The diagnostic value of Ki-67 and repp86 in distinguishing between benign and malignant mesothelial proliferations

Immunochemistry in the work-up of mesothelioma and its differential diagnosis and mimickers

The differential diagnosis in cellular effusions with cytological atypia often includes malignant mesothelioma (MM), reactive mesothelial proliferation, and malignancies of metastatic origin, particularly carcinomas. The International Reporting System for Serous Fluid recently established guidelines for reporting MM. In conjunction with the cytomorphologic evaluation, the role of immunochemistry (IC) was emphasized as a very useful tool in the workup of serous fluids, especially with the availability of novel markers. Utilizing a panel of markers, IC allows the characterization of the cells, whether mesothelial or not, and when mesothelial origin is established, IC can frequently assist in delineating its benign or malignant nature. IC can also confirm metastatic disease, allowing the identification of the primary origin in most cases. This review summarizes the current status of IC and its role in the diagnosis of MM and its differential diagnosis in serous fluids.

Can BAP1 expression loss in mesothelial cells be an indicator of malignancy?

Background

Malignant mesothelioma is a highly aggressive tumor that can be confused with a benign mesothelial lesion, especially cytomorphologic lesions. BRCA1-associated protein 1 (BAP1) acts as a tumor suppressor. In this study, we aim to investigate the value of BAP1 staining of malignant mesothelioma cases with expression loss and diagnosis in cell block and biopsy tissue.

Methods

Between January 2009 and March 2017, 64 mesotheliomas, 117 reactive mesothelial hyperplasias, and 20 fibrinous pleuritis/pericarditis were diagnosed with morphologic and immunohistochemical findings in our pathology clinic and were included in the study. Formalin-fixed, paraffin-embedded tissues were immunohistochemically examined for BAP1. Inflammatory and stromal cells were used as positive internal controls. BAP1 was assessed for nuclear staining in mesothelial cells.

Results

Examinations of the relationship between patient diagnosis and BAP1 biopsy status showed that the BAP1 loss rate (76.6%) was significantly higher in malignant mesothelioma cases than in other benign diseases (0%) (p<.001). Sensitivity and specificity were 76.56% and 100%, respectively, for biopsy tissue from malignant mesothelioma. Sensitivity and specificity were both 100% for BAP1 test on cell block tissue. Furthermore, the consistency between BAP1 cell block and biopsy results was excellent (ĸ=0.90) and the correlation was significant (p<.001).

Conclusions

This study shows that BAP1 expression loss in both cytology and biopsy tissue in biopsy-confirmed malignant mesothelioma cases is an essential parameter for malignant mesothelioma diagnosis.

Diagnostic value of GLUT-1 in distinguishing malignant mesothelioma from reactive mesothelial cells: a meta-analysis

Background: Previous studies have demonstrated the diagnostic value of glucose transporter 1 (GLUT-1) to distinguish malignant mesothelioma (MM) from reactive mesothelial cells (RMC), but the results are inconsistent. The purpose of this meta-analysis is to investigate the diagnostic accuracy of GLUT-1 in distinguishing MM from RMC.Methods: A systematical search was conducted until May 2019 in PubMed, Medline, Embase and the Cochrane Library. The revised tool for the quality assessment of diagnostic accuracy studies (QUADAS-2) was used to assess the quality of the eligible studies. The Stata15 and Review Manager5.3 software programmes were used to perform the meta-analysis.Results: A total of 24 studies, including 969 MM patients and 1080 RMC individuals were explored in the meta-analysis. The summary assessments revealed that the pooled sensitivity was 0.73 (95% CI, 0.62-0.81) and the pooled specificity was 0.95 (95% CI, 0.91-0.98). The area under the summary ROC curve (AUC) was 0.93 (95% CI: 0.91-0.95).Conclusions: GLUT-1 is highly accurate to distinguish MM from RMC.

Utility of Survivin, BAP1, and Ki-67 immunohistochemistry in distinguishing epithelioid mesothelioma from reactive mesothelial hyperplasia

Histological distinction between epithelioid mesothelioma (EM) and reactive mesothelial hyperplasia (RMH) can be challenging. The aim of this study was to assess the diagnostic utility of Survivin, Ki-67, and loss of BRCA1-associated protein 1 (BAP1) expressions in distinguishing EM from RMH using immunohistochemistry. Formalin-fixed, paraffin-embedded specimens from 78 cases of EM and 80 cases of RMH were immunohistochemically examined for Survivin, BAP1, and Ki-67. In addition, receiver operating characteristic curve analyses were performed to establish the cut-off values for Survivin and Ki-67 labelling indices. Survivin (cut-off value: 5%) had 67.7% sensitivity and 100% specificity, while Ki-67 (cut-off value: 10%) had 85.1% sensitivity and 87.5% specificity, and BAP1 had 66.2% sensitivity and 100% specificity for the differentiation of EM from RMH. Among the combinations of two markers, the combination of Survivin and BAP1 (Survivin-positive and/or BAP1-loss finding) had the highest diagnostic accuracy (sensitivity: 89.8%; specificity: 100%; accuracy: 95.3%). We recommend using the combination of Survivin and BAP1 to distinguish EM from RMH.

Overexpressed genes in malignant pleural mesothelioma: implications in clinical management

Malignant pleural mesothelioma (MPM) is a very aggressive cancer poorly responsive to current therapies. MPM patients have a very poor prognosis with a median survival of less than one year from the onset of symptoms. The biomarkers proposed so far do not lead to a sufficiently early diagnosis for a radical treatment of the disease. Thus, the finding of novel diagnostic and prognostic biomarkers and therapeutic targets is needed. Gene overexpression has been frequently associated with a malignant phenotype in several cancer types; therefore the identification of overexpressed genes may lead to the detection of novel prognostic or diagnostic marker and to the development of novel therapeutic approaches, based on their inhibition. In the last years, several overexpressed genes have been identified in MPM through gene expression profiling techniques: among them it has been found a group of 51 genes that resulted overexpressed in more than one independent study, revealing their consistency among studies. This article reviews the clinical implications of confirmed overexpressed genes in MPM described so far in literature.

Molecular markers and new diagnostic methods to differentiate malignant from benign mesothelial pleural proliferations: a literature review

Malignant pleural mesothelioma (MPM) is an aggressive tumor associated with asbestos exposure. Histopathological analysis of pleural tissues is the gold standard for diagnosis; however, it can be difficult to differentiate malignant from benign pleural lesions. The purpose of this review is to describe the most important biomarkers and new diagnostic tools suggested for this differential diagnosis. There are many studies concerning the separation between MPM and benign pleural proliferations from both pleural tissues or effusions; most of them are based on the evaluation of one or few biomarkers by immunohistochemistry (IHC) or enzyme-linked immunosorbent assays (ELISAs), whereas others focused on the identification of MPM signatures given by microRNA (miRNA) or gene expression profiles as well as on the combination of molecular data and classification algorithms. None of the reported biomarkers showed adequate diagnostic accuracy, except for p16 [evaluated by fluorescent in situ hybridization (FISH)] and BAP1 (evaluated by IHC), both biomarkers are recommended by the International Mesothelioma Interest Group guidelines for histological and cytological diagnosis. BAP1 and p16 showed a specificity of 100% in discerning malignant from benign lesions because they are exclusively unexpressed or deleted in MPM. However, their sensitivity, even when used together, is not completely sufficient, and absence of their alterations cannot confirm the benign nature of the lesion. Recently, the availability of new techniques and increasing knowledge regarding MPM genetics led to the definition of some molecular panels, including genes or miRNAs specifically deregulated in MPM, that are extremely valuable for differential diagnosis. Moreover, the development of classification algorithms is facilitating the application of molecular data for clinical practice. Data regarding new diagnostic tools and MPM signatures are absolutely promising; however, before their application in clinical practice, a prospective validation is necessary, as these approaches could surely improve the differential diagnosis between malignant and benign pleural lesions.

The legacy of the F344 rat as a cancer bioassay model (a retrospective summary of three common F344 rat neoplasms)

The Fischer 344 (F344) rat was used by the National Toxicology Program (NTP) for over 5 decades for toxicity and carcinogenicity studies. However, in 2006, the NTP decided to switch to a different rat stock due largely to high background control incidences of Leydig cell tumors (LCTs) and mononuclear cell leukemia (MNCL), also known as large granular lymphocytic (LGL) leukemia. In the current review, we aim (1) to provide a summary of NTP bioassays with treatment-associated effects involving MNCL and LCTs in addition to male F344-specific tunica vaginalis mesothelioma (TVM); (2) to describe important pathobiological differences between these F344 rat tumor responses and similar target tissue-tumor response in humans; and (3) to present the NTP reasons for switching away from the F344 rat. We show that due to the highly variable background incidence of F344 MNCL, more reliance on historical control data than is usual for most tumor responses is warranted to evaluate potential effect of any chemical treatment in this rat strain. The high spontaneous incidence of LCTs in the testes of male F344 rats has made this tumor endpoint of little practical use in identifying potential testicular carcinogenic responses. TVM responses in F344 rats have a biological plausible relationship to LCTs unlike TVM in humans. Given their high spontaneous background incidence and species-specific biology, we contend that MNCL and LCT, along with TVM responses, in F344 rat carcinogenicity studies are inappropriate tumor types for human health risk assessment and lack relevance in predicting human carcinogenicity.

TPX2: of spindle assembly, DNA damage response, and cancer

For more than 15 years, TPX2 has been studied as a factor critical for mitosis and spindle assembly. These functions of TPX2 are attributed to its Ran-regulated microtubule-associated protein properties and to its control of the Aurora A kinase. Overexpressed in cancers, TPX2 is being established as marker for the diagnosis and prognosis of malignancies. During interphase, TPX2 resides preferentially in the nucleus where its function had remained elusive until recently. The latest finding that TPX2 plays a role in amplification of the DNA damage response, combined with the characterization of TPX2 knockout mice, open new perspectives to understand the biology of this protein. This review provides an historic overview of the discovery of TPX2 and summarizes its cytoskeletal and signaling roles with relevance to cancer therapies. Finally, the review aims to reconcile discrepancies between the experimental and pathological effects of TPX2 overexpression and advances new roles for compartmentalized TPX2.

CD146 and insulin-like growth factor 2 mRNA-binding protein 3 predict prognosis of asbestos-induced rat mesothelioma

Malignant mesothelioma (MM), which is associated with asbestos exposure, is one of the most deadly tumors in humans. Early MM is concealed in the serosal cavities and lacks specific clinical symptoms. For better treatment, early detection and prognostic markers are necessary. Recently, CD146 and insulin-like growth factor 2 mRNA-binding protein 3 (IMP3) were reported as possible positive markers of MM to distinguish from reactive mesothelia in humans. However, their application on MM of different species and its impact on survival remain to be elucidated. To disclose the utility of these molecules as early detection and prognostic markers of MM, we injected chrysotile or crocidolite intraperitoneally to rats, thus obtaining 26 peritoneal MM and establishing 11 cell lines. We immunostained CD146 and IMP3 using paraffin-embedded tissues and cell blocks and found CD146 and IMP3 expression in 58% (15/26) and 65% (17/26) of MM, respectively, but not in reactive mesothelia. There was no significant difference in both immunostainings for overexpression among the three histological subtypes of MM and the expression of CD146 and IMP3 was proportionally associated. Furthermore, the overexpression of CD146 and/or IMP3 was proportionally correlated with shortened survival. These results suggest that CD146 and IMP3 are useful diagnostic and prognostic markers of MM.

Mitotic Stress and Chromosomal Instability in Cancer: The Case for TPX2

Cell cycle deregulation is a common motif in human cancer, and multiple therapeutic strategies are aimed to prevent tumor cell proliferation. Whereas most current therapies are designed to arrest cell cycle progression either in G1/S or in mitosis, new proposals include targeting the intrinsic chromosomal instability (CIN, an increased rate of gain or losses of chromosomes during cell division) or aneuploidy (a genomic composition that differs from diploid) that many tumor cells display. Why tumors cells are chromosomally unstable or aneuploid and what are the consequences of these alterations are not completely clear at present. Several mitotic regulators are overexpressed as a consequence of oncogenic alterations, and they are likely to alter the proper regulation of chromosome segregation in cancer cells. In this review, we propose the relevance of TPX2, a mitotic regulator involved in the formation of the mitotic spindle, in oncogene-induced mitotic stress. This protein, as well as its partner Aurora-A, is frequently overexpressed in human cancer, and its deregulation may participate not only in chromosome numeric aberrations but also in other forms of genomic instability in cancer cells.

Automated Quantification of MART1-Verified Ki67 Indices by Digital Image Analysis in Melanocytic Lesions

Context.—The proliferation marker Ki67 is an important diagnostic and prognostic aid in surgical pathology. However, manual quantification in a counting frame to accurately establish the proliferation rate (Ki67 index) is cumbersome and time-consuming. Instead, digital image analysis of Ki67/MART1 double stains may provide fast and novel index computations for entire tumor sections.

Objectives.—To design and compare image analysis protocols that compute Ki67 indices of Ki67/MART1 double stains, to compare automated indices with previously published manual indices, and to compare the total number of proliferating cells (mimicking a Ki67 single stain) with the number of MART1-verified proliferating cells.

Design.—Whole slide images were captured from 48 melanomas and 77 nevi stained with an immunohistochemical cocktail against Ki67 and MART1. Ki67 indices were determined by digital image analysis and different equations based on number or area.

Results.—The differences between mean indices of melanomas and nevi were significant (P &lt; .001) in all index computations. Number-based image analysis of lesions with more than 250 melanocytic cells misclassified 1 of 42 melanomas and 4 of 53 nevi, numbers comparable with manual counting. Automated indices were significantly higher than manual indices, as were indices of mimicked Ki67 single stains compared with MART1-verified Ki67 indices (P &lt; .001).

Conclusions.—Ki67 indices established by digital image analysis of Ki67/MART1 double stains demonstrated excellent abilities to discriminate melanomas from nevi with diagnostic performances equal to manually performed indices. Testing different definitions of the automated MART1-verified Ki67 index, no single definition stood out; thus, a variety of definitions may be used.

Immunohistochemistry and Molecular Diagnostics of Pleural Malignant Mesothelioma

Context.—The pathologic approach to pleural-based lesions is stepwise and uses morphologic assessment, correlated with clinical and imaging data supplemented by immunohistochemistry (IHC), and more recently, molecular tests, as an aid for 2 main diagnostic problems: malignant mesothelioma (MM) versus other malignant tumors and malignant versus reactive mesothelial proliferations.

Objective.—To present the current knowledge regarding IHC and molecular tests with respect to MM diagnosis, and in particular, the differentiation of the epithelioid type of MM from carcinoma metastatic to the pleural cavity.

Data Sources.—A review of immunohistochemical features of 286 consecutive MMs from 459 cases of pleural pathology, diagnosed during routine practice from 2003 to 2009. A survey of biomedical journal literature from MedLine/PubMed (US National Library of Medicine) focused on MM and associated tissue-based diagnostic IHC markers and molecular tests.

Conclusions.—The search for a single diagnostic marker of MM has so far been discouraging, given the biologic and phenotypic tumor heterogeneity of MM. The use of antibody panels has gained unanimous acceptance especially in the differential diagnosis between MM and metastatic carcinoma, whereas the usefulness of IHC is more limited when dealing with spindle cell malignancies or distinguishing malignant from reactive mesothelium. A great degree of interlaboratory variability in antibody combinations and clone selection within diagnostic panels still exists. Current investigations aim at selecting the most suitable and cost-effective combination of antibodies by using novel statistical approaches for assessing diagnostic performance beyond the traditional measures of sensitivity and specificity.

Recombinant lysyl oxidase propeptide protein inhibits growth and promotes apoptosis of pre-existing murine breast cancer xenografts

Lysyl oxidase propeptide (LOX-PP) ectopic overexpression inhibits the growth of cancer xenografts. Here the ability and mode of action of purified recombinant LOX-PP (rLOX-PP) protein to inhibit the growth of pre-existing xenografts was determined. Experimental approaches employed were direct intratumoral injection (i.t.) of rLOX-PP protein into murine breast cancer NF639 xenografts, and application of a slow release formulation of rLOX-PP implanted adjacent to tumors in NCR nu/nu mice (n = 10). Tumors were monitored for growth, and after sacrifice were subjected to immunohistochemical and Western blot analyses for several markers of proliferation, apoptosis, and for rLOX-PP itself. Direct i.t. injection of rLOX-PP significantly reduced tumor volume on days 20, 22 and 25 and tumor weight at harvest on day 25 by 30% compared to control. Implantation of beads preloaded with 35 micrograms rLOX-PP (n = 10) in vivo reduced tumor volume and weight at sacrifice when compared to empty beads (p<0.05). A 30% reduction of tumor volume on days 22 and 25 (p<0.05) and final tumor weight on day 25 (p<0.05) were observed with a reduced tumor growth rate of 60% after implantation. rLOX-PP significantly reduced the expression of proliferation markers and Erk1/2 MAP kinase activation, while prominent increases in apoptosis markers were observed. rLOX-PP was detected by immunohistochemistry in harvested rLOX-PP tumors, but not in controls. Data provide pre-clinical findings that support proof of principle for the therapeutic anti-cancer potential of rLOX-PP protein formulations.

A review of transcriptome studies combined with data mining reveals novel potential markers of malignant pleural mesothelioma

Malignant pleural mesothelioma (MPM), a cancer of the serosal pleural cavities, is one of the most aggressive human tumors. In order to identify genes crucial for the onset and progression of MPM, we performed an extensive literature review focused on transcriptome studies (RTS). In this kind of studies a great number of transcripts are analyzed without formulating any a priori hypothesis, thus preventing any bias coming from previously established knowledge that could lead to an over-representation of specific genes. Each study was thoroughly analyzed paying particular attention to: (i) the employed microarray platform, (ii) the number and type of samples, (iii) the fold-change, and (iv) the statistical significance of deregulated genes. We also performed data mining (DM) on MPM using three different tools (Coremine, SNPs3D, and GeneProspector). Results from RTS and DM were compared in order to restrict the number of genes potentially deregulated in MPM. Our main requirement for a gene to be a "mesothelioma gene" (MG) is to be reproducibly deregulated among independent studies and confirmed by DM. A list of MGs was thus produced, including PTGS2, BIRC5, ASS1, JUNB, MCM2, AURKA, FGF2, MKI67, CAV1, SFRP1, CCNB1, CDK4, and MSLN that might represent potential novel biomarkers or therapeutic targets for MPM. Moreover, it was found a sub-group of MGs including ASS1, JUNB, PTGS2, EEF2, SULF1, TOP2A, AURKA, BIRC5, CAV1, IFITM1, PCNA, and PKM2 that could explain, at least in part, the mechanisms of resistance to cisplatin, one first-line chemotherapeutic drug used for the disease. Finally, the pathway analysis showed that co-regulation networks related to the cross-talk between MPM and its micro-environment, in particular involving the adhesion molecules, integrins, and cytokines, might have an important role in MPM. Future studies are warranted to better characterize the role played by these genes in MPM.

Adenomatoid tumors of the female and male genital tract. A comparative clinicopathologic and immunohistochemical analysis of 47 cases emphasizing their site-specific morphologic diversity

Adenomatoid tumors (ATs) are uncommon benign mesothelial tumors with a predilection for the genital tract. We reviewed 47 ATs diagnosed at our institutions during 10-year period. Thirty tumors (64%) originated in the female (21 uterine, 8 tubal, and 1 ovarian) and 17 (36%) in the male (9 epididymal and 8 testicular) genital tract. The median age for females and males was 47.5 and 51 years, respectively. While 83% of tumors in females were incidental findings in resections for unrelated diseases, 94% of male lesions presented as clinical masses leading to surgery. The median size was 2, 1, and 0.5 cm for uterine, epididymo-testicular, and tubo-ovarian lesions, respectively. Architecturally, the microcystic/angiomatoid pattern was the most frequent (32/47; 68%), followed by combined microcystic/trabecular (26/47; 55%) and retiform/adenoid (15/47; 32%) pattern. The trabecular/solid (6%) and macrocystic (4%) patterns were uncommon. However, 57% of cases revealed ≥2 growth patterns. Taken by anatomic site, 20 of 21 uterine cases were at least focally microcystic but none was retiform. In contrast, the retiform pattern dominated in male genital tract tumors (12/17; 71%). Immunohistochemistry showed expression of calretinin (36/36) and D2-40 (30/30) and lack of CD34 (0/30) and PAX8 (0/32). GLUT-1 was expressed in 11/11 male genital tract tumors but in none of the microcystic uterine lesions. Estrogen and progesterone receptor expression was weak and focal (two and three uterine cases, respectively). None stained for the androgen receptor. Our study illustrates the great site-specific morphological diversity of ATs emphasizing their wide site-dependent differential diagnosis.

Mesothelioma of the tunica vaginalis: a series of eight cases with uncertain malignant potential

Well-differentiated papillary mesotheliomas have rarely been reported to involve the testis tunica vaginalis. While the classic histology of the originally described well-differentiated papillary mesotheliomas consisted of papillae lined by a single layer of bland cuboidal cells, more complex architectural patterns have been described. This report details our experience with eight paratesticular mesotheliomas that span the histological spectrum from classic well-differentiated papillary mesotheliomas to those with more complex patterns. We attempt to determine whether there are prognostic or immunohistochemical differences between these lesions and 28 diffuse malignant mesotheliomas. All cases had papillary/tubulopapillary in addition to more complex architectures (cribriform, condensed) and none showed evidence of invasion. Mitotic figures were present in seven cases and averaged 2.1 mitosis per 50 h.p.f. All cases showed <or=1% of positive cells staining for Ki-67 and p53, and only one case stained positive for GLUT-1 in comparison with 50% positivity for GLUT-1 in diffuse mesotheliomas. Of five patients with follow-up information of more than 1 year, three patients were alive at 2, 3 and 9 years, and two died of unknown causes at 5 and 47 years, suggesting that these lesions behave more indolently than typical malignant mesotheliomas of the tunica vaginalis. If these cases were diffuse malignant mesotheliomas, one would have expected at least some of the patients to have died of disease within a couple of years after diagnosis. However, detailed and long-term follow-up were not sufficiently available to reach definitive conclusions on the true biological behavior of these tumors. The morphological continuum noted between our cases and classic well-differentiated papillary mesotheliomas, combined with their immunohistochemical profile and indolent behavior, contrasts with that of diffuse malignant mesotheliomas. We propose that these tunica vaginalis mesotheliomas, which are more complex than the classic well-differentiated papillary mesotheliomas, and yet are not overtly histologically malignant, be classified as 'mesotheliomas of uncertain malignant potential'.

Benign and Malignant Mesothelial Proliferation

Malignant mesothelioma (MM) is a rare primary malignant tumor of the surface serosal cells. The diagnosis of MM is challenging with a broad differential diagnosis. For many decades, studies have focused on distinguishing MM from other types of cancer; however, benign mesothelial cell hyperplasia, especially in small biopsies, has emerged as a major problem. The features of pleural lesions are somewhat different from peritoneal diseases, and this article primarily focuses on pleural diseases. Thorough interpretation and correlation of clinical, radiologic, and pathologic findings are essential for a correct diagnosis.

Predicting neuroendocrine tumor (carcinoid) neoplasia using gene expression profiling and supervised machine learning

BACKGROUND

A more accurate taxonomy of small intestinal (SI) neuroendocrine tumors (NETs) is necessary to accurately predict tumor behavior and prognosis and to define therapeutic strategy. In this study, the authors identified a panel of such markers that have been implicated in tumorigenicity, metastasis, and hormone production and hypothesized that transcript levels of the genes melanoma antigen family D2 (MAGE-D2), metastasis-associated 1 (MTA1), nucleosome assembly protein 1-like (NAP1L1), Ki-67 (a marker of proliferation), survivin, frizzled homolog 7 (FZD7), the Kiss1 metastasis suppressor (Kiss1), neuropilin 2 (NRP2), and chromogranin A (CgA) could be used to define primary SI NETs and to predict the development of metastases.

METHODS

Seventy-three clinically and World Health Organization pathologically classified NET samples (primary tumor, n = 44 samples; liver metastases, n = 29 samples) and 30 normal human enterochromaffin (EC) cell preparations were analyzed using real-time polymerase chain reaction. Transcript levels were normalized to 3 NET housekeeping genes (asparagine-linked glycosylation 9 or ALG9, transcription factor CP2 or TFCP2, and zinc finger protein 410 or ZNF410) using geNorm analysis. A predictive gene-based model was constructed using supervised learning algorithms from the transcript expression levels.

RESULTS

Primary SI NETs could be differentiated from normal human EC cell preparations with 100% specificity and 92% sensitivity. Well differentiated NETs (WDNETs), well differentiated neuroendocrine carcinomas, and poorly differentiated NETs (PDNETs) were classified with a specificity of 78%, 78%, and 71%, respectively; whereas poorly differentiated neuroendocrine carcinomas were misclassified as either WDNETs or PDNETs. Metastases were predicted in all cases with 100% sensitivity and specificity.

CONCLUSIONS

The current results indicated that gene expression profiling and supervised machine learning can be used to classify SI NET subtypes and accurately predict metastasis. The authors believe that the application of this technique will facilitate accurate molecular pathologic delineation of NET disease, better define its extent, facilitate the assessment of prognosis, and provide a guide for the identification of appropriate strategies for individualized patient treatment.