Identification of Circulating Serum miRNAs as Novel Biomarkers in Pancreatic Cancer Using a Penalized Algorithm

Abstract

Pancreatic cancer (PC) is difficult to detect in the early stages; thus, identifying specific and sensitive biomarkers for PC diagnosis is crucial, especially in the case of early-stage tumors. Circulating microRNAs are promising non-invasive biomarkers. Therefore, we aimed to identify non-invasive miRNA biomarkers and build a model for PC diagnosis. For the training model, blood serum samples from 63 PC patients and 63 control subjects were used. We selected 39 miRNA markers using a smoothly clipped absolute deviation-based penalized support vector machine and built a PC diagnosis model. From the double cross-validation, the average test AUC was 0.98. We validated the diagnosis model using independent samples from 25 PC patients and 81 patients with intrahepatic cholangiocarcinoma (ICC) and compared the results with those obtained from the diagnosis using carbohydrate antigen 19-9. For the markers miR-155-5p, miR-4284, miR-346, miR-7145-5p, miR-5100, miR-661, miR-22-3p, miR-4486, let-7b-5p, and miR-4703-5p, we conducted quantitative reverse transcription PCR using samples from 17 independent PC patients, 8 ICC patients, and 8 healthy individuals. Differential expression was observed in samples from PC patients. The diagnosis model based on the identified markers showed high sensitivity and specificity for PC detection and is potentially useful for early PC diagnosis.

Dimensionality reduction by sparse orthogonal projection with applications to miRNA expression analysis and cancer prediction

“…al . in [12], i.e., the data discussed in point (3) of section A.5. In figure 4, we present an ROC plot comparison using SOP, OPLS, FR, EFC, SPCA, and DROP-D.…”

“…This study introduced the dimension reduction methodology, SOP. In section 2, we tested the effectiveness of SOP, and the methods from the literature, on synthetic data, and real miRNA expression data publicly available in the literature [19, 16, 14, 12], with the goal to distinguish control patients from those with cancer. We introduced a novel, synthetic “spin top” data set in section 2.1, designed for testing dimensionality reduction methods which use orthogonal projections.…”

“…Typically, in cancer prediction using miRNA expression, binary classification problems are considered [20, 19, 16, 14, 12, 2], and methods such as OPLS and EFC are applied to separate the two classes. While the majority of the classifications conducted in this paper were binary (i.e., the number of classes n c = 2), in section 2.4 we also tested the effectiveness of SOP on a multi-classification problem, where n c = 3, and the goal was to separate control patients from those with breast cancer and advanced breast cancer simultaneously.…”

“…As the number of case samples for Wilms tumors is low, we do not consider the Wilms tumor cases in our analyses. So, in total, we consider 13 different diseases from the Keller data set (i.e., the 14 provided minus the Wilms tumor samples). miRNA expression data of [12] collected by Lee et. al .…”

“…This data is comprised of 88 patients with Pancreatic Cancer (PC), and 19 healthy controls. In [12], the authors combine the 19 healthy patients with 10 cholelithiasis patients to form a larger control set of 29 patients for use in the PC classifications. We use the same control set here, and aim to separate the controls from the PC patients.…”