and nevi using gene expression microarray signatures and formalin-fixed and paraffin-embedded tissue', Modern Pathology, vol. 22, no. 4, pp. 538-546. https://doi.org/10.1038/modpathol.2009.8

Molecular classification of melanomas and nevi using gene expression microarray signatures and formalin-fixed and paraffin-embedded tissue. / Koh, Stephen S.; Opel, Michael L.; Wei, Jia Perng J.; Yau, Kenneth; Shah, Rashmi; Gorre, Mercedes E.; Whitman, Eric; Shitabata, Paul K.; Tao, Yong; Cochran, Alistair J.; Abrishami, Payam; Binder, Scott W.

In: Modern Pathology, Vol. 22, No. 4, 01.04.2009, p. 538-546.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Molecular classification of melanomas and nevi using gene expression microarray signatures and formalin-fixed and paraffin-embedded tissue

AU - Koh, Stephen S.

AU - Opel, Michael L.

AU - Wei, Jia Perng J.

AU - Yau, Kenneth

AU - Shah, Rashmi

AU - Gorre, Mercedes E.

AU - Whitman, Eric

AU - Shitabata, Paul K.

AU - Tao, Yong

AU - Cochran, Alistair J.

AU - Abrishami, Payam

AU - Binder, Scott W.

PY - 2009/4/1

Y1 - 2009/4/1

N2 - Melanoma may be difficult to identify histologically and relatively high rates of misdiagnosis leads to many malpractice claims. Currently separation of melanomas from nevi is based primarily on light microscopic interpretation of hematoxylin and eosin stained sections with limited assistance from immunohistology. To increase the accuracy of discrimination of benign and malignant melanocytic lesions we identified DNA microarray-derived gene expression profiles of different melanocytic lesions and evaluated the performance of these gene signatures as molecular diagnostic tools in the molecular classification and separation of melanomas and nevi. Melanocyte-derived cells were isolated by laser capture microdissection from 165 formalin-fixed and paraffin-embedded melanocytic nevi and melanoma tissue sections. RNA was isolated, amplified, labeled, and hybridized to a custom DNA microarray. In all 120 samples were used to identify differentially expressed genes and generate a gene expression classifier capable of distinguishing between melanomas and nevi. These classifiers were tested by the leave-one-out method and in a blinded study. RT-PCR verified the results. Unsupervised hierarchical clustering identified two distinct lesional groups that closely correlated with the histopathologically identified melanomas and nevi. Analysis of gene expression levels identified 36 significant differentially expressed genes. In comparison with nevi, melanomas expressed higher levels of genes promoting signal transduction, transcription, and cell growth. In contrast, expression of L1CAM (homolog) was reduced in melanomas relative to nevi. Genes differentially expressed in melanomas and nevi, on the basis of molecular signal, sub classified a group of unknown melanocytic lesions as melanomas or nevi and had high concordance rates with histopathology. Gene signatures established using DNA microarray gene expression profiling can distinguish melanomas from nevi, indicating the feasibility of using molecular classification as a supplement to standard histology. Our successful use of a standard formalin-fixed and paraffin-embedded tissue further supports the practicability of combining molecular diagnostic testing with histopathology in evaluation of difficult melanocytic lesions.

AB - Melanoma may be difficult to identify histologically and relatively high rates of misdiagnosis leads to many malpractice claims. Currently separation of melanomas from nevi is based primarily on light microscopic interpretation of hematoxylin and eosin stained sections with limited assistance from immunohistology. To increase the accuracy of discrimination of benign and malignant melanocytic lesions we identified DNA microarray-derived gene expression profiles of different melanocytic lesions and evaluated the performance of these gene signatures as molecular diagnostic tools in the molecular classification and separation of melanomas and nevi. Melanocyte-derived cells were isolated by laser capture microdissection from 165 formalin-fixed and paraffin-embedded melanocytic nevi and melanoma tissue sections. RNA was isolated, amplified, labeled, and hybridized to a custom DNA microarray. In all 120 samples were used to identify differentially expressed genes and generate a gene expression classifier capable of distinguishing between melanomas and nevi. These classifiers were tested by the leave-one-out method and in a blinded study. RT-PCR verified the results. Unsupervised hierarchical clustering identified two distinct lesional groups that closely correlated with the histopathologically identified melanomas and nevi. Analysis of gene expression levels identified 36 significant differentially expressed genes. In comparison with nevi, melanomas expressed higher levels of genes promoting signal transduction, transcription, and cell growth. In contrast, expression of L1CAM (homolog) was reduced in melanomas relative to nevi. Genes differentially expressed in melanomas and nevi, on the basis of molecular signal, sub classified a group of unknown melanocytic lesions as melanomas or nevi and had high concordance rates with histopathology. Gene signatures established using DNA microarray gene expression profiling can distinguish melanomas from nevi, indicating the feasibility of using molecular classification as a supplement to standard histology. Our successful use of a standard formalin-fixed and paraffin-embedded tissue further supports the practicability of combining molecular diagnostic testing with histopathology in evaluation of difficult melanocytic lesions.

UR - http://www.scopus.com/inward/record.url?scp=63949087383&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=63949087383&partnerID=8YFLogxK

U2 - 10.1038/modpathol.2009.8

DO - 10.1038/modpathol.2009.8

M3 - Article

VL - 22

SP - 538

EP - 546

JO - Modern Pathology

JF - Modern Pathology

SN - 0893-3952

IS - 4

ER -