COLI-NET: Fully Automated COVID-19 Lung and Infection Pneumonia Lesion Detection and Segmentation from Chest CT Images

  1. Isaac Shiri
  2. Hossein Arabi
  3. Yazdan Salimi
  4. Amir Hossein Sanaat
  5. Azadeh Akhavanalaf
  6. Ghasem Hajianfar
  7. Dariush Askari
  8. Shakiba Moradi
  9. Zahra Mansouri
  10. Masoumeh Pakbin
  11. Saleh Sandoughdaran
  12. Hamid Abdollahi
  13. Amir Reza Radmard
  14. Kiara Rezaei-Kalantari
  15. Mostafa Ghelich Oghli
  16. Habib Zaidi
1 evaluations Published on
Read the full article Related papers
This article on Sciety

Abstract

Background

We present a deep learning (DL)-based automated whole lung and COVID-19 pneumonia infectious lesions (COLI-Net) detection and segmentation from chest CT images.

Methods

We prepared 2358 (347’259, 2D slices) and 180 (17341, 2D slices) volumetric CT images along with their corresponding manual segmentation of lungs and lesions, respectively, in the framework of a multi-center/multi-scanner study. All images were cropped, resized and the intensity values clipped and normalized. A residual network (ResNet) with non-square Dice loss function built upon TensorFlow was employed. The accuracy of lung and COVID-19 lesions segmentation was evaluated on an external RT-PCR positive COVID-19 dataset (7’333, 2D slices) collected at five different centers. To evaluate the segmentation performance, we calculated different quantitative metrics, including radiomic features.

Results

The mean Dice coefficients were 0.98±0.011 (95% CI, 0.98-0.99) and 0.91±0.038 (95% CI, 0.90-0.91) for lung and lesions segmentation, respectively. The mean relative Hounsfield unit differences were 0.03±0.84% (95% CI, −0.12 – 0.18) and −0.18±3.4% (95% CI, −0.8 - 0.44) for the lung and lesions, respectively. The relative volume difference for lung and lesions were 0.38±1.2% (95% CI, 0.16-0.59) and 0.81±6.6% (95% CI, −0.39-2), respectively. Most radiomic features had a mean relative error less than 5% with the highest mean relative error achieved for the lung for theRangefirst-order feature (- 6.95%) andleast axis lengthshape feature (8.68%) for lesions.

Conclusion

We set out to develop an automated deep learning-guided three-dimensional whole lung and infected regions segmentation in COVID-19 patients in order to develop fast, consistent, robust and human error immune framework for lung and pneumonia lesion detection and quantification.

Related articles

Related articles are currently not available for this article.