The pathogenesis of amyotrophic lateral sclerosis (ALS) critically involves white matter (WM) degeneration. This study aimed to employ differential tractography to investigate WM microstructural alterations in ALS and their associations with disease severity. This study enrolled 40 ALS patients and 66 healthy controls who underwent diffusion MRI. Quantitative anisotropy (QA), a diffusion metric reflecting axonal density, was used to assess microstructural alterations in WM tracts. Differential tractography was applied to identify WM tracts affected by significant reductions in QA, thereby mapping the distribution of axonal injury in ALS. Correlational tractography was employed to assess the association between WM tract damage and clinical disease severity. Differential tractography revealed a relatively consistent pattern of WM tracts exhibiting decreased QA, primarily involving the bilateral corticospinal tracts (CST), bilateral superior corticostriatal tracts, bilateral medial lemnisci, body and forceps minor of the corpus callosum, left parietal corticopontine tract, and left anterior thalamic radiation. The most significant and extensive decline in QA was observed in bilateral CST. Positive correlations were observed between ALSFRS-R decline and QA reduction in several WM tracts, particularly the left CST, which correlated most closely with clinical progression, as well as the body and forceps major of the corpus callosum, the left parietal corticopontine tract, and the left corticobulbar tract. Our findings confirmed that axonal injury and myelination were pivotal pathophysiological processes in ALS. The prominent axonal impairment in the CST, coupled with its strong correlation with clinical severity, suggested that CST integrity measurement might serve as a favorable imaging biomarker for ALS. • Differential tractography enables direct visualization of white matter tracts with significant neuronal damage in ALS by identifying specific segments of affected fiber tracts. • Affected tracts exhibited significantly reduced axonal density, with the bilateral CSTs showing the most prominent and extensive decline. • Axonal damage in several WM tracts, particularly the left corticospinal tract and corpus callosum, correlated with disease progression, supporting the potential of the tractography-based measurement as an imaging marker for monitoring ALS.
Technetium-99 m pyrophosphate (Tc-99 m PYP) planar scintigraphy is widely used for evaluating transthyretin cardiac amyloidosis (ATTR-CM). However, interpreting equivocal Perugini Grade 1 cardiac uptake remains a significant clinical challenge due to overlapping blood-pool activity and low signal-to-noise ratio. We aimed to develop and validate a multimodal deep learning framework to predict expert SPECT/CT adjudicated Perugini categories from planar imaging and pre-procedural clinical variables, with particular attention to Grade 1 examinations. In this multicenter study utilizing data from the Taiwan Multicenter TTCA-PYP Registry (N = 438), we developed a multimodal neural network. The framework processes physician-verified cardiac planar ROIs (32 × 128 pixels) with a truncated Wide-ResNet50-2 encoder and 23 routine clinical variables with a frozen TabPFN embedding, aligned by a CLIP-style contrastive loss and fused by cross-attention, to predict SPECT/CT-adjudicated Perugini categories (Grade 0, Grade 1, Grades 2–3). Genetic testing status and study design were not inputs. Evaluation used 10 repetitions of a hospital- and grade-stratified assessment set with 10-fold cross-validation, decision curve analysis, and leave-one-center-out testing of the two centers with all grades (FEMH, NTUH). Overall accuracy was higher with the multimodal model than with imaging only (0.826 vs. 0.751; paired difference + 0.075, 95
To evaluate the diagnostic accuracy of histogram-based texture analysis on high-resolution computed tomography (HRCT) images for differentiating middle ear cholesteatoma from non-cholesteatomatous lesions. This retrospective study included 96 patients (48 with surgically or histopathologically confirmed cholesteatoma and 48 with non-cholesteatomatous middle ear pathologies) who underwent temporal bone HRCT between 2020 and 2026. Three-dimensional volume of interest (VOI) segmentation was manually performed on axial sections, and histogram-based radiomic features were extracted using LIFEx software. A multivariable binary logistic regression model was constructed using the backward elimination method to identify independent predictors. Diagnostic performance was evaluated using receiver operating characteristic (ROC) curve analysis. Internal validation was performed using five-fold cross-validation, with variable selection repeated independently within each training fold. Inter-observer reliability was assessed via the intraclass correlation coefficient (ICC). The groups were age- and sex-matched. The predictive model incorporated four independent histogram features: Intensity Histogram Mean Absolute Deviation and Minimum Grey Level (both significantly lower in the cholesteatoma group), alongside Intensity Histogram Kurtosis and Range (both significantly higher in the cholesteatoma group). The model demonstrated favorable calibration (Hosmer–Lemeshow p = 0.947) and achieved an AUC of 0.912. Following five-fold cross-validation, the model maintained favorable diagnostic performance, with an AUC of 0.905 and an accuracy of 82.0
The prognosis for glioblastoma(GBM) remains dismal. An accurate assessment of prognostic stratification is crucial for guiding clinical decisions. We aimed to establish and validate a multimodal model for individualized and accurate survival risk stratification in patients with GBM. We collected a multimodal dataset comprising 194 consecutively enrolled IDH-wildtype GBM patients. First, we performed analysis of clinical features using Cox regression, resulting in a clinical signature model. Second, we use ResNet101, DenseNet121, and Inception_v3 deep learning neural networks, extraction radiomics features, formed deep learning radiomics signature (DL-Rad_Sig), we also extracted pathomics features based on H E-stained whole slide images (WSIs), resulting in the pathomics signature (Path_Sig). The full cohort was split into training (136 patients) and test (58 patients) sets via stratified random sampling based on 1-year overall survival status. LASSO-Cox regression with 10-fold cross-validation within training set was adopted for feature screening. Finally, a multimodal model was constructed to predict overall survival (OS) of GBM patients. The clinical features performed suboptimal in prognosis analysis. DL-Rad_Sig model, the c-index values of 0.852 and 0.721 for the train and test cohort. Six pathomics features were used develop Path_Sig modeling, resulting in c-index values of 0.821 and 0.762 in the train and test cohort. And we compared the fusion of the three signatures with each individual signature. We found that the combined model exhibited a better C-index, the C-index of 0.887 and 0.791 for OS in train and test cohort. Kaplan-Meier survival curves of the combined model nomogram show the OS rate difference was statistically significant between the high- and low-risk patients. The ROC curve of the combined model nomogram at predicts 1-, 2-, and 3-year OS produced AUC of 0.871, 0.725, and 0.833, respectively. Our internally validated multimodal nomogram integrating clinical, deep radiomics and weakly-supervised pathomics signatures achieves favorable OS risk stratification for single-center IDH-wildtype glioblastoma patients, though external multi-center validation is still required to verify generalizability. Not applicable. Both deep learning radiomics and pathomics models achieve effective prognostic stratification for GBM patients. Pathomics features extracted from H E-stained WSIs are significantly associated with clinical outcomes in GBM. The multimodal fusion nomogram serves as a reliable and non-invasive tool for prognostic evaluation of GBM patients. This study only performed single-center internal validation without independent external cohorts, and patch-level pathomics models show obvious training-set overfitting which needs cautious interpretation.
To evaluate the performance of a Channel Dense Dense U-Generative Adversarial Network (CDDU-GAN), in improving the image quality of four-fold accelerated Prostate-specific membrane antigen (PSMA) PET acquisitions across different radiotracers and hybrid imaging platforms. This study enrolled 84 patients who underwent either 18F-PSMA or 68Ga-PSMA PET imaging on PET/CT or PET/MR systems. Standard-dose list-mode data (120s/bed for PET/CT, 240s/bed for PET/MR) were retrospectively subsampled to the first quarter of the original counts to simulate a rapid-scan protocol. The CDDU-GAN was trained to transform these low-count images into standard-dose equivalent acquisitions. Objective image quality was quantified using Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index (SSIM). Subjective image quality was independently assessed by two blinded nuclear medicine physicians using a 5-point Likert scale for overall image quality, lesion conspicuity, and diagnostic confidence. Quantitative accuracy was evaluated via linear regression and concordance correlation coefficient (CCC) of SUVmax in 84 lesions. The CDDU-GAN significantly improved objective image quality across the entire cohort, increasing the mean PSNR from 46.8 ± 7.5 to 48.2 ± 6.6 (p < 0.0001) and SSIM from 0.980 ± 0.026 to 0.986 ± 0.020 (p < 0.0001). Performance gains were substantially larger for the PET/CT cohort (ΔPSNR: 2.91 ± 0.70 vs. 0.47 ± 1.08 dB, p < 0.0001) than for the PET/MR cohort. Diagnostic confidence improved significantly in all four cohorts (p = 0.031–0.042). SUVmax demonstrated stronger linear agreement between CDDU-GAN and standard-dose (y = 1.024x + 0.22, R² = 0.985, CCC = 0.992) compared with fast-scan (R² = 0.944, CCC = 0.972). The CDDU-GAN framework effectively improves image quality and SUV accuracy in four-fold accelerated PSMA PET imaging across the tracers and scanners evaluated in this study, although the magnitude of benefit varied substantially across platforms. These findings provide preliminary evidence for the feasibility of deep learning-based acceleration in PSMA PET. External validation with larger, multi-center cohorts and task-based assessments of lesion detection are required prior to clinical adoption.
