Artificial intelligence-based algorithms in detection and 3D reconstruction of lung nodules on chest computed tomography scans

Gavrilov P.V.

St. Petersburg Research Institute of Phthisiopulmonology

Roitberg P.G.

CareMentorAI;
JSC «Medicine»

Blinov D.S.

CareMentorAI

Goldin M.G.

CareMentorAI

Blinova E.V.

I.M. Sechenov First Moscow State Medical University

Leontiev V.S.

F.I. Inozemtzev Municipsl Clinical Hospital

Kamishanskaya I.G.

St. Petersburg State University

Dorofeev A.A.

CareMentorAI

Cheremisin V.M.

St. Petersburg State University;
St. Mary Municipal Hospital

Novokhat’ko Yu.A.

CareMentorAI

Sushkov E.V.

Moscow Municipal Clinical Hospital No.40

Shmatok D.O.

N.P. Ogarev National Research Mordovia State University

Sokolov A.I.

Penza State University

Artificial intelligence-based algorithms in detection and 3D reconstruction of lung nodules on chest computed tomography scans

Authors:

Gavrilov P.V., Roitberg P.G., Blinov D.S., Goldin M.G., Blinova E.V., Leontiev V.S., Kamishanskaya I.G., Dorofeev A.A., Cheremisin V.M., Novokhat’ko Yu.A., Sushkov E.V., Shmatok D.O., Sokolov A.I.

More about the authors

Journal: Russian Journal of Operative Surgery and Clinical Anatomy. 2021;5(3): 15‑22

DOI: 10.17116/operhirurg2021503115

Downloaded: 13

Download PDF (RU)

Contact the author

Table of contents

To cite this article:

Gavrilov PV, Roitberg PG, Blinov DS, et al. . Artificial intelligence-based algorithms in detection and 3D reconstruction of lung nodules on chest computed tomography scans. Russian Journal of Operative Surgery and Clinical Anatomy. 2021;5(3):15‑22. (In Russ.)
https://doi.org/10.17116/operhirurg2021503115

Close metadata

OBJECTIVE

To develop and explore the functionality and diagnostic accuracy of a computer vision system based on artificial intelligence algorithms for the detection of solid and subsolid nodules in chest computed tomographic cases.

MATERIAL AND METHODS

The model of a computer vision system was carried out using architectures of deep convolutional neural networks; 1367 impersonal full-slice computed tomographic examinations of the chest were used for training and testing of the system.

RESULTS

The developed model with high accuracy that is corresponding to AUC ROC 0.9127, predicts the probability of malignant findings in the tomographic case at threshold values on the network output neurons in the range from 0.25 to 0.30; correctly segments the pulmonary fields and lung nodules with a size of 5 mm and above; carries out volumetric reconstruction of the lungs with contrast visualization of findings. An efficient and functional integration scheme has been created that allows the model to be introduced into a radiologist’s workflow.

CONCLUSION

A computer vision system based on algorithms of convolutional artificial neural networks makes it possible to detect effectively lung nodules on tomographic scans of the chest, to visualize them in the form of volumetric reconstruction of the lungs. The developed integration model is a convenient tool for embedding a neural network service into the workflow of a radiology department.

Keywords:

computed tomography

lungs

nodules

neural networks

artificial intelligence

algorithm

anatomical 3D reconstruction

Authors:

Gavrilov P.V.

St. Petersburg Research Institute of Phthisiopulmonology

Roitberg P.G.

CareMentorAI;
JSC «Medicine»

Blinov D.S.

CareMentorAI

Goldin M.G.

CareMentorAI

Blinova E.V.

I.M. Sechenov First Moscow State Medical University

Leontiev V.S.

F.I. Inozemtzev Municipsl Clinical Hospital

Kamishanskaya I.G.

St. Petersburg State University

Dorofeev A.A.

CareMentorAI

Cheremisin V.M.

St. Petersburg State University;
St. Mary Municipal Hospital

Novokhat’ko Yu.A.

CareMentorAI

Sushkov E.V.

Moscow Municipal Clinical Hospital No.40

Shmatok D.O.

N.P. Ogarev National Research Mordovia State University

Sokolov A.I.

Penza State University

Received:

17.03.2021

Accepted:

21.05.2021

List of references:

Waite S, Scott J, Gale J, Fuchs T, Kolla S, Reede D. Interpretative error in radiology. AJR. 2017;208:739-749. https://doi.org/10.2214/ajr.16.16963
Del Ciello A, Franchi D, Contegiacomo A, Cicchetti G, Bonomo L, Larici AR. Missed lung cancer: when, where, and why? Diagn Interv Radiol. 2017;23(2):118-126. https://doi.org/10.5152/dir.2016.16187
Er O, Yumusak N, Temurtas F. Chest diseases diagnosis using artificial neural networks. Expert Sys Appl. 2010;37(12):7648-7655. https://doi.org/10.1016/j.eswa.2010.04.078
Litjens G, Kooi T, Bejnordi EB? et al. A survey on deep learning in medical image analysis. Med Image Analys. 2017;42:60-88. https://doi.org/10.1016/j.media.2017.07.005
Abiyev RH, Ma’aitah MKS. Deep convolutional networks for chest diseases detection. J Healthcare Engin. 2018;2018:1-12. https://doi.org/10.1155/2018/4168538
Lassen BC, Jacobs C, Kuhnigk JM, van Ginneken B, van Rikxoort EM. Robust semi-automatic segmentation of pulmonary subsolid nodules in chest computed tomography scans. Physics Med Biol. 2015;60:1307. https://doi.org/10.1088/0031-9155/60/3/1307
Гаврилов П.В., Щеткин К.С., Залялов Р.М. Нейросетевая модель детекции признаков поражения легких, ассоциированных с COVID-19, на аксиальных срезах нативной компьютерной томографии грудной клетки. Медицинский альянс. 2020;8(2):6-13.
Ronneberger O, Fischer P, Brox T. U-Net: Convolutional Networks for Biomedical Image Segmentation. Arxiv Lib. 2015;15:1-8. https://doi.org/10.1007/978-3-319-24574-4_28
He K, Zhang X, Ren S. Deep Residual Learning for Image Recognition. IEEE Comput. Vis Pattern Recognit. 2016;10:1-4. https://doi.org/10.1109/cvpr.2016.90
Glantz SA. Primer of Biostatistics, 7th ed. McGraw-Hill Education, USA.

Close metadata