MALDI‐Imaging for Classification of Epithelial Ovarian Cancer Histotypes from a Tissue Microarray Using Machine Learning Methods
Oliver KleinFrederic KanterHagen KulbePaul JankCarsten DenkertGrit NebrichWolfgang SchmittZhiyang WuCatarina Alisa KunzeJalid SehouliSilvia Darb‐EsfahaniElena Ioana BraicuJan LellmannHerbert ThieleEliane T. Taube
54
Citation
40
Reference
10
Related Paper
Citation Trend
Abstract:
Precise histological classification of epithelial ovarian cancer (EOC) has immanent diagnostic and therapeutic consequences, but remains challenging in histological routine. The aim of this pilot study is to examine the potential of matrix-assisted laser desorption/ionization (MALDI) imaging mass spectrometry in combination with machine learning methods to classify EOC histological subtypes from tissue microarray.Formalin-fixed-paraffin-embedded tissue of 20 patients with ovarian clear-cell, 14 low-grade serous, 19 high-grade serous ovarian carcinomas, and 14 serous borderline tumors are analyzed using MALDI-Imaging. Classifications are computed by linear discriminant analysis (LDA), support vector machines with linear (SVM-lin) and radial basis function kernels (SVM-rbf), a neural network (NN), and a convolutional neural network (CNN).MALDI-Imaging and machine learning methods result in classification of EOC histotypes with mean accuracy of 80% for LDA, 80% SVM-lin, 74% SVM-rbf, 83% NN, and 85% CNN. Based on sensitivity (69-100%) and specificity (90-99%), CCN and NN are most suited to EOC classification.The pilot study demonstrates the potential of MALDI-Imaging derived proteomic classifiers in combination with machine learning algorithms to discriminate EOC histotypes. Applications may support the development of new prognostic parameters in the assessment of EOC.Keywords:
Tissue microarray
We describe an ectopic ovary in a stillborn female. To our knowledge, this is the first report of an extra ovary in an infant. This case prompted a review of ectopic ovarian tissue, which is known by a variety of terms, the most common being accessory ovary and supernumerary ovary. We suggest that (1) many of the past cases should be classified as ovarian implants rather than true embryologically derived ectopic tissue; and (2) the terms accessory ovary and supernumerary ovary are imprecise and should be modified.
Cite
Citations (91)
Combining cDNA microarray with RNA i n situ hybridization on frozen tissue microarray to identify novel candidate oncogenes and provide possible theoretical basis for early diagnosis and treatment of ovarian c ancer.cDNA microarrays were used to seek significantly expressed genes in 3t ypes of ovarian tumors(serous borderline ovarian tumors,s erous ovarian cancers,and endometrioid ovarian carcinoma s).RNA in situ hybridization on frozen tissue microarray was used to confirm the fi nding from cDNA microarrays.In the study of cDNA microarray,28genes were over-expr essed and 18genes were under-expressed in 70%ovarian tumors.Interferon induced transme mbrane protein 1(IITP1)was further validated by RNA in situ hybridization on frozen tissue microarray.[Conclusion]The methods through combining cDNA microarray with RNA in situ hybridiz ation on frozen tissue microarray,i s an ideal choice for identifying novel oncogenes.The ge nes identified in this study might be the new candidate oncogenes of ovarian cancer.
Tissue microarray
Gene chip analysis
Cite
Citations (0)
Objective:To search for the most adequate duration and conditions of tissue culture to reduce the immunogenicity and therefore raise survival rate of the cultivated tissue after transplantation.Methods:The changes of morphology and enzymatic activity of the cultivated ovary tissue were observed before and after cultivating for 1d、2d and 3d respectively.Results:Ovarian enzyme activity appeared significant differences as compared with unculfured ovary within two day's culture.Conclusion:Under the same cultural conditions,the most appropriate cultural duration for the rat's ovary tissue is within two days.
Cite
Citations (0)
Vitrification
Cite
Citations (52)
Ovarian cancer is the leading cause of death among the gynecological malignancy mainly due to lacking of effective early diagnostic methods. To identify novel candidate genes and further explore their clinical significance of epithelial ovarian tumors, we developed a new method in our laboratory by combining cDNA microarray with RNA in situ hybridization in frozen tissue microarray.cDNA microarrays were used to seek differentially expressed genes among 3 subtypes of ovarian tumors (serous borderline ovarian tumors, serous ovarian cancers, and endometrioid ovarian carcinomas). RNA in situ hybridization in frozen tissue microarray was used to further confirm the findings from cDNA microarrays.In the study of cDNA microarray, 40 genes and ESTs showed significant differential expression between low and high-malignancy, as well as serous and endometrioid carcinomas. EPHB6, PTPRF, GFER, ERG25, PLRP1, FLJ22060, and WISP2 were further validated by RNA in situ hybridization in tissue microarray.cDNA microarray combined with RNA in situ hybridization in frozen tissue microarray is an ideal choice for identifying novel oncogenes. EPHB6, PTPRF, GFER, ERG25, PLRP1, FLJ22060, and WISP2 might become the new candidate oncogenes for epithelial ovarian cancer.
Tissue microarray
Ovarian tumor
Cite
Citations (5)
To study the value of screening hereditary nonpolyposis colorectal cancer (HNPCC) kindreds by detecting the expressions of hMLH1/hMSH2 with tissue microarray.A tissue microarray with 22 colorectal cancers from HNPCC families and 15 sporadic colorectal cancers was established, and the expressions of hMLH1/hMSH2 were detected by immunohistochemistry (IHC).The expressions of hMLH1 or hMSH2 were negative in 15 of 22 HNPCC and 1 of 15 sporadic colorectal cancers in routine IHC. The expressions of hMLH1 or hMSH2 were negative in 17 of 22 HNPCC and 2 of 15 sporadic colorectal cancers in tissue microarray. The examination of hMSH2 expression yielded same results between routine IHC and tissue microarray. There were no difference on the hMLH1 expressions between routine IHC and tissue microarray.Tissue microarray is a high-throughput way to detect the expressions of hMLH1/hMSH2 and is applicable to screen HNPCC kindreds.
Tissue microarray
Microsatellite Instability
Cite
Citations (0)
Objective: To detect the feasibility of tissue microarray so that a new method could be provided for research. Methods: Two tissue microarray blocks were manufactured and stained on the sections cutting from them by HE,immunohistochemical and fluorescence in situ hybridization(FISH) methods were manufactured and compared with the normal sections. Results: The cores on slides were basically accordant with requirement of research, and there were no significant differences between the results of tissue microarray and normal sections that were dealed with immunohistochemical method. Conclusion: Tissue microarray has characters of high-throughput, economy and low rate of experimental error; it might be a new technique in scientific researches in the future.
Tissue microarray
Cite
Citations (0)
BACKGROUND & OBJECTIVE Ovarian cancer is the leading cause of death among the gynecological malignancy mainly due to lacking of effective early diagnostic methods. To identify novel candidate genes and further explore their clinical significance of epithelial ovarian tumors, we developed a new method in our laboratory by combining cDNA microarray with RNA in situ hybridization in frozen tissue microarray. METHODS cDNA microarrays were used to seek differentially expressed genes among 3 subtypes of ovarian tumors (serous borderline ovarian tumors, serous ovarian cancers, and endometrioid ovarian carcinomas). RNA in situ hybridization in frozen tissue microarray was used to further confirm the findings from cDNA microarrays. RESULTS In the study of cDNA microarray, 40 genes and ESTs showed significant differential expression between low and high-malignancy, as well as serous and endometrioid carcinomas. EPHB6, PTPRF, GFER, ERG25, PLRP1, FLJ22060, and WISP2 were further validated by RNA in situ hybridization in tissue microarray. CONCLUSIONS cDNA microarray combined with RNA in situ hybridization in frozen tissue microarray is an ideal choice for identifying novel oncogenes. EPHB6, PTPRF, GFER, ERG25, PLRP1, FLJ22060, and WISP2 might become the new candidate oncogenes for epithelial ovarian cancer.
Tissue microarray
Gene chip analysis
Ovarian tumor
Cite
Citations (4)
Objective To exploethe significance of each type of tissue microarrary on the examine of associated genes production in breast carcinoma.Methods Tissue microarray technology and immunohistochemistry were used to detect the expression of ER,PR,nm23,Her-2 and p53 in 31 cases of breast carcinoma and compared theresult with the traditional pathological investigation.Results The difference was no found detween the immunohistochemistric result by 1.0 mm tissue microarray and by 0.5 mm tissue microarray.2 cases of ER and 1 of PR were negative by traditional pathological technology,but were positive by tissue microarray.1 case of nm23 and 1 of Her-2 were positive by traditional pathological technology,but were negative by tissue microarray.On the expressional strength of ER,PR,nm23,Her-2 and p53,the nigligible difference was found detween by traditional pathological technology and by tissue microarray in some cases.Conclusion On the positive rate of ER,PR,nm23,Her-2 and p53,the significant difference was no found detween the result with the traditional pathological investigation and with the tissue microarray(P0.05).It has important pracyical significance and broad application prospect in pathology.
Tissue microarray
Gene chip analysis
Breast carcinoma
Clinical Significance
Cite
Citations (0)
Tissue microarray
Microarray databases
Cite
Citations (39)