In recent years, the application of deep learning (DL) technology in the thyroid field has shown exponential growth, greatly promoting innovation in thyroid disease research. As the most common malignant tumor of the endocrine system, the precise diagnosis and treatment of thyroid cancer has been a key focus of clinical research. This article systematically reviews the latest research progress in DL research for the diagnosis and treatment of thyroid malignancies, focusing on the breakthrough application of advanced models such as convolutional neural networks (CNNs), long short-term memory networks (LSTMs), and generative adversarial networks (GANs) in key areas such as ultrasound images analysis for thyroid nodules, automatic classification of pathological images, and assessment of extrathyroidal extension. Furthermore, the review highlights the great potential of DL techniques in the development of individualized treatment planning and prognosis prediction. In addition, it analyzes the technical bottlenecks and clinical challenges faced by current DL applications in thyroid cancer diagnosis and treatment and looks ahead to future directions for development. The aim of this review is to provide the latest research insights for clinical practitioners, promote further improvements in the precision diagnosis and treatment system for thyroid cancer, and ultimately achieve better diagnostic and therapeutic outcomes for thyroid cancer patients.

Most artificial intelligence (AI) models for thyroid nodules are designed to screen for malignancy to guide further interventions; however, these models have not yet been fully implemented in clinical practice. This study aimed to evaluate AI in real clinical settings for identifying potentially benign thyroid nodules initially deemed to be at risk for malignancy by radiologists, reducing unnecessary fine needle aspiration (FNA) and optimizing management. We retrospectively collected a validation cohort of thyroid nodules that had undergone FNA. These nodules were initially assessed as "suspicious for malignancy" by radiologists based on ultrasound features, following standard clinical practice, which prompted further FNA procedures. Ultrasound images of these nodules were re-evaluated using a deep learning-based AI system, and its diagnostic performance was assessed in terms of correct identification of benign nodules and error identification of malignant nodules. Performance metrics such as sensitivity, specificity, and the area under the receiver operating characteristic curve were calculated. In addition, a separate comparison cohort was retrospectively assembled to compare the AI system's ability to correctly identify benign thyroid nodules with that of radiologists. The validation cohort comprised 4572 thyroid nodules (benign: n=3134, 68.5%; malignant: n=1438, 31.5%). AI correctly identified 2719 (86.8% among benign nodules) and reduced unnecessary FNAs from 68.5% (3134/4572) to 9.1% (415/4572). However, 123 malignant nodules (8.6% of malignant cases) were mistakenly identified as benign, with the majority of these being of low or intermediate suspicion. In the comparison cohort, AI successfully identified 81.4% (96/118) of benign nodules. It outperformed junior and senior radiologists, who identified only 40% and 55%, respectively. The area under the curve (AUC) for the AI model was 0.88 (95% CI 0.85-0.91), demonstrating a superior AUC compared with that of the junior radiologists (AUC=0.43, 95% CI 0.36-0.50; P=.002) and senior radiologists (AUC=0.63, 95% CI 0.55-0.70; P=.003). Compared with radiologists, AI can better serve as a "goalkeeper" in reducing unnecessary FNAs by identifying benign nodules that are initially assessed as malignant by radiologists. However, active surveillance is still necessary for all these nodules since a very small number of low-aggressiveness malignant nodules may be mistakenly identified.

The purpose of this study was to create and validate an ultrasound-based graph convolutional network (US-based GCN) model for the prediction of axillary lymph node metastasis (ALNM) in patients with breast cancer. A total of 820 eligible patients with breast cancer who underwent preoperative breast ultrasonography (US) between April 2016 and June 2022 were retrospectively enrolled. The training cohort consisted of 621 patients, whereas validation cohort 1 included 112 patients, and validation cohort 2 included 87 patients. A US-based GCN model was built using US deep learning features. In validation cohort 1, the US-based GCN model performed satisfactorily, with an AUC of 0.88 and an accuracy of 0.76. In validation cohort 2, the US-based GCN model performed satisfactorily, with an AUC of 0.84 and an accuracy of 0.75. This approach has the potential to help guide optimal ALNM management in breast cancer patients, particularly by preventing overtreatment. In conclusion, we developed a US-based GCN model to assess the ALN status of breast cancer patients prior to surgery. The US-based GCN model can provide a possible noninvasive method for detecting ALNM and aid in clinical decision-making. High-level evidence for clinical use in later studies is anticipated to be obtained through prospective studies.

Automated cardiac interpretation in resource-constrained settings (RCS) is often hindered by poor-quality echocardiographic imaging, limiting the effectiveness of downstream diagnostic models. While super-resolution (SR) techniques have shown promise in enhancing magnetic resonance imaging (MRI) and computed tomography (CT) scans, their application to echocardiography-a widely accessible but noise-prone modality-remains underexplored. In this work, we investigate the potential of deep learning-based SR to improve classification accuracy on low-quality 2D echocardiograms. Using the publicly available CAMUS dataset, we stratify samples by image quality and evaluate two clinically relevant tasks of varying complexity: a relatively simple Two-Chamber vs. Four-Chamber (2CH vs. 4CH) view classification and a more complex End-Diastole vs. End-Systole (ED vs. ES) phase classification. We apply two widely used SR models-Super-Resolution Generative Adversarial Network (SRGAN) and Super-Resolution Residual Network (SRResNet), to enhance poor-quality images and observe significant gains in performance metric-particularly with SRResNet, which also offers computational efficiency. Our findings demonstrate that SR can effectively recover diagnostic value in degraded echo scans, making it a viable tool for AI-assisted care in RCS, achieving more with less.

Accurate fetal biometric measurements, such as abdominal circumference, play a vital role in prenatal care. However, obtaining high-quality ultrasound images for these measurements heavily depends on the expertise of sonographers, posing a significant challenge in low-income countries due to the scarcity of trained personnel. To address this issue, we leverage FetalCLIP, a vision-language model pretrained on a curated dataset of over 210,000 fetal ultrasound image-caption pairs, to perform automated fetal ultrasound image quality assessment (IQA) on blind-sweep ultrasound data. We introduce FetalCLIP$_{CLS}$, an IQA model adapted from FetalCLIP using Low-Rank Adaptation (LoRA), and evaluate it on the ACOUSLIC-AI dataset against six CNN and Transformer baselines. FetalCLIP$_{CLS}$ achieves the highest F1 score of 0.757. Moreover, we show that an adapted segmentation model, when repurposed for classification, further improves performance, achieving an F1 score of 0.771. Our work demonstrates how parameter-efficient fine-tuning of fetal ultrasound foundation models can enable task-specific adaptations, advancing prenatal care in resource-limited settings. The experimental code is available at: https://github.com/donglihe-hub/FetalCLIP-IQA.

Glioblastoma multiforme (GBM) is the most aggressive type of brain cancer, making effective treatments essential to improve patient survival. To advance the understanding of GBM and develop more effective therapies, preclinical studies commonly use mouse models due to their genetic and physiological similarities to humans. In particular, the GL261 mouse glioma model is employed for its reproducible tumor growth and ability to mimic key aspects of human gliomas. Ultrasound imaging is a valuable modality in preclinical studies, offering real-time, non-invasive tumor monitoring and facilitating treatment response assessment. Furthermore, its potential therapeutic applications, such as in tumor ablation, expand its utility in preclinical studies. However, real-time segmentation of GL261 tumors during surgery introduces significant complexities, such as precise tumor boundary delineation and maintaining processing efficiency. Automated segmentation offers a solution, but its success relies on high-quality datasets with precise labeling. Our study introduces the first publicly available ultrasound dataset specifically developed to improve tumor segmentation in GL261 glioblastomas, providing 1,856 annotated images to support AI model development in preclinical research. This dataset bridges preclinical insights and clinical practice, laying the foundation for developing more accurate and effective tumor resection techniques.

Ultrasound is widely used in clinical care, yet standard deep learning methods often struggle with full video analysis due to non-standardized acquisition and operator bias. We offer a new perspective on ultrasound video analysis through implicit neural representations (INRs). We build on Functa, an INR framework in which each image is represented by a modulation vector that conditions a shared neural network. However, its extension to the temporal domain of medical videos remains unexplored. To address this gap, we propose VidFuncta, a novel framework that leverages Functa to encode variable-length ultrasound videos into compact, time-resolved representations. VidFuncta disentangles each video into a static video-specific vector and a sequence of time-dependent modulation vectors, capturing both temporal dynamics and dataset-level redundancies. Our method outperforms 2D and 3D baselines on video reconstruction and enables downstream tasks to directly operate on the learned 1D modulation vectors. We validate VidFuncta on three public ultrasound video datasets -- cardiac, lung, and breast -- and evaluate its downstream performance on ejection fraction prediction, B-line detection, and breast lesion classification. These results highlight the potential of VidFuncta as a generalizable and efficient representation framework for ultrasound videos. Our code is publicly available under https://github.com/JuliaWolleb/VidFuncta_public.

Diagnosis of prostate cancer requires histopathology of tissue samples. Following an MRI to identify suspicious areas, a biopsy is performed under ultrasound (US) guidance. In existing assistance systems, 3D US information is generally available (taken before the biopsy session and/or in between samplings). However, without registration between 2D images and 3D volumes, the urologist must rely on cognitive navigation. This work introduces a deep learning model to track the orientation of real-time US slices relative to a reference 3D US volume using only image and volume data. The dataset comprises 515 3D US volumes collected from 51 patients during routine transperineal biopsy. To generate 2D images streams, volumes are resampled to simulate three degrees of freedom rotational movements around the rectal entrance. The proposed model comprises two ResNet-based sub-modules to address the symmetry ambiguity arising from complex out-of-plane movement of the probe. The first sub-module predicts the unsigned relative orientation between consecutive slices, while the second leverages a custom similarity model and a spatial context volume to determine the sign of this relative orientation. From the sub-modules predictions, slices orientations along the navigated trajectory can then be derived in real-time. Results demonstrate that registration error remains below 2.5 mm in 92% of cases over a 5-second trajectory, and 80% over a 25-second trajectory. These findings show that accurate, sensorless 2D/3D US registration given a spatial context is achievable with limited drift over extended navigation. This highlights the potential of AI-driven biopsy assistance to increase the accuracy of freehand biopsy.

The rising global prevalence of gout necessitates advancements in diagnostic methodologies. Ultrasonographic imaging of the foot has become an important diagnostic modality for gout because of its non-invasiveness, cost-effectiveness, and real-time imaging capabilities. This study aims to develop and validate a deep learning-based artificial intelligence (AI) model for automated gout diagnosis using ultrasound images. In this study, ultrasound images were primarily acquired at the first metatarsophalangeal joint (MTP1) from 598 cases in two institutions: 520 from Institution 1 and 78 from Institution 2. From Institution 1's dataset, 66% of cases were randomly allocated for model training, while the remaining 34% constitute the internal test set. The dataset from Institution 2 served as an independent external validation cohort. A novel deep learning model integrating a patch-wise attention mechanism and multi-scale feature extraction was developed to enhance the detection of subtle sonographic features and optimize diagnostic performance. The proposed model demonstrated robust diagnostic efficacy, achieving an accuracy of 87.88%, a sensitivity of 87.85%, a specificity of 87.93%, and an area under the curve (AUC) of 93.43%. Additionally, the model generates interpretable visual heatmaps to localize gout-related pathological features, thereby facilitating interpretation for clinical decision-making. In this paper, a deep learning-based artificial intelligence (AI) model was developed for the automated detection of gout using ultrasound images, which achieved better performance than other models. Furthermore, the features highlighted by the model align closely with expert assessments, demonstrating its potential to assist in the ultrasound-based diagnosis of gout.

Periventricular-intraventricular haemorrhage (IVH) is the most prevalent type of neonatal intracranial haemorrhage. It is especially threatening to preterm infants, in whom it is associated with significant morbidity and mortality. Cranial ultrasound has become an important means of screening periventricular IVH in infants. The integration of artificial intelligence with neonatal ultrasound is promising for enhancing diagnostic accuracy, reducing physician workload, and consequently improving periventricular IVH outcomes. The study investigated whether deep learning-based analysis of the cranial ultrasound images of infants could detect and grade periventricular IVH. This multicentre observational study included 1,060 cases and healthy controls from two hospitals. The retrospective modelling dataset encompassed 773 participants from January 2020 to July 2023, while the prospective two-centre validation dataset included 287 participants from August 2023 to January 2024. The periventricular IVH net model, a deep learning model incorporating the convolutional block attention module mechanism, was developed. The model's effectiveness was assessed by randomly dividing the retrospective data into training and validation sets, followed by independent validation with the prospective two-centre data. To evaluate the model, we measured its recall, precision, accuracy, F1-score, and area under the curve (AUC). The regions of interest (ROI) that influenced the detection by the deep learning model were visualised in significance maps, and the t-distributed stochastic neighbour embedding (t-SNE) algorithm was used to visualise the clustering of model detection parameters. The final retrospective dataset included 773 participants (mean (standard deviation (SD)) gestational age, 32.7 (4.69) weeks; mean (SD) weight, 1,862.60 (855.49) g). For the retrospective data, the model's AUC was 0.99 (95% confidence interval (CI), 0.98-0.99), precision was 0.92 (0.89-0.95), recall was 0.93 (0.89-0.95), and F1-score was 0.93 (0.90-0.95). For the prospective two-centre validation data, the model's AUC was 0.961 (95% CI, 0.94-0.98) and accuracy was 0.89 (95% CI, 0.86-0.92). The two-centre prospective validation results of the periventricular IVH net model demonstrated its tremendous potential for paediatric clinical applications. Combining artificial intelligence with paediatric ultrasound can enhance the accuracy and efficiency of periventricular IVH diagnosis, especially in primary hospitals or community hospitals.