Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2022 Mar 24;8(4):88.
doi: 10.3390/jimaging8040088.

YOLOv4-Based CNN Model versus Nested Contours Algorithm in the Suspicious Lesion Detection on the Mammography Image: A Direct Comparison in the Real Clinical Settings

Alexey Kolchev  1   2   3   4 , Dmitry Pasynkov  1   5 , Ivan Egoshin  1 , Ivan Kliouchkin  6 , Olga Pasynkova  1 , Dmitrii Tumakov  4
Affiliations

YOLOv4-Based CNN Model versus Nested Contours Algorithm in the Suspicious Lesion Detection on the Mammography Image: A Direct Comparison in the Real Clinical Settings

Alexey Kolchev et al. J Imaging. .

Abstract

Background: We directly compared the mammography image processing results obtained with the help of the YOLOv4 convolutional neural network (CNN) model versus those obtained with the help of the NCA-based nested contours algorithm model.

Method: We used 1080 images to train the YOLOv4, plus 100 images with proven breast cancer (BC) and 100 images with proven absence of BC to test both models.

Results: the rates of true-positive, false-positive and false-negative outcomes were 60, 10 and 40, respectively, for YOLOv4, and 93, 63 and 7, respectively, for NCA. The sensitivities for the YOLOv4 and the NCA were comparable to each other for star-like lesions, masses with unclear borders, round- or oval-shaped masses with clear borders and partly visualized masses. On the contrary, the NCA was superior to the YOLOv4 in the case of asymmetric density and of changes invisible on the dense parenchyma background. Radiologists changed their earlier decisions in six cases per 100 for NCA. YOLOv4 outputs did not influence the radiologists' decisions.

Conclusions: in our set, NCA clinically significantly surpasses YOLOv4.

Keywords: YOLOv4; breast cancer; convolutional neural network; mammography; nested contours algorithm.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Figures

Figure 1
Figure 1
The YOLOv4 architecture with DarkNet framework.
Figure 2
Figure 2
The result of the source image pre-processing. (A): source image; (B): after pre-processing.
Figure 3
Figure 3
The results of the YOLOv4 training. The red line—mean Average Precision (mAP). The blue line–error graph (Loss).
Figure 4
Figure 4
Star-like lesion (arrow). (A): Source image; (B): YOLOv4 outcome; (C): NCA outcome.
Figure 5
Figure 5
Mass with unclear border (arrows). (A): Source image; (B): YOLOv4 outcome; (C): NCA outcome. lesion (arrow).
Figure 6
Figure 6
Round- or oval-shaped mass with clear border (arrow). (A): Source image; (B): YOLOv4 outcome; (C): NCA outcome.
Figure 7
Figure 7
Partly visualized mass (arrow). (A): Source image; (B): YOLOv4 outcome; (C): NCA outcome.
Figure 8
Figure 8
Asymmetric density (arrows). (A): Source image; (B): NCA outcome. The YOLOv4 did not mark the lesion.
Figure 9
Figure 9
Changes poorly visible or invisible on the dense parenchyma background (arrow). (A): Source image; (B): NCA outcome. The YOLOv4 did not mark the lesion.
Figure 10
Figure 10
Confusion matrixes for (A): YOLOv4-based method and (B): NCA-based method. TP—the model detected a lesion where it actually exists; FP—the model detected a lesion where it actually does not exist; FN—the model did not detect the lesion, where it actually exists.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Bray F., Ferlay J., Soerjomataram I., Siegel R.L., Torre L.A., Jemal A. Global Cancer Statistics 2018: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. Cancer J. Clin. 2018;68:394–424. doi: 10.3322/caac.21492. - DOI - PubMed
    1. Seely J.M., Alhassan T. Screening for breast cancer in 2018—What should we be doing today? Curr. Oncol. 2018;25:S115–S124. doi: 10.3747/co.25.3770. - DOI - PMC - PubMed
    1. Lauby-Secretan B., Scoccianti C., Loomis D., Benbrahim-Tallaa L., Bouvard V., Bianchini F., Straif K. Breast-cancer screening —Viewpoint of the IARC Working Group. N. Engl. J. Med. 2015;372:2353–2358. doi: 10.1056/NEJMsr1504363. - DOI - PubMed
    1. Thigpen D., Kappler A., Brem R. The Role of Ultrasound in Screening Dense Breasts—A Review of the Literature and Practical Solutions for Implementation. Diagnostics. 2018;8:20. doi: 10.3390/diagnostics8010020. - DOI - PubMed
    1. Weigel S., Heindel W., Heidrich J., Hense H.W., Heidinger O. Digital mammography screening: Sensitivity of the programme dependent on breast density. Eur. Radiol. 2017;27:2744–2751. doi: 10.1007/s00330-016-4636-4. - DOI - PubMed

LinkOut - more resources