Background: High-grade soft tissue sarcomas are a heterogeneous, complex group of aggressive malignant tumors showing mesenchymal differentiation. Recently, soft tissue sarcomas have increasingly been classified on the basis of underlying genetic alterations; however, the role of aberrant DNA methylation in these tumors is not well understood and, consequently, the usefulness of methylation-based classification is unclear. Results: We used the Infinium HumanMethylation27 platform to profile DNA methylation in 80 primary, untreated high-grade soft tissue sarcomas, representing eight relevant subtypes, two non-neoplastic fat samples and 14 representative sarcoma cell lines. The primary samples were partitioned into seven stable clusters. A classification algorithm identified 216 CpG sites, mapping to 246 genes, showing different degrees of DNA methylation between these seven groups. The differences between the clusters were best represented by a set of eight CpG sites located in the genes SPEG, NNAT, FBLN2, PYROXD2, ZNF217, COL14A1, DMRT2 and CDKN2A. By integrating DNA methylation and mRNA expression data, we identified 27 genes showing negative and three genes showing positive correlation. Compared with non-neoplastic fat, NNAT showed DNA hypomethylation and inverse gene expression in myxoid liposarcomas, and DNA hypermethylation and inverse gene expression in dedifferentiated and pleomorphic liposarcomas. Recovery of NNAT in a hypermethylated myxoid liposarcoma cell line decreased cell migration and viability. Conclusions: Our analysis represents the first comprehensive integration of DNA methylation and transcriptional data in primary high-grade soft tissue sarcomas. We propose novel biomarkers and genes relevant for pathogenesis, including NNAT as a potential tumor suppressor in myxoid liposarcomas.

Recently, recurrent point mutations in the telomerase reverse transcriptase (TERT) promoter region have been found in many human cancers, leading to a new transcription factor binding site, increased induction of TERT and subsequently to telomere maintenance. We determined the prevalence of TERT promoter mutations in soft tissue sarcomas of 341 patients comprising 16 entities and in 16 sarcoma cell lines covering 7 different soft tissue sarcoma types. The sarcoma tissue samples were collected from the archives of the Institute of Pathology, University of Heidelberg and were composed of 39 myxoid liposarcomas (MLS), 61 dedifferentiated liposarcomas, 15 pleomorphic liposarcomas, 27 leiomyosarcomas, 25 synovial sarcomas (SS), 35 malignant peripheral nerve sheath tumors (MPNST), 40 undifferentiated pleomorphic sarcomas, 17 myxofibrosarcomas, 9 low grade fibromyxoid sarcomas, 10 cases of dermatofibrosarcoma protuberans, 31 solitary fibrous tumors (SFT), 8 extraskeletal myxoid chondrosarcomas, 9 angiosarcomas, 6 alveolar soft part sarcomas, 5 clear cell sarcomas and 4 epithelioid sarcomas. Sarcoma cell lines were obtained from the raising laboratories. A 193 bp fragment of the TERT promoter region covering the hot-spot mutations C228T and C250T was amplified, and direct sequencing of the PCR products was performed. TERT promoter mutations were detected in 36/341 sarcomas. They were highly recurrent in MLS (29/39; 74%) and were in the present MLS series not associated with the phenotype (myxoid vs. round cell variant), tumor grade, tumor site and patients’ median age or gender. In the remaining cases, TERT promoter mutations were found only in 7/302 sarcoma samples and confined to SFTs (4/31; 13%), MPNSTs (2/35; 6%), and SSs (1/25; 4%). Within the collection of sarcoma cell lines examined, TERT promoter mutations were detected in two MLS and in one of three MPNST cell lines. TERT promoter mutations are frequent in MLSs including their round cell variants, representing the most prevalent mutation identified in this sarcoma entity to date, and in a minor fraction of SFTs, MPNSTs and SSs. The majority of sarcomas are devoid of TERT promoter hotspot mutations. These data suggest that telomere maintenance through increased expression of telomerase plays an important role in the pathogenesis especially of MLS.

The diagnosis and the subtyping of non-Hodgkin lymphoma (NHL) are challenging and require expert knowledge, great experience, thorough morphological analysis, and often additional expensive immunohistological and molecular methods. As these requirements are not always available, supplemental methods supporting morphological-based decision making and potentially entity subtyping are required. Deep learning methods have been shown to classify histopathological images with high accuracy, but data on NHL subtyping are limited. After annotation of histopathological whole-slide images and image patch extraction, we trained and optimized an EfficientNet convolutional neuronal network algorithm on 84,139 image patches from 629 patients and evaluated its potential to classify tumor-free reference lymph nodes, nodal small lymphocytic lymphoma/chronic lymphocytic leukemia, and nodal diffuse large B-cell lymphoma. The optimized algorithm achieved an accuracy of 95.56% on an independent test set including 16,960 image patches from 125 patients after the application of quality controls. Automatic classification of NHL is possible with high accuracy using deep learning on histopathological images and routine diagnostic applications should be pursued.

Histone 3.3 (H3.3) hotspot mutations in bone tumors occur in the vast majority of giant cell tumors of bone (GCTBs; 96%), chondroblastomas (95%) and in a few cases of osteosarcomas. However, clinical presentation, histopathological features, and additional molecular characteristics of H3.3 mutant osteosarcomas are largely unknown. In this multicentre, retrospective study, a total of 106 conventional high-grade osteosarcomas, across all age groups were re-examined for hotspot mutations in the H3.3 coding genes H3F3A and H3F3B. H3.3 mutant osteosarcomas were re-evaluated in a multidisciplinary manner and analyzed for genome-wide DNA-methylation patterns and DNA copy number aberrations alongside H3.3 wild-type osteosarcomas and H3F3A G34W/L mutant GCTBs. Six osteosarcomas (6/106) carried H3F3A hotspot mutations. No mutations were found in H3F3B. All patients with H3F3A mutant osteosarcoma were older than 30 years with a median age of 65 years. Copy number aberrations that are commonly encountered in high-grade osteosarcomas also occurred in H3F3A mutant osteosarcomas. Unlike a single osteosarcoma with a H3F3A K27M mutation, the DNA methylation profiles of H3F3A G34W/R mutant osteosarcomas were clearly different from H3.3 wild-type osteosarcomas, but more closely related to GCTBs. The most differentially methylated promoters between H3F3A G34W/R mutant and H3.3 wild-type osteosarcomas were in KLLN/PTEN (p < 0.00005) and HIST1H2BB (p < 0.0005). H3.3 mutations in osteosarcomas may occur in H3F3A at mutational hotspots. They are overall rare, but become more frequent in osteosarcoma patients older than 30 years. Osteosarcomas carrying H3F3A G34W/R mutations are associated with epigenetic dysregulation of KLLN/PTEN and HIST1H2BB.

Atypical fibroxanthomas (AFX) and pleomorphic dermal sarcomas (PDS) are lesions of the skin with overlapping histologic features and unspecific molecular traits. PDS behaves aggressive compared to AFX. Thus, a precise delineation, although challenging in some instances, is relevant. We examined the value of DNA-methylation profiling and copy number analysis for separating these tumors. DNA-methylation data were generated from 17 AFX and 15 PDS using the Illumina EPIC array. These were compared with DNA-methylation data generated from 196 tumors encompassing potential histologic mimics like cutaneous squamous carcinomas (cSCC; n = 19), basal cell carcinomas (n = 10), melanoma metastases originating from the skin (n = 11), leiomyosarcomas (n = 11), angiosarcomas of the skin and soft tissue (n = 11), malignant peripheral nerve sheath tumors (n = 19), dermatofibrosarcomas protuberans (n = 13), extraskeletal myxoid chondrosarcomas (n = 9), myxoid liposarcomas (n = 14), schwannomas (n = 10), neurofibromas (n = 21), alveolar (n = 19) and embryonal (n = 17) rhabdomyosarcomas as well as undifferentiated pleomorphic sarcomas (n = 12). DNA-methylation profiling did not separate AFX from PDS. The DNA-methylation profiles of the other cases, however, were distinct from AFX/PDS. They reliably assigned to subtype-specific DNA-methylation clusters, although overlap occurred between some AFX/PDS and cSCC. Copy number profiling revealed alterations in a similar frequency and distribution between AFX and PDS. They involved losses of 9p (22/32) and 13q (25/32). Gains frequently involved 8q (8/32). Notably, a homozygous deletion of CDKN2A was more frequent in PDS (6/15) than in AFX (2/17), whereas amplifications were non-recurrent and overall rare (5/32). Our findings support the concept that AFX and PDS belong to a common tumor spectrum. We could demonstrate the diagnostic value of DNA-methylation profiling to delineating AFX/PDS from potential mimics. However, the assessment of certain histologic features remains crucial for separating PDS from AFX.