Colorectal cancer (CRC) is the third cause of cancer-related death and the fourth most frequently diagnosed cancer across the globe. The objective of this study is to obtain novel and effective diagnostic markers to enrich CRC diagnosis methods. Herein, exosomal miRNA expression data of CRC and normal blood were subjected to XGBoost algorithm, and 5 miRNAs related to CRC diagnosis were primarily confirmed. Then multilayer perceptron (MLP) classifiers were constructed based on different subsets. Via integrated feature selection (IFS), we noticed that the MLP classifier constructed by the first four miRNAs (miR-654-5p, miR-126, miR-10b, and miR-144) had the highest Matthews correlation coefficient (MCC). Subsequently, principal component analysis (PCA) for dimensionality reduction was performed on samples based on the miR-654-5p, miR-126, miR-10b, and miR-144 expression data. The signature based on these four feature miRNAs, as the analysis indicated, could effectively distinguish CRC samples from normal samples. Further, we extracted the exosomes from clinical blood samples and applied qRT-PCR analysis, which revealed that the expression of these four feature miRNAs was in the trend of that in the test set. Collectively, these four feature miRNAs might be tumor biomarkers in the serum, and our study offers innovative thinking on early-stage CRC diagnosis.

In this study, we are going to investigate the effect of nano carbon combined with ex vitro anatomical sorting on the detection rate of lymph nodes (LNs) in gastric cancer (GC) along with the analysis of the correlation between LNs detection rate and patients’ prognosis. The clinical data of patients undergoing radical gastrectomy in Shanxi Bethune Hospital, Shanxi Academy of Medical Sciences, Tongji Shanxi Hospital, Third Hospital of Shanxi Medical University from January 2018 to January 2019 were examined retrospectively. According to whether they adopt nano carbon tracing and specimen sorting method, patients were divided into nano carbon and control groups. The respective rate of detection and correlation of total and positive LNs, respectively, clinical treatment, tumor marker level, and long-term prognosis were matched between these groups. At the same time, the effects of the nano carbon tracer on the detection of total and positive LNs were evaluated. In nano carbon group, more LN specimens could be detected, and the number of positive LNs increased significantly. In addition, in patients with different infiltration stages and LN substations, more LNs could be detected in the nano carbon group for examination, and the detection rate of LNs with diameter less than 5 mm was also more. Furthermore, LNs (preferably positive in number) were correlated positively with the attained LNs number. Otherwise, the use of nano carbon suspension could better label LNs in each substation, especially N1 station, and improve micro-LN detection rate. At the same time, the positive metastasis rate in black-stained LNs was higher (31.67% vs. 13.51%). In relation to the clinical prognosis, CEA’s level, i.e., CA199 and CA125, in the nano carbon group is controlled more effectively. Their condition was not easy to progress and relapse, and their mortality was further reduced. As a result, nano carbon, coupled with ex vitro anatomical sorting, may considerably enhance the detection rate of total and positive LNs, thereby improving the accuracy of clinical staging in GC patients, which has a good influence on their long-term prognosis.