Journal of Medical Imaging最新文献

Purpose: Visualization of subcortical gray matter is essential in neuroscience and clinical practice, particularly for disease understanding and surgical planning. Although multi-inversion time (multi-TI) $T_1$ -weighted ( $T_1$ -w) magnetic resonance (MR) imaging improves visualization, it is only acquired in specific clinical settings and not available in common public MR datasets.

Approach: We present SyMTIC (synthetic multi-TI contrasts), a deep learning method that generates synthetic multi-TI images using routinely acquired $T_1$ -w, $T_2$ -weighted ( $T_2$ -w), and fluid-attenuated inversion recovery (FLAIR) images. Our approach combines image translation via deep neural networks with imaging physics to estimate longitudinal relaxation time ( $T1$ ) and proton density ( $\rho$ ) maps. These maps are then used to compute multi-TI images with arbitrary inversion times.

Results: SyMTIC was trained using paired magnetization prepared rapid acquisition with gradient echo (MPRAGE) and fast gray matter acquisition T1 inversion recovery (FGATIR) images along with $T_2$ -w and FLAIR images. It accurately synthesized multi-TI images from standard clinical inputs, achieving image quality comparable to that from explicitly acquired multi-TI data. The synthetic images, especially for TI values between 400 to 800 ms, enhanced visualization of subcortical structures and improved segmentation of thalamic nuclei.

Conclusion: SyMTIC enables robust generation of high-quality multi-TI images from routine MR contrasts. When paired with the HACA3 algorithm, it generalizes well to varied clinical datasets, including those without FLAIR or $T_2$ -w images and unknown parameters, offering a practical solution for improving brain MR image visualization and analysis.

Purpose: Dynamic focusing of received ultrasound signals, or beamforming, is foundational for ultrasound imaging. Conventionally, it requires arrays of ultrasound sensors to estimate where sound came from using time-of-flight (TOF) measurements. We demonstrate passive beamforming with a single biaxial sensor and accurate passive acoustic mapping with two biaxial sensors using only direction of arrival (DOA) information.

Approach: We introduce two single-element biaxial beamforming algorithms and four biaxial image reconstruction algorithms for a two-element biaxial piezoceramic transducer array. Imaging of a hemispherical acoustic source is characterized in an acoustic scanning tank within the region $-30.29\mathrm{mm}$ $\le x\le$ 29.94 mm and 50.11 mm $\le z\le$ 90.45 mm relative to the center of the array. Imaging performance is contrasted with delay, sum, and integrate (DSAI) and delay, multiply, sum, and integrate (DMSAI) algorithms.

Results: Single-element biaxial beamforming can identify DOA with a median error (± interquartile range) of $0.36\pm 0.63\mathrm{deg}$ and median full-width half-prominence of $7.3\pm 8.6\mathrm{deg}$ . Using both array elements, DOA-only images demonstrate overall median localization error of 6.41 mm (lateral: 1.02 mm, axial: 5.85 mm, signal-to-noise ratio (SNR): 15.37) and DOA + TOF images demonstrate overall median error of 6.91 mm (lateral: 1.69 mm, axial: 6.11 mm, SNR: 18.37).

Conclusions: To the best of our knowledge, we provide the first demonstration of single-element beamforming using a single stationary piezoceramic and the first demonstration of passive ultrasound imaging without the use of TOF information. These results enable simpler, smaller, more cost-effective arrays for passive ultrasound imaging.

Purpose: Breast density (BD) and background parenchymal enhancement (BPE) are important imaging biomarkers for breast cancer (BC) risk. We aim to evaluate longitudinal changes in quantitative BD and BPE in high-risk women undergoing dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI), focusing on the effects of age and transition into menopause.

Approach: A retrospective cohort study analyzed 834 high-risk women undergoing breast DCE-MRI for screening between 2005 and 2020. Quantitative BD and BPE were derived using deep-learning segmentation. Linear mixed-effects models assessed longitudinal changes and the effects of age, menopausal status, weeks since the last menstrual period (LMP-wks), body mass index (BMI), and hormone replacement therapy (HRT) on these imaging biomarkers.

Results: BD decreased with age across all menopausal stages, whereas BPE declined with age in postmenopausal women but remained stable in premenopausal women. HRT elevated BPE in postmenopausal women. Perimenopausal women exhibited decreases in both BD and BPE during the menopausal transition, though cross-sectional age at menopause had no significant effect on either measure. Fibroglandular tissue was positively associated with BPE in perimenopausal women.

Conclusions: We highlight the dynamic impact of menopause on BD and BPE and correlate well with the known relationship between risk and age at menopause. These findings advance the understanding of imaging biomarkers in high-risk populations and may contribute to the development of improved risk assessment leading to personalized chemoprevention and BC screening recommendations.

Purpose: We propose and evaluate multimodal deep learning (DL) approaches that combine ultrasound (US) B-mode and Nakagami parametric images for breast tumor classification. It is hypothesized that integrating tissue brightness information from B-mode images with scattering properties from Nakagami images will enhance diagnostic performance compared with single-input approaches.

Approach: An EfficientNetV2B0 network was used to develop multimodal DL frameworks that took as input (i) numerical two-dimensional (2D) maps or (ii) rendered red-green-blue (RGB) representations of both B-mode and Nakagami data. The diagnostic performance of these frameworks was compared with single-input counterparts using 831 US acquisitions from 264 patients. In addition, gradient-weighted class activation mapping was applied to evaluate diagnostically relevant information utilized by the different networks.

Results: The multimodal architectures demonstrated significantly higher area under the receiver operating characteristic curve (AUC) values ( $p<0.05$ ) than their monomodal counterparts, achieving an average improvement of 10.75%. In addition, the multimodal networks incorporated, on average, 15.70% more diagnostically relevant tissue information. Among the multimodal models, those using RGB representations as input outperformed those that utilized 2D numerical data maps ( $p<0.05$ ). The top-performing multimodal architecture achieved a mean AUC of 0.896 [95% confidence interval (CI): 0.813 to 0.959] when performance was assessed at the image level and 0.848 (95% CI: 0.755 to 0.903) when assessed at the lesion level.

Conclusions: Incorporating B-mode and Nakagami information together in a multimodal DL framework improved classification outcomes and increased the amount of diagnostically relevant information accessed by networks, highlighting the potential for automating and standardizing US breast cancer diagnostics to enhance clinical outcomes.

Purpose: Breast cancer may persist within milk ducts (ductal carcinoma in situ, DCIS) or advance into surrounding breast tissue (invasive ductal carcinoma, IDC). Occasionally, invasiveness in cancer may be underestimated during biopsy, leading to adjustments in the treatment plan based on unexpected surgical findings. Artificial intelligence/computer-aided diagnosis (AI/CADx) techniques in medical imaging may have the potential to predict whether a lesion is purely DCIS or exhibits a mixture of IDC and DCIS components, serving as a valuable supplement to biopsy findings. To enhance the evaluation of AI/CADx performance, assessing variability on a lesion-by-lesion basis via likelihood assurance measures could add value.

Approach: We evaluated the performance in the task of distinguishing between pure DCIS and mixed IDC/DCIS breast cancers using computer-extracted radiomic features from dynamic contrast-enhanced magnetic resonance imaging using 0.632+ bootstrapping methods (2000 folds) on 550 lesions (135 pure DCIS, 415 mixed IDC/DCIS). Lesion-based likelihood assurance was measured using a sureness metric based on the 95% confidence interval of the classifier output for each lesion.

Results: The median and 95% CI of the 0.632+-corrected area under the receiver operating characteristic curve for the task of classifying lesions as pure DCIS or mixed IDC/DCIS were 0.81 [0.75, 0.86]. The sureness metric varied across the dataset with a range of 0.0002 (low sureness) to 0.96 (high sureness), with combinations of high and low classifier output and high and low sureness for some lesions.

Conclusions: Sureness metrics can provide additional insights into the ability of CADx algorithms to pre-operatively predict whether a lesion is invasive.

Purpose: Automated segmentation and classification of the cell nuclei in microscopic images is crucial for disease diagnosis and tissue microenvironment analysis. Given the difficulties in acquiring large labeled datasets for supervised learning, semi-supervised methods offer alternatives by utilizing unlabeled data alongside labeled data. Effective semi-supervised methods to address the challenges of extremely limited labeled data or diverse datasets with varying numbers and types of annotations remain under-explored.

Approach: Unlike other semi-supervised learning methods that iteratively use labeled and unlabeled data for model training, we introduce a semi-supervised learning framework that combines a latent diffusion model (LDM) with a transformer-based decoder, allowing for independent usage of unlabeled data to optimize their contribution to model training. The model is trained based on a sequential training strategy. LDM is trained in an unsupervised manner on diverse datasets, independent of cell nuclei types, thereby expanding the training data and enhancing training performance. The pre-trained LDM serves as a powerful feature extractor to support the transformer-based decoder's supervised training on limited labeled data and improve final segmentation performance. In addition, the paper explores a collaborative learning strategy to enhance segmentation performance on out-of-distribution (OOD) data.

Results: Extensive experiments conducted on four diverse datasets demonstrated that the proposed framework significantly outperformed other semi-supervised and supervised methods for both in-distribution and OOD cases. Through collaborative learning with supervised methods, diffusion model and transformer decoder-based segmentation (DTSeg) achieved consistent performance across varying cell types and different amounts of labeled data.

Conclusions: The proposed DTSeg framework addresses cell nuclei segmentation under limited labeled data by integrating unsupervised LDM training on diverse unlabeled datasets. Collaborative learning demonstrated effectiveness in enhancing the generalization capability of DTSeg to achieve superior results across diverse datasets and cases. Furthermore, the method supports multi-channel inputs and demonstrates strong generalization to both in-distribution and OOD scenarios.

Purpose: Recent advances in deep learning have led to robust automated tools for segmentation of abdominal computed tomography (CT). Meanwhile, segmentation of magnetic resonance imaging (MRI) is substantially more challenging due to the inherent signal variability and the increased effort required for annotating training datasets. Hence, existing approaches are trained on limited sets of MRI sequences, which might limit their generalizability.

Approach: To characterize the landscape of MRI abdominal segmentation tools, we present a comprehensive benchmarking of three state-of-the-art and open-source models: MRSegmentator, MRISegmentator-Abdomen, and TotalSegmentator MRI. As these models are trained using labor-intensive manual annotation cycles, we also introduce and evaluate ABDSynth, a SynthSeg-based model purely trained on widely available CT segmentations (no real images). We assess accuracy and generalizability by leveraging three public datasets (not seen by any of the evaluated methods during their training), which span all major manufacturers, five MRI sequences, as well as a variety of subject conditions, voxel resolutions, and fields-of-view.

Results: Our results reveal that MRSegmentator achieves the best performance and is most generalizable. By contrast, ABDSynth yields slightly less accurate results, but its relaxed requirements in training data make it an alternative when the annotation budget is limited.

Conclusions: We perform benchmarking of four open-source models for abdominal MR segmentation on three datasets and demonstrate that models trained on real, heterogeneous, multimodal data yield the best overall performance. We provide evaluation code and datasets for future benchmarking at https://github.com/deepakri201/AbdoBench.

Purpose: Accurate pixel-level segmentation is essential for medical image analysis, particularly in assisting diagnosis and treatment planning. However, fully supervised learning methods rely heavily on high-quality annotated data, which are often scarce due to the high cost of manual labeling, privacy concerns, and limited availability. We aim to reduce reliance on precise annotations and improve segmentation performance under weak supervision.

Approach: We propose scribble position and temporal contrast learning (SPTCL), an innovative segmentation method that combines contrastive learning with weak supervision. Our method leverages the spatial continuity in 3D medical image volumes and the anatomical similarities across different cardiac phases to construct a contrastive learning task for robust feature representation from unlabeled data. To enhance the feature extraction capabilities, we employ a pre-trained encoder, which is initially trained on the ACDC dataset using contrastive learning to capture robust feature representations. This pre-trained encoder is then transferred to a weakly supervised segmentation network with a dual-branch decoder for further fine-tuning on the task. The predictions from both branches are fused to generate refined pseudo-labels, which are iteratively used to guide network training with only coarse scribble annotations.

Results: Experiments on the ACDC dataset show that SPTCL outperforms existing models, achieving a Dice coefficient of 90.5%, with a 2.5% improvement over the baseline and a 1.7% improvement over the latest model. Furthermore, SPTCL reduces training time by $\sim 33\%$ .

Conclusions: SPTCL effectively addresses the challenges of limited annotation in medical image segmentation by uniting contrastive learning with weak supervision. It demonstrates strong potential for practical deployment in clinical settings where high-quality labels are difficult to obtain.

Purpose: Coronary artery disease is the leading global cause of mortality. Automated detection and scoring of calcified plaques can help cardiovascular risk assessment. We propose a deep learning method for automatic detection and scoring of coronary artery calcified plaques on noncontrast CT scans.

Approach: We utilized five datasets from one internal and four external tertiary care institutions, three of them with manually annotated plaques. A coronary artery calcified plaque detection model was developed using the state-of-the-art nnU-Net deep learning framework, incorporating simultaneous segmentation of the aorta, heart, and lungs to reduce false positives. The training data consisted of 641 noncontrast CT scans from three labeled datasets, representing diverse vascular disease etiologies. Agatston scores were automatically computed to quantify plaque burden. The model was tested on 160 labeled CT scans and compared with a previous detection method. In addition, Agatston scores were correlated with patient demographics and clinical outcomes using two unlabeled datasets.

Results: The predicted and reference Agatston scores demonstrated a strong correlation ( $r^2=0.973$ ), with a precision of 89.3%, recall of 89.1%, and an average Dice score of $75.0\pm 16.0\%$ on the labeled testing datasets. The stratified four Agatston groups achieved 92.0% accuracy and a Cohen's Kappa of 0.913. In the unlabeled datasets, Agatston groups showed significant correlations with the Framingham risk score, cardiovascular disease, heart failure, cancer status, fragility fracture risk, smoking, and age, whereas remaining consistent across race and scanner types.

Conclusions: Coronary artery plaques were accurately detected and segmented using the proposed nnU-Net-based method on noncontrast CT scans. The Agatston-score-based plaque burden assessment facilitates cardiovascular risk stratification, enabling opportunistic screening and population-based studies.

Purpose: To avoid over-treatment of prostate cancer patients following screening for elevated prostate-specific antigen levels, keeping patients on active surveillance has been suggested as an alternative to radical treatment. This means recurring visits for patients with low-grade cancer to monitor progression. Our aim was to develop an artificial intelligence-based model that can identify high-risk patients in a cohort of prostate cancer patients on active surveillance.

Approach: We have developed a multiple instance learning-based framework for predicting the longitudinal outcomes for prostate cancer patients on active surveillance. Our models were trained only on whole-slide images with patient-level labels without using explicit Gleason grades. We employed the UNI-2 foundation model and the well-established attention-based multiple instance learning approach. We further evaluated our models by fitting Cox proportional hazards models and testing them on an external dataset.

Results: With this approach, we achieved an average area under the receiver operator characteristic curve of 0.958 (95% CI, 0.957 to 0.959). Fitting Cox models to the predicted probabilities achieved a $C$ -index of 0.824 and a hazard ratio of 2.32. However, all models showed a large drop in performance when evaluated on an external dataset.

Conclusion: We show that avoiding Gleason grades is beneficial for longitudinal outcome prediction of prostate cancer. Our results suggest that benign prostate tissue contains prognostic information. However, before our models could be used clinically, much more work remains to improve the generalization.