Circulating miRNAs and Machine Learning for Lateralizing Primary Aldosteronism

Abstract

Background: Distinguishing between unilateral and bilateral primary aldosteronism, a major cause of secondary hypertension, is crucial due to different treatment approaches. While adrenal venous sampling is the gold standard, its invasiveness, limited availability, and often difficult interpretation pose challenges. This study explores the utility of circulating microRNAs (miRNAs) and machine learning in distinguishing between unilateral and bilateral forms of primary aldosteronism.

Methods: MiRNA profiling was conducted on plasma samples from 18 patients with primary aldosteronism taken during adrenal venous sampling on an Illumina MiSeq platform. Bioinformatics and machine learning identified 9 miRNAs for validation by reverse transcription real-time quantitative polymerase chain reaction. Validation was performed on a cohort consisting of 108 patients with known subdifferentiation. A 30-patient subset of the validation cohort involved both adrenal venous sampling and peripheral, the rest only peripheral samples. A neural network model was used for feature selection and comparison between adrenal venous sampling and peripheral samples, while a deep-learning model was used for classification.

Results: Our model identified 10 miRNA combinations achieving >85% accuracy in distinguishing unilateral primary aldosteronism and bilateral adrenal hyperplasia on a 30-sample subset, while also confirming the suitability of peripheral samples for analysis. The best model, involving 6 miRNAs, achieved an area under curve of 87.1%. Deep learning resulted in 100% accuracy on the subset and 90.9% sensitivity and 81.8% specificity on all 108 samples, with an area under curve of 86.7%.

Conclusions: Machine learning analysis of circulating miRNAs offers a minimally invasive alternative for primary aldosteronism lateralization. Early identification of bilateral adrenal hyperplasia could expedite treatment initiation without the need for further localization, benefiting both patients and health care providers.

Keywords: deep learning; hyperaldosteronism; hypertension; microRNAs.

Figure 1.

Volcano-plot resulting from the analysis by the DESeq2 algorithm. MicroRNAs (miRNAs) overexpressed in unilateral primary aldosteronism (UPA) compared with bilateral adrenal hyperplasia (BAH) are on the left, while miRNAs overexpressed in BAH compared with UPA are on the right side of the diagram, and are coded with blue color if the significance is <0.001 (7 dots on the plot) and with red, if the significance is <0.001 and the logarithmic fold change is larger than 2 (40 dots on the plot).

Figure 2.

Scatter-plots, presenting the results of the reverse transcription real-time quantitative polymerase chain reaction measurements from the 6 microRNAs included in the best model, mean±SD of −ΔCt values. A, hsa-miR-146-5p, hsa-miR-24-3p (B), hsa-miR-130b-3p (C), hsa-miR-99b-5p (D), hsa-miR-151a-3p (E), and hsa-miR-199a-3p (F). BAH indicates bilateral adrenal hyperplasia; and UPA, unilateral primary aldosteronism.

Figure 3.

The receiver operating characteristic (ROC) curves of the machine learning algorithms on peripheral blood samples. A, Neural network model on the 30-sample subset. B, Neural network models performance on the whole validation cohort (n=108). C, Deep-learning model on the 30-sample subset. D, Deep-learning model on the validation cohort (n=108). AUC indicates area under the curve.