Deep Learning Fuzzy Inference: An Interpretable Model for Detecting Indirect Immunofluorescence Patterns Associated with Nasopharyngeal Cancer.

The detection of serum Epstein-Barr virus antibodies by immunofluorescence assay (IFA) is considered the gold standard screening test for nasopharyngeal cancer (NPC) in high-risk populations. Given the high survival rate after early detection in asymptomatic patients, compared to the poor prognosis in patients with late-stage NPC, screening using IFA has tremendous potential for saving lives in the general population. However, IFA requires visual interpretation of cellular staining patterns by trained pathology staff, making it labor intensive and hence nonscalable. In this study, an automated fuzzy inference (FI) system achieved high agreement with a human IFA expert in identifying cellular patterns associated with NPC (κ = 0.82). The integration of a deep learning module into FI further improved the performance of FI (κ = 0.90) and reduced the number of uncertain cases that required manual evaluation. The performance of the resulting hybrid model, termed deep learning FI (DeLFI), was then evaluated with a separate testing set of clinical samples. In this clinical validation, DeLFI outperformed human evaluation on the area under the curve (0.926 versus 0.821) and closely matched human performance on Youden J index (0.81 versus 0.80). Data from this study indicate that the combination of deep learning with FI in DeLFI has the potential to improve the scalability and accuracy of NPC detection.

The microdissected gene expression landscape of nasopharyngeal cancer reveals vulnerabilities in FGF and noncanonical NF-κB signaling.

Nasopharyngeal cancer (NPC) is an Epstein-Barr virus (EBV)-positive epithelial malignancy with an extensive inflammatory infiltrate. Traditional RNA-sequencing techniques uncovered only microenvironment signatures, while the gene expression of the tumor epithelial compartment has remained a mystery. Here, we use Smart-3SEQ to prepare transcriptome-wide gene expression profiles from microdissected NPC tumors, dysplasia, and normal controls. We describe changes in biological pathways across the normal to tumor spectrum and show that fibroblast growth factor (FGF) ligands are overexpressed in NPC tumors, while negative regulators of FGF signaling, including SPRY1, SPRY2, and LGALS3, are down-regulated early in carcinogenesis. Within the NF-κB signaling pathway, the critical noncanonical transcription factors, RELB and NFKB2, are enriched in the majority of NPC tumors. We confirm the responsiveness of EBV-positive NPC cell lines to targeted inhibition of these pathways, reflecting the heterogeneity in NPC patient tumors. Our data comprehensively describe the gene expression landscape of NPC and unravel the mysteries of receptor tyrosine kinase and NF-κB pathways in NPC.

Epstein-Barr virus-induced ectopic CD137 expression helps nasopharyngeal carcinoma to escape immune surveillance and enables targeting by chimeric antigen receptors.

Non-keratinizing nasopharyngeal carcinoma (NPC) is a malignancy with a poor prognosis for relapsing patients and those with metastatic disease. Here, we identify a novel disease mechanism of NPC which may be its Achilles' heel that makes it susceptible to immunotherapy. CD137 is a potent costimulatory receptor on activated T cells, and CD137 agonists strongly enhance anti-tumor immune responses. A negative feedback mechanism prevents overstimulation by transferring CD137 from T cells to CD137 ligand (CD137L)-expressing antigen presenting cells (APC) during cognate interaction, upon which the CD137-CD137L complex is internalized and degraded. We found ectopic expression of CD137 on 42 of 122 (34.4%) NPC cases, and that CD137 is induced by the Epstein-Barr virus latent membrane protein (LMP) 1. CD137 expression enables NPC to hijack the inbuilt negative feedback mechanism to downregulate the costimulatory CD137L on APC, facilitating its escape from immune surveillance. Further, the ectopically expressed CD137 signals into NPC cells via the p38-MAPK pathway, and induces the expression of IL-6, IL-8 and Laminin γ2. As much as ectopic CD137 expression may support the growth and spread of NPC, it may be a target for its immunotherapeutic elimination. Natural killer cells that express a CD137-specific chimeric antigen receptor induce death in CD137+ NPC cells, in vitro, and in vivo in a murine xenograft model. These data identify a novel immune escape mechanism of NPC, and lay the foundation for an urgently needed immunotherapeutic approach for NPC.

Digital CRISPR-based method for the rapid detection and absolute quantification of nucleic acids.

Rapid diagnostics of adventitious agents in biopharmaceutical/cell manufacturing release testing and the fight against viral infection have become critical. Quantitative real-time PCR and CRISPR-based methods rapidly detect DNA/RNA in 1 h but suffer from inter-site variability. Absolute quantification of DNA/RNA by methods such as digital PCR reduce this variability but are currently too slow for wider application. Here, we report a RApid DIgital Crispr Approach (RADICA) for absolute quantification of nucleic acids in 40-60 min. Using SARS-CoV-2 as a proof-of-concept target, RADICA allows for absolute quantification with a linear dynamic range of 0.6-2027 copies/µL (R2 value > 0.99), high accuracy and low variability, no cross-reactivity to similar targets, and high tolerance to human background DNA. RADICA's versatility is validated against other targets such as Epstein-Barr virus (EBV) from human B cells and patients' serum. RADICA can accurately detect and absolutely quantify EBV DNA with similar dynamic range of 0.5-2100 copies/µL (R2 value > 0.98) in 1 h without thermal cycling, providing a 4-fold faster alternative to digital PCR-based detection. RADICA therefore enables rapid and sensitive absolute quantification of nucleic acids which can be widely applied across clinical, research, and biomanufacturing areas.