International Journal of Biomedical Imaging最新文献

We present a deep learning-based method that corrects motion artifacts and thus accelerates data acquisition and reconstruction of magnetic resonance images. The novel model, the Motion Artifact Correction by Swin Network (MACS-Net), uses a Swin transformer layer as the fundamental block and the Unet architecture as the neural network backbone. We employ a hierarchical transformer with shifted windows to extract multiscale contextual features during encoding. A new dual upsampling technique is employed to enhance the spatial resolutions of feature maps in the Swin transformer-based decoder layer. A raw magnetic resonance imaging dataset is used for network training and testing; the data contain various motion artifacts with ground truth images of the same subjects. The results were compared to six state-of-the-art MRI image motion correction methods using two types of motions. When motions were brief (within 5 s), the method reduced the average normalized root mean square error (NRMSE) from 45.25% to 17.51%, increased the mean structural similarity index measure (SSIM) from 79.43% to 91.72%, and increased the peak signal-to-noise ratio (PSNR) from 18.24 to 26.57 dB. Similarly, when motions were extended from 5 to 10 s, our approach decreased the average NRMSE from 60.30% to 21.04%, improved the mean SSIM from 33.86% to 90.33%, and increased the PSNR from 15.64 to 24.99 dB. The anatomical structures of the corrected images and the motion-free brain data were similar.

Brain tumors are critical neurological ailments caused by uncontrolled cell growth in the brain or skull, often leading to death. An increasing patient longevity rate requires prompt detection; however, the complexities of brain tissue make early diagnosis challenging. Hence, automated tools are necessary to aid healthcare professionals. This study is particularly aimed at improving the efficacy of computerized brain tumor detection in a clinical setting through a deep learning model. Hence, a novel thresholding-based MRI image segmentation approach with a transfer learning model based on contour (ContourTL-Net) is suggested to facilitate the clinical detection of brain malignancies at an initial phase. The model utilizes contour-based analysis, which is critical for object detection, precise segmentation, and capturing subtle variations in tumor morphology. The model employs a VGG-16 architecture priorly trained on the "ImageNet" collection for feature extraction and categorization. The model is designed to utilize its ten nontrainable and three trainable convolutional layers and three dropout layers. The proposed ContourTL-Net model is evaluated on two benchmark datasets in four ways, among which an unseen case is considered as the clinical aspect. Validating a deep learning model on unseen data is crucial to determine the model's generalization capability, domain adaptation, robustness, and real-world applicability. Here, the presented model's outcomes demonstrate a highly accurate classification of the unseen data, achieving a perfect sensitivity and negative predictive value (NPV) of 100%, 98.60% specificity, 99.12% precision, 99.56% F1-score, and 99.46% accuracy. Additionally, the outcomes of the suggested model are compared with state-of-the-art methodologies to further enhance its effectiveness. The proposed solution outperforms the existing solutions in both seen and unseen data, with the potential to significantly improve brain tumor detection efficiency and accuracy, leading to earlier diagnoses and improved patient outcomes.

A challenge in accurately identifying and classifying left ventricular hypertrophy (LVH) is distinguishing it from hypertrophic cardiomyopathy (HCM) and Fabry disease. The reliance on imaging techniques often requires the expertise of multiple specialists, including cardiologists, radiologists, and geneticists. This variability in the interpretation and classification of LVH leads to inconsistent diagnoses. LVH, HCM, and Fabry cardiomyopathy can be differentiated using T1 mapping on cardiac magnetic resonance imaging (MRI). However, differentiation between HCM and Fabry cardiomyopathy using echocardiography or MRI cine images is challenging for cardiologists. Our proposed system named the MRI short-axis view left ventricular hypertrophy classifier (MSLVHC) is a high-accuracy standardized imaging classification model developed using AI and trained on MRI short-axis (SAX) view cine images to distinguish between HCM and Fabry disease. The model achieved impressive performance, with an F1-score of 0.846, an accuracy of 0.909, and an AUC of 0.914 when tested on the Taipei Veterans General Hospital (TVGH) dataset. Additionally, a single-blinding study and external testing using data from the Taichung Veterans General Hospital (TCVGH) demonstrated the reliability and effectiveness of the model, achieving an F1-score of 0.727, an accuracy of 0.806, and an AUC of 0.918, demonstrating the model's reliability and usefulness. This AI model holds promise as a valuable tool for assisting specialists in diagnosing LVH diseases.

Purpose: This study is aimed at evaluating the efficacy of the gradient-spin echo- (GraSE-) based short tau inversion recovery (STIR) sequence (GraSE-STIR) in cardiovascular magnetic resonance (CMR) imaging compared to the conventional turbo spin echo- (TSE-) based STIR sequence, specifically focusing on image quality, specific absorption rate (SAR), and image acquisition time.

Methods: In a prospective study, we examined forty-four normal volunteers and seventeen patients referred for CMR imaging using conventional STIR and GraSE-STIR techniques. Signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), image quality, T₂ signal intensity (SI) ratio, SAR, and image acquisition time were compared between both sequences.

Results: GraSE-STIR showed significant improvements in image quality (4.15 ± 0.8 vs. 3.34 ± 0.9, p = 0.024) and cardiac motion artifact reduction (7 vs. 18 out of 53, p = 0.038) compared to conventional STIR. Furthermore, the acquisition time (27.17 ± 3.53 vs. 36.9 ± 4.08 seconds, p = 0.041) and the local torso SAR (<13% vs. <17%, p = 0.047) were significantly lower for GraSE-STIR compared to conventional STIR in short-axis plan. However, no significant differences were shown in T₂ SI ratio (p = 0.141), SNR (p = 0.093), CNR (p = 0.068), and SAR (p = 0.071) between these two sequences.

Conclusions: GraSE-STIR offers notable advantages over conventional STIR sequence, with improved image quality, reduced motion artifacts, and shorter acquisition times. These findings highlight the potential of GraSE-STIR as a valuable technique for routine clinical CMR imaging.

Background: MRI is an important tool for accurate detection and targeted biopsy of prostate lesions. However, the imaging appearances of some prostate cancers are similar to those of the surrounding normal tissue on MRI, which are referred to as MRI-invisible prostate cancers (MIPCas). The detection of MIPCas remains challenging and requires extensive systematic biopsy for identification. In this study, we developed a weakly supervised UNet (WSUNet) to detect MIPCas.

Methods: The study included 777 patients (training set: 600; testing set: 177), all of them underwent comprehensive prostate biopsies using an MRI-ultrasound fusion system. MIPCas were identified in MRI based on the Gleason grade (≥7) from known systematic biopsy results.

Results: The WSUNet model underwent validation through systematic biopsy in the testing set with an AUC of 0.764 (95% CI: 0.728-0.798). Furthermore, WSUNet exhibited a statistically significant precision improvement of 91.3% (p < 0.01) over conventional systematic biopsy methods in the testing set. This improvement resulted in a substantial 47.6% (p < 0.01) decrease in unnecessary biopsy needles, while maintaining the same number of positively identified cores as in the original systematic biopsy.

Conclusions: In conclusion, the proposed WSUNet could effectively detect MIPCas, thereby reducing unnecessary biopsies.

Background: Artificial intelligence (AI) applications are rapidly advancing in the field of medical imaging. This study is aimed at investigating the perception and knowledge of radiographers towards artificial intelligence.

Methods: An online survey employing Google Forms consisting of 20 questions regarding the radiographers' perception of AI. The questionnaire was divided into two parts. The first part consisted of demographic information as well as whether the participants think AI should be part of medical training, their previous knowledge of the technologies used in AI, and whether they prefer to receive training on AI. The second part of the questionnaire consisted of two fields. The first one consisted of 16 questions regarding radiographers' perception of AI applications in radiology. Descriptive analysis and logistic regression analysis were used to evaluate the effect of gender on the items of the questionnaire.

Results: Familiarity with AI was low, with only 52 out of 100 respondents (52%) reporting good familiarity with AI. Many participants considered AI useful in the medical field (74%). The findings of the study demonstrate that nearly most of the participants (98%) believed that AI should be integrated into university education, with 87% of the respondents preferring to receive training on AI, with some already having prior knowledge of AI used in technologies. The logistic regression analysis indicated a significant association between male gender and experience within the range of 23-27 years with the degree of familiarity with AI technology, exhibiting respective odds ratios of 1.89 (COR = 1.89) and 1.87 (COR = 1.87).

Conclusions: This study suggests that medical practices have a favorable attitude towards AI in the radiology field. Most participants surveyed believed that AI should be part of radiography education. AI training programs for undergraduate and postgraduate radiographers may be necessary to prepare them for AI tools in radiology development.

Superresolution, structured illumination microscopy (SIM) is an ideal modality for imaging live cells due to its relatively high speed and low photon-induced damage to the cells. The rate-limiting step in observing a superresolution image in SIM is often the reconstruction speed of the algorithm used to form a single image from as many as nine raw images. Reconstruction algorithms impose a significant computing burden due to an intricate workflow and a large number of often complex calculations to produce the final image. Further adding to the computing burden is that the code, even within the MATLAB environment, can be inefficiently written by microscopists who are noncomputer science researchers. In addition, they do not take into consideration the processing power of the graphics processing unit (GPU) of the computer. To address these issues, we present simple but efficient approaches to first revise MATLAB code, followed by conversion to GPU-optimized code. When combined with cost-effective, high-performance GPU-enabled computers, a 4- to 500-fold improvement in algorithm execution speed is observed as shown for the image denoising Hessian-SIM algorithm. Importantly, the improved algorithm produces images identical in quality to the original.

Background: The deterministic fiber tracking method has the advantage of high computational efficiency and good repeatability, making it suitable for the noninvasive estimation of brain structural connectivity in clinical fields. To address the issue of the current classical deterministic method tending to deviate in the tracking direction in the region of crossing fiber region, in this paper, we propose an adaptive correction-based deterministic white matter fiber tracking method, named FTACTD.

Methods: The proposed FTACTD method can accurately track white matter fibers by adaptively adjusting the deflection direction strategy based on the tensor matrix and the input fiber direction of adjacent voxels. The degree of correction direction changes adaptively according to the shape of the diffusion tensor, mimicking the actual tracking deflection angle and direction. Furthermore, both forward and reverse tracking techniques are employed to track the entire fiber. The effectiveness of the proposed method is validated and quantified using both simulated and real brain datasets. Various indicators such as invalid bundles (IB), valid bundles (VB), invalid connections (IC), no connections (NC), and valid connections (VC) are utilized to assess the performance of the proposed method on simulated data and real diffusion-weighted imaging (DWI) data.

Results: The experimental results of the simulated data show that the FTACTD method tracks outperform existing methods, achieving the highest number of VB with a total of 13 bundles. Additionally, it identifies the least number of incorrect fiber bundles, with only 32 bundles identified as wrong. Compared to the FACT method, the FTACTD method reduces the number of NC by 36.38%. In terms of VC, the FTACTD method surpasses even the best performing SD_Stream method among deterministic methods by 1.64%. Extensive in vivo experiments demonstrate the superiority of the proposed method in terms of tracking more accurate and complete fiber paths, resulting in improved continuity.

Conclusion: The FTACTD method proposed in this study indicates superior tracking results and provides a methodological basis for the investigating, diagnosis, and treatment of brain disorders associated with white matter fiber deficits and abnormalities.

Skin cancer is a significant health concern worldwide, and early and accurate diagnosis plays a crucial role in improving patient outcomes. In recent years, deep learning models have shown remarkable success in various computer vision tasks, including image classification. In this research study, we introduce an approach for skin cancer classification using vision transformer, a state-of-the-art deep learning architecture that has demonstrated exceptional performance in diverse image analysis tasks. The study utilizes the HAM10000 dataset; a publicly available dataset comprising 10,015 skin lesion images classified into two categories: benign (6705 images) and malignant (3310 images). This dataset consists of high-resolution images captured using dermatoscopes and carefully annotated by expert dermatologists. Preprocessing techniques, such as normalization and augmentation, are applied to enhance the robustness and generalization of the model. The vision transformer architecture is adapted to the skin cancer classification task. The model leverages the self-attention mechanism to capture intricate spatial dependencies and long-range dependencies within the images, enabling it to effectively learn relevant features for accurate classification. Segment Anything Model (SAM) is employed to segment the cancerous areas from the images; achieving an IOU of 96.01% and Dice coefficient of 98.14% and then various pretrained models are used for classification using vision transformer architecture. Extensive experiments and evaluations are conducted to assess the performance of our approach. The results demonstrate the superiority of the vision transformer model over traditional deep learning architectures in skin cancer classification in general with some exceptions. Upon experimenting on six different models, ViT-Google, ViT-MAE, ViT-ResNet50, ViT-VAN, ViT-BEiT, and ViT-DiT, we found out that the ML approach achieves 96.15% accuracy using Google's ViT patch-32 model with a low false negative ratio on the test dataset, showcasing its potential as an effective tool for aiding dermatologists in the diagnosis of skin cancer.