The causal role of somatic mutation and its interrelationship with gene expression profile during tumor development has already been observed, which plays a major role to decide the cancer grades and overall survival. Accurate and robust prediction of tumor grades and patients’ overall survival are important for prognosis, risk factors identification and betterment of the treatment strategy, especially for highly lethal tumors, like gliomas. Here, with the help of more accurate and widely used machine learning‐based approaches, we propose an integrative computational pipeline that incorporates somatic mutations and gene expression profile for survival and grade prediction of glioma patients and simultaneously relates it to the drugs to be administered. This study gives us a clear understanding that the same drug is not effective for the treatment of same grade of cancer if the gene mutations are different. The alteration in a specific gene plays a very important role in tumor progression and should also be considered for the selection of appropriate drugs. This proposed framework includes all the necessary factors required for enhancement of therapeutic designs and could be useful for clinicians in determining an accurate and personalized treatment strategy for individual patients suffering from different life threatening diseases.

中文翻译：

---

预测存活率和药物对具有体细胞基因突变的癌症患者的作用：基于AI的方法

已经观察到了体细胞突变的因果作用及其在肿瘤发展过程中与基因表达谱的相互关系，这在决定癌症的等级和总体生存中起着重要的作用。准确，可靠地预测肿瘤的分级和患者的总体存活率对于预后，危险因素的识别和治疗策略的改善非常重要，特别是对于高致死性肿瘤（例如神经胶质瘤）而言。在这里，借助更准确和广泛使用的基于机器学习的方法，我们提出了一个集成的计算管道，该管道将体细胞突变和基因表达谱相结合，用于神经胶质瘤患者的生存和等级预测，同时将其与所用药物联系起来。这项研究使我们清楚地认识到，如果基因突变不同，则相同的药物对相同等级的癌症无效。特定基因的改变在肿瘤进展中起着非常重要的作用，在选择合适的药物时也应考虑到这一点。该提议的框架包括增强治疗设计所需的所有必要因素，并且对于临床医生为患有不同生命威胁疾病的个体患者确定准确和个性化的治疗策略可能有用。