BackgroundmiR-182 is one of the most significantly up-regulated miRNAs in hepatocellular carcinoma (HCC). Metastasis suppressor 1 (MTSS1), one target gene of miR-182, plays an important role in the metastasis of cancers. However, it remains unclear what role does function and mechanism of miR-182 and MTSS1play in HCC.MethodsmiR-182 expression was tested in 86 cases of paired HCC and normal tissues by real-time PCR and the relationships between miR-182 expression and clinicopathological parameters were analyzed. The expression of MTSS1 was evaluated by immunohistochemistry and western blot in the above tissues and its correlation with miR-182 expression was analyzed. Moreover, western blot and invasion assays were performed after transfection of pre-miR-182 or anti-miR-182 to HCC cell lines. In addition, luciferase assays was performed to confirm the regulation of miR-182 on MTSS1.ResultsCompared with normal tissue, miR-182 was up-regulated and MTSS1 was down-regulated in HCC tissues. Moreover, the over-expression of miR-182 was correlated with intrahepatic metastasis (p = 0.034) and poor prognosis (p = 0.039) of HCC patients. There was a negative correlation between miR-182 and MTSS1 expression in both HCC tissues (r = −0.673, p < 0.01) and HCC cell lines (r = −0.931, p = 0.021). Furthermore, the up-regulation of miR-182 resulted in the down-regulation of MTSS1 and increased invasive potential of HUH-1, and reverse results were also confirmed when the expression of miR-182 was inhibited. In addition, the results of the luciferase assay demonstrated the targeted regulation of miR-182 on MTSS1.ConclusionsmiR-182 could promote metastasis of HCC and inhibit the expression of MTSS1. miR-182 and MTSS1 are potential prognostic markers and/or therapeutic targets in HCC.

No abstract

Motivation Tumor mutational burden (TMB) is an indicator of the efficacy and prognosis of immune checkpoint therapy in colorectal cancer (CRC). Cancer patients with high TMB (TMB_H) values tend to benefit from immunotherapy, whereas those with low TMB (TMB_L) values tend to be not. Though whole-exome sequencing (WES) is considered the gold standard for determining TMB, it is difficult to be applied in clinical practice due to its high cost. There are also a few DNA panel-based methods to estimate TMB; however, their detection cost is also high, and the associated wet-lab experiments usually take days, which emphasize the need for faster and cheaper alternatives. Methods In this study, we propose a multi-modal deep learning model based on a residual network (ResNet) and multi-modal compact bilinear pooling to predict TMB status (i.e., TMB_H or TMB_L) directly from histopathological images and clinical data. We applied the model to CRC data from The Cancer Genome Atlas and compared it with four other popular methods, namely, ResNet18, ResNet50, VGG19, and AlexNet. We tested different TMB thresholds, namely, percentiles of 10%, 14.3%, 15%, 16.3%, 20%, 30% and 50%, to differentiate TMB_H and TMB_L. Results For the percentile 14.3% (i.e., TMB value 20) and ResNet18, our model achieved an area under the receiver operating characteristic curve of 0.817 after five-fold cross-validation, which was better than that of other compared models. In addition, we also found that TMB values were significantly associated with tumor stage and N and M stages. Our study shows that deep learning models can predict TMB status from histopathological images and clinical information only, which is worth clinical application.