Colin Jacobs1, Eva M van Rikxoort, Ernst Th Scholten, Pim A de Jong, Mathias Prokop, Cornelia Schaefer-Prokop, Bram van Ginneken. 1. From the *Department of Radiology and Nuclear Medicine, Radboud University Medical Center, Nijmegen, the Netherlands; †Fraunhofer MEVIS, Bremen, Germany; ‡Department of Radiology, Kennemer Gasthuis, Haarlem; §Department of Radiology, University Medical Center Utrecht, Utrecht; and ∥Department of Radiology, Meander Medical Center, Amersfoort, the Netherlands.
Abstract
OBJECTIVES: The purpose of this study was to develop and validate a computer-aided diagnosis (CAD) tool for automatic classification of pulmonary nodules seen on low-dose computed tomography into solid, part-solid, and non-solid. MATERIALS AND METHODS:Study lesions were randomly selected from 2 sites participating in the Dutch-Belgian NELSON lung cancer screening trial. On the basis of the annotations made by the screening radiologists, 50 part-solid and 50 non-solid pulmonary nodules with a diameter between 5 and 30 mm were randomly selected from the 2 sites. For each unique nodule, 1 low-dose chest computed tomographic scan was randomly selected, in which the nodule was visible. In addition, 50 solid nodules in the same size range were randomly selected. A completely automatic 3-dimensional segmentation-based classification system was developed, which analyzes the pulmonary nodule, extracting intensity-, texture-, and segmentation-based features to perform a statistical classification. In addition to the nodule classification by the screening radiologists, an independent rating of all nodules by 3 experienced thoracic radiologists was performed. Performance of CAD was evaluated by comparing the agreement between CAD and human experts and among human experts using the Cohen κ statistics. RESULTS: Pairwise agreement for the differentiation between solid, part-solid, and non-solid nodules between CAD and each of the human experts had a κ range between 0.54 and 0.72. The interobserver agreement among the human experts was in the same range (κ range, 0.56-0.81). CONCLUSIONS: A novel automated classification tool for pulmonary nodules achieved good agreement with the human experts, yielding κ values in the same range as the interobserver agreement. Computer-aided diagnosis may aid radiologists in selecting the appropriate workup for pulmonary nodules.
RCT Entities:
OBJECTIVES: The purpose of this study was to develop and validate a computer-aided diagnosis (CAD) tool for automatic classification of pulmonary nodules seen on low-dose computed tomography into solid, part-solid, and non-solid. MATERIALS AND METHODS: Study lesions were randomly selected from 2 sites participating in the Dutch-Belgian NELSON lung cancer screening trial. On the basis of the annotations made by the screening radiologists, 50 part-solid and 50 non-solid pulmonary nodules with a diameter between 5 and 30 mm were randomly selected from the 2 sites. For each unique nodule, 1 low-dose chest computed tomographic scan was randomly selected, in which the nodule was visible. In addition, 50 solid nodules in the same size range were randomly selected. A completely automatic 3-dimensional segmentation-based classification system was developed, which analyzes the pulmonary nodule, extracting intensity-, texture-, and segmentation-based features to perform a statistical classification. In addition to the nodule classification by the screening radiologists, an independent rating of all nodules by 3 experienced thoracic radiologists was performed. Performance of CAD was evaluated by comparing the agreement between CAD and human experts and among human experts using the Cohen κ statistics. RESULTS: Pairwise agreement for the differentiation between solid, part-solid, and non-solid nodules between CAD and each of the human experts had a κ range between 0.54 and 0.72. The interobserver agreement among the human experts was in the same range (κ range, 0.56-0.81). CONCLUSIONS: A novel automated classification tool for pulmonary nodules achieved good agreement with the human experts, yielding κ values in the same range as the interobserver agreement. Computer-aided diagnosis may aid radiologists in selecting the appropriate workup for pulmonary nodules.
Authors: Samuel G Armato; Karen Drukker; Feng Li; Lubomir Hadjiiski; Georgia D Tourassi; Roger M Engelmann; Maryellen L Giger; George Redmond; Keyvan Farahani; Justin S Kirby; Laurence P Clarke Journal: J Med Imaging (Bellingham) Date: 2016-12-19
Authors: Finbar Foley; Srinivasan Rajagopalan; Sushravya M Raghunath; Jennifer M Boland; Ronald A Karwoski; Fabien Maldonado; Brian J Bartholmai; Tobias Peikert Journal: Semin Thorac Cardiovasc Surg Date: 2016-01-08
Authors: Douglas E Wood; Ella A Kazerooni; Scott L Baum; George A Eapen; David S Ettinger; Lifang Hou; David M Jackman; Donald Klippenstein; Rohit Kumar; Rudy P Lackner; Lorriana E Leard; Inga T Lennes; Ann N C Leung; Samir S Makani; Pierre P Massion; Peter Mazzone; Robert E Merritt; Bryan F Meyers; David E Midthun; Sudhakar Pipavath; Christie Pratt; Chakravarthy Reddy; Mary E Reid; Arnold J Rotter; Peter B Sachs; Matthew B Schabath; Mark L Schiebler; Betty C Tong; William D Travis; Benjamin Wei; Stephen C Yang; Kristina M Gregory; Miranda Hughes Journal: J Natl Compr Canc Netw Date: 2018-04 Impact factor: 11.908
Authors: Francesco Ciompi; Kaman Chung; Sarah J van Riel; Arnaud Arindra Adiyoso Setio; Paul K Gerke; Colin Jacobs; Ernst Th Scholten; Cornelia Schaefer-Prokop; Mathilde M W Wille; Alfonso Marchianò; Ugo Pastorino; Mathias Prokop; Bram van Ginneken Journal: Sci Rep Date: 2017-04-19 Impact factor: 4.379
Authors: Jean-Paul Charbonnier; Kaman Chung; Ernst T Scholten; Eva M van Rikxoort; Colin Jacobs; Nicola Sverzellati; Mario Silva; Ugo Pastorino; Bram van Ginneken; Francesco Ciompi Journal: Sci Rep Date: 2018-01-12 Impact factor: 4.379
Authors: Kyung Hee Lee; Kyung Won Lee; Ji Hoon Park; Kyunghwa Han; Jihang Kim; Sang Min Lee; Chang Min Park Journal: Korean J Radiol Date: 2018-04-06 Impact factor: 3.500