PhD Scientific Days 2022

Clinical Medicine V. (Poster discussion will take place in the Aula during the Coffee Break)

Establishment of Novel Tissue microRNA Combinations for the Differential Diagnosis of Adrenocortical Tumors Based on Machine Learning Methods

Text of the abstract

Adrenal tumors are common, occuring in 5-7% of the population. Adrenocortical carcinoma (ACC) is rare (0.7-2/million/year) and it has a five-year survival of less than 30% in advanced stages. The histological differentiation of benign and malignant adrenocortical tumors is challenging.

We explored the diagnostic utility of multiple microRNAs in various combinations as markers of adrenocortical malignancy by using artificial intelligence methods, based on machine learning and neural networks.

63 formalin-fixed, paraffin-embedded adrenocortical tissues were studied. The discovery cohort included 10 adrenocortical adenoma (ACA) and 10 ACC samples. An independent validation cohort encompassed another 21 ACC and 22 ACA samples. Based on literature data, 16 microRNAs shown to be differentially expressed were included. MicroRNA expression was studied by TaqMan RT-qPCR. RNU48 and cel-miR-39 was used as internal and external controls, respectively. The relevance of microRNAs for the classification of ACA and ACC samples was determined by the random forest classification method. The possibility of automatic classification of samples into ACA or ACC groups was tested by machine learning methods. Only models with more than 90% classification capability were selected for RT-qPCR validation and subsequent artificial intelligence-based classification. The best performing microRNA combinations were selected by 90-10% random learner-tester cross validation. 24 microRNA models were included in the validation performed in a blind manner.

Hsa-miR-195, hsa-miR-375, hsa-miR-483_3p, hsa-miR-483_5p and hsa-miR-503 were the best 5 microRNAs to correctly classify the previously unkown samples. The following three, best performing statistical models were selected out of the former microRNAs: hsa-miR-210 + hsa-miR-483-5p + hsa-miR-503, hsa-miR-210 + hsa-miR-375 + hsa-miR-503 and hsa-miR-195 + hsa-miR-210 + hsa-miR-503 with sensitivity and specificity of more than 90%. The diagnostic performance of these three models was undoubtedly superior over that of individual microRNAs.

Three microRNA combinations with exceptional diagnostic performance were established using artificial intelligence-based methods. These biomarker combinations can assist histological studies, and their use in small amount preoperative biopsy samples might also serve diagnosis.

ÚNKP-21-3; (NKFIH) K134215