Modern Pathology最新文献

Tumors resembling tenosynovial giant cell tumor (TGCT) but additionally forming chondroid matrix are rare and most often involve the temporomandibular joint (TMJ). We studied 21 tumors consisting of synoviocytes (large, eosinophilic mononuclear cells containing hemosiderin) and chondroid matrix to better understand these unusual neoplasms. The tumors occurred in 10 males and 11 females, in the age group of 31 to 80 years (median, 50 years) and involved the TMJ region (16), extremities (4), and spine (1). As in conventional TGCT, all were composed of synoviocytes, small histiocytes, foamy macrophages, siderophages, and osteoclast-like giant cells in variably hyalinized background. Expansile nodules of large, moderately atypical synoviocytes were present, in addition to “chondroblastoma-like,” “chondroma-like,” or “phosphaturic mesenchymal tumor-like” calcified matrix. The synoviocytes expressed clusterin (17/19) and less often desmin (3/15). The tumors were frequently CSF1 positive by chromogenic in situ hybridization (8/13) but at best weakly positive for CSF1 by immunohistochemistry (0/3). Background small histiocytes were CD163 positive (12/12). All were FGF23 negative (0/10). Cells within lacunae showed a synoviocytic phenotype (clusterin positive; S100 protein and ERG negative). RNA-Seq was successful in 13 cases; fusions were present in 7 tumors, including FN1::TEK (5 cases); FN1::PRG4 (2 cases); and MALAT1::FN1, PDGFRA::USP35, and TIMP3::ZCCHC7 (1 case each). Three tumors contained more than 1 fusion (FN1::PRG4 with TIMP3::ZCCHC7, FN1::TEK with FN1::PRG4, and FN1::TEK with MALAT1::FN1). Clinical follow-up (17 patients; median follow-up duration 38 months; range 4-173 months) showed 13 (76%) to be alive without evidence of disease and 4 (24%) to be alive with persistent/recurrent local disease. No metastases or deaths from disease were observed. We conclude that these unusual tumors represent a distinct category of synoviocytic neoplasia, which we term “chondroid synoviocytic neoplasm,” rather than simply ordinary TGCT with cartilage. Despite potentially worrisome morphologic features, they appear to behave in at most a locally aggressive fashion.

Acral fibrochondromyxoid tumor (AFCMT) is a recently described likely benign mesenchymal neoplasm arising in the distal extremities with distinctive histologic features and a recurrent THBS1::ADGRF5 fusion. We studied an additional 37 cases of AFCMT and expanded on the so-far reported clinicopathologic and molecular findings. Tumors occurred in 21 females and 16 males, ranging in age from 17 to 78 years (median age: 47), and solely involved the hands (24/37, 65%) or feet (13/37, 35%). Histologic examination revealed well-delineated uni- or multinodular tumors with prominent vasculature-rich septa and bland, chondrocyte-like tumor cells set within abundant chondromyxoid stroma. Immunohistochemical studies showed that tumor cells were positive for CD34 (25/27; 93%) and ERG (27/27; 100%), whereas negative for S100 protein (0/31). Molecular analysis revealed evidence of a THBS1::ADGRF5 fusion in 17 of 19 (89%) successfully tested tumors. Clinical follow-up was available in 8 cases (median: 97 months), with multiple local recurrences in 1 case at 276, 312, and 360 months. We conclude that AFCMT is a distinct entity with reproducible morphologic, immunohistochemical, and molecular genetic features that should be differentiated from other similar appearing acral mesenchymal neoplasms.

Squamoid eccrine ductal carcinoma is a rare infiltrative tumor with morphologic features intermediate between squamous cell carcinoma (SCC) and sweat gland carcinomas such as microcystic adnexal carcinoma. Although currently classified as a sweat gland carcinoma, it has been debated whether squamoid eccrine ductal carcinoma is better classified as a variant of SCC. Furthermore, therapeutic options for patients with advanced disease are lacking. Here, we describe clinicopathologic features of a cohort of 15 squamoid eccrine ductal carcinomas from 14 unique patients, with next-generation sequencing DNA profiling for 12 cases. UV signature mutations were the dominant signature in the majority of cases. TP53 mutations were the most highly recurrent specific gene alteration, followed by mutations in NOTCH genes. Recurrent mutations in driver oncogenes were not identified. By unsupervised comparison of global transcriptome profiles in squamoid eccrine ductal carcinoma (n = 7) to SCC (n = 10), porocarcinoma (n = 4), and microcystic adnexal carcinoma (n = 4), squamoid eccrine ductal carcinomas displayed an intermediate phenotype between SCC and sweat gland tumors. Squamoid eccrine ductal carcinoma displayed significantly higher expression of 364 genes (including certain eccrine markers) and significantly lower expression of 525 genes compared with other groups. Our findings support the classification of squamoid eccrine ductal carcinoma as a carcinoma with intermediate features between SCC and sweat gland carcinoma.

Spread through air spaces (STAS), an important prognostic indicator included in the 2015 World Health Organization classification, is defined as micropapillary, solid, and/or single tumor cell clusters beyond the edge of the main mass and distinct from processing artifacts. This study aimed to assess the interresponder agreement on current STAS criteria vs artifacts, identify discrepancies, and compare responses between pulmonary and general pathologists. A multiple-choice online questionnaire illustrating multiple criteria for STAS vs artifacts was available internationally for 6 days to Pulmonary Pathology Society members, thoracic pathology course attendees, and International Association for the Study of Lung Cancer pathology committee members. Additional 4 questions gathered demographic and practice setting information. One hundred thirty-six unique responses were analyzed. The majority were from North America and Europe (42.6% and 30.2%), practicing pulmonary pathology (70.6%) in academia (64.7%), and with >20 years of experience (31.6%). Excluding trainees, the greatest overall agreement was in defining solid and micropapillary tumor clusters of STAS located ≥3 alveolar spaces from the main tumor edge (91.5%) and recognizing strips of ciliated cells as artifacts (97.7%). Lesser agreement on STAS was evident when tumor cell clusters were immediately adjacent to the tumor edge, a single tumor cell cluster was present at the tissue edge, tumor cell clusters were jagged edged, or tumor cell clusters were admixed with ciliated cell strips (artifacts). There was no significant difference in agreements on STAS for multiple criteria between pulmonary and general pathologists. Significant interresponder agreement on STAS vs artifacts was achieved only for a few criteria. To improve the reproducibility of STAS vs artifacts, areas of lesser agreement require further clarification.

Despite recent advances, the adoption of computer vision methods into clinical and commercial applications has been hampered by the limited availability of accurate ground truth tissue annotations required to train robust supervised models. Generating such ground truth can be accelerated by annotating tissue molecularly using immunofluorescence (IF) staining and mapping these annotations to a post-IF hematoxylin and eosin (H&E) (terminal H&E) stain. Mapping the annotations between IF and terminal H&E increases both the scale and accuracy by which ground truth could be generated. However, discrepancies between terminal H&E and conventional H&E caused by IF tissue processing have limited this implementation. We sought to overcome this challenge and achieve compatibility between these parallel modalities using synthetic image generation, in which a cycle-consistent generative adversarial network was applied to transfer the appearance of conventional H&E such that it emulates terminal H&E. These synthetic emulations allowed us to train a deep learning model for the segmentation of epithelium in terminal H&E that could be validated against the IF staining of epithelial-based cytokeratins. The combination of this segmentation model with the cycle-consistent generative adversarial network stain transfer model enabled performative epithelium segmentation in conventional H&E images. The approach demonstrates that the training of accurate segmentation models for the breadth of conventional H&E data can be executed free of human expert annotations by leveraging molecular annotation strategies such as IF, so long as the tissue impacts of the molecular annotation protocol are captured by generative models that can be deployed prior to the segmentation process.

Many pancreatic neuroendocrine tumors (PanNETs) fall into 2 major prognostic subtypes based on DAXX/ATRX-induced alternative lengthening of telomerase phenotype and alpha- and beta-cell-like epigenomic profiles. However, these PanNETs are still flanked by other PanNETs that do not fit into either subtype. Furthermore, despite advanced genotyping, PanNETs are generally not well-characterized in terms of their histologic and hormonal phenotypes. We aimed to identify new subgroups of PanNETs by extending the currently used transcription factor signatures and investigating their correlation with histologic, hormonal, molecular, and prognostic findings. One hundred eighty-five PanNETs (nonfunctioning 165 and functioning 20), resected between 1996 and 2023, were classified into 5 subgroups (A1, A2, B, C, and D) by cluster analysis based on ARX, PDX1, islet-1 (ISL1), and CDX2 expressions and correlated with trabecular vs solid histology, expression of insulin, glucagon, polypeptide (PP), somatostatin, serotonin, gastrin, calcitonin, adrenocorticotropic hormone (ACTH), DAXX/ATRX, MEN1, and alternative lengthening of telomerase status by fluorescence in situ hybridization, and disease-free survival. A1 (46%, ARX+/ISL1+/PDX1−/CDX2−) and A2 (15%, ARX+/ISL1+/PDX1+/CDX2−) showed trabecular histology and glucagon/PP expression, with A2 also showing gastrin expression. B (18%, PDX1+/ISL1+/ARX−/CDX2−) showed solid histology, insulin, and somatostatin expression (P < .001). It included all insulinomas and had the best outcome (P < .01). C (15%, ARX−/PDX1−/ISL1−/CDX2−) showed solid histology and frequent expression of serotonin, calcitonin, and ACTH. D (5%, PDX1+/CDX2+/ISL1−/ARX−) showed solid histology, expressed ACTH/serotonin, and was an independent poor prognosticator (P < .01). Differential expressions of ARX, PDX1, ISL1, and CDX2 stratified PanNETs into 5 subgroups with different histologies, hormone expressions, and outcomes. Subgroups A1 and A2 resembled the alpha-cell-like type, and subgroup B, the beta-cell-like type. Subgroup C with almost no transcription factor signature was unclear in cell lineage, whereas the PDX+/CDX2+ signature of subgroup D suggested a pancreatic/intestinal cell lineage. Assigning PanNETs to the subgroups may help establish the diagnosis, predict the outcome, and guide the treatment.

Alveolar rhabdomyosarcoma (ARMS) with FOXO1 gene rearrangements is an aggressive pediatric rhabdomyosarcoma subtype that is prognostically distinct from embryonal rhabdomyosarcoma and fusion-negative ARMS. Here, we report 2 cases of ARMS with PAX3::MAML3 fusions. The tumors arose in an infant and an adolescent as stage IV metastatic disease (by Children’s Oncology Group staging system). Histologically, both cases were small round blue cell tumors arranged in vague nests and solid sheets that were diffusely positive for desmin and myogenin. By methylation profiling and unsupervised clustering analysis, the tumors clustered with ARMS with classic FOXO1 rearrangements and ARMS with variant PAX3::NCOA1/INO80D fusions, but not with biphenotypic sinonasal sarcoma (BSNS) with PAX3::MAML3/NCOA2/FOXO1/YAP1 fusions nor with other small round blue cell tumors, including embryonal rhabdomyosarcoma. The differentially methylated genes between ARMS and BSNS were highly enriched in genes involved in myogenesis, and 21% of these genes overlap with target genes of the PAX3::FOXO1 fusion transcription factor. On follow-up after initiation of vincristine/actinomycin/cyclophosphamide chemotherapy, the tumors showed partial and complete clinical responses, consistent with typical upfront chemotherapy responsiveness of ARMS with the classic FOXO1 rearrangement. We conclude that PAX3::MAML3 is a novel variant fusion of ARMS, which displays a methylation signature distinct from BSNS despite sharing similar PAX3 fusions. These findings highlight the utility of methylation profiling in classifying ARMS with noncanonical fusions.