Comparative effectiveness of convolutional neural network (CNN) and recurrent neural network (RNN) architectures for radiology text report classification

被引:152
作者
Banerjee, Imon [1 ]
Ling, Yuan [2 ]
Chen, Matthew C. [3 ]
Hasan, Sadid A. [2 ]
Langlotz, Curtis P. [3 ]
Moradzadeh, Nathaniel [3 ]
Chapman, Brian [4 ]
Amrhein, Timothy [5 ]
Mong, David [6 ]
Rubin, Daniel L. [1 ,3 ]
Farri, Oladimeji [2 ]
Lungren, Matthew P. [3 ]
机构
[1] Stanford Univ, Sch Med, Dept Biomed Data Sci, Stanford, CA 94305 USA
[2] Philips Res North Amer, Artificial Intelligence Lab, Cambridge, MA USA
[3] Stanford Univ, Sch Med, Dept Radiol, Stanford, CA 94305 USA
[4] Univ Utah, Med Ctr, Dept Bioinformat, Salt Lake City, UT 84112 USA
[5] Duke Univ, Sch Med, Dept Neuroradiol, Durham, NC 27706 USA
[6] Children Hosp Colorado, Dept Radiol, Aurora, CO USA
关键词
Convolutional neural network (CNN); Recurrent neural network (RNN); Pulmonary embolism; Text report classification; Radiology report analysis; CT PULMONARY ANGIOGRAPHY; SPINE MRI; INFORMATION; FREQUENCY; MARKER;
D O I
10.1016/j.artmed.2018.11.004
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
This paper explores cutting-edge deep learning methods for information extraction from medical imaging free text reports at a multi-institutional scale and compares them to the state-of-the-art domain-specific rule-based system - PEFinder and traditional machine learning methods - SVM and Adaboost. We proposed two distinct deep learning models - (i) CNN Word - Glove, and (ii) Domain phrase attention-based hierarchical recurrent neural network (DPA-HNN), for synthesizing information on pulmonary emboli (PE) from over 7370 clinical thoracic computed tomography (CT) free-text radiology reports collected from four major healthcare centers. Our proposed DPA-HNN model encodes domain-dependent phrases into an attention mechanism and represents a radiology report through a hierarchical RNN structure composed of word-level, sentence-level and document level representations. Experimental results suggest that the performance of the deep learning models that are trained on a single institutional dataset, are better than rule-based PEFinder on our multi-institutional test sets. The best F1 score for the presence of PE in an adult patient population was 0.99 (DPA-HNN) and for a pediatrics population was 0.99 (HNN) which shows that the deep learning models being trained on adult data, demonstrated generalizability to pediatrics population with comparable accuracy. Our work suggests feasibility of broader usage of neural network models in automated classification of multi-institutional imaging text reports for a variety of applications including evaluation of imaging utilization, imaging yield, clinical decision support tools, and as part of automated classification of large corpus for medical imaging deep learning work.
引用
收藏
页码:79 / 88
页数:10
相关论文
共 54 条
[1]   Physician Self-Referral and Imaging Use Appropriateness: Negative Cervical Spine MRI Frequency as an Assessment Metric [J].
Amrhein, T. J. ;
Paxton, B. E. ;
Lungren, M. P. ;
Befera, N. T. ;
Collins, H. R. ;
Yurko, C. ;
Eastwood, J. D. ;
Kilani, R. K. .
AMERICAN JOURNAL OF NEURORADIOLOGY, 2014, 35 (12) :2248-2253
[2]  
Anavi Y, 2015, IEEE ENG MED BIO, P2940, DOI 10.1109/EMBC.2015.7319008
[3]  
[Anonymous], COMPUT MED IMAGING G
[4]  
Bakthula R, 2014, INT J COMPUT APPL, V104
[5]  
Banerjee I, 2017, AMIA 2017 ANN S WASH
[6]   LEARNING LONG-TERM DEPENDENCIES WITH GRADIENT DESCENT IS DIFFICULT [J].
BENGIO, Y ;
SIMARD, P ;
FRASCONI, P .
IEEE TRANSACTIONS ON NEURAL NETWORKS, 1994, 5 (02) :157-166
[7]   Natural Language Processing Technologies in Radiology Research and Clinical Applications [J].
Cai, Tianrun ;
Giannopoulos, Andreas A. ;
Yu, Sheng ;
Kelil, Tatiana ;
Ripley, Beth ;
Kumamaru, Kanako K. ;
Rybicki, Frank J. ;
Mitsouras, Dimitrios .
RADIOGRAPHICS, 2016, 36 (01) :176-191
[8]   Document-level classification of CT pulmonary angiography reports based on an extension of the ConText algorithm [J].
Chapman, Brian E. ;
Lee, Sean ;
Kang, Hyunseok Peter ;
Chapman, Wendy W. .
JOURNAL OF BIOMEDICAL INFORMATICS, 2011, 44 (05) :728-737
[9]  
Chen Chi-hau., 2014, Computer vision in medical imaging, V2
[10]  
Cho J., 2015, ARXIV151106348