Understanding the spatial heterogeneity of tumours and its links to disease initiation and progression is a cornerstone of cancer biology. Presently, histopathology workflows heavily rely on hematoxylin and eosin and serial immunohistochemistry staining, a cumbersome, tissue-exhaustive process that results in non-aligned tissue images. We propose the VirtualMultiplexer, a generative artificial intelligence toolkit that effectively synthesizes multiplexed immunohistochemistry images for several antibody markers (namely AR, NKX3.1, CD44, CD146, p53 and ERG) from only an input hematoxylin and eosin image. The VirtualMultiplexer captures biologically relevant staining patterns across tissue scales without requiring consecutive tissue sections, image registration or extensive expert annotations. Thorough qualitative and quantitative assessment indicates that the VirtualMultiplexer achieves rapid, robust and precise generation of virtually multiplexed imaging datasets of high staining quality that are indistinguishable from the real ones. The VirtualMultiplexer is successfully transferred across tissue scales and patient cohorts with no need for model fine-tuning. Crucially, the virtually multiplexed images enabled training a graph transformer that simultaneously learns from the joint spatial distribution of several proteins to predict clinically relevant endpoints. We observe that this multiplexed learning scheme was able to greatly improve clinical prediction, as corroborated across several downstream tasks, independent patient cohorts and cancer types. Our results showcase the clinical relevance of artificial intelligence-assisted multiplexed tumour imaging, accelerating histopathology workflows and cancer biology.
Diagnosis of salivary gland neoplasms is often challenging due to their high morphological diversity and overlaps. Several recurrent molecular alterations have been described recently, which can serve as powerful diagnostic tools and potential therapeutic targets (e.g. NTRK or RET fusions). However, current sequential molecular testing can be expensive and time consuming. In order to facilitate the diagnosis of salivary gland neoplasms, we designed an all-in-one RNA-based next generation sequencing panel suitable for the detection of mutations, fusions and gene expression levels (including NR4A3) of 27 genes involved in salivary gland neoplasms. Here we present the validation of the "SalvGlandDx" panel on FFPE histological specimen including fine needle aspiration (FNA) cell block material, against the standard methods currently used at our institution. In a second part we describe selected unique cases in which the SalvGlandDx panel allowed proper diagnosis and new insights into special molecular characteristics of selected salivary gland tumors. We characterize a unique salivary gland adenocarcinoma harboring a ZCCHC7-NTRK2 fusion, a highly uncommon spindle cell and pseudoangiomatoid adenoid-cystic carcinoma with MYBL1-NFIB fusion, and a purely oncocytic mucoepidermoid carcinoma, whereas diagnosis could be made by detection of a CRTC3-MAML2 rearrangement on the cell block specimen of the FNA. Further, a rare case of a SS18-ZBTB7A rearranged low-grade adenocarcinoma previously described as potential spectrum of microsecretory adenocarcinoma, is reported. In addition, features of six cases within the spectrum of polymorphous adenocarcinoma / cribriform adenocarcinoma of salivary gland including PRKD1 p.E710D mutations and novel fusions involving PRKAR2A-PRKD1, SNX9-PRKD1 and ATL2-PRKD3, are described.
Pancreatic cancer is characterized by extensive resistance to conventional therapies, making clinical management a challenge. Here we map the epigenetic dependencies of cancer stem cells, cells that preferentially evade therapy and drive progression, and identify SWI/SNF complex member SMARCD3 as a regulator of pancreatic cancer cells. Although SWI/SNF subunits often act as tumor suppressors, we show that SMARCD3 is amplified in cancer, enriched in pancreatic cancer stem cells and upregulated in the human disease. Diverse genetic mouse models of pancreatic cancer and stage-specific Smarcd3 deletion reveal that Smarcd3 loss preferentially impacts established tumors, improving survival especially in context of chemotherapy. Mechanistically, SMARCD3 acts with FOXA1 to control lipid and fatty acid metabolism, programs associated with therapy resistance and poor prognosis in cancer. These data identify SMARCD3 as an epigenetic modulator responsible for establishing the metabolic landscape in aggressive pancreatic cancer cells and a potential target for new therapies.
Acute graft-versus-host disease (aGVHD) is a severe complication of allogeneic hematopoietic stem cell transplantation, driven by alloreactive donor T cells in the gut. However, the roles of additional donor and host cells in this process are not fully understood. We conducted multiplexed imaging to generate comprehensive spatial maps of healthy and GVHD-afflicted duodena at single-cell resolution. We profiled 59 biopsies from patients with varying gastrointestinal (GI) aGVHD manifestations and 18 healthy controls and measured the spatial expression of 40 proteins, designated to delineate distinct epithelial cell populations of the gut, stromal cells, and immune cells. Normal duodenum was stereotypical across individuals, showing organization and zonation of epithelial, stromal, and immune cells. Patients with aGVHD exhibited increased fibrosis, alterations in crypt morphology, loss of Paneth cells, and accumulation of endocrine cells in the crypts. Homeostatic immune organization was disrupted, with a prominent reduction in immunoglobulin A–secreting plasma cells. CD8 T cells were enriched only in a subset of patients, whereas others exhibited noncanonical enrichments of macrophages and neutrophils. Immune composition was correlated with the time after transplantation, with shared dynamics observed across individuals. Host cells dominated the plasma and T cell compartments in the gut for extended periods of time after transplantation, suggesting that additional players may drive aGVHD across individuals in addition to donor-derived T cells. This spatial atlas of healthy and GI aGVHD uncovers epithelial and immune dynamics across individuals, offering insights into disease pathophysiology and potential clinical applications.
Background: In recent years, the list of tumor entities in the sinonasal tract has significantly expanded, requiring advanced diagnostic testing. We report the case of a 32-year-old patient with an unusual NUT carcinoma originating in the maxillary sinus, which showed extensive well-differentiated, papillary squamous morphology, similar to the spectrum of the recently described DEK::AFF2 fusion-associated carcinoma. Methods: We performed immunohistochemical and molecular studies including EBV- and HPV-testing, as well as DNA/RNA next generation sequencing. Results: The tumor showed predominantly exophytic papillary growth with mature squamous differentiation. An additional component harbored atypical, less differentiated basaloid tumor cells with infiltration of the adjacent stroma. Conspicuous inflammation was evident. There was no evidence of HPV DNA or EBV RNA. Next-generation sequencing revealed a NUT::NSD3 gene fusion corresponding to (“speckled-type”) immunopositivity of NUT in the tumor cells. Conclusions: We describe a NUT::NSD3 gene fusion-associated NUT carcinoma of the sinonasal tract with a deceptively well-differentiated papillary growth pattern, thus expanding the morphological spectrum of this typically poorly differentiated neoplasm.