Artificial Neural Networks for differential diagnosis of breast lesions in MR-Mammography: A systematic approach addressing the influence of network architecture on diagnostic performance using a large clinical database

详细信息	查看全文 | 推荐本文 |

• 作者：Matthias Dietzel^a ; ¹ ; ^{dietzelmatthias2@hotmail.com} ; Pascal A.T. Baltzer^a ; ¹ ; Andreas Dietzel^b ; Ramy Zoubi^a ; Tobias Grö ; schel^a ; Hartmut P. Burmeister^a ; Martin Bogdan^b ; Werner A. Kaiser^a ; ^c
• 关键词：Neoplasms-Primary ; Artificial Neural Networks ; Breast ; Magnetic resonance imaging ; Pathology ; Diagnostic imaging ; Computer aided diagnosis ; Lesion characterization
• 刊名：European Journal of Radiology
• 出版年：2012
• 期刊代码：93_0720048x
• 类别：med
• 出版时间：July, 2012
• 卷：81
• 期：7
• 页码：1508-1513
• 文件大小：319 K

摘要

Rationale and objectives

Differential diagnosis of lesions in MR-Mammography (MRM) remains a complex task. The aim of this MRM study was to design and to test robustness of Artificial Neural Network architectures to predict malignancy using a large clinical database.

Materials and methods

For this IRB-approved investigation standardized protocols and study design were applied (T1w-FLASH; 0.1 mmol/kgBW Gd-DTPA; T2w-TSE; histological verification after MRM). All lesions were evaluated by two experienced (>500 MRM) radiologists in consensus. In every lesion, 18 previously published descriptors were assessed and documented in the database.

An Artificial Neural Network (ANN) was developed to process this database (The-MathWorks/Inc., feed-forward-architecture/resilient back-propagation-algorithm). All 18 descriptors were set as input variables, whereas histological results (malignant vs. benign) was defined as classification variable. Initially, the ANN was optimized in terms of 鈥淭raining Epochs鈥?(TE), 鈥淗idden Layers鈥?(HL), 鈥淟earning Rate鈥?(LR) and 鈥淣eurons鈥?(N). Robustness of the ANN was addressed by repeated evaluation cycles (n: 9) with receiver operating characteristics (ROC) analysis of the results applying 4-fold Cross Validation. The best network architecture was identified comparing the corresponding Area under the ROC curve (AUC).

Results

Histopathology revealed 436 benign and 648 malignant lesions. Enhancing the level of complexity could not increase diagnostic accuracy of the network (P: n.s.). The optimized ANN architecture (TE: 20, HL: 1, N: 5, LR: 1.2) was accurate (mean-AUC 0.888; P: <0.001) and robust (CI: 0.885-0.892; range: 0.880-0.898).

Conclusion

The optimized neural network showed robust performance and high diagnostic accuracy for prediction of malignancy on unknown data.