基于机器学习的 CT 鉴别胃 schwannomas 与胃肠道间质瘤。

Differentiation of gastric schwannomas from gastrointestinal stromal tumors by CT using machine learning.

机构信息

Department of Radiology, Tongde Hospital of Zhejiang Province, Number 234, Road Gucui, Hangzhou, Zhejiang, China.

Department of Radiology, The First Affliated Hospital of Bengbu Medical College, Bengbu, Anhui, China.

出版信息

Abdom Radiol (NY). 2021 May;46(5):1773-1782. doi: 10.1007/s00261-020-02797-9. Epub 2020 Oct 20.

DOI:10.1007/s00261-020-02797-9

PMID:33083871

Abstract

OBJECTIVE

To identify schwannomas from gastrointestinal stromal tumors (GISTs) by CT features using Logistic Regression (LR), Decision Trees (DT), Random Forest (RF), and Gradient Boosting Decision Tree (GBDT).

METHODS

This study enrolled 49 patients with schwannomas and 139 with GISTs proven by pathology. CT features with P < 0.1 derived from univariate analysis were inputted to four models. Five machine learning (ML) versions, multivariate analysis, and radiologists' subjective diagnostic performance were compared to evaluate diagnosis performance of all the traditional and advanced methods.

RESULTS

The CT features with P < 0.1 were as follows: (1) CT attenuation value of unenhancement phase (CTU), (2) portal venous enhancement (CTV), (3) degree of enhancement in the portal venous phase (DEPP), (4) CT attenuation value of portal venous phase minus arterial phase (CTV-CTA), (5) enhanced potentiality (EP), (6) location, (7) contour, (8) growth pattern, (9) necrosis, (10) surface ulceration, (11) enlarged lymph node (LN). LR (M1), RF, DT, and GBDT models contained all of the above 11 variables, while LR (M2) was developed using six most predictive variables derived from (M1). LR (M2) model with AUC of 0.967 in test dataset was thought to be optimal model in differentiating the two tumors. Location in gastric body, exophytic and mixed growth pattern, lack of necrosis and surface ulceration, enlarged lymph nodes, and larger EP were the most important CT features suggestive of schwannomas.

CONCLUSION

LR (M2) provided the optimal diagnostic potency among other ML versions, multivariate analysis, and radiologists' performance on differentiation of schwannomas from GISTs.

摘要

目的

利用 Logistic 回归（LR）、决策树（DT）、随机森林（RF）和梯度提升决策树（GBDT），通过 CT 特征从胃肠道间质瘤（GISTs）中识别神经鞘瘤。

方法

本研究纳入了 49 例经病理证实为神经鞘瘤和 139 例 GISTs 的患者。单变量分析得出的 P<0.1 的 CT 特征被输入到四个模型中。比较了五种机器学习（ML）版本、多变量分析和放射科医生的主观诊断性能，以评估所有传统和先进方法的诊断性能。

结果

单变量分析中 P<0.1 的 CT 特征包括：（1）平扫期 CT 衰减值（CTU），（2）门静脉强化（CTV），（3）门静脉期强化程度（DEPP），（4）门静脉期与动脉期 CT 值差值（CTV-CTA），（5）强化潜能（EP），（6）位置，（7）轮廓，（8）生长方式，（9）坏死，（10）表面溃疡，（11）增大的淋巴结（LN）。LR（M1）、RF、DT 和 GBDT 模型均包含上述 11 个变量，而 LR（M2）则是使用从（M1）中得出的 6 个最具预测性的变量开发的。在测试数据集的 AUC 为 0.967 的 LR（M2）模型被认为是区分两种肿瘤的最佳模型。胃体的位置、外生性和混合性生长方式、无坏死和表面溃疡、增大的淋巴结和较大的 EP 是提示神经鞘瘤的最重要 CT 特征。