DW-UNet: Loss Balance under Local-Patch for 3D Infection Segmentation from COVID-19 CT Images

Chen, Cheng; Zhou, Jiancang; Zhou, Kangneng; Wang, Zhiliang; Xiao, Ruoxiu

doi:10.3390/diagnostics11111942

Cited by 9 publications

(3 citation statements)

References 41 publications

(46 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…COVID-19 pneumonia is characterized by a mixed morphology of consolidations and rounded-shaped ground glass opacities [ 16 , 17 ]. Importantly, these lesions present a diffuse patchy distribution and can affect anterior and posterior lung areas [ 17 ].…”

Section: Discussionmentioning

confidence: 99%

Prone Positioning Decreases Inhomogeneity and Improves Dorsal Compliance in Invasively Ventilated Spontaneously Breathing COVID-19 Patients—A Study Using Electrical Impedance Tomography

et al. 2022

View full text Add to dashboard Cite

Background: We studied prone positioning effects on lung aeration in spontaneously breathing invasively ventilated patients with coronavirus disease 2019 (COVID-19). Methods: changes in lung aeration were studied prospectively by electrical impedance tomography (EIT) from before to after placing the patient prone, and back to supine. Mixed effect models with a random intercept and only fixed effects were used to evaluate changes in lung aeration. Results: fifteen spontaneously breathing invasively ventilated patients were enrolled, and remained prone for a median of 19 [17 to 21] hours. At 16 h the global inhomogeneity index was lower. At 2 h, there were neither changes in dorsal nor in ventral compliance; after 16 h, only dorsal compliance (βFe +18.9 [95% Confidence interval (CI): 9.1 to 28.8]) and dorsal end-expiratory lung impedance (EELI) were increased (βFe, +252 [95% CI: 13 to 496]); at 2 and 16 h, dorsal silent spaces was unchanged (βFe, –4.6 [95% CI: –12.3 to +3.2]). The observed changes induced by prone positioning disappeared after turning patients back to supine. Conclusions: in this cohort of spontaneously breathing invasively ventilated COVID-19 patients, prone positioning decreased inhomogeneity, increased lung volumes, and improved dorsal compliance.

show abstract

Section: Discussionmentioning

confidence: 99%

Prone Positioning Decreases Inhomogeneity and Improves Dorsal Compliance in Invasively Ventilated Spontaneously Breathing COVID-19 Patients—A Study Using Electrical Impedance Tomography

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Salehi [13] proposed the Tversky loss, which is a generalization of the Dice loss, allowing for better control of the tradeoff between false positives and false negatives. This loss function has been employed in various medical image segmentation tasks, including the segmentation of COVID-19 lesions in CT scans [14] . The combination of multiple loss functions, such as the Dice loss with traditional cross-entropy loss, has also been explored to optimize segmentation performance [15].…”

Section: Related Workmentioning

confidence: 99%

Comprehensive exploration of UNet adaptations for improved COVID-19 CT segmentation

2023

Sixth International Conference on Computer Information Science and Application Technology (CISAT 2023)

View full text Add to dashboard Cite

The COVID-19 pandemic has led to a widespread need for accurate and efficient medical image analysis, particularly in the context of lung CT scans. In this paper, we present a comprehensive exploration of UNet adaptations for COVID-19 CT segmentation, focusing on the impact of different loss functions and modifications to the UNet's Convolutional Neural Network (CNN) architecture. Our experimental results demonstrate the efficacy of these approaches in improving segmentation performance. We provide a thorough comparison of various loss functions and UNet modifications, highlighting the most effective combinations for robust and accurate COVID-19 CT segmentation.

show abstract

“…In this study, we aimed to compare the efficacy of mandatory and opportunistic routine CT scans for the early detection of lung cancer. During the COVID-19 pandemic, there were many studies on the role of CT imaging in the diagnosis of lung diseases [ 15 , 16 , 17 ], but there are few studies on early screening of lung cancer.…”

Section: Introductionmentioning

confidence: 99%

Diagnostic Efficacy of CT Examination on Early Detection of Lung Cancer during Pandemic of COVID-19

et al. 2022

View full text Add to dashboard Cite

Background: Since the outbreak of COVID-19 in 2020, routine CT examination was recommended to hospitalized patients at some hospitals and discovered lung cancer patients at an early stage. This study aimed to investigate the detection efficacy of routine CT examination on early diagnosis of lung cancer, especially on pathological characteristics. Methods: The epidemic of COVID-19 outbreak in January 2020 in China, and routine CT examination was recommended to hospitalized patients in June 2020 and ended in July 2021. Based on the time points, we compared the diagnosis efficacy between three periods: pre-period, peri-period, and the period of routine CT examination. Results: During the period of routine CT examination, more early stages of lung cancer were detected and the tumor size was reduced to 2.14 cm from 3.21 cm at pre-period (p = 0.03). The proportion of lung adenocarcinoma and early stage adenocarcinoma was increased by 12% and 30% in the period of routine CT examination, with referral to the pre-period of CT examination (p < 0.05). A total of 61% of diagnosed patients had the wild type of TP53 gene during the period of routine CT examination, compared to 45% of patients at the pre-period of CT examination (p = 0.001). The median Ki-67 index was 15% among patients diagnosed at the period of routine CT examination and increased to 35% at the pre-period of CT examination (p < 0.001). The period of routine CT examination was associated with a 78% higher probability of detecting an early stage of adenocarcinoma (OR = 1.78, 95%CI 1.03, 3.08) but no significant association was observed for squamous cell carcinoma. From the pre-period to the period of routine CT examination, the proportion of female patients and non-smoking patients increased by 57% and 44%, respectively (p < 0.001). Conclusion: Routine CT examination could detect more lung cancer at an early stage, especially for adenocarcinoma, and detect patients with less aggressive features. Further studies were warranted to confirm the findings.

show abstract

DW-UNet: Loss Balance under Local-Patch for 3D Infection Segmentation from COVID-19 CT Images

Cited by 9 publications

References 41 publications

Prone Positioning Decreases Inhomogeneity and Improves Dorsal Compliance in Invasively Ventilated Spontaneously Breathing COVID-19 Patients—A Study Using Electrical Impedance Tomography

Prone Positioning Decreases Inhomogeneity and Improves Dorsal Compliance in Invasively Ventilated Spontaneously Breathing COVID-19 Patients—A Study Using Electrical Impedance Tomography

Comprehensive exploration of UNet adaptations for improved COVID-19 CT segmentation

Diagnostic Efficacy of CT Examination on Early Detection of Lung Cancer during Pandemic of COVID-19

Contact Info

Product

Resources

About