Novel miRNA set for screening gastric cancer with high sensitivity: an ROC analysis of 6,997 samples.

AIMS: Cancer screening is essential for reducing GC-associated deaths. Nonetheless, implementing population-based screening strategies poses challenges. In this study, a comprehensive receiver operating characteristic (ROC) analysis was conducted on 2565 microRNAs (miRNAs) from 6997 samples to identify a classifier for GC screening.

METHOD: First, batch effects were corrected across the five Gene Expression Omnibus Series (GSE) datasets. Second, a predictive model was established using the incidence method. Finally, this model was validated using randomly selected datasets, all datasets after batch effect removal, and each of the datasets separately.

RESULTS

Six miRNAs, namely miR-1290, miR-5100, miR-1343-3p, miR-8073, miR-4706, and miR-4787-3p were identified, with an area under the ROC curve (AUC) of 0.990 ± 0.002, 0.993 ± 0.002, 0.996 ± 0.004, 0.978 ± 0.010, 0.957 ± 0.015, and 0.969 ± 0.015, respectively. The miRNA set obtained by combining the six miRNAs yielded an AUC of 0.997 ± 0.001, which was higher than that of the six individual miRNAs (p < 0.001). Validation across the total dataset and five GSE datasets (GSE106817, GSE112264, GSE113486, GSE113740, and GSE164174) yielded AUCs of 0.997, 0.999, 1.000, 1.000, 0.996, and 0.994, respectively.

CONCLUSION

The novel miRNA set comprising miR-1290, miR-5100, miR-1343-3p, miR-8073, miR-4706, and miR-4787-3p holds promise as a diagnostic classifier for GC screening.