Journal of Medical Imaging最新文献

Purpose: Whole slide image (WSI) classification relies on multiple instance learning (MIL) with spatial patch features, but current methods struggle to capture global dependencies due to the immense size of WSIs and the local nature of patch embeddings. This limitation hinders the modeling of coarse structures essential for robust diagnostic prediction.

Approach: We propose Fourier Transform Multiple Instance Learning (FFT-MIL), a framework that augments MIL with a frequency-domain branch to provide compact global context. Low-frequency crops are extracted from WSIs via the Fast Fourier Transform and processed through a modular FFT-Block composed of convolutional layers and Min-Max normalization to mitigate the high variance of frequency data. The learned global frequency feature is fused with spatial patch features through lightweight integration strategies, enabling compatibility with diverse MIL architectures.

Results: FFT-MIL was evaluated across six state-of-the-art MIL methods on three public datasets (BRACS, LUAD, and IMP). Integration of the FFT-Block improved macro $F1$ scores by an average of 3.51% and area under the curve by 1.51%, demonstrating consistent gains across architectures and datasets.

Conclusions: FFT-MIL establishes frequency-domain learning as an effective and efficient mechanism for capturing global dependencies in WSI classification, complementing spatial features and advancing the scalability and accuracy of MIL-based computational pathology. The source code is publicly available at https://github.com/irulenot/FFT-MIL.

Purpose: The purpose of this study was to assess the variations in shear-wave speed (SWS) in individual thenar muscles under varied pinch forces in healthy adults. It was hypothesized that (1) SWS would vary among the individual thenar muscles, and (2) there would be an increase in SWS with increased pinch force.

Approach: Thirteen healthy participants' dominant hands were imaged using an ultrasound probe aligned longitudinally along the muscle fibers of the abductor pollicis brevis (APB), opponens pollicis (OPP), and flexor pollicis brevis (FPB). The SWS of each muscle was derived. Each participant completed trials consisting of randomly ordered pinch forces at 0, 10, and 20 N, 10% of maximum pinch force (MPF), and 20% MPF.

Results: The SWSs varied significantly among individual thenar muscles ( $p<0.01$ ) under absolute ( $p<0.01$ ) and relative forces ( $p<0.05$ ). There was a significant increase in SWS as the force increased from 0 to 20 N in the APB ( $p<0.001$ ) and OPP ( $p<0.001$ ), and not in the FPB ( $p=0.873$ ). There was a significant increase in SWS as the force increased from 0 to 20% MPF in the APB ( $p=0.005$ ), and not in the OPP ( $p=0.586$ ) or the FPB ( $p=0.984$ ).

Conclusions: The SWS of the APB and OPP increased as force increased and was different among the thenar muscles. This suggests SWS evaluations may be an appropriate method for evaluating muscles under tension, or different voluntary force conditions, specifically for the APB and OPP muscles.

Purpose: Weakly supervised learning (WSL) is widely used for histopathological image analysis by modeling images as sets of fixed-size patches and utilizing image-level diagnoses as weak labels. However, in multiclass classification scenarios, patches corresponding to a wide spectrum of diagnostic categories can co-exist in a single image, complicating the learning process. We aim to address label uncertainty in such multiclass settings.

Approach: We propose a two-branch architecture and a complementary training strategy to improve patch-based WSL. One branch estimates patch-level class likelihoods, whereas the other predicts per-class patch relevance weights. These outputs are combined into image-level class predictions via a relevance-weighted sum of per-patch class likelihoods. To further improve performance, we introduce a multilabel augmentation strategy that forms new training samples by combining patch sets and labels from pairs of images, resulting in multilabel samples that enrich the training set by increasing the chance of having more patches that are relevant to the augmented label sets.

Results: We evaluate our method on two challenging multiclass breast histopathology datasets for region of interest classification. The proposed architecture and training strategy outperform conventional weakly supervised methods, demonstrating improved classification accuracy and robustness, particularly in underrepresented classes.

Conclusions: The proposed architecture effectively models the complex relationship between image-level labels and patch-level content in multiclass histopathological image analysis. Combined with the image-level multilabel augmentation strategy, it improves learning under label uncertainty. These contributions hold potential for more accurate and scalable diagnostic support systems in digital pathology.

Purpose: We aim to investigate the performance of dedicated breast computed tomography (CT) for the detection of soft-tissue lesions and compare it to the detection of microcalcification clusters using cascaded systems analysis, with the intent of identifying which lesion type should be used for system optimization.

Approach: Signal and noise were propagated through the imaging chain using a cascaded systems model to obtain the modulation transfer function and noise power spectrum. Two imaging tasks were considered: a soft-tissue mass lesion modeled as a disk of 4 mm diameter and a cluster of microcalcifications modeled as calcium carbonate spheres of $220\mu m$ diameter. Detectability indices using three numerical observer models were obtained for various scintillator thicknesses and acquisition conditions at a fixed 4.5 mGy mean glandular dose.

Results: Detectability index trends are reversed between soft-tissue lesion and microcalcification cluster for the range of X-ray tube voltages and filtrations studied, indicating a potential need for compromise. However, for each of the 150 combinations studied (6 kV settings $\times 5$ Cu filter thicknesses $\times 5$ CsI:Tl scintillator thicknesses) and for each of the three numerical observer models, the detectability index for soft-tissue lesions always exceeded the microcalcification cluster.

Conclusion: When the lesion type is unknown, such as during breast cancer screening, it is more appropriate to optimize the system parameters for the task of detecting a microcalcification cluster, as the detectability index for the soft-tissue lesion exceeded that for the microcalcification cluster for all conditions investigated.

Purpose: X-ray detector noise decomposition and normalized noise power spectrum (NNPS) are two metrics proposed in the European Guidelines for the quality control (QC) of digital mammography (DM) systems. We aim to examine the reproducibility of these metrics in longitudinal testing and the relevance of the limiting values in the Guidelines and produce device-specific performance data.

Approach: Semiannual QC data for 55 DM systems (16 models, 6 manufacturers) were retrieved from our medical physics archive, giving a total of 455 QC tests covering a period of 5 years (10 tests). Average values of the detector response function fit coefficients, the electronic, quantum and structure noise coefficients, and the NNPS data were analyzed across DM models and longitudinally for a given device. The fraction of quantum noise at the clinical detector air kerma ( ${\mathrm{DAK}}_{50\mathrm{mm}}$ ) level was determined, along with the longitudinal change in NNPS.

Results: Coefficient of variation for the NNPS at $2.0{\mathrm{mm}}^{-1}$ was 0.04, averaged over all systems. Quantum noise evaluated at ${\mathrm{DAK}}_{50\mathrm{mm}}$ was the largest noise fraction for all devices studied, ranging from 61.2% to 98.2%. Electronic noise was generally lower for the latest X-ray detectors. Consequently, the fraction of quantum noise at ${\mathrm{DAK}}_{50\mathrm{mm}}$ has improved from 6.2% to 28.0% and corresponds to a broader quantum noise-limited range.

Conclusion: The evaluated noise metrics were reproducible, identified changes in X-ray detector performance, and have a useful role to play in QC testing. The average values have application as reference performance data.

Purpose: We aim to analyze higher-order textural components of digital breast tomosynthesis (DBT) images to quantify differences in the appearance of breast parenchyma produced by different vendors.

Approach: We included consecutive women who had normal screening DBT exams in January 2018 from a GE system and in adjacent years from Hologic systems. Laplacian fractional entropy (LFE), as a measure of non-Gaussian statistical properties of breast tissue texture, was calculated from for-presentation Craniocaudal (CC) view DBT slices and synthetic mammograms (SMs) through frequency-based filtering with Gabor filters, which were considered mathematical models for human visual response to image textures. The LFE values were compared within and across subjects and vendors along with secondary parameters (laterality, year-to-year, modality, and breast density) via two-way analysis of variance (ANOVA) tests using frequency as one of the two independent variables, and a $P$ -value $<0.05$ was considered statistically significant.

Results: A total of 8529 CC view DBT slices and SM images from 73 screening exams in 25 women were analyzed. Significant differences in LFE were observed for different frequencies ( $P<0.001$ ) and across vendors (GE versus Hologic DBT: $P<0.001$ , GE versus Hologic SM: $P<0.001$ ).

Conclusion: Significant differences in perception of breast parenchyma textures among two DBT vendors were demonstrated via higher-order non-Gaussian statistical properties. This finding extends previously observed differences in anatomical noise power spectra in DBT images and provides quantitative evidence to support caution in across-vendor comparative reading and will be beneficial to facilitate future development of vendor-neutral artificial intelligence algorithms for breast cancer screening.

Purpose: The purpose is to retrospectively investigate how the addition of prior and concurrent mammograms affects wide-angle digital breast tomosynthesis (DBT) screening false-positive recall rates, malignancy scoring, and recall agreement.

Approach: A total of 200 cases were selected from the Malmö Breast Tomosynthesis Screening Trial. They consist of 150 recalled cases [30 true positives (TPs), 120 false positives (FPs), and 50 healthy, non-recalled true-negative (TN) cases]. The positive cases were categorized based on being recalled by either DBT, digital mammography (DM), or both. Each case had DBT, synthetic mammography (SM), and DM (prior screening round) images. Five radiologists participated in a reading study where detection, risk of malignancy, and recall were assessed. They read each case twice, once using only DBT and once using DBT together with SM and DM priors.

Results: The results showed a significant reduction in recall rates for all FP categories, as well as for the TN cases, when adding SM and prior DM to DBT. This resulted also in a significant increase in recall agreement for these categories, with more of the negative cases being recalled by few or no readers. These categories were overall rated as appearing more malignant in the DBT reading arm. For the TP categories, there was a significant decrease in recalls for DM-recalled cancers ( $p=0.047$ ), but no significant difference for DBT-recalled cancers ( $p=0.063$ ), or DBT/DM-recalled cancers ( $p=0.208$ ).

Conclusions: Similar to the documented effect of priors in DM screening, we suggest that added two-dimensional priors improve the specificity of DBT screening but may reduce the sensitivity.

Purpose: Breast cancer is one of the leading causes of cancer-related deaths among women, and digital mammography plays a key role in screening and early detection. The radiation dose on mammographic exams directly influences image quality and radiologists' performance. We evaluate the impact of an image restoration pipeline-designed to simulate higher dose acquisitions-on the detectability of microcalcifications of various sizes in mammograms acquired at different radiation doses.

Approach: The restoration pipeline denoises the image using a Poisson-Gaussian noise model, combining it with the noisy image to achieve a signal-to-noise ratio comparable with an acquisition at twice the original dose. We created a database of images using a physical breast phantom at doses ranging from 50% to 200% of the standard dose. Clustered microcalcifications were computationally inserted into the phantom images. The channelized Hotelling observer was employed in a four-alternative forced-choice to evaluate the detectability of microcalcifications across different sizes and exposure levels.

Results: The restoration of low-dose images acquired at $\sim 75\%$ of the standard dose resulted in detectability levels comparable with those of images acquired at the standard dose. Moreover, images restored at the standard dose demonstrated detectability similar to those acquired at 160% of the nominal radiation dose, with no statistically significant differences.

Conclusions: We demonstrate the potential of an image restoration pipeline to simulate higher quality mammography images. The results indicate that reducing noise through denoising and restoration impacts the detectability of microcalcifications. This method improves image quality without hardware modifications or additional radiation exposure.

Purpose: Virtual imaging trials (VITs) are of interest for regulatory evaluation because they enable faster and more cost-effective evaluation of new imaging technologies than patient clinical trials. Our purpose is to develop a hybrid VIT methodology for breast computed tomography (CT) applications and use it to investigate microcalcification detectability.

Approach: Ray tracing was used to generate projection images of clusters of five microcalcifications which varied in diameter, chemical composition, and density. These simulated projection images were added to patient projection images acquired with the fourth-generation breast CT scanner from UC Davis (Doheny) and reconstructed using the Feldkamp filtered backprojection algorithm with varying apodization kernels. Volumes of interest and maximum intensity projections were extracted from the reconstructed volumes. Human observers (HOs) and deep learning model observers (DLMOs) were used to detect calcification clusters, and receiver operating characteristic curve analysis was used to analyze detection performance.

Results: DLMO detected 0.18-mm type I calcifications with AUC = 0.80 and 0.21 mm calcifications with $\mathrm{AUC}=0.99$ . HO performance was inferior to deep learning model observer performance, but both HO and DLMO detected 0.21-mm type I calcifications with $\mathrm{AUC}>0.90$ and 0.24-mm type I calcifications with near-perfect performance. Microcalcification clusters embedded in adipose tissue were more conspicuous than clusters embedded in fibroglandular tissue. There was superior detection performance for clusters located anteriorly within the breast compared with clusters located posteriorly within the breast.

Conclusions: A hybrid approach for virtual imaging trials shows promise for the assessment of imaging systems across a broad range of parameters.

Purpose: Data-driven harmonization can mitigate systematic confounding signals across imaging cohorts caused by variance in scanners and acquisition protocols. As diffusion magnetic resonance imaging data are often acquired with different hardware and software, harmonization is essential for integrating these scattered datasets into a cohesive analysis for improved statistical power. Large-scale, multi-site studies for Alzheimer's disease (AD), a neurodegenerative condition characterized by high data variability and complex pathology, pose the challenge of both site-based and biological variation.

Approach: We learn lower-dimensional representations of structural connectivity invariant to imaging cohort, geographical location, scanner, and acquisition factors. We design a conditional variational autoencoder that creates latent representations with minimal information about imaging factors and maximal information related to patient cognitive status. With this model, we consolidate 9 cohorts and 35 unique imaging acquisitions (for a total of 38 imaging "sites") into a cohesive dataset of 6956 persons (16.4% with mild cognitive impairment and 10.7% with AD) imaged for 1 to 16 sessions for a total of 11,927 diffusion-weighted imaging sessions.

Results: These site-invariant representations successfully remove significant ( $p<0.05$ ) site effects in 12 network connectivity measures of interest and enhance the prediction of cognitive diagnosis (from 68% accuracy to 73% accuracy).

Conclusions: The proposed model yields reproducible precision across 15 data configurations. This approach demonstrates the effectiveness of representation learning in enhancing biological signals by mitigating acquisition-specific confounding factors in neuroimaging studies.