Le Duy Mai Huyen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Le Duy Mai Huyen has a total of 28 co-authors. Here are the top co-authors by number of shared publications:
- Chien Cong Phan
Department of Diagnostic Imaging, University Medical Center, Ho Chi Minh City, VNM
2 shared publications - Le Huu Hanh Nhi
Cardiovascular Research Laboratories, Interventional Cardiology Department, Methodist Hospital, Merrillville, USA
2 shared publications - Le Huu Nhat Minh
Section of Cardiovascular Medicine, Department of Internal Medicine, Yale School of Medicine, New Haven, USA
2 shared publications - Le Nguyen Gia Hy
Department of Radiology, University of Medicine and Pharmacy at Ho Chi Minh City, Ho Chi Minh City, VNM
2 shared publications - Nguyen Hoang Nam
Department of Radiology, University of Medicine and Pharmacy at Ho Chi Minh City, Ho Chi Minh City, VNM
2 shared publications - Vo Tan Duc
Department of Diagnostic Imaging, University Medical Center, Ho Chi Minh City, VNM
2 shared publications
Latest Publications
- Open Access
CAAF-ResUNet: Adaptive Attention Fusion with Boundary-Aware Loss for Lung Nodule Segmentation
Background and Objectives: The accurate segmentation of pulmonary nodules in computed tomography (CT) remains a critical yet challenging task due to variations in nodule size, shape, and boundary ambiguity. This study proposes CAAF-ResUNet (Context-Aware Adaptive Attention Fusion ResUNet), a novel deep learning model designed to address these challenges through adaptive feature fusion and edge-sensitive learning. Materials and Methods: Central to our approach is the Adaptive Attention Controller (AAC), which dynamically adjusts the contribution of channel and position attention based on contextual features in each input. To further enhance boundary localization, we incorporate three complementary boundary-aware loss functions: Sobel, Laplacian, and Hausdorff. Results: An extensive evaluation of two benchmark datasets demonstrates the superiority of the proposed model, achieving Dice scores of 90.88% on LUNA16 and 85.92% on LIDC-IDRI, both exceeding prior state-of-the-art methods. A clinical validation of a dataset comprising 804 CT slices from 35 patients at the University Medical Center of Ho Chi Minh City confirmed the model’s practical reliability, yielding a Dice score of 95.34% and a notably low Miss Rate of 4.60% under the Hausdorff loss configuration. Conclusions: These results establish CAAF-ResUNet as a robust and clinically viable solution for pulmonary nodule segmentation, offering enhanced boundary precision and minimized false negatives, two critical properties in early-stage lung cancer diagnosis and radiological decision support.
Open AccessCAAF-ResUNet: Adaptive Attention Fusion with Boundary-Aware Loss for Lung Nodule Segmentation
Background and Objectives: The accurate segmentation of pulmonary nodules in computed tomography (CT) remains a critical yet challenging task due to variations in nodule size, shape, and boundary ambiguity. This study proposes CAAF-ResUNet (Context-Aware Adaptive Attention Fusion ResUNet), a novel deep learning model designed to address these challenges through adaptive feature fusion and edge-sensitive learning. Materials and Methods: Central to our approach is the Adaptive Attention Controller (AAC), which dynamically adjusts the contribution of channel and position attention based on contextual features in each input. To further enhance boundary localization, we incorporate three complementary boundary-aware loss functions: Sobel, Laplacian, and Hausdorff. Results: An extensive evaluation of two benchmark datasets demonstrates the superiority of the proposed model, achieving Dice scores of 90.88% on LUNA16 and 85.92% on LIDC-IDRI, both exceeding prior state-of-the-art methods. A clinical validation of a dataset comprising 804 CT slices from 35 patients at the University Medical Center of Ho Chi Minh City confirmed the model’s practical reliability, yielding a Dice score of 95.34% and a notably low Miss Rate of 4.60% under the Hausdorff loss configuration. Conclusions: These results establish CAAF-ResUNet as a robust and clinically viable solution for pulmonary nodule segmentation, offering enhanced boundary precision and minimized false negatives, two critical properties in early-stage lung cancer diagnosis and radiological decision support.
- Open Access
- 9 May 2022
- Published by Springer Nature in Cureus
- Vol. 14 (5) , e24864
Correlation of Chest X-Ray Scores in SARS-CoV-2 Patients With the Clinical Severity Classification and the Quick COVID-19 Severity Index
This study aimed to assess the role of chest X-ray (CXR) scoring methods and their correlations with the clinical severity categories and the Quick COVID-19 Severity Index (qCSI). We conducted a retrospective study of 159 COVID-19 patients who were diagnosed and treated at the University Medical Center between July and September 2021. Chest X-ray findings were evaluated, and severity scores were calculated using the modified CXR (mCXR), Radiographic Assessment of Lung Edema (RALE), and Brixia scoring systems. The three scores were then compared to the clinical severity categories and the qCSI using Spearman's correlation coefficient. Overall, 159 patients (63 males and 96 females) (mean age: 58.3 ± 15.7 years) were included. The correlation coefficients between the mCXR score and the Brixia and RALE scores were 0.9438 and 0.9450, respectively. The correlation coefficient between the RALE and Brixia scores was marginally higher, at 0.9625. The correlation coefficients between the qCSI and the Brixia, RALE, and mCXR scores were 0.7298, 0.7408, and 0.7156, respectively. The significant difference in the mean values of the three CXR scores between asymptomatic, mild, moderate, severe, and critical groups was also noted. There were strong correlations between the three CXR scores and the clinical severity classification and the qCSI.
- 9 May 2022
- Published by Springer Nature in Cureus
- Vol. 14 (5) , e24864
Open AccessCorrelation of Chest X-Ray Scores in SARS-CoV-2 Patients With the Clinical Severity Classification and the Quick COVID-19 Severity Index
This study aimed to assess the role of chest X-ray (CXR) scoring methods and their correlations with the clinical severity categories and the Quick COVID-19 Severity Index (qCSI). We conducted a retrospective study of 159 COVID-19 patients who were diagnosed and treated at the University Medical Center between July and September 2021. Chest X-ray findings were evaluated, and severity scores were calculated using the modified CXR (mCXR), Radiographic Assessment of Lung Edema (RALE), and Brixia scoring systems. The three scores were then compared to the clinical severity categories and the qCSI using Spearman's correlation coefficient. Overall, 159 patients (63 males and 96 females) (mean age: 58.3 ± 15.7 years) were included. The correlation coefficients between the mCXR score and the Brixia and RALE scores were 0.9438 and 0.9450, respectively. The correlation coefficient between the RALE and Brixia scores was marginally higher, at 0.9625. The correlation coefficients between the qCSI and the Brixia, RALE, and mCXR scores were 0.7298, 0.7408, and 0.7156, respectively. The significant difference in the mean values of the three CXR scores between asymptomatic, mild, moderate, severe, and critical groups was also noted. There were strong correlations between the three CXR scores and the clinical severity classification and the qCSI.
- Open Access
- 17 January 2022
- Published by Springer Nature in Cureus
- Vol. 14 (1) , e21347
Deep Learning Model With Convolutional Neural Network for Detecting and Segmenting Hepatocellular Carcinoma in CT: A Preliminary Study
Hepatocellular carcinoma (HCC) is one of the most common malignancies in the world. Early detection and accurate diagnosis of HCC play an important role in patient management. This study aimed to develop a convolutional neural network-based model to identify and segment HCC lesions utilizing dynamic contrast agent-enhanced computed tomography (CT). This retrospective study used CT image sets of histopathology-confirmed hepatocellular carcinoma over three phases (arterial, venous, and delayed). The proposed convolutional neural network (CNN) segmentation method was based on the U-Net architecture and trained using the domain adaptation technique. The proposed method was evaluated using 115 liver masses of 110 patients (87 men and 23 women; mean age, 56.9 years ± 11.9 (SD); mean mass size, 6.0 cm ± 3.6). The sensitivity for identifying HCC of the model and Dice score for segmentation of liver masses between radiologists and the CNN model were calculated for the test set. The sensitivity for HCC identification of the model was 100%. The median Dice score for HCC segmenting between radiologists and the CNN model was 0.81 for the test set. Deep learning with CNN had high performance in the identification and segmentation of HCC on dynamic CT.
- 17 January 2022
- Published by Springer Nature in Cureus
- Vol. 14 (1) , e21347
Open AccessDeep Learning Model With Convolutional Neural Network for Detecting and Segmenting Hepatocellular Carcinoma in CT: A Preliminary Study
Hepatocellular carcinoma (HCC) is one of the most common malignancies in the world. Early detection and accurate diagnosis of HCC play an important role in patient management. This study aimed to develop a convolutional neural network-based model to identify and segment HCC lesions utilizing dynamic contrast agent-enhanced computed tomography (CT). This retrospective study used CT image sets of histopathology-confirmed hepatocellular carcinoma over three phases (arterial, venous, and delayed). The proposed convolutional neural network (CNN) segmentation method was based on the U-Net architecture and trained using the domain adaptation technique. The proposed method was evaluated using 115 liver masses of 110 patients (87 men and 23 women; mean age, 56.9 years ± 11.9 (SD); mean mass size, 6.0 cm ± 3.6). The sensitivity for identifying HCC of the model and Dice score for segmentation of liver masses between radiologists and the CNN model were calculated for the test set. The sensitivity for HCC identification of the model was 100%. The median Dice score for HCC segmenting between radiologists and the CNN model was 0.81 for the test set. Deep learning with CNN had high performance in the identification and segmentation of HCC on dynamic CT.