RESUMO
Liquid biopsies offer minimally invasive diagnosis and monitoring of cancer disease. This biosource is often analyzed using sequencing, which generates highly complex data that can be used using machine learning tools. Nevertheless, validating the clinical applications of such methods is challenging. It requires: (a) using data from many patients; (b) verifying potential bias concerning sample collection; and (c) adding interpretability to the model. In this work, we have used RNA sequencing data of tumor-educated platelets (TEPs) and performed a binary classification (cancer vs. no-cancer). First, we compiled a large-scale dataset with more than a thousand donors. Further, we used different convolutional neural networks (CNNs) and boosting methods to evaluate the classifier performance. We have obtained an impressive result of 0.96 area under the curve. We then identified different clusters of splice variants using expert knowledge from the Kyoto Encyclopedia of Genes and Genomes (KEGG). Employing boosting algorithms, we identified the features with the highest predictive power. Finally, we tested the robustness of the models using test data from novel hospitals. Notably, we did not observe any decrease in model performance. Our work proves the great potential of using TEP data for cancer patient classification and opens the avenue for profound cancer diagnostics.
RESUMO
INTRODUCTION: Evaluation of the differential value of the first thyroglobulin (Tg) concentration, measured after thyroidectomy (Tx) but before thyroid remnant ablation, in patients with differentiated thyroid carcinoma (DTC) as a marker of either metastases or residual cancer (M). MATERIAL AND METHODS: Data from 517 patients with DTC after Tx, with follow-up > 1.5 year were analysed retrospectively. Patients in whom either the course of the disease was unclear or interference in the Tg test was possible (a-TgAb [+], Tg recovery < 80%) were excluded from the study. Finally, the data from 247 patients were evaluated (age: 14-79 years; 223 women, 24 men). The results of TSH, thyroid radioiodine uptake (T(up24)), thyroid remnant volume (V) and Tg in patients with diagnosed M (group M1; n = 35) were compared with the same parameters in patients with remission > 1.5 year (group M0; n = 212). The area under the ROC curve was calculated. The clinical decision limit of Tg level to be suggestive of metastases was determined by means of efficiency curve. RESULTS: Groups M0 and M1 did not differ from each other with respect to TSH concentration (median 49.7 mIU/l vs 44.3; p = 0.16) or thyroid remnant volume (1.4 vs 1.1 ml; p = 0.79). However, they did differ with respect to T(up24) (7.6 vs 3.2%; p = 0.01) and Tg (4.5 vs 96.7 ng/ml; p = 0.000000). Area under ROC for Tg was 0.78 +/- 0.05 (mean +/- s.e.m.). The decision limit of Tg for suspected M was determined at 38.1 ng/ml, Tg sensitivity was 0.57 (95%CI 0.39-0.74) and specificity 0.96 (95%CI 0.92-0.98). CONCLUSIONS: First thyroglobulin concentration, determined after thyroidectomy but before other treatment, is higher in patients with metastatic DTC than in patients without such metastases. This indicates that Tg level may be used as an early marker of either residual or metastatic DTC (even if thyroid remnants are present).