| 张斌,刘弋,胡宗举.人工神经网络在胃癌预后预测中的应用[J].安徽医药,2021,25(5):987-991. |
| 人工神经网络在胃癌预后预测中的应用 |
| Application of artificial neural networks in prognosis of gastric cancer |
| |
| DOI：10.3969/j.issn.1009-6469.2021.05.035 |
| 中文关键词: 胃肿瘤 人工智能 人工神经网络 预后 |
| 英文关键词: Stomach neoplasms Artificial intelligence Artificial neural network Prognosis |
| 基金项目: |
|
| 摘要点击次数: 1437 |
| 全文下载次数: 427 |
| 中文摘要: |
| 目的应用人工神经网络( ANN)技术建立胃癌预后预测模型，并与传统 logistic回归模型相比较，探讨其临床意义。方法回顾性分析安徽省阜阳市人民医院 2011年 1月至 2016年 12月收治的 152例经内镜活检或术后病理证实的胃癌病人的临床及病理资料，单因素生存分析筛选出有统计学意义的指标，用人工神经网络技术建立胃癌的预后模型对胃癌术后生存时间进行预测，同时与传统 logistic回归预测做对比。结果癌胚抗原(CEA)水平、 Borrmann分型、浸润深度、淋巴结转移、分化程度、 TNM分期、手术方式、神经脉管浸润与胃癌病人的预后有关， ANN对胃癌病人的 5年生存率预测准确率 85.19%，敏感度 88.24%，特异度 80%；二元 logistic回归对胃癌病人的 5年生存率预测准确率 74.07%，敏感度 82.35%，特异度 60%；ANN模型在不同 TNM分期中的预测准确率 77.78%，敏感度 40%，特异度 92.31%。结论人工神经网络作为预测胃癌术后生存时间的一种新的方法，具有良好的应用前景。 |
| 英文摘要: |
| Objective Artificial neural network (ANN) technology was used to establish a prognostic prediction model for gastriccancer, and compared with the traditional logistic regression model to explore its clinical significance.Methods The clinical data of 152 cases which were confirmed to be gastric cancer after endoscopic biopsy and postoperative pathologic from January,2011 to December,2016 in Fuyang people′hospital of Anhui Province were analyzed retrospectively, Screening out statistically significant indicatorsby single factor survival analysis, To build a prediction model of gastric cancer prognosis by artificial neural network technology, compared with the traditional logistic regression model.Results CEA level, Borrmann typing,depth of invasion,lymph node metastasis,degree of differentiation,TNM stage,operation method and neurovascular invasion are related to the prognosis of gastric cancer patients,The accuracy,sensitivity and specificity of 5-year survival rate of gastric cancer patients predicted by artificial neural network was85.19%,88.24%,80%;the accuracy,sensitivity and specificity of binary logistic regression was 74.07%,82.35%,60%;the accuracy,sensitivity and specificity of artificial neural network model in different TNM stages was 77.78%,40%,92.31%.Conclusion As a new method to predict the survival time of gastric cancer, ANN has a good application prospect. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
