Automatic detection of lung nodules in computed tomography images: training and validation of algorithms using public research databases

被引:17
作者
Camarlinghi, Niccol [1 ,2 ]
机构
[1] Univ Pisa, Dipartimento Fis, Pisa, Italy
[2] Ist Nazl Fis Nucl, Pisa, Italy
关键词
SELECTIVE ENHANCEMENT FILTERS; AIDED DETECTION; CANCER STATISTICS; PULMONARY NODULES; CT IMAGES; SEGMENTATION; METHODOLOGY; SCANS;
D O I
10.1140/epjp/i2013-13110-5
中图分类号
O4 [物理学];
学科分类号
0702 ;
摘要
Lung cancer is one of the main public health issues in developed countries. Lung cancer typically manifests itself as non-calcified pulmonary nodules that can be detected reading lung Computed Tomography (CT) images. To assist radiologists in reading images, researchers started, a decade ago, the development of Computer Aided Detection (CAD) methods capable of detecting lung nodules. In this work, a CAD composed of two CAD subprocedures is presented: , devoted to the identification of parenchymal nodules, and , devoted to the identification of the nodules attached to the pleura surface. Both CADs are an upgrade of two methods previously presented as Voxel Based Neural Approach CAD . The novelty of this paper consists in the massive training using the public research Lung International Database Consortium (LIDC) database and on the implementation of new features for classification with respect to the original VBNA method. Finally, the proposed CAD is blindly validated on the ANODE09 dataset. The result of the validation is a score of 0.393, which corresponds to the average sensitivity of the CAD computed at seven predefined false positive rates: 1/8, 1/4, 1/2, 1, 2, 4, and 8 FP/CT.
引用
收藏
页数:21
相关论文
共 36 条
[1]   Reduced Lung-Cancer Mortality with Low-Dose Computed Tomographic Screening [J].
Aberle, Denise R. ;
Adams, Amanda M. ;
Berg, Christine D. ;
Black, William C. ;
Clapp, Jonathan D. ;
Fagerstrom, Richard M. ;
Gareen, Ilana F. ;
Gatsonis, Constantine ;
Marcus, Pamela M. ;
Sicks, JoRean D. .
NEW ENGLAND JOURNAL OF MEDICINE, 2011, 365 (05) :395-409
[2]  
[Anonymous], MED IMAGING 2009 COM
[3]  
[Anonymous], LUNG CT SEGMENTATION
[4]  
[Anonymous], INT J COMPUT ASSIST
[5]  
[Anonymous], 2001, Neural Networks: A Comprehensive Foundation
[6]  
[Anonymous], P SPIE MED IM C
[7]  
[Anonymous], ANN ONCOL
[8]   Combination of computer-aided detection algorithms for automatic lung nodule identification [J].
Camarlinghi, Niccolo ;
Gori, Ilaria ;
Retico, Alessandra ;
Bellotti, Roberto ;
Bosco, Paolo ;
Cerello, Piergiorgio ;
Gargano, Gianfranco ;
Lopez Torres, Ernesto ;
Megna, Rosario ;
Peccarisi, Marco ;
Fantacci, Maria Evelina .
INTERNATIONAL JOURNAL OF COMPUTER ASSISTED RADIOLOGY AND SURGERY, 2012, 7 (03) :455-464
[9]   FREE-RESPONSE METHODOLOGY - ALTERNATE ANALYSIS AND A NEW OBSERVER-PERFORMANCE EXPERIMENT [J].
CHAKRABORTY, DP ;
WINTER, LHL .
RADIOLOGY, 1990, 174 (03) :873-881
[10]   Detection of pulmonary nodules at spiral CT: comparison of maximum intensity projection sliding slabs and single-image reporting [J].
Diederich, S ;
Lentschig, MG ;
Overbeck, TR ;
Wormanns, D ;
Heindel, W .
EUROPEAN RADIOLOGY, 2001, 11 (08) :1345-1350