Baishya Nirupam Konwar, Baishya Kangkana
Department of Radiology, Fakhruddin Ali Ahmed Medical College & Hospital, Assam, India.
Department of Electrical Engineering, Assam Engineering College, Assam, India.
Discov Oncol. 2024 Dec 27;15(1):844. doi: 10.1007/s12672-024-01720-8.
Radiomics is a method that extracts many features from medical images using various algorithms. Medical nomograms are graphical representations of statistical predictive models that produce a likelihood of a clinical event for a specific individual based on biological and clinical data. The radiomic nomogram was first introduced in 2016 to study the integration of specific radiomic characteristics with clinically significant risk factors for patients with colorectal cancer lymph node metastases. Thereby it gained momentum and made its way into different domains of breast, liver, and head and neck cancer. Deep learning-based radiomics which automatically generates and extracts significant features from the input data using various neural network architectures, along with the generation and usage of nomograms are the latest developments in the application of radiomics for the diagnosis of gall bladder carcinoma. Although radiomics has demonstrated encouraging outcomes in the diagnosis of gall bladder carcinoma, but most of the studies conducted suffer from a lack of external validation cohorts, smaller sample sizes, and paucity of prospective utility in routine clinical settings.
放射组学是一种使用各种算法从医学图像中提取许多特征的方法。医学列线图是统计预测模型的图形表示，它根据生物学和临床数据为特定个体产生临床事件的可能性。放射组学列线图于2016年首次引入，用于研究特定放射组学特征与结直肠癌淋巴结转移患者临床显著风险因素的整合。由此它获得了发展动力，并进入了乳腺癌、肝癌和头颈癌的不同领域。基于深度学习的放射组学利用各种神经网络架构自动从输入数据中生成和提取显著特征，以及列线图的生成和使用，是放射组学在胆囊癌诊断应用中的最新进展。尽管放射组学在胆囊癌诊断中已显示出令人鼓舞的结果，但大多数进行的研究存在缺乏外部验证队列、样本量较小以及在常规临床环境中前瞻性实用性不足的问题。
