PleThora: Pleural effusion and thoracic cavity segmentations in diseased lungs for benchmarking chest CT processing pipelines

Kiser, Kendall J.; Ahmed, Sara; Stieb, Sonja; Mohamed, Abdallah S.R.; Elhalawani, Hesham; Park, Peter Y. S.; Doyle, Nathan; Wang, Brandon J.; Barman, Arko; Zhao, Li; Zheng, Weili; Fuller, C.D.; Giancardo, Luca

doi:10.1002/mp.14807

“…For lung segmentation training, we used 2298 chest CT images (328205, 2D slices) with different pathologies from different centers, including 800 normal subjects without any lung abnormalities from Iran Center#1 (81347, 2D slices), 400 images of non-small cell lung carcinoma patients from Cancer Imaging Archive (TCIA) (38–40) (48568, 2D slices), 200 non-COVID-19 pneumonia (49465, 2D slices) and 898 (148825, 2D slices) RT-PCR positive COVID-19 patients from Iran Center#2. All images were manually segmented by experienced radiologists to delineate the lungs.…”

Section: Methodsmentioning

confidence: 99%

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

Shiri

¹

,

Arabi

²

,

Salimi

³

et al. 2021

Preprint

View full text Add to dashboard Cite

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 ( 347259, 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 (7333, 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 the Range first-order feature (-6.95%) and least axis length shape 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. Keywords: X-ray CT, COVID-19, pneumonia, deep learning, segmentation.

show abstract

“…More specifically, we exploit the public datasets summarized in [36] to construct a real-world OSS task. Source data includes 30 CT scans which are randomly selected from NSCLC left and right lung segmentation dataset [49,50,51] (with CC BY-NC license). Target data consists of 20 COVID-19 CT scans and 10 non-infected CT scans.…”

Section: B2 Experiments On Covid-19 Infection Segmentation In Ct Scansmentioning

confidence: 99%

KRADA: Known-region-aware Domain Alignment for Open-set Domain Adaptation in Semantic Segmentation

Chen¹,

Liu²,

Chen³

et al. 2021

Preprint

0

View full text Add to dashboard Cite

In semantic segmentation, we aim to train a pixel-level classifier to assign category labels to all pixels in an image, where labeled training images and unlabeled test images are from the same distribution and share the same label set. However, in an open world, the unlabeled test images probably contain unknown categories and have different distributions from the labeled images. Hence, in this paper, we consider a new, more realistic, and more challenging problem setting where the pixel-level classifier has to be trained with labeled images and unlabeled open-world images-we name it open world semantic segmentation (OSS). In OSS, the trained classifier is expected to identify unknown-class pixels and classify known-class pixels well. To solve OSS, we first investigate which distribution that unknown-class pixels obey. Then, motivated by the goodness-of-fit test, we use statistical measurements to show how a pixel fits the distribution of an unknown class and select highly-fitted pixels to form the unknown region in each image. Eventually, we propose an end-to-end learning framework, known-region-aware domain alignment (KRADA), to distinguish unknown classes while aligning distributions of known classes in labeled and unlabeled open-world images. The effectiveness of KRADA has been verified on two synthetic tasks and one COVID-19 segmentation task.

show abstract

“…A giant thoracic tumor is a kind of rare tumor clinically [1][2][3]. Because of its complicated pathological types and prognosis, there is no unified standard for the diagnosis and treatment, and surgical resection becomes the major clinical treatment means [4][5][6][7].…”

Section: Introductionmentioning

confidence: 99%

Impact Evaluation of Comfort Care on the Prognosis of Giant Thoracic Tumor Patients with Computerized Tomography Imaging under Intelligent Reconstruction Algorithm

Yan

¹

2022

Scientific Programming

0

View full text Add to dashboard Cite

The reconstruction algorithm based on the network of generative adversarial and contextual coding (RANGC) was proposed in this study to analyze the impacts on the prognosis of patients with giant thoracic tumors with CT imaging under artificial intelligence algorithms. The algorithms of Feldkamp–Davis–Kress (FDK) and the generative adversarial network (GAN) were introduced. The patients were divided into the test group with comfort care and the control group with conventional care. Three sets of indicators below were also compared between the patients in two groups, including the pain level and complication incidence, the self-rating anxiety scale (SAS), the self-rating depression scale (SDS), and the patients’ satisfaction and average duration of hospital stay. When the scanning range was [0°, 89°], the peak signal-to-noise ratio (PSNR) and structural similarity index measurement (SSIM) under the RANGC algorithm were 45.6 dB and 0.99, respectively. When the scanning range changed to [0°, 119°], the PSNR and SSIM were 39.21 dB and 0.98, respectively. The results were significantly higher than those under the FDK and GAN algorithms, and the difference was obviously of statistical significance P < 0.05 . The average postoperative pain level of the patients in the control group was 3.12 points, and the postoperative complication incidence was 36.13%, while those of the test group patients were 2.27 points and 20.02%, respectively, which was greatly lower than those of the control group patients, and such a difference was of statistical significance P < 0.05 . There was no statistically significant difference in the SDS and SAS scores between the patients in the two groups before surgery. However, the SAS and SDS scores of the test group patients were 41.23 and 43.25, respectively, after surgery, which are obviously lower than those of the control group patients, with a statistically significant difference P < 0.05 . The average duration of hospital stay of the test group patients was 6.31 days, which was lower than that of the control group patients, with a statistically significant difference P < 0.05 . The overall satisfaction of the test group patients was 83.33%, which was remarkably higher than that of the control group patients, and the difference had statistical significance P < 0.05 . All these showed that the performance of the RANGC algorithm was relatively better, and comfort care did good to improve the negative mood, satisfaction, and life quality of patients after surgery.

show abstract

PleThora: Pleural effusion and thoracic cavity segmentations in diseased lungs for benchmarking chest CT processing pipelines

Abstract: Published in Medical Physics, Vol. 47, No. 11

Cited by 13 publications

References 30 publications

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

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

KRADA: Known-region-aware Domain Alignment for Open-set Domain Adaptation in Semantic Segmentation

Impact Evaluation of Comfort Care on the Prognosis of Giant Thoracic Tumor Patients with Computerized Tomography Imaging under Intelligent Reconstruction Algorithm

Contact Info

Product

Resources

About