A survey on deep learning in medical ultrasound imaging

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Around 2017, applications of deep learning in beamforming began to appear in publications, and interest in this area has been increasing ever since.

“Around 2017, applications of deep learning in beamforming began to appear in publications [,], and the interest in this area has been increasing ever since.”

• cites [,]Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

More research is focused on the application of deep learning in plane wave imaging.

“More research is focused on the application of deep learning in plane wave imaging [–].”

• cites [–]Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Deep learning has brought a revolutionary change in ultrasound beamforming, significantly enhancing image quality and improving computational efficiency.

“Deep learning has brought a revolutionary change in ultrasound beamforming, significantly enhancing image quality and improving computational efficiency.”

May combine two claims.

Deep learning models have shown promising results in enhancing image resolution, reducing speckle noise, improving contrast, and performing advanced tasks like tissue characterization and acoustic aberration correction in ultrasound beamforming.

“They have shown promising results in enhancing image resolution, reducing speckle noise, improving contrast, and even performing advanced tasks like tissue characterization and acoustic aberration correction.”

May combine two claims.

Rzaz et al.'s DeepBreastCancerNet achieved a classification accuracy of 99.35%, outperforming several state-of-the-art deep learning models.

“The proposed model achieved a classification accuracy of 99.35%, outperforming several state-of-the-art deep learning models.”

May combine two claims.

CohereNet is faster than SLSC and has high generality.

“As described in [], the CohereNet is faster than SLSC, and this network also has high generality.”

• cites []Not checked yet

  Not checked in the free run.

May combine two claims.

Xu et al.'s proposed segmentation method achieved Accuracy, Precision, Recall, and F1-measure all exceeding 80%.

“Accuracy, Precision, Recall, and F1-measure all exceeded 80%.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Deep learning helps build more detailed 3D and 4D imaging from 2D ultrasound images.

“Deep learning helps build more detailed 3D and 4D imaging from 2D ultrasound images, providing a more comprehensive picture of the patient’s condition.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The application of AI in medical ultrasound analysis predates that of beamforming.

“From Ref. [–], it can be seen that the application of AI in medical ultrasound analysis predates that of beamforming.”

• cites [–]Not checked yet

  Not checked in the free run.

Classified as probably not a substantial claim. May combine two claims.

These review articles mainly focus on the principles of ultrasound imaging and do not discuss clinical applications.

“However, these articles mainly focus on the principles of ultrasound imaging and do not discuss clinical applications.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Deep learning algorithms can be used to predict clinical outcomes or progression of a disease based on ultrasound imaging data.

“Deep learning algorithms can be used to predict clinical outcomes or progression of a disease based on ultrasound imaging data.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Medical ultrasound analysis mainly includes segmentation, classification, registration, and localization.

“Medical ultrasound analysis mainly includes segmentation, classification, registration, and localization [,].”

• cites [,]Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Ultrasonography is a primary diagnostic method for thyroid diseases.

“Ultrasonography is also a primary diagnostic method for thyroid diseases [].”

• cites []Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

CAMUS is the largest publicly available and fully annotated dataset for echocardiographic assessment, containing data from 500 patients.

“CAMUS is the largest publicly available and fully annotated dataset for echocardiographic assessment, containing data from 500 patients.”

May combine two claims.

The automated segmentation method can offer objective references for radiologists, aiding in breast cancer diagnosis and breast density assessments.

“The automated segmentation method can offer objective references for radiologists, aiding in breast cancer diagnosis and breast density assessments.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The fast evaluation speed of the software supports the feasibility of real-time image analysis during ultrasound examinations.

“The fast evaluation speed of the software supports the feasibility of real-time image analysis during ultrasound examinations.”

Probably not a substantial claim: may not pass S4 (based).

Deep learning models can be trained to automatically detect and segment lesions in ultrasound images.

“Deep learning models can be trained to automatically detect and segment lesions in ultrasound images.”

Probably not a substantial claim: may not pass S3 (load-bearing).

There are also some articles that provide reviews from the perspective of image analysis and clinical practices.

“There are also some articles that provide reviews from the perspective of image analysis and clinical practices.”

Probably not a substantial claim: may not pass S4 (based).

The application of deep learning in breast ultrasound can effectively assist radiologists or clinicians in diagnosis.

“The application of deep learning in breast ultrasound can effectively assist radiologists or clinicians in diagnosis.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Some open-source datasets, such as BUSI, have promoted the widespread application of deep learning.

“Some open-source datasets, such as Breast Ultrasound Images Dataset (BUSI) [], have also promoted the widespread application of deep learning.”

• cites []Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The ultrasound community organized a challenge to encourage researchers to engage in deep learning research.

“In addition, the ultrasound community also organized a challenge to encourage researchers to engage in deep learning research [,].”

• cites [,]Not checked yet

  Not checked in the free run.

May combine two claims.

Jaccard similarity index for mass segmentation reached 85.1%, outperforming previous methods.

“Jaccard similarity index (JSI) for mass segmentation reached 85.1%, outperforming previous methods.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Automated detection and segmentation by deep learning reduces the workload for radiologists and increases accuracy, as human interpretation can be subjective and variable.

“This reduces the workload for radiologists and increases accuracy, as human interpretation can be subjective and variable.”

May combine two claims.

The DLS achieved diagnostic accuracy comparable to radiologists and performed better than a medical student with no prior experience.

“The DLS achieved diagnostic accuracy comparable to radiologists and performed better than a medical student with no prior experience.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Deep learning is widely applied in cardiac ultrasound.

“Deep learning is also widely applied in cardiac ultrasound.”

May combine two claims.

The deep learning and feature fusion method has potential to provide reliable support for radiologists, enhancing early detection and treatment planning.

“It highlights the potential of the method to provide reliable support for radiologists, enhancing early detection and treatment planning.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Breast ultrasound imaging is commonly utilized to detect potential breast diseases.

“Breast ultrasound imaging is commonly utilized to detect potential breast diseases [].”

• cites []Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

There are already some review articles about deep learning in medical ultrasound imaging.

“There are already some review articles about deep learning in medical ultrasound imaging.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

With the development of AI, especially deep learning, it has promoted the evolution of ultrasound breast imaging.

“With the development of AI, especially the advent of deep learning technologies, it has also promoted the evolution of ultrasound breast imaging.”

May combine two claims.

EchoNet could identify local cardiac structures, estimate cardiac function, and predict systemic phenotypes like age, sex, weight, and height with significant accuracy.

“This model could identify local cardiac structures, estimate cardiac function, and predict systemic phenotypes like age, sex, weight, and height with significant accuracy.”

May combine two claims.

The left atrium segmentation method boosts accuracy and ensures generalizability of the model across different devices.

“This method not only boosts accuracy but also ensures the generalizability of the model across different devices, making it a promising tool for clinical use.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

There are many applications of deep learning in prostate cancer detection and prostate segmentation.

“There are also many applications of deep learning in prostate cancer detection [–] and prostate segmentation [–].”

• cites [–]Not checked yet

  Not checked in the free run.

Classified as probably not a substantial claim.

Deep learning models can be trained to classify diseases based on ultrasound images.

“They can also be trained to classify diseases based on ultrasound images.”

May combine two claims.

Qian et al.'s system could facilitate the adoption of ultrasound in breast cancer screening, particularly beneficial for women with dense breasts where mammography is less effective.

“It could facilitate the adoption of ultrasound in breast cancer screening, particularly beneficial for women with dense breasts where mammography is less effective.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

In ultrasound fetal imaging, deep learning plays an increasingly important role.

“In ultrasound fetal imaging, deep learning also plays an increasingly important role[–].”

• cites [–]Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

There is a constant emergence of deep learning research in ultrasound brain imaging.

“In addition, there is a constant emergence of deep learning research in ultrasound brain imaging [–].”

• cites [–]Not checked yet

  Not checked in the free run.

Classified as probably not a substantial claim. May combine two claims.

Deep learning plays a significant role in ultrasound clinical applications as it enhances the efficiency and accuracy of diagnosis, reducing human errors.

“Deep learning plays a significant role in ultrasound clinical applications as it enhances the efficiency and accuracy of diagnosis, reducing human errors and paving the way for more sophisticated applications.”

May combine two claims.

Karimi et al.'s method achieved a Dice score of 93.9% ± 3.5% and a Hausdorff distance of 2.7 ± 2.3 mm, outperforming several other methods.

“The method achieved a Dice score of 93.9% ± 3.5% and a Hausdorff distance of 2.7 ± 2.3 mm, outperforming several other methods and demonstrating its effectiveness in reducing the likelihood of large segmentation errors.”

Classified as probably not a substantial claim.

Transformer models and their self-attention mechanisms have introduced new opportunities for enhancing ultrasound image analysis.

“On the other hand, the advent of Transformer models and their self-attention mechanisms [] has introduced new opportunities for enhancing ultrasound image analysis.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Transformers utilize a self-attention mechanism that allows the model to weigh the importance of different input elements dynamically.

“Transformers, originally designed for natural language processing, utilize a self-attention mechanism that allows the model to weigh the importance of different input elements dynamically [].”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

CNNs have been the backbone of medical image analysis for years.

“CNNs have been the backbone of medical image analysis for years.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The Vector Self-Attention Layer (VSAL) performs long-range spatial and channel-wise reasoning simultaneously.

“Another segmentation method specific to ultrasound images is the integration of a Vector Self-Attention Layer (VSAL) [], which performs long-range spatial and channel-wise reasoning simultaneously.”

May combine two claims.

The brain tumor segmentation method's effectiveness was assessed on a dataset of 33 patients, showing promising results in computational time and accuracy.

“The method’s effectiveness was assessed on a dataset of 33 patients, showing promising results in terms of computational time and accuracy, indicating its potential utility in supporting neurosurgeons during brain tumor resections.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Yang et al. propose a framework for generating high-quality medical reports from ultrasound images using an adaptive multimodal attention network (AMAnet).

“Yang et al. propose a novel framework for generating high-quality medical reports from ultrasound images using an adaptive multimodal attention network (AMAnet).”

Probably not a substantial claim: may not pass S3 (load-bearing).

Networks such as U-Net and its variants have been particularly successful in medical image segmentation tasks due to their encoder-decoder architectures, which capture detailed spatial hierarchies.

“Networks such as U-Net[] and its variants [–] have been particularly successful in medical image segmentation tasks due to their encoder-decoder architectures, which capture detailed spatial hierarchies.”

May combine two claims.

Di Ianni and Airan's method significantly reduces the amount of data required for imaging while maintaining image quality.

“The method significantly reduces the amount of data required for imaging while maintaining image quality, using a CNN to reconstruct power Doppler images from sparsely sampled ultrasound data.”

Classified as probably not a substantial claim.

Mean squared error validation loss is not a reliable predictor for image quality.

“it points out that mean squared error (MSE) validation loss is not a reliable predictor for image quality.”

Probably not a substantial claim: may not pass S3 (load-bearing).

GPA-TUNet integrates Group Parallel Axial Attention (GPA) with Transformers to enhance both local and global feature extraction.

“the GPA-TUNet[] model integrates Group Parallel Axial Attention (GPA) with Transformers to enhance both local and global feature extraction.”

May combine two claims.

VSAL can be seamlessly integrated into existing CNN architectures, enhancing their performance by adding the benefits of self-attention.

“This layer can be seamlessly integrated into existing CNN architectures, enhancing their performance by adding the benefits of self-attention.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Studies have shown that Transformer-based models significantly improve the accuracy of ultrasound image segmentation tasks.

“Studies [] have shown that Transformer-based models significantly improve the accuracy of ultrasound image segmentation tasks.”

Classified as probably not a substantial claim. May combine two claims.

AI models automating echocardiogram interpretation could streamline clinical workflows, provide preliminary interpretations in regions lacking specialized cardiologists, and offer insights into phenotypes challenging for human evaluation.

“By automating echocardiogram interpretation, such AI models could streamline clinical workflows, provide preliminary interpretations in regions lacking specialized cardiologists, and offer insights into phenotypes challenging for human evaluation.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A larger number of architectures are based on CNNs.

“It can be seen from our previous review that a larger number of architectures are based on the CNNs.”

The quote only partly states this claim.

LSTM networks outperformed other models in identifying cancerous tissues.

“demonstrating that LSTM networks outperformed other models in identifying cancerous tissues.”

Classified as probably not a substantial claim. May combine two claims.

VSAL is designed to maintain translational equivariance and accommodate multi-scale inputs, which are critical for handling the variability in ultrasound images.

“VSAL is designed to maintain translational equivariance and accommodate multi-scale inputs, which are critical for handling the variability in ultrasound images.”

May combine two claims.

The pooled AUROC increased from 0.837 to 0.875 with ThyNet assistance for image reviews and from 0.862 to 0.873 in a clinical setting involving image and video reviews.

“The pooled AUROC increased from 0.837 to 0.875 with ThyNet assistance for image reviews and from 0.862 to 0.873 in a clinical setting involving image and video reviews.”

May combine two claims.

Significant discriminative features could be captured within the first half of the TeUS sequence, suggesting a potential reduction in data acquisition time for clinical applications.

“This analysis revealed that significant discriminative features could be captured within the first half of the TeUS sequence, suggesting a potential reduction in data acquisition time for clinical applications.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The self-attention mechanism operates by creating attention scores between all pairs of input elements, which in the context of images correspond to pixels or features.

“The self-attention mechanism operates by creating attention scores between all pairs of input elements, which in the context of images, correspond to pixels or features.”

May combine two claims.

Milletari et al.'s method is designed to adapt to different imaging modalities, showing effectiveness in MRI and transcranial ultrasound volumes.

“The method is particularly designed to adapt to different imaging modalities, showing effectiveness in MRI and transcranial ultrasound volumes.”

May combine two claims.

The GPA-TUNet hybrid approach leverages the strengths of Transformers in capturing long-range dependencies and the efficiency of GPA in highlighting local information.

“This hybrid approach leverages the strengths of Transformers in capturing long-range dependencies and the efficiency of GPA in highlighting local information.”

May combine two claims.

LSTM networks achieved an AUC of 0.96, with sensitivity, specificity, and accuracy rates of 0.76, 0.98, and 0.93, respectively.

“LSTM networks achieved an area under the curve (AUC) of 0.96, with sensitivity, specificity, and accuracy rates of 0.76, 0.98, and 0.93, respectively”

May combine two claims.

TransUNet integrates CNNs and Transformers into a unified framework, where CNNs extract initial feature maps from medical images and Transformers encode these features into tokenized patches to capture global context.

“For instance, TransUNet [] architecture integrates CNNs and Transformers into a unified framework, where CNNs are employed to extract initial feature maps from medical images, and Transformers encode these features into tokenized patches to capture global context.”

May combine two claims.

Novice-operated, AI-assisted echocardiograms were of diagnostic quality in a high percentage of cases, closely aligning with the quality of scans performed by experienced sonographers.

“Results indicated that the novice-operated, AI-assisted echocardiograms were of diagnostic quality in a high percentage of cases, closely aligning with the quality of scans performed by experienced sonographers.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

CNNs excel in extracting local features through convolutional layers, pooling, and activation functions.

“They excel in extracting local features through convolutional layers, pooling, and activation functions.”

May combine two claims.

The CNN + LSTM model is more versatile and generalizable, making it a superior choice for medical ultrasound tomography applications.

“The findings suggest that the CNN + LSTM model is more versatile and generalizable, making it a superior choice for medical ultrasound tomography applications.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Deep learning within an IIoT framework presents a promising approach to enhancing the efficiency and reliability of fetal ultrasound analysis, particularly in resource-constrained environments.

“The use of deep learning within the IIoT framework presents a promising approach to enhancing the efficiency and reliability of fetal ultrasound analysis, particularly in resource-constrained environments.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The integration of U-net with transformer has become a new direction for current research.

“The integration of U-net with transformer has also become a new direction for current research [–].”

May combine two claims.

The ThyNet-assisted strategy significantly decreased the percentage of unnecessary FNAs from 61.9% to 35.2% and reduced the rate of missed malignancies from 18.9% to 17.0%.

“The ThyNet-assisted strategy significantly decreased the percentage of unnecessary FNAs from 61.9% to 35.2%, while also reducing the rate of missed malignancies from 18.9% to 17.0%.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The dynamic weighting capability of self-attention is particularly beneficial for medical image analysis, where different regions of an image may hold varying levels of significance for accurate diagnosis.

“This capability is particularly beneficial for medical image analysis [,,–], where different regions of an image may hold varying levels of significance for accurate diagnosis.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Vienneau et al.'s novel training scheme can potentially increase the maximum achievable image quality for ultrasound beamforming using DNNs.

“The authors suggest that their novel training scheme can potentially increase the maximum achievable image quality for ultrasound beamforming using DNNs.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Deep learning can improve the quality of ultrasound images by denoising, enhancing edges and contrast, and reconstructing details more finely.

“Deep learning can improve the quality of ultrasound images by denoising, enhancing edges and contrast, and reconstructing details more finely.”

May combine two claims.

On the private dataset, the model achieved 95.2% accuracy in AS detection and 78.1% in severity classification.

“On the private dataset, the model achieved 95.2% accuracy in AS detection and 78.1% in severity classification.”

May combine two claims.

Deep learning models can be trained on large datasets of ultrasound images to learn optimal beamforming parameters for different imaging conditions, resulting in better image quality compared to traditional beamforming techniques that use preset parameters.

“Deep learning models can be trained on large datasets of ultrasound images to learn optimal beamforming parameters for different imaging conditions. This can result in better image quality compared to traditional beamforming techniques that use preset parameters.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Deep learning can align the spatial orientation and adjust the pixel intensity of multiple images from different sources, times, directions, or modalities to increase the effective sample size and reduce non-biological differences.

“Deep learning can align the spatial orientation and adjust the pixel intensity of multiple images from different sources, times, directions, or modalities to increase the effective sample size and reduce non-biological differences.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Deep learning can help identify and reduce artifacts in ultrasound images, such as speckle noise, which can interfere with the clarity of the images and the diagnosis.

“Deep learning can help identify and reduce artifacts in ultrasound images, such as speckle noise, which can interfere with the clarity of the images and the diagnosis.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Models like AMAnet incorporate adaptive attention mechanisms that dynamically control the focus on visual features and language model memories, enabling the model to generate coherent and contextually appropriate reports.

“Models like AMAnet[] incorporate adaptive attention mechanisms that dynamically control the focus on visual features and language model memories. This enables the model to generate coherent and contextually appropriate reports, enhancing its utility in clinical settings.”

May combine two claims.

By providing accurate real-time force feedback, TransVFS can help reduce the risk of tissue damage and improve the precision of biopsy procedures, thereby enhancing patient outcomes.

“By providing accurate real-time force feedback, TransVFS can help reduce the risk of tissue damage and improve the precision of biopsy procedures, thereby enhancing patient outcomes.”

• AssumptionRests on an assumption: Performance improvements reported in the reviewed studies on specific datasets or experimental settings generalize to broader clinical ultrasound imaging populations and settings.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Deep learning can assist doctors in diagnosing various diseases, including tumors, inflammations, and injuries, and has been successfully used in the diagnosis of diseases such as lung cancer and breast cancer.

“Deep learning can assist doctors in diagnosing various diseases, including tumors, inflammations, injuries, etc. For example, it has been successfully used in the diagnosis of many diseases such as lung cancer and breast cancer.”

Probably not a substantial claim: may not pass S4 (based).

The ability of Transformers to capture long-range dependencies and model global context enhances the accuracy and reliability of medical image segmentation tasks.

“The ability of Transformers to capture long-range dependencies and model global context enhances the accuracy and reliability of medical image segmentation tasks.”

May combine two claims.

Deep learning can play a crucial role in the advancement of beamforming algorithms by enhancing image quality, reducing noise, and improving the overall efficiency of medical ultrasound imaging.

“In summary, deep learning can play a crucial role in the advancement of beamforming algorithms by enhancing image quality, reducing noise, and improving the overall efficiency of medical ultrasound imaging.”

May combine two claims.

The model was tested on both private and public datasets, demonstrating state-of-the-art accuracy in AS detection and severity classification.

“The model was tested on both private and public datasets, demonstrating state-of-the-art accuracy in AS detection and severity classification.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Neural networks, especially CNNs, are highly efficient at automatically extracting relevant features from ultrasound data, which can be employed to improve the spatial and contrast resolution of images.

“Neural networks, especially CNNs, are highly efficient at automatically extracting relevant features from ultrasound data. These features can then be employed to improve the spatial and contrast resolution of the images.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Wang et al. present a method for enhancing the safety and efficiency of robot-assisted prostate biopsy through advanced force sensing techniques.

“Wang et al.[] presents a groundbreaking method for enhancing the safety and efficiency of robot-assisted prostate biopsy through advanced force sensing techniques.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Intelligent calculations of medical imaging based on deep learning can automatically learn excellent feature expressions from large sample data.

“Intelligent calculations of medical imaging based on deep learning can automatically learn excellent feature expressions from large sample data.”

May combine two claims.

The proposed Hybrid Transformer UNet (H-TUNet) integrates both intra-frame and inter-frame features through a combination of 2D and 3D Transformer UNets, significantly improving segmentation accuracy and efficiency.

“The proposed Hybrid Transformer UNet (H-TUNet) integrates both intra-frame and inter-frame features through a combination of 2D and 3D Transformer UNets, significantly improving segmentation accuracy and efficiency.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

For both adaptive beamforming and SLSC beamforming algorithms, the use of deep learning can reduce computational complexity and enhance efficiency.

“For both adaptive beamforming and SLSC beamforming algorithms, the use of deep learning can reduce computational complexity and enhance efficiency.”

Classified as probably not a substantial claim. May combine two claims.

Transformers can capture long-range dependencies and relationships across the entire image, which is essential for accurately interpreting ultrasound images that may contain complex structures and subtle differences.

“Transformers can capture long-range dependencies and relationships across the entire image, which is essential for accurately interpreting ultrasound images that may contain complex structures and subtle differences.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Ahmadi et al. integrate a spatio-temporal architecture that combines anatomical features and the motion of the aortic valve to accurately classify AS severity.

“Ahmadi et al.[] integrate a spatio-temporal architecture that combines anatomical features and the motion of the aortic valve to accurately classify AS severity.”

Classified as probably not a substantial claim.

The sentinel gate in AMAnet dynamically controls the attention level on visual features and language model memories.

“The sentinel gate is a pivotal element in the AMAnet framework, designed to dynamically control the attention level on visual features and language model memories.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The spatial attention mechanism of AMAnet focuses on relevant regions of the ultrasound images, ensuring the model captures essential visual details.

“The spatial attention mechanism focuses on the relevant regions of the ultrasound images, ensuring that the model captures essential visual details.”

May combine two claims.

Hybrid models such as GPA-TUNet combine the strengths of CNNs and Transformers, using CNN layers for initial feature extraction and Transformers for global context modeling, leading to superior performance in segmentation tasks, particularly for images with large axial spans.

“Hybrid models, such as GPA-TUNet[], combine the strengths of CNNs and Transformers, using CNN layers for initial feature extraction and Transformers for global context modeling. This integration leads to superior performance in segmentation tasks, particularly for images with large axial spans.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Deep learning technology can achieve precise segmentation of target structures in medical images, which helps to improve the speed and accuracy of medical image analysis.

“Deep learning technology can achieve precise segmentation of target structures in medical images, which helps to improve the speed and accuracy of medical image analysis.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Methods like TransVFS introduce factorized spatio-temporal processing strategies that significantly reduce computational complexity, making them suitable for real-time force-sensing applications in clinical settings.

“Methods like TransVFS[,] introduce factorized spatio-temporal processing strategies that significantly reduce computational complexity, making them suitable for real-time force-sensing applications in clinical settings.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Through the use of deep learning, more advanced beamforming algorithms, such as synthetic aperture and plane wave imaging, can be optimized for better resolution and frame rates.

“Through the use of deep learning, more advanced beamforming algorithms, such as synthetic aperture and plane wave imaging, can be optimized for better resolution and frame rates.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Deep learning models can increase the dynamic range of images, making it possible to display both high and low signal areas in the same image, and enhance resolution.

“Deep learning models can increase the dynamic range of images, making it possible to display both high and low signal areas in the same image, and enhance resolution.”

May combine two claims.

TransVFS includes efficient local-global attention modules that reduce the computational burden associated with processing 4D spatio-temporal data, making the method suitable for real-time force-sensing applications in clinical settings.

“The architecture includes efficient local–global attention modules that reduce the computational burden associated with processing 4D spatio-temporal data. This makes the method suitable for real-time force-sensing applications in clinical settings.”

May combine two claims.

The incorporation of semantic features and the adaptive attention mechanism in AMAnet contribute to the model's superior performance, highlighting its potential for practical clinical applications.

“The incorporation of semantic features and the adaptive attention mechanism contribute to the model’s superior performance, highlighting its potential for practical clinical applications.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Deep learning models can optimize beamforming algorithms based on different patient characteristics and scanning conditions to achieve more personalized imaging.

“Deep learning models can optimize beamforming algorithms based on different patient characteristics and scanning conditions to achieve more personalized imaging.”

Probably not a substantial claim: may not pass S4 (based).

Chi et al. propose a unified framework that combines 2D and 3D Transformer-UNets into a single end-to-end network.

“Chi et al. [] propose a unified framework that combines the 2D and 3D Transformer-UNets into a single end-to-end network.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Deep learning has significant potential in improving the performance of beamforming algorithms in medical ultrasound.

“Deep learning, as an advanced machine learning technique, has significant potential in improving the performance of beamforming algorithms in medical ultrasound.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Deep learning can identify and reduce artifacts in ultrasound imaging, such as sidelobe contamination and Doppler artifacts.

“It can also identify and reduce artifacts in ultrasound imaging, such as sidelobe contamination and Doppler artifacts.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Transformers address the limitations of CNNs by incorporating self-attention mechanisms that consider the entire input sequence or image simultaneously.

“Transformers address the limitations of CNNs by incorporating self-attention mechanisms that consider the entire input sequence (or image) simultaneously.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Deep learning can significantly reduce the computational time required for beamforming, making real-time imaging more feasible and efficient.

“Deep learning can significantly reduce the computational time required for beamforming, making real-time imaging more feasible and efficient.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Accelerating the beamforming computational process through deep learning models could significantly reduce the time required to acquire high-quality ultrasound images.

“Accelerating the beamforming computational process through deep learning models could significantly reduce the time required to acquire high-quality ultrasound images.”

May combine two claims.

On the public TMED-2 dataset, the model achieved 91.5% accuracy in AS detection and 83.8% in severity classification.

“On the public TMED-2 dataset, the model achieved 91.5% accuracy in AS detection and 83.8% in severity classification.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The integration of Transformer models and attention mechanisms into ultrasound image analysis represents a significant advancement over traditional CNN-based approaches.

“The integration of Transformer models and attention mechanisms into ultrasound image analysis represents a significant advancement over traditional CNN-based approaches.”

• AssumptionRests on an assumption: Performance improvements reported in the reviewed studies on specific datasets or experimental settings generalize to broader clinical ultrasound imaging populations and settings.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

By dynamically adjusting the attention weights, Transformers can focus on the most relevant parts of the image, enhancing feature extraction and reducing the impact of irrelevant or noisy regions.

“By dynamically adjusting the attention weights, Transformers can focus on the most relevant parts of the image, enhancing feature extraction and reducing the impact of irrelevant or noisy regions.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The semantic attention mechanism of AMAnet predicts crucial local properties, such as boundary conditions and tumor morphology, using a multi-label classification network.

“Meanwhile, the semantic attention mechanism predicts crucial local properties, such as boundary conditions and tumor morphology, by using a multi-label classification network.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

AI can help in organizing and managing vast amounts of ultrasound data, making it easier for healthcare professionals to access and retrieve patient information when needed.

“AI also can help in organizing and managing vast amounts of ultrasound data, making it easier for healthcare professionals to access and retrieve patient information when needed.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

With the continuous development of large language models (LLMs), AI technology based on these models can also be applied to medicine.

“With the continuous development of large language models (LLMs), AI technology based on these models can also be applied to medicine [–].”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

LLMs can be integrated with AI and machine learning algorithms aimed at identifying and diagnosing diseases from ultrasound imagery, and can provide detailed explanations or feedback based on the AI's findings in a way that is understandable for the healthcare provider.

“LLMs can also be integrated with AI and machine learning algorithms aimed at identifying and diagnosing diseases from ultrasound imagery. The LLM can then provide detailed explanations or feedback based on the AI’s findings in a way that is understandable for the healthcare provider.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

AI can generate preliminary reports based on the ultrasound data, highlighting key findings and suggesting possible diagnoses, which can expedite the review process by healthcare professionals.

“AI can generate preliminary reports based on the ultrasound data, highlighting key findings and suggesting possible diagnoses. This can expedite the review process by healthcare professionals.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

LLM-generated preliminary reports will not only save time but also reduce the workload of radiologists.

“The reports will not only save time but also reduce the workload of radiologists.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

The SAM demonstrates impressive zero-shot performance across various tasks, often matching or exceeding fully supervised methods.

“The SAM demonstrates impressive zero-shot performance across various tasks, often matching or exceeding fully supervised methods.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Deep learning can enhance image quality and diagnostic efficiency and promote the development of personalized medicine and precision medicine.

“It can not only enhance image quality and diagnostic efficiency but also promote the development of personalized medicine and precision medicine.”

• AssumptionRests on an assumption: Improvements in deep learning model performance for ultrasound image analysis will translate into improved clinical outcomes and patient care.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Deep learning can be used for rapid reconstruction of three-dimensional and four-dimensional data, providing clinicians with a more comprehensive view.

“Deep learning can be used for rapid reconstruction of three-dimensional and four-dimensional data, providing clinicians with a more comprehensive view.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

LLMs can assist in data collection, research conduction, and generating insights from large bodies of medical texts or research papers, offering valuable contributions to the field of medical ultrasound imaging.

“LLMs can assist in data collection, research conduction, and generating insights from large bodies of medical texts or research papers, offering valuable contributions to the field of medical ultrasound imaging.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The SAM's zero-shot performance indicates the model's generalizability and potential applicability to a wide range of segmentation challenges.

“This indicates the model’s generalizability and potential applicability to a wide range of segmentation challenges.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

LLMs can generate descriptive annotations for the images based on the features identified through image processing techniques.

“LLMs can generate descriptive annotations for the images based on the features identified through image processing techniques.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

The integration of AI in portable ultrasound devices with remote servers is helpful.

“The integration of AI in portable ultrasound devices with remote servers is also helpful.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Numerous studies have adopted the SAM in medical image segmentation.

“Numerous studies have adopted the SAM in the medical image segmentation [,].”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

Deep learning can be used to automatically determine the optimal beamforming parameters, simplifying clinical operations and reducing the workload of physicians.

“Deep learning can be used to automatically determine the optimal beamforming parameters, simplifying clinical operations and reducing the workload of physicians.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

By working through complex medical language and jargon, LLMs can translate these into more patient-friendly language, helping patients better understand their medical condition and the significance of their ultrasound results.

“By working through complex medical language and jargon, LLMs can translate these into more patient-friendly language. This helps patients better understand their medical condition and the significance of their ultrasound results.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

AI algorithms can analyze ultrasound images in real-time, helping to identify patterns, anomalies, or specific conditions, which can assist healthcare professionals in making more accurate and faster diagnoses.

“AI algorithms can analyze ultrasound images in real-time, helping to identify patterns, anomalies, or specific conditions. This can assist healthcare professionals in making more accurate and faster diagnoses.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

LLMs can be used to generate preliminary reports of ultrasound imaging by analyzing the textual descriptions provided by the sonographer or the data obtained from the ultrasound device.

“It can be used to generate preliminary reports of ultrasound imaging by analyzing the textual descriptions provided by the sonographer or the data obtained from the ultrasound device.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Based on reviewed studies, LLMs can play a significant role in medical ultrasound imaging.

“Based on these studies, it can be seen that LLMs can play a significant role in medical ultrasound imaging.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

By continuously learning from clinical data, deep learning models can improve their performance over time, enhancing the accuracy and reliability of beamforming technology.

“By continuously learning from clinical data, deep learning models can improve their performance over time, enhancing the accuracy and reliability of beamforming technology.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

AI can enable remote monitoring of patients, analyzing ultrasound data transmitted to the remote server and alerting healthcare professionals to any concerning changes or findings that require immediate attention.

“AI can enable remote monitoring of patients, analyzing ultrasound data transmitted to the remote server and alerting healthcare professionals to any concerning changes or findings that require immediate attention.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Beamforming algorithms integrated with deep learning can assist in real-time detection of lesions and measurement of biomarkers, providing more diagnostic information.

“Beamforming algorithms integrated with deep learning can assist in real-time detection of lesions and measurement of biomarkers, providing more diagnostic information.”