Background: In the era of precision oncology, tumor genomic information is incorporated in clinical decision making to personalize treatment strategies; however, drawing clinically relevant conclusions from tumor molecular profiles are not straightforward if more druggable drivers or targets are identified, more drugs are linked to the same genetic alteration or evidence are conflicting. Here, we present the analysis of 704 lung cancer profiles using our proprietary algorithm and the personalized treatment protocols established by the rule-based artificial intelligence software, the RealTime Oncology Treatment Calculator. Methods: Lung cancer samples (469 adenocarcinomas, 115 squamous cell carcinomas, 57 small cell carcinomas, 63 other histology) were profiled using next generation sequencing of 50-58-600 cancer genes, fluorescent in situ hybridization, and immunohistochemistry. Variants were classified by the Molecular Treatment Calculator algorithm, which dynamically aggregates and ranks all relevant scientific and clinical evidence to find the most efficient drug for the best target of the strongest driver in the given tumor. Results: 1008 different genetic variants were identified. 417 (41%) variants were classified as drivers and 49 (5%) as non-drivers. No direct evidence was available for 541 (54%) of the alterations, 460 (85%) of which were classified as drivers by the algorithm based on the aggregated evidence related to the cancer gene itself or other mutations in the same gene. 354 (50.3%) patients had more than one driver mutations, where the algorithmic ranking of drivers and related compounds was particularly warranted. The algorithm linked molecular targets to 799 (90%) of the driver variants. Conclusions: Molecular treatment calculator algorithm is an efficient tool to classify tumor genetic variants and link them to molecular targets and drugs based on the curated, continuously expanding evidence database and the more than 20 000 rules of the RealTime Oncology Treatment Calculator. Calculator results represent optimal input for Molecular Tumor Boards to design personalized treatment strategies.

The aim: to study the impact of Type1 Diabetes mellitus (DM) on systolic function of left ventricle (LV) of young patients without cardiovascular disease (CVD) and identify factors associated with dysfunction of global longitudinal systolic deformation. Young patients with Type1 DM (N=71) and without CVD were included in the study. Mean age was 28,7 years, 57% men, glycated hemoglobin 9,9%, body mass index 23,4 kg/m2, and diabetes duration 6,84 [0,5; 24], NT-proBNP 62,62 pg/ml, LV EF 61,7%. Treadmill test was conducted to all patients in order to exclude coronary disease.