In cancer research, genomic survival analysis is commonly used to identify potential targets or subtyping models. However, there is limited survival information available for some non-tumor diseases, such as hypertension. In this manuscript, we propose a novel approach to address this issue by leveraging survival data from tumors to identify malignant drivers in genomic data for non-tumor diseases. Our approach begins by identifying common genes that may be implicated in both tumor and non-tumor diseases. We calculate similarities between feature genes using co-occurrence analysis in PubMed and also gene expression levels accopanied with disease stage information in transcriptome data. To assess the similarities and rank the genes, we have developed a customized metric. From the higher-ranking genes, we select potential malignant drivers based on calculated hazard ratios greater than 1 and log-rank p values less than 0.05. The validation was applied on 5 different diseases and afterwards the predicted malignant drivers were validated by accordance with reported information, gene ontology and protein-protein interaction analysis. Predicted log-rank p value and hazard ratio on known markers for target disease were also compared. The high accordance for our prediction with public information for all 5 diseases confirmed that our methods were effective in predicting malignant drivers and consequently gaining clinical and research significance.