Human-multimodal deep learning collaboration in 'precise' diagnosis of lupus erythematosus subtypes and similar skin diseases.

BACKGROUND: Lupus erythematosus (LE) is a spectrum of autoimmune diseases. Due to the complexity of cutaneous LE (CLE), clinical skin image-based artificial intelligence is still experiencing difficulties in distinguishing subtypes of LE. OBJECTIVES: We aim to develop a multimodal deep learning system (MMDLS) for human-AI collaboration in diagnosis of LE subtypes. METHODS: This is a multi-centre study based on 25 institutions across China to assist in diagnosis of LE subtypes, other eight similar skin diseases and healthy subjects. In total, 446 cases with 800 clinical skin images, 3786 multicolor-immunohistochemistry (multi-IHC) images and clinical data were collected, and EfficientNet-B3 and ResNet-18 were utilized in this study. RESULTS: In the multi-classification task, the overall performance of MMDLS on 13 skin conditions is much higher than single or dual modals (Sen = 0.8288, Spe = 0.9852, Pre = 0.8518, AUC = 0.9844). Further, the MMDLS-based diagnostic-support help improves the accuracy of dermatologists from 66.88% ± 6.94% to 81.25% ± 4.23% (p = 0.0004). CONCLUSIONS: These results highlight the benefit of human-MMDLS collaborated framework in telemedicine by assisting dermatologists and rheumatologists in the differential diagnosis of LE subtypes and similar skin diseases.

The expression profiles of microRNAs in Kaposi's sarcoma.

Kaposi's sarcoma (KS) is a multicentric angioproliferative tumor of mesenchymal origin. The molecular and biologic aspects of KS are not fully understood. MicroRNAs are non-protein-coding small RNAs in the size range 19-25 nucleotides (nt) that play important roles in biological processes, including cellular differentiation, proliferation, and death. We performed a miRNA microarray analysis by detecting six paired KS and matched adjacent healthy tissues using the 7th generation of miRCURY(TM) LNA Array (v.18.0) (Exiqon) containing 3100 capture probes. We selected 10 significant differentially expressed miRNAs, which were confirmed by quantitative real-time reverse transcription polymerase chain reaction (qRT-PCR) in 18 paired KS and matched adjacent healthy tissue specimens. We also investigated the associations between clinical features and miRNA expression. Among the 3100 human miRNA probes in the microarrays, we identified 170 differentially expressed miRNAs (69 upregulated and 101 downregulated miRNAs) in KS versus adjacent healthy tissues. Among the most significantly upregulated miRNAs were miR-126-3p, miR-199a-3p, miR-16-5p, and the 13 KSHV-related miRNAs. The most significantly downregulated miRNAs included miR-125b-1-3p and miR-1183. Eight upregulated miRNAs, miR-181b-5p, miR-199a-3p, miR-15a-5p, miR-126-3p, miR-1297, kshv-miR-k12-12-3p, kshv-miR-k12-1-5p, and miR-16-5p, and two downregulated miRNAs, miR-125b-1-3p and miR-1183, were confirmed by qRT-PCR in 18 paired KS samples. The qRT-PCR results for 10 miRNAs were consistent with our microarray results. The miR-125b-1-3p and miR-16-5p had statistically significant associations with HHV-8 and HIV infections in KS. The results of miRNA profiling showed that KS appears to have unique expression patterns when compared with paired adjacent healthy tissues, suggesting that deregulation of miRNAs plays an important role in the progression of KS. These differentially expressed miRNAs may provide novel diagnostic and prognostic tools.

MicroRNA-126 regulates the phosphatidylinositol-3 kinase (PI3K)/protein kinase B (AKT) pathway in SLK cells in vitro and the expression of its pathway members in Kaposi's sarcoma tissue.

In vitro, microRNA-126 (miR-126) inhibits SLK cell proliferation, inhibits the cell cycle, induces cell apoptosis, and reduces cell invasiveness. Double luciferase assays have shown that phosphatidylinositol-3 kinase (PI3K) is the miR-126 target in SLK cells. We aimed to investigate the influence of miR-126 on the phosphate and tension homology deleted on chromosome ten (PTEN)/PI3K/protein kinase B (AKT) pathway members in SLK cells and to determine the expression of these pathway members in Kaposi's sarcoma (KS). The mimic and inhibitor of miR-126 were transfected into SLK cells and PTEN and AKT1 expression was assayed in SLK cells by real-time quantitative PCR and western blotting. PTEN, AKT1, phosphorylated (P)-PTEN, and phosphorylated (P)-AKT expression in KS and paraneoplastic skin were assayed by immunohistochemistry. AKT1 expression was downregulated in SLK cells that overexpressed miR-126, while there was no significant difference in PTEN expression between SLK cells overexpressing miR-126 and those in which its expression was knocked down. PTEN and AKT1 were expressed in KS and paraneoplastic skin but P-AKT was not. Interestingly, P-PTEN was not expressed in paraneoplastic skin but it was expressed in 90% of KS biopsies (P < .05). P-PTEN expression was also significantly higher in visceral than in cutaneous KS (P = .01) and was higher in indoor than in outdoor workers (P = .018). In vitro, miR-126 negatively regulated AKT1 expression but no regulation of PTEN expression was evident. Results indicated that in KS, PTEN is activated and may therefore be a potential therapeutic target for KS. In addition, these results also indicate that sunlight may not be the cause of KS.