Automated segmentation of brain metastases with deep learning: a multi-center, randomized crossover, multi-reader evaluation study.

BACKGROUND: Artificial intelligence has been proposed for brain metastasis (BM) segmentation but it has not been fully clinically validated. The aim of this study was to develop and evaluate a system for BM segmentation. METHODS: A deep-learning-based BM segmentation system (BMSS) was developed using contrast-enhanced MR images from 488 patients with 10,338 brain metastases. A randomized crossover, multi-reader study was then conducted to evaluate the performance of the BMSS for BM segmentation using data prospectively collected from 50 patients with 203 metastases at five centers. Five radiology residents and five attending radiologists were randomly assigned to contour the same prospective set in assisted and unassisted modes. Aided and unaided Dice similarity coefficients (DSCs) and contouring times per lesion were compared. RESULTS: The BMSS alone yielded a median DSC of 0.91 (95% confidence interval, 0.90-0.92) in the multi-center set and showed comparable performance between the internal and external sets (p = 0.67). With BMSS assistance, the readers increased the median DSC from 0.87 (0.87-0.88) to 0.92 (0.92-0.92) (p < 0.001) with a median time saving of 42% (40-45%) per lesion. Resident readers showed a greater improvement than attending readers in contouring accuracy (improved median DSC, 0.05 [0.05-0.05] vs. 0.03 [0.03-0.03]; p < 0.001), but a similar time reduction (reduced median time, 44% [40-47%] vs. 40% [37-44%]; p = 0.92) with BMSS assistance. CONCLUSIONS: The BMSS can be optimally applied to improve the efficiency of brain metastasis delineation in clinical practice.

Carotid Vessel Wall Segmentation Through Domain Aligner, Topological Learning, and Segment Anything Model for Sparse Annotation in MR Images.

Medical image analysis poses significant challenges due to limited availability of clinical data, which is crucial for training accurate models. This limitation is further compounded by the specialized and labor-intensive nature of the data annotation process. For example, despite the popularity of computed tomography angiography (CTA) in diagnosing atherosclerosis with an abundance of annotated datasets, magnetic resonance (MR) images stand out with better visualization for soft plaque and vessel wall characterization. However, the higher cost and limited accessibility of MR, as well as time-consuming nature of manual labeling, contribute to fewer annotated datasets. To address these issues, we formulate a multi-modal transfer learning network, named MT-Net, designed to learn from unpaired CTA and sparsely-annotated MR data. Additionally, we harness the Segment Anything Model (SAM) to synthesize additional MR annotations, enriching the training process. Specifically, our method first segments vessel lumen regions followed by precise characterization of carotid artery vessel walls, thereby ensuring both segmentation accuracy and clinical relevance. Validation of our method involved rigorous experimentation on publicly available datasets from COSMOS and CARE-II challenge, demonstrating its superior performance compared to existing state-of-the-art techniques.

FEFA: Frequency Enhanced Multi-Modal MRI Reconstruction with Deep Feature Alignment.

Integrating complementary information from multiple magnetic resonance imaging (MRI) modalities is often necessary to make accurate and reliable diagnostic decisions. However, the different acquisition speeds of these modalities mean that obtaining information can be time consuming and require significant effort. Reference-based MRI reconstruction aims to accelerate slower, under-sampled imaging modalities, such as T2-modality, by utilizing redundant information from faster, fully sampled modalities, such as T1-modality. Unfortunately, spatial misalignment between different modalities often negatively impacts the final results. To address this issue, we propose FEFA, which consists of cascading FEFA blocks. The FEFA block first aligns and fuses the two modalities at the feature level. The combined features are then filtered in the frequency domain to enhance the important features while simultaneously suppressing the less essential ones, thereby ensuring accurate reconstruction. Furthermore, we emphasize the advantages of combining the reconstruction results from multiple cascaded blocks, which also contributes to stabilizing the training process. Compared to existing registration-then-reconstruction and cross-attention-based approaches, our method is end-to-end trainable without requiring additional supervision, extensive parameters, or heavy computation. Experiments on the public fastMRI, IXI and in-house datasets demonstrate that our approach is effective across various under-sampling patterns and ratios. Our code will be available at: https://github.com/chenxm12394/FEFA.

Evaluating Large Language Models for Automated Reporting and Data Systems Categorization: Cross-Sectional Study.

BACKGROUND: Large language models show promise for improving radiology workflows, but their performance on structured radiological tasks such as Reporting and Data Systems (RADS) categorization remains unexplored. OBJECTIVE: This study aims to evaluate 3 large language model chatbots-Claude-2, GPT-3.5, and GPT-4-on assigning RADS categories to radiology reports and assess the impact of different prompting strategies. METHODS: This cross-sectional study compared 3 chatbots using 30 radiology reports (10 per RADS criteria), using a 3-level prompting strategy: zero-shot, few-shot, and guideline PDF-informed prompts. The cases were grounded in Liver Imaging Reporting & Data System (LI-RADS) version 2018, Lung CT (computed tomography) Screening Reporting & Data System (Lung-RADS) version 2022, and Ovarian-Adnexal Reporting & Data System (O-RADS) magnetic resonance imaging, meticulously prepared by board-certified radiologists. Each report underwent 6 assessments. Two blinded reviewers assessed the chatbots' response at patient-level RADS categorization and overall ratings. The agreement across repetitions was assessed using Fleiss κ. RESULTS: Claude-2 achieved the highest accuracy in overall ratings with few-shot prompts and guideline PDFs (prompt-2), attaining 57% (17/30) average accuracy over 6 runs and 50% (15/30) accuracy with k-pass voting. Without prompt engineering, all chatbots performed poorly. The introduction of a structured exemplar prompt (prompt-1) increased the accuracy of overall ratings for all chatbots. Providing prompt-2 further improved Claude-2's performance, an enhancement not replicated by GPT-4. The interrun agreement was substantial for Claude-2 (k=0.66 for overall rating and k=0.69 for RADS categorization), fair for GPT-4 (k=0.39 for both), and fair for GPT-3.5 (k=0.21 for overall rating and k=0.39 for RADS categorization). All chatbots showed significantly higher accuracy with LI-RADS version 2018 than with Lung-RADS version 2022 and O-RADS (P<.05); with prompt-2, Claude-2 achieved the highest overall rating accuracy of 75% (45/60) in LI-RADS version 2018. CONCLUSIONS: When equipped with structured prompts and guideline PDFs, Claude-2 demonstrated potential in assigning RADS categories to radiology cases according to established criteria such as LI-RADS version 2018. However, the current generation of chatbots lags in accurately categorizing cases based on more recent RADS criteria.

Transferring Adult-like Phase Images for Robust Multi-view Isointense Infant Brain Segmentation.

Accurate tissue segmentation of infant brain in magnetic resonance (MR) images is crucial for charting early brain development and identifying biomarkers. Due to ongoing myelination and maturation, in the isointense phase (6-9 months of age), the gray and white matters of infant brain exhibit similar intensity levels in MR images, posing significant challenges for tissue segmentation. Meanwhile, in the adult-like phase around 12 months of age, the MR images show high tissue contrast and can be easily segmented. In this paper, we propose to effectively exploit adult-like phase images to achieve robustmulti-view isointense infant brain segmentation. Specifically, in one way, we transfer adult-like phase images to the isointense view, which have similar tissue contrast as the isointense phase images, and use the transferred images to train an isointense-view segmentation network. On the other way, we transfer isointense phase images to the adult-like view, which have enhanced tissue contrast, for training a segmentation network in the adult-like view. The segmentation networks of different views form a multi-path architecture that performs multi-view learning to further boost the segmentation performance. Since anatomy-preserving style transfer is key to the downstream segmentation task, we develop a Disentangled Cycle-consistent Adversarial Network (DCAN) with strong regularization terms to accurately transfer realistic tissue contrast between isointense and adult-like phase images while still maintaining their structural consistency. Experiments on both NDAR and iSeg-2019 datasets demonstrate a significant superior performance of our method over the state-of-the-art methods.

Prior Knowledge-guided Triple-Domain Transformer-GAN for Direct PET Reconstruction from Low-Count Sinograms.

To obtain high-quality positron emission tomography (PET) images while minimizing radiation exposure, numerous methods have been dedicated to acquiring standard-count PET (SPET) from low-count PET (LPET). However, current methods have failed to take full advantage of the different emphasized information from multiple domains, i.e., the sinogram, image, and frequency domains, resulting in the loss of crucial details. Meanwhile, they overlook the unique inner-structure of the sinograms, thereby failing to fully capture its structural characteristics and relationships. To alleviate these problems, in this paper, we proposed a prior knowledge-guided transformer-GAN that unites triple domains of sinogram, image, and frequency to directly reconstruct SPET images from LPET sinograms, namely PK-TriDo. Our PK-TriDo consists of a Sinogram Inner-Structure-based Denoising Transformer (SISD-Former) to denoise the input LPET sinogram, a Frequency-adapted Image Reconstruction Transformer (FaIR-Former) to reconstruct high-quality SPET images from the denoised sinograms guided by the image domain prior knowledge, and an Adversarial Network (AdvNet) to further enhance the reconstruction quality via adversarial training. Specifically tailored for the PET imaging mechanism, we injected a sinogram embedding module that partitions the sinograms by rows and columns to obtain 1D sequences of angles and distances to faithfully preserve the inner-structure of the sinograms. Moreover, to mitigate high-frequency distortions and enhance reconstruction details, we integrated global-local frequency parsers (GLFPs) into FaIR-Former to calibrate the distributions and proportions of different frequency bands, thus compelling the network to preserve high-frequency details. Evaluations on three datasets with different dose levels and imaging scenarios demonstrated that our PK-TriDo outperforms the state-of-the-art methods.

Detail-preserving image warping by enforcing smooth image sampling.

Multi-phase dynamic contrast-enhanced magnetic resonance imaging image registration makes a substantial contribution to medical image analysis. However, existing methods (e.g., VoxelMorph, CycleMorph) often encounter the problem of image information misalignment in deformable registration tasks, posing challenges to the practical application. To address this issue, we propose a novel smooth image sampling method to align full organic information to realize detail-preserving image warping. In this paper, we clarify that the phenomenon about image information mismatch is attributed to imbalanced sampling. Then, a sampling frequency map constructed by sampling frequency estimators is utilized to instruct smooth sampling by reducing the spatial gradient and discrepancy between all-ones matrix and sampling frequency map. In addition, our estimator determines the sampling frequency of a grid voxel in the moving image by aggregating the sum of interpolation weights from warped non-grid sampling points in its vicinity and vectorially constructs sampling frequency map through projection and scatteration. We evaluate the effectiveness of our approach through experiments on two in-house datasets. The results showcase that our method preserves nearly complete details with ideal registration accuracy compared with several state-of-the-art registration methods. Additionally, our method exhibits a statistically significant difference in the regularity of the registration field compared to other methods, at a significance level of p < 0.05. Our code will be released at https://github.com/QingRui-Sha/SFM.

Spatial and Modal Optimal Transport for Fast Cross-Modal MRI Reconstruction.

Multi-modal magnetic resonance imaging (MRI) plays a crucial role in comprehensive disease diagnosis in clinical medicine. However, acquiring certain modalities, such as T2-weighted images (T2WIs), is time-consuming and prone to be with motion artifacts. It negatively impacts subsequent multi-modal image analysis. To address this issue, we propose an end-to-end deep learning framework that utilizes T1-weighted images (T1WIs) as auxiliary modalities to expedite T2WIs' acquisitions. While image pre-processing is capable of mitigating misalignment, improper parameter selection leads to adverse pre-processing effects, requiring iterative experimentation and adjustment. To overcome this shortage, we employ Optimal Transport (OT) to synthesize T2WIs by aligning T1WIs and performing cross-modal synthesis, effectively mitigating spatial misalignment effects. Furthermore, we adopt an alternating iteration framework between the reconstruction task and the cross-modal synthesis task to optimize the final results. Then, we prove that the reconstructed T2WIs and the synthetic T2WIs become closer on the T2 image manifold with iterations increasing, and further illustrate that the improved reconstruction result enhances the synthesis process, whereas the enhanced synthesis result improves the reconstruction process. Finally, experimental results from FastMRI and internal datasets confirm the effectiveness of our method, demonstrating significant improvements in image reconstruction quality even at low sampling rates.

Real-time diagnosis of intracerebral hemorrhage by generating dual-energy CT from single-energy CT.

Real-time diagnosis of intracerebral hemorrhage after thrombectomy is crucial for follow-up treatment. However, this is difficult to achieve with standard single-energy CT (SECT) due to similar CT values of blood and contrast agents under a single energy spectrum. In contrast, dual-energy CT (DECT) scanners employ two different energy spectra, which allows for real-time differentiation between hemorrhage and contrast extravasation based on energy-related attenuation characteristics. Unfortunately, DECT scanners are not as widely used as SECT scanners due to their high costs. To address this dilemma, in this paper, we generate pseudo DECT images from a SECT image for real-time diagnosis of hemorrhage. More specifically, we propose a SECT-to-DECT Transformer-based Generative Adversarial Network (SDTGAN), which is a 3D transformer-based multi-task learning framework equipped with a shared attention mechanism. In this way, SDTGAN can be guided to focus more on high-density areas (crucial for hemorrhage diagnosis) during the generation. Meanwhile, the introduced multi-task learning strategy and the shared attention mechanism also enable SDTGAN to model dependencies between interconnected generation tasks, improving generation performance while significantly reducing model parameters and computational complexity. In the experiments, we approximate real SECT images using mixed 120kV images from DECT data to address the issue of not being able to obtain the true paired DECT and SECT data. Extensive experiments demonstrate that SDTGAN can generate DECT images better than state-of-the-art methods. The code of our implementation is available at https://github.com/jiang-cw/SDTGAN.

A domain knowledge-based interpretable deep learning system for improving clinical breast ultrasound diagnosis.

BACKGROUND: Though deep learning has consistently demonstrated advantages in the automatic interpretation of breast ultrasound images, its black-box nature hinders potential interactions with radiologists, posing obstacles for clinical deployment. METHODS: We proposed a domain knowledge-based interpretable deep learning system for improving breast cancer risk prediction via paired multimodal ultrasound images. The deep learning system was developed on 4320 multimodal breast ultrasound images of 1440 biopsy-confirmed lesions from 1348 prospectively enrolled patients across two hospitals between August 2019 and December 2022. The lesions were allocated to 70% training cohort, 10% validation cohort, and 20% test cohort based on case recruitment date. RESULTS: Here, we show that the interpretable deep learning system can predict breast cancer risk as accurately as experienced radiologists, with an area under the receiver operating characteristic curve of 0.902 (95% confidence interval = 0.882 - 0.921), sensitivity of 75.2%, and specificity of 91.8% on the test cohort. With the aid of the deep learning system, particularly its inherent explainable features, junior radiologists tend to achieve better clinical outcomes, while senior radiologists experience increased confidence levels. Multimodal ultrasound images augmented with domain knowledge-based reasoning cues enable an effective human-machine collaboration at a high level of prediction performance. CONCLUSIONS: Such a clinically applicable deep learning system may be incorporated into future breast cancer screening and support assisted or second-read workflows.

Breast cancer is one of the most common cancers, and finding it early can greatly improve patients' chances of survival and recovery. We create a tool based on artificial intelligence (AI)­whereby computer software learns to perform tasks that normally require human thinking­called MUP-Net. MUP-Net can analyze medical images to predict a patient's risk of having breast cancer. To make this AI tool usable in clinical practice, we enabled doctors to see the reasoning behind the AI's predictions by visualizing the key image features it analyzed. We showed that our AI tool not only makes doctors more confident in their diagnosis but also helps them make better decisions, especially for less experienced doctors. With further testing, our AI tool may help clinicians to diagnose breast cancer more accurately and quickly, potentially improving patient outcomes.

Attention-Based MultiOffset Deep Learning Reconstruction of Chemical Exchange Saturation Transfer (AMO-CEST) MRI.

One challenge of chemical exchange saturation transfer (CEST) magnetic resonance imaging (MRI) is the long scan time due to multiple acquisitions of images at different saturation frequency offsets. k-space under-sampling strategy is commonly used to accelerate MRI acquisition, while this could introduce artifacts and reduce signal-to-noise ratio (SNR). To accelerate CEST-MRI acquisition while maintaining suitable image quality, we proposed an attention-based multioffset deep learning reconstruction network (AMO-CEST) with a multiple radial k-space sampling strategy for CEST-MRI. The AMO-CEST also contains dilated convolution to enlarge the receptive field and data consistency module to preserve the sampled k-space data. We evaluated the proposed method on a mouse brain dataset containing 5760 CEST images acquired at a pre-clinical 3T MRI scanner. Quantitative results demonstrated that AMO-CEST showed obvious improvement over zero-filling method with a PSNR enhancement of 11 dB, a SSIM enhancement of 0.15, and a NMSE decrease of 4.37×10-2 in three acquisition orientations. Compared with other deep learning-based models, AMO-CEST showed visual and quantitative improvements in images from three different orientations. We also extracted molecular contrast maps, including the amide proton transfer (APT) and the relayed nuclear Overhauser enhancement (rNOE). The results demonstrated that the CEST contrast maps derived from the CEST images of AMO-CEST were comparable to those derived from the original high-resolution CEST images. The proposed AMO-CEST can efficiently reconstruct high-quality CEST images from under-sampled k-space data and thus has the potential to accelerate CEST-MRI acquisition.

ChatCAD+: Towards a Universal and Reliable Interactive CAD using LLMs.

The integration of Computer-Aided Diagnosis (CAD) with Large Language Models (LLMs) presents a promising frontier in clinical applications, notably in automating diagnostic processes akin to those performed by radiologists and providing consultations similar to a virtual family doctor. Despite the promising potential of this integration, current works face at least two limitations: (1) From the perspective of a radiologist, existing studies typically have a restricted scope of applicable imaging domains, failing to meet the diagnostic needs of different patients. Also, the insufficient diagnostic capability of LLMs further undermine the quality and reliability of the generated medical reports. (2) Current LLMs lack the requisite depth in medical expertise, rendering them less effective as virtual family doctors due to the potential unreliability of the advice provided during patient consultations. To address these limitations, we introduce ChatCAD+, to be universal and reliable. Specifically, it is featured by two main modules: (1) Reliable Report Generation and (2) Reliable Interaction. The Reliable Report Generation module is capable of interpreting medical images from diverse domains and generate high-quality medical reports via our proposed hierarchical in-context learning. Concurrently, the interaction module leverages up-to-date information from reputable medical websites to provide reliable medical advice. Together, these designed modules synergize to closely align with the expertise of human medical professionals, offering enhanced consistency and reliability for interpretation and advice. The source code is available at GitHub.

Adaptive and Iterative Learning with Multi-perspective Regularizations for Metal Artifact Reduction.

Metal artifact reduction (MAR) is important for clinical diagnosis with CT images. The existing state-of-the-art deep learning methods usually suppress metal artifacts in sinogram or image domains or both. However, their performance is limited by the inherent characteristics of the two domains, i.e., the errors introduced by local manipulations in the sinogram domain would propagate throughout the whole image during backprojection and lead to serious secondary artifacts, while it is difficult to distinguish artifacts from actual image features in the image domain. To alleviate these limitations, this study analyzes the desirable properties of wavelet transform in-depth and proposes to perform MAR in the wavelet domain. First, wavelet transform yields components that possess spatial correspondence with the image, thereby preventing the spread of local errors to avoid secondary artifacts. Second, using wavelet transform could facilitate identification of artifacts from image since metal artifacts are mainly high-frequency signals. Taking these advantages of the wavelet transform, this paper decomposes an image into multiple wavelet components and introduces multi-perspective regularizations into the proposed MAR model. To improve the transparency and validity of the model, all the modules in the proposed MAR model are designed to reflect their mathematical meanings. In addition, an adaptive wavelet module is also utilized to enhance the flexibility of the model. To optimize the model, an iterative algorithm is developed. The evaluation on both synthetic and real clinical datasets consistently confirms the superior performance of the proposed method over the competing methods.

Unified Model for Children's Brain Image Segmentation with Co-Registration Framework Guided by Longitudinal MRI.

Accurate segmentation of brain structures is crucial for analyzing longitudinal changes in children's brains. However, existing methods are mostly based on models established at a single time-point due to difficulty in obtaining annotated data and dynamic variation of tissue intensity. The main problem with such approaches is that, when conducting longitudinal analysis, images from different time points are segmented by different models, leading to significant variation in estimating development trends. In this paper, we propose a novel unified model with co-registration framework to segment children's brain images covering neonates to preschoolers, which is formulated as two stages. First, to overcome the shortage of annotated data, we propose building gold-standard segmentation with co-registration framework guided by longitudinal data. Second, we construct a unified segmentation model tailored to brain images at 0-6 years old through the introduction of a convolutional network (named SE-VB-Net), which combines our previously proposed VB-Net with Squeeze-and-Excitation (SE) block. Moreover, different from existing methods that only require both T1- and T2-weighted MR images as inputs, our designed model also allows a single T1-weighted MR image as input. The proposed method is evaluated on the main dataset (320 longitudinal subjects with average 2 time-points) and two external datasets (10 cases with 6-month-old and 40 cases with 20-45 weeks, respectively). Results demonstrate that our proposed method achieves a high performance (>92%), even over a single time-point. This means that it is suitable for brain image analysis with large appearance variation, and largely broadens the application scenarios.

Constructing hierarchical attentive functional brain networks for early AD diagnosis.

Analyzing functional brain networks (FBN) with deep learning has demonstrated great potential for brain disorder diagnosis. The conventional construction of FBN is typically conducted at a single scale with a predefined brain region atlas. However, numerous studies have identified that the structure and function of the brain are hierarchically organized in nature. This urges the need of representing FBN in a hierarchical manner for more effective analysis of the complementary diagnostic insights at different scales. To this end, this paper proposes to build hierarchical FBNs adaptively within the Transformer framework. Specifically, a sparse attention-based node-merging module is designed to work alongside the conventional network feature extraction modules in each layer. The proposed module generates coarser nodes for further FBN construction and analysis by combining fine-grained nodes. By stacking multiple such layers, a hierarchical representation of FBN can be adaptively learned in an end-to-end manner. The hierarchical structure can not only integrate the complementary information from multiscale FBN for joint analysis, but also reduce the model complexity due to decreasing node sizes. Moreover, this paper argues that the nodes defined by the existing atlases are not necessarily the optimal starting level to build FBN hierarchy and exploring finer nodes may further enrich the FBN representation. In this regard, each predefined node in an atlas is split into multiple sub-nodes, overcoming the scale limitation of the existing atlases. Extensive experiments conducted on various data sets consistently demonstrate the superior performance of the proposed method over the competing methods.

Multimodal Brain Tumor Segmentation Boosted by Monomodal Normal Brain Images.

Many deep learning based methods have been proposed for brain tumor segmentation. Most studies focus on deep network internal structure to improve the segmentation accuracy, while valuable external information, such as normal brain appearance, is often ignored. Inspired by the fact that radiologists often screen lesion regions with normal appearance as reference in mind, in this paper, we propose a novel deep framework for brain tumor segmentation, where normal brain images are adopted as reference to compare with tumor brain images in a learned feature space. In this way, features at tumor regions, i.e., tumor-related features, can be highlighted and enhanced for accurate tumor segmentation. It is known that routine tumor brain images are multimodal, while normal brain images are often monomodal. This causes the feature comparison a big issue, i.e., multimodal vs. monomodal. To this end, we present a new feature alignment module (FAM) to make the feature distribution of monomodal normal brain images consistent/inconsistent with multimodal tumor brain images at normal/tumor regions, making the feature comparison effective. Both public (BraTS2022) and in-house tumor brain image datasets are used to evaluate our framework. Experimental results demonstrate that for both datasets, our framework can effectively improve the segmentation accuracy and outperforms the state-of-the-art segmentation methods. Codes are available at https://github.com/hb-liu/Normal-Brain-Boost-Tumor-Segmentation.

Multi-Modal Modality-Masked Diffusion Network for Brain MRI Synthesis With Random Modality Missing.

Synthesis of unavailable imaging modalities from available ones can generate modality-specific complementary information and enable multi-modality based medical images diagnosis or treatment. Existing generative methods for medical image synthesis are usually based on cross-modal translation between acquired and missing modalities. These methods are usually dedicated to specific missing modality and perform synthesis in one shot, which cannot deal with varying number of missing modalities flexibly and construct the mapping across modalities effectively. To address the above issues, in this paper, we propose a unified Multi-modal Modality-masked Diffusion Network (M2DN), tackling multi-modal synthesis from the perspective of "progressive whole-modality inpainting", instead of "cross-modal translation". Specifically, our M2DN considers the missing modalities as random noise and takes all the modalities as a unity in each reverse diffusion step. The proposed joint synthesis scheme performs synthesis for the missing modalities and self-reconstruction for the available ones, which not only enables synthesis for arbitrary missing scenarios, but also facilitates the construction of common latent space and enhances the model representation ability. Besides, we introduce a modality-mask scheme to encode availability status of each incoming modality explicitly in a binary mask, which is adopted as condition for the diffusion model to further enhance the synthesis performance of our M2DN for arbitrary missing scenarios. We carry out experiments on two public brain MRI datasets for synthesis and downstream segmentation tasks. Experimental results demonstrate that our M2DN outperforms the state-of-the-art models significantly and shows great generalizability for arbitrary missing modalities.

Dysfunctions of multiscale dynamic brain functional networks in subjective cognitive decline.

Subjective cognitive decline is potentially the earliest symptom of Alzheimer's disease, whose objective neurological basis remains elusive. To explore the potential biomarkers for subjective cognitive decline, we developed a novel deep learning method based on multiscale dynamical brain functional networks to identify subjective cognitive declines. We retrospectively constructed an internal data set (with 112 subjective cognitive decline and 64 healthy control subjects) to develop and internally validate the deep learning model. Conventional deep learning methods based on static and dynamic brain functional networks are compared. After the model is established, we prospectively collect an external data set (26 subjective cognitive decline and 12 healthy control subjects) for testing. Meanwhile, our method provides monitoring of the transitions between normal and abnormal (subjective cognitive decline-related) dynamical functional network states. The features of abnormal dynamical functional network states are quantified by network and variability metrics and associated with individual cognitions. Our method achieves an area under the receiver operating characteristic curve of 0.807 ± 0.046 in the internal validation data set and of 0.707 (P = 0.007) in the external testing data set, which shows improvements compared to conventional methods. The method further suggests that, at the local level, the abnormal dynamical functional network states are characterized by decreased connectivity strength and increased connectivity variability at different spatial scales. At the network level, the abnormal states are featured by scale-specifically altered modularity and all-scale decreased efficiency. Low tendencies to stay in abnormal states and high state transition variabilities are significantly associated with high general, language and executive functions. Overall, our work supports the deficits in multiscale brain dynamical functional networks detected by the deep learning method as reliable and meaningful neural alternation underpinning subjective cognitive decline.

ScribFormer: Transformer Makes CNN Work Better for Scribble-Based Medical Image Segmentation.

Most recent scribble-supervised segmentation methods commonly adopt a CNN framework with an encoder-decoder architecture. Despite its multiple benefits, this framework generally can only capture small-range feature dependency for the convolutional layer with the local receptive field, which makes it difficult to learn global shape information from the limited information provided by scribble annotations. To address this issue, this paper proposes a new CNN-Transformer hybrid solution for scribble-supervised medical image segmentation called ScribFormer. The proposed ScribFormer model has a triple-branch structure, i.e., the hybrid of a CNN branch, a Transformer branch, and an attention-guided class activation map (ACAM) branch. Specifically, the CNN branch collaborates with the Transformer branch to fuse the local features learned from CNN with the global representations obtained from Transformer, which can effectively overcome limitations of existing scribble-supervised segmentation methods. Furthermore, the ACAM branch assists in unifying the shallow convolution features and the deep convolution features to improve model's performance further. Extensive experiments on two public datasets and one private dataset show that our ScribFormer has superior performance over the state-of-the-art scribble-supervised segmentation methods, and achieves even better results than the fully-supervised segmentation methods. The code is released at https://github.com/HUANGLIZI/ScribFormer.

A deep learning system for predicting time to progression of diabetic retinopathy.

Diabetic retinopathy (DR) is the leading cause of preventable blindness worldwide. The risk of DR progression is highly variable among different individuals, making it difficult to predict risk and personalize screening intervals. We developed and validated a deep learning system (DeepDR Plus) to predict time to DR progression within 5 years solely from fundus images. First, we used 717,308 fundus images from 179,327 participants with diabetes to pretrain the system. Subsequently, we trained and validated the system with a multiethnic dataset comprising 118,868 images from 29,868 participants with diabetes. For predicting time to DR progression, the system achieved concordance indexes of 0.754-0.846 and integrated Brier scores of 0.153-0.241 for all times up to 5 years. Furthermore, we validated the system in real-world cohorts of participants with diabetes. The integration with clinical workflow could potentially extend the mean screening interval from 12 months to 31.97 months, and the percentage of participants recommended to be screened at 1-5 years was 30.62%, 20.00%, 19.63%, 11.85% and 17.89%, respectively, while delayed detection of progression to vision-threatening DR was 0.18%. Altogether, the DeepDR Plus system could predict individualized risk and time to DR progression over 5 years, potentially allowing personalized screening intervals.