分子病理應用在唾液腺透明性細胞癌和唾液腺類乳腺分泌性癌的診斷及病理型態學與免疫染色上的新發現

Abstract

近年來在唾液腺腫瘤的分子病理致病機轉上有許多重大的成果，過去以型態學特徵來分類的腫瘤都紛紛被發現出帶有特定的染色體轉位與基因突變，利用分子病理找出這些具腫瘤特異性的染色體轉位與基因突變，讓我們能夠更精確的診斷唾液腺腫瘤，並進一步比較不同腫瘤間型態學及免疫染色上的異同。本論文主要研究唾液腺透明性細胞癌和唾液腺類乳腺分泌性癌，以及SOX10免疫染色在唾液腺腫瘤診斷上的應用。 唾液腺透明性細胞癌帶有特定的EWSR1基因轉位，它在型態學上與較為常見的粘液表皮樣癌有重疊的地方會造成診斷的困擾，粘液表皮樣癌則有特定的MAML2基因轉位，實驗上先用螢光原位雜交法將49 例粘液表皮樣癌的案例分成有MAML2基因轉位(32例)與無MAML2基因轉位(17例)兩群，再去檢驗無MAML2基因轉位群是否帶有EWSR1基因轉位，發現其中3例是有EWSR1基因轉位，因此應該歸類為透明性細胞癌而非原本的粘液表皮樣癌。進一步比較有透明性細胞的粘液表皮樣癌與透明性細胞癌(8例)後發現，透明性細胞、粘液、與玻璃樣變基質在這兩種唾液腺腫瘤均可見到因而不具特異性，反而是細胞質多的表皮樣細胞與黏液細胞才是真正有特異性的型態學特徵。 唾液腺類乳腺分泌性癌帶有特定的ETV6基因轉位，在未發現它的ETV6基因轉位前，過去常常被診斷為腺細胞癌，因此過去對腺細胞癌的病理學研究實際上是混雜了兩種不同的腫瘤後得出的結果，有需要修正。腺細胞癌過去認為有四種型態學結構，分別是微囊性、實心性、乳突-囊性、濾泡性。實驗上先用螢光原位雜交法發現過去21例診斷為腺細胞癌中有6例實際上是有ETV6基因轉位的唾液腺類乳腺分泌性癌。進一步比較腺細胞癌(21例)與唾液腺類乳腺分泌性癌(14例)後發現，乳突-囊性結構是唾液腺類乳腺分泌性癌而非腺細胞癌的特徵，乳突-囊性結構不但在診斷上有重要意義，更修正了過去對腺細胞癌的錯誤認知。 SOX10免疫染色在軟組織病理上有重要應用，它會表現在黑色素瘤、許旺氏細胞腫瘤、及肌上皮細胞腫瘤。很特別的是在正常唾液腺的腺細胞及小管腔細胞也會表現SOX10，而大的管腔細胞則不表現，本研究主要是評估SOX10在各種不同唾液腺腫瘤的表現，發現SOX10會表現於腺細胞癌、腺樣囊性癌、類乳腺分泌性癌、上皮-肌上皮細胞癌、唾液腺母細胞癌、基底細胞腫瘤、及多形性腺瘤上，SOX10免疫染色可應用在唾液腺腫瘤診斷上。

In recent years there are great advances in the molecular pathology of salivary gland tumors. Many tumor-specific molecular changes including translocations and point mutations were identified. Using these tumor-specific genetic changes we can now classify salivary gland tumors more precisely and their unique hsitopathologic features can be more clearly defined. Hyalinizing clear cell carcinoma (HCCC) is a rare salivary gland tumor with a specific EWSR1-ATF1 fusion gene and can have mucin production. Mucoepidermoid carcinoma (MEC) with a clear cell component is its morphologic mimic. Using MAML2 fluorescence in situ hybridization (FISH), a total of 49 MEC cases were separated into MAML2-fusion-positive (32 cases) and MAML2-fusion-negative groups (17 cases). This study used EWSR1 FISH to investigate MAML2-fusion-negative cases to identify previously unrecognized HCCC. Among 17 MAML2-fusion-negative cases, 3 had rearrangement of EWSR1 gene and reclassified as HCCC. Including 5 previously diagnosed HCCC cases, these 8 HCCC cases had a male-to-female ratio of 1:7, and most (7/8) tumors arose from oral minor salivary glands in oral cavity (tongue base and palate). EWSR1-ATF1 fusion was confirmed by FISH in all 8 HCCC cases. The histologic features between genetically confirmed HCCC and MEC were compared. HCCC was significantly associated with minor salivary gland involvement, a discrepancy between low-grade cytology and intermediate to high-grade histology using the MEC grading system, and absence of both epidermoid cells with abundant cytoplasm and goblet cells lining cysts or forming clusters. Clear cells and a hyalinized stroma were not specific for HCCC. HCCC may be erroneously classified as MEC as clear cell may be the minor histologic component and mucin production is not uncommon. Previously diagnosed MEC cases should be reevaluated, especially those arising from minor salivary glands or without MAML2 fusion. Careful histologic evaluation with supporting molecular testing can facilitate pathologic diagnoses. MASC has a specific ETV6-NTRK3 translocation and morphologically overlaps with AciCC. AciCC is generally thought to have four histologic patterns including microcystic, solid, papillary-cystic, and follicular patterns before the recognition of MASC. The aim of this study was to evaluate histologic patterns in these two neoplasms through comprehensive histologic subtyping. Using FISH, we identified 14 cases of MASC and 21 cases of AciCC. We used comprehensive histologic subtyping to provide a semiquantitive assessment of histologic patterns in each tumor and performed immunohistochemical analyses including S100/vimentin/ mammaglobin/DOG1. In these analyses, MASC often presented papillary-cystic patterns without a solid component; however, this histologic feature rarely appeared in AciCC. An immunohistochemical panel of S100/mammaglobin/DOG1 is useful for differential diagnosis. Papillary-cystic pattern was previously considered to be one of the four major patterns associated with AciCC; however, it is actually a rare presentation in this type of carcinoma and appears to be more characteristic of MASC. Comprehensive subtyping of histologic patterns is a useful screening method prior to initiation of molecular testing. Transcription factor, SRY-related HMG-box 10 (SOX10), is an important marker for melanocytic, schwannian, myoepithelial, and some salivary gland tumors. The aim of this study was to investigate SOX10 expression more thoroughly in the salivary gland neoplasms, including mammary analogue secretory carcinoma and hyalinizing clear cell carcinoma harboring specific genetic rearrangements. A new rabbit monoclonal anti-SOX10 antibody (clone EP268) was used to examine SOX10 expression in 14 different types of salivary gland tumors. We found that acinic cell carcinoma (AciCC), adenoid cystic carcinoma, mammary analogue secretory carcinoma (MASC), epithelial-myoepithelial carcinoma, low-grade salivary duct carcinoma, sialoblastoma, basal cell adenocarcinoma, basal cell adenoma, and pleomorphic adenoma were SOX10-positive. Salivary duct carcinoma, lymphoepithelial carcinoma, hyalinizing clear cell carcinoma, and oncocytoma were SOX10-negative. Earlier, mucoepidermoid carcinoma (MEC) was considered as SOX10-negative tumor. This study identified a subgroup of SOX10-positive MEC cases with characteristic polygonal epithelial cells, pale-to-eosinophilic cytoplasm, and colloid-like dense eosinophilic material. Our data show SOX10 expression can be observed in salivary gland tumors with either one of the 4 cell types: acinic cells, cuboidal ductal cells with low grade cytologic features, basaloid cells, and myoepithelial cells. In this article we thoroughly evaluated SOX10 expression in salivary gland tumors. SOX10 is useful in the differential diagnosis between myoepithelial carcinoma with clear cell features and hyalinizing clear cell carcinoma. It can also be used to discriminate low-grade salivary duct carcinoma from high-grade ones.