Faizan Fuzail¹, Muhammad Mehran Mouzam²*, Aamna Bibi², Saud Ur Rehman³ and Tayba Batool⁴

¹Allama Iqbal Medical College Lahore, Department of Medicine and Allied, Pakistan
²University of Agriculture Faisalabad, Department of Medicine and Surgery, Pakistan
³Dera Gazi Khan Medical College, Department of Medicine and Surgery, Pakistan
⁴Government College Women University Fsd, department of Computer Science, Pakistan

*Corresponding Author: Muhammad Mehran Mouzam, University of Agriculture Faisalabad, Department of Medicine and Surgery, Pakistan.

Received: August 09, 2022; Published: November 28, 2022

Abstract

Background: Lately, however, deep learning reconstruction (DLR) technologies have emerged as a viable technical option for reducing radiation dose because they can effectively reconstruct images developed at a low radiation dose to create clear and interpretable images for diagnostic and clinical use.

Objectives: This study explores the possibility of using DLR algorithms in healthcare imaging by comparing image quality between vendor-specific DLR algorithms like TrueFidelityTM and vendor-non-specific DLRs to improve the diagnostic quality of ultra-low-dose CT scans in lung imaging.

Methods: The study involved studying past CT scans of 50 patients who had undergone CT scans between January and February of 2021. Two reconstructed images for each patient were submitted to experienced radiographers from whom reconstruction details had been hidden for qualitative assessment. The assessors rated the images on a scale of 1 to 4 for the noise, resolution, and distortion properties and gave their preferred choice between the two images for each case. Quantitatively, one experienced radiographer assessed the signal-to-noise ratio (SNR) and Edge-Rise-Distance (ERD) for each image.

Results: Non-tuberculous lung diseases, precisely, were the main, accounting for 62% (n = 31). The other conditions included atelectasis (12%, n = 6), pneumonia including COVID-19 (18%, N = 9), and active tuberculosis (8%, n = 4). For subjective noise, TrueFidelity^TM scored higher than ClariCT.AI. On the qualitative noise assessment scale, the former scored 3.72, whereas the latter scored an average of 3.22. On resolution, whereas TrueFidelity^TM had a score of 3.66, ClariCT.AI recorded an average score of 3.49. In terms of Image distortion, TrueFidelity^TM had a score of 3.46, while ClariCT.AI recorded an average score of 3.51. the average preference rate for TrueFidelity^TM was 72%, while for ClariCT.AI was 28%. Quantitatively, whereas the SNR for TrueFidelity^TM was 22.65 ± 2.84, that for ClariCT.AI was 25.95 ± 5.82. While the ERD for TrueFidelity^TM was 0.97 ± 0.19, that for ClariCT.AI was 1.48 ± 0.19.

This study confirmed that vendor-specific DLR algorithms were generally more effective at delivering quality images, confirming the need to develop more specific DLR for the different CT scanners available in the market.

 Keywords: Iterative Reconstruction (IR); Signal-To-Noise Ratio (SNR); Edge-Rise-Distance (ERD)

References

1. Singh R., et al. “Image quality and lesion detection on deep learning reconstruction and iterative reconstruction of Submillisievert chest and abdominal CT”. American Journal of Roentgenology3 (2020): 566-573.
2. Nagayama Y., et al. “Radiation dose reduction at pediatric CT: Use of low tube voltage and iterative reconstruction”. RadioGraphics 5 (2018): 1421-1440.
3. Park S B. “Advances in deep learning for computed tomography denoising”. World Journal of Clinical Cases26 (2021): 7614-7619.
4. Sui X., et al. “Detection and size measurements of pulmonary nodules in ultra-low-dose CT with iterative reconstruction compared to low dose CT”. European Journal of Radiology3 (2016): 564-570.
5. Makaju S., et al. “Lung cancer detection using CT scan images”. Procedia Computer Science 125 (2018): 107-114.
6. Ardila D., et al. “End-to-end lung cancer screening with three-dimensional deep learning on low-dose chest computed tomography”. Nature Medicine6 (2019): 954-961.
7. Zanon M., et al. “Early detection of lung cancer using ultra-low-dose computed tomography in coronary CT angiography scans among patients with suspected coronary heart disease”. Lung Cancer 114 (2017): 1-5.
8. Ahn CK., et al. “A deep learning-enabled iterative reconstruction of ultra-low-dose CT: Use of synthetic sinogram-based noise simulation technique”. Medical Imaging 2018: Physics of Medical Imaging (2018).
9. Nam JG., et al. “Image quality of ultralow-dose chest CT using deep learning techniques: Potential superiority of vendor-agnostic post-processing over vendor-specific techniques”. European Radiology 7 (2021): 5139-5147.
10. Jensen C T., et al. “Image quality assessment of abdominal CT by use of new deep learning image reconstruction: Initial experience”. American Journal of Roentgenology1 (2020): 50-57.
11. Morozov S., et al. “MosMedData: Chest CT scans with COVID-19 related findings dataset” (2020).
12. Rawashdeh M A and Saade C. “Radiation dose reduction considerations and imaging patterns of ground glass opacities in coronavirus: Risk of over exposure in computed tomography”. La Radiologia Medica3 (2020): 380-387.
13. Jensen C T., et al. “Image quality assessment of abdominal CT by use of new deep learning image reconstruction: Initial experience”. American Journal of Roentgenology1 (2020): 50-57.
14. Putman R., et al. “Imaging patterns are associated with interstitial lung abnormality progression and mortality. D103”. Idiopathic Interstitial Pneumonias: Natural History and Prognosis (2019).
15. Hatabu H., et al. “Interstitial lung abnormality: Recognition and perspectives”. Radiology1 (2019): 1-3.
16. Radpour A., et al. “COVID-19 evaluation by low-dose high resolution CT scans protocol”. Academic Radiology6 (2020): 901.
17. Kim J H., et al. “Validation of deep-learning image reconstruction for low-dose chest computed tomography scan: Emphasis on image quality and noise”. Korean Journal of Radiology1 (2021): 131.
18. Hata A., et al. “The image quality of deep-learning image reconstruction of chest CT images on a mediastinal window setting”. Clinical Radiology2 (2021): 155.e15-155.e23.
19. Willemink M J and Noël P B. “The evolution of image reconstruction for CT—from filtered back projection to artificial intelligence”. European Radiology5 (2018): 2185-2195.
20. Weis M., et al. “Radiation dose comparison between 70 kVp and 100 kVp with spectral beam shaping for non–contrast-enhanced pediatric chest computed tomography”. Investigative Radiology3 (2017): 155-162.
21. Gordic S., et al. “Ultralow-dose chest computed tomography for pulmonary nodule detection”. Investigative Radiology7 (2014): 465-473.
22. Messerli M., et al. “Ultralow dose CT for pulmonary nodule detection with chest x-ray equivalent dose – a prospective intra-individual comparative study”. European Radiology8 (2017): 3290-3299.
23. Hata A., et al. “Ultra-low-dose chest computed tomography for interstitial lung disease using model-based iterative reconstruction with or without the lung setting”. Medicine22 (2019): e15936.
24. Svahn T M., et al. “Dose estimation of ultra-low-dose chest CT to different sized adult patients”. European Radiology8 (2018): 4315-4323.
25. Ernst C W., et al. “Pulmonary disease in cystic fibrosis: Assessment with chest CT at chest radiography dose levels”. Radiology2 (2014): 597-605.

Citation

Citation: Muhammad Mehran Mouzam., et al. “Differences in Image Quality Between Different Deep Learning Algorithms in Chest CT Scans”.Acta Scientific Medical Sciences 6.12 (2022): 114-120.

Copyright

Copyright: © 2022 Muhammad Mehran Mouzam., et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.