Functional magnetic resonance imaging (fMRI) has opened new frontiers in neuroscience by instrumentally driving our understanding of brain function and development. Despite its substantial successes, fMRI studies persistently encounter obstacles stemming from inherent, unavoidable physiological confounds. The adverse effects of these confounds are especially noticeable with higher magnetic fields, which have been gaining momentum in fMRI experiments. This review focuses on the four major physiological confounds impacting fMRI studies: low-frequency fluctuations in both breathing depth and rate, low-frequency fluctuations in the heart rate, thoracic movements, and cardiac pulsatility. Over the past three decades, numerous correction techniques have emerged to address these challenges. Correction methods have effectively enhanced the detection of task-activated voxels and minimized the occurrence of false positives and false negatives in functional connectivity studies. While confound correction methods have merit, they also have certain limitations. For instance, model-based approaches require externally recorded physiological data that is often unavailable in fMRI studies. Methods reliant on independent component analysis, on the other hand, need prior knowledge about the number of components. Machine learning techniques, although showing potential, are still in the early stages of development and require additional validation. This article reviews the mechanics of physiological confound correction methods, scrutinizes their performance and limitations, and discusses their impact on fMRI studies.

Accurate quantitative assessment using gadolinium-contrast magnetic resonance imaging (MRI) is crucial in therapy planning, surveillance and prognostic assessment of primary central nervous system lymphoma (PCNSL). The present study aimed to develop a multimodal artificial intelligence deep learning segmentation model to address the challenges associated with traditional 2D measurements and manual volume assessments in MRI. Data from 49 pathologically-confirmed patients with PCNSL from six Chinese medical centers were analyzed, and regions of interest were manually segmented on contrast-enhanced T1-weighted and T2-weighted MRI scans for each patient, followed by fully automated voxel-wise segmentation of tumor components using a 3-dimenstional convolutional deep neural network. Furthermore, the efficiency of the model was evaluated using practical indicators and its consistency and accuracy was compared with traditional methods. The performance of the models were assessed using the Dice similarity coefficient (DSC). The Mann-Whitney U test was used to compare continuous clinical variables and the χ² test was used for comparisons between categorical clinical variables. T1WI sequences exhibited the optimal performance (training dice: 0.923, testing dice: 0.830, outer validation dice: 0.801), while T2WI showed a relatively poor performance (training dice of 0.761, a testing dice of 0.647, and an outer validation dice of 0.643. In conclusion, the automatic multi-sequences MRI segmentation model for PCNSL in the present study displayed high spatial overlap ratio and similar tumor volume with routine manual segmentation, indicating its significant potential.

This study aimed to develop and evaluate a non-invasive XGBoost-based machine learning model using radiomic features extracted from pre-treatment CT images to differentiate grade 4 renal cell carcinoma (RCC) from lower-grade tumours. A total of 102 RCC patients who underwent contrast-enhanced CT scans were included in the analysis. Radiomic features were extracted, and a two-step feature selection methodology was applied to identify the most relevant features for classification. The XGBoost model demonstrated high performance in both training (AUC = 0.87) and testing (AUC = 0.92) sets, with no significant difference between the two (p = 0.521). The model also exhibited high sensitivity, specificity, positive predictive value, and negative predictive value. The selected radiomic features captured both the distribution of intensity values and spatial relationships, which may provide valuable insights for personalized treatment decision-making. Our findings suggest that the XGBoost model has the potential to be integrated into clinical workflows to facilitate personalized adjuvant immunotherapy decision-making, ultimately improving patient outcomes. Further research is needed to validate the model in larger, multicentre cohorts and explore the potential of combining radiomic features with other clinical and molecular data.

The reliance on computationally intensive U-Net and Transformer architectures significantly limits their accessibility in low-resource environments, creating a technological divide that hinders global healthcare equity, especially in medical diagnostics and treatment planning. This divide is most pronounced in low- and middle-income countries, primary care facilities, and conflict zones. We introduced MED-NCA, Neural Cellular Automata (NCA) based segmentation models characterized by their low parameter count, robust performance, and inherent quality control mechanisms. These features drastically lower the barriers to high-quality medical image analysis in resource-constrained settings, allowing the models to run efficiently on hardware as minimal as a Raspberry Pi or a smartphone. Building upon the foundation laid by MED-NCA, this paper extends its validation across eight distinct anatomies, including the hippocampus and prostate (MRI, 3D), liver and spleen (CT, 3D), heart and lung (X-ray, 2D), breast tumor (Ultrasound, 2D), and skin lesion (Image, 2D). Our comprehensive evaluation demonstrates the broad applicability and effectiveness of MED-NCA in various medical imaging contexts, matching the performance of two magnitudes larger UNet models. Additionally, we introduce NCA-VIS, a visualization tool that gives insight into the inference process of MED-NCA and allows users to test its robustness by applying various artifacts. This combination of efficiency, broad applicability, and enhanced interpretability makes MED-NCA a transformative solution for medical image analysis, fostering greater global healthcare equity by making advanced diagnostics accessible in even the most resource-limited environments.

To assess the image quality and biventricular function utilizing a free-breathing artificial intelligence cine method with motion correction (FB AI MOCO). A total of 72 participants (mean age 38.3 ± 15.4 years, 40 males) prospectively enrolled in this single-center, cross-sectional study underwent cine scans using standard breath-holding (BH) cine and FB AI MOCO cine at 3.0 Tesla. The image quality of the cine images was evaluated with a 5-point Ordinal Likert scale based on blood-pool to myocardium contrast, endocardial edge definition, and artifacts, and overall quality score was calculated by the equal weight average of all three criteria, apparent signal to noise ratio (aSNR), estimated contrast to noise ratio (eCNR) were assessed. Biventricular functional parameters including Left Ventricular (LV), Right Ventricular (RV) End-Diastolic Volume (EDV), End-Systolic Volume (ESV), Stroke Volume (SV), Ejection Fraction (EF), and LV End-Diastolic Mass (LVEDM) were also assessed. Comparison between two sequences was assessed using paired t-test and Wilcoxon signed-rank test, correlation using Pearson correlation. The agreement of quantitative parameters was assessed using intraclass correlation coefficient (ICC) and Bland-Altman analysis. P < 0.05 was statistically significant. The total acquisition time of the entire stack for FB AI MOCO cine (14.7 s ± 1.9 s) was notably shorter than that for standard BH cine (82.6 s ± 11.9 s, P < 0.001). The aSNR between FB AI MOCO cine and standard BH cine has no significantly difference (76.7 ± 20.7 vs. 79.8 ± 20.7, P = 0.193). The eCNR of FB AI MOCO cine was higher than standard BH cine (191.6 ± 54.0 vs. 155.8 ± 68.4, P < 0.001), as was the scores of blood-pool to myocardium contrast (4.6 ± 0.5 vs. 4.4 ± 0.6, P = 0.003). Qualitative scores including endocardial edge definition (4.2 ± 0.5 vs. 4.3 ± 0.7, P = 0.123), artifact presence (4.3 ± 0.6 vs. 4.1 ± 0.8, P = 0.085), and overall image quality (4.4 ± 0.4 vs. 4.3 ± 0.6, P = 0.448), showed no significant differences between the two methods. Representative RV and LV functional parameters - including RVEDV (102.2 (86.4, 120.4) ml vs. 104.0 (88.5, 120.3) ml, P = 0.294), RVEF (31.0 ± 11.1 % vs. 31.2 ± 11.0 %, P = 0.570), and LVEDV (106.2 (86.7, 131.3) ml vs. 105.8 (84.4, 130.3) ml, P = 0.450) - also did not differ significantly between the two methods. Strong correlations (r > 0.900) and excellent agreement (ICC > 0.900) were found for all biventricular functional parameters between the two sequences. In subgroups with reduced LVEF (<50 %, n = 24) or elevated heart rate (≥80  bpm, n = 17), no significant differences were observed in any biventricular functional metrics (P > 0.05 for all) between the two sequences. In comparison to multiple BH cine, the FB AI MOCO cine achieved comparable image quality and biventricular functional parameters with shorter scan times, suggesting its promising potential for clinical applications.

Amide Proton Transfer (APT) technique is a novel functional MRI technique that enables quantification of protein metabolism, but its wide application is largely limited in clinical settings by its long acquisition time. One way to reduce the scanning time is to obtain fewer frequency offset images during image acquisition. However, sparse frequency offset images are not inadequate to fit the z-spectral, a curve essential to quantifying the APT effect, which might compromise its quantification. In our study, we develop a deep learning-based model that allows for reconstructing dense frequency offsets from sparse ones, potentially reducing scanning time. We propose to leverage time-series convolution to extract both short and long-range spatial and frequency features of the APT imaging sequence. Our proposed model outperforms other seq2seq models, achieving superior reconstruction with a peak signal-to-noise ratio of 45.8 (95% confidence interval (CI): [44.9 46.7]), and a structural similarity index of 0.989 (95% CI:[0.987 0.993]) for the tumor region. We have integrated a weighted layer into our model to evaluate the impact of individual frequency offset on the reconstruction process. The weights assigned to the frequency offset at ±6.5 ppm, 0 ppm, and 3.5 ppm demonstrate higher significance as learned by the model. Experimental results demonstrate that our proposed model effectively reconstructs dense frequency offsets (n = 29, from 7 to -7 with 0.5 ppm as an interval) from data with 21 frequency offsets, reducing scanning time by 25%. This work presents a method for shortening the APT imaging acquisition time, offering potential guidance for parameter settings in APT imaging and serving as a valuable reference for clinicians.

The recent popularity of foundation models and the pre-train-and-adapt paradigm, where a large-scale model is transferred to downstream tasks, is gaining attention for volumetric medical image segmentation. However, current transfer learning strategies devoted to full fine-tuning for transfer learning may require significant resources and yield sub-optimal results when the labeled data of the target task is scarce. This makes its applicability in real clinical settings challenging since these institutions are usually constrained on data and computational resources to develop proprietary solutions. To address this challenge, we formalize Few-Shot Efficient Fine-Tuning (FSEFT), a novel and realistic scenario for adapting medical image segmentation foundation models. This setting considers the key role of both data- and parameter-efficiency during adaptation. Building on a foundation model pre-trained on open-access CT organ segmentation sources, we propose leveraging Parameter-Efficient Fine-Tuning and black-box Adapters to address such challenges. Furthermore, novel efficient adaptation methodologies are introduced in this work, which include Spatial black-box Adapters that are more appropriate for dense prediction tasks and constrained transductive inference, leveraging task-specific prior knowledge. Our comprehensive transfer learning experiments confirm the suitability of foundation models in medical image segmentation and unveil the limitations of popular fine-tuning strategies in few-shot scenarios. The project code is available: https://github.com/jusiro/fewshot-finetuning.

The present study aimed to develop and validate a fusion model based on multi-phase contrast-enhanced computed tomography (CECT) radiomics features combined with clinical features to preoperatively predict the expression levels of Ki-67 in patients with gastric cancer (GC). A total of 164 patients with GC who underwent surgical treatment at our hospital between September 2015 and September 2023 were retrospectively included and were randomly divided into a training set (n=114) and a testing set (n=50). Using Pyradiomics, radiomics features were extracted from multi-phase CECT images and were combined with significant clinical features through various machine learning algorithms [support vector machine (SVM), random forest (RandomForest), K-nearest neighbors (KNN), LightGBM and XGBoost] to build a fusion model. Receiver operating characteristic, area under the curve (AUC), calibration curve and decision curve analysis (DCA) were used to evaluate, validate and compare the predictive performance and clinical utility of the model. Among the three single-phase models, for the arterial phase model, the SVM radiomics model had the highest AUC value in the training set, which was 0.697; and the RandomForest radiomics model had the highest AUC value in the testing set, which was 0.658. For the venous phase model, the SVM radiomics model had the highest AUC value in the training set, which was 0.783; and the LightGBM radiomics model had the highest AUC value in the testing set, which was 0.747. For the delayed phase model, the KNN radiomics model had the highest AUC value in the training set, which was 0.772; and the SVM radiomics model had the highest AUC in the testing set, which was 0.719. The clinical feature model had the lowest AUC values in both the training set and the testing set, which were 0.614 and 0.520, respectively. Notably, the multi-phase model and the fusion model, which were constructed by combining the clinical features and the multi-phase features, demonstrated excellent discriminative performance, with the fusion model achieving AUC values of 0.933 and 0.817 in the training and testing sets, thus outperforming other models (DeLong test, both P<0.05). The calibration curve showed that the fusion model had goodness of fit (Hosmer-Lemeshow test, >0.5 in the training and validation sets). The DCA showed that the net benefit of the fusion model in identifying high expression of Ki-67 was improved compared with that of other models. Furthermore, the fusion model achieved an AUC value of 0.805 in the external validation data from The Cancer Imaging Archive. In conclusion, the fusion model established in the present study was revealed to have excellent performance and is expected to serve as a non-invasive tool for predicting Ki-67 status and guiding clinical treatment.

Osteoarthritis is a degenerative condition that affects bones and cartilage, often leading to structural changes, including osteophyte formation, bone density loss, and the narrowing of joint spaces. Over time, this process may disrupt the glenohumeral (GH) joint functionality, requiring a targeted treatment. Various options are available to restore joint functions, ranging from conservative management to surgical interventions, depending on the severity of the condition. This work introduces an innovative deep learning framework to process shoulder CT scans. It features the semantic segmentation of the proximal humerus and scapula, the 3D reconstruction of bone surfaces, the identification of the GH joint region, and the staging of three common osteoarthritic-related conditions: osteophyte formation (OS), GH space reduction (JS), and humeroscapular alignment (HSA). Each condition was stratified into multiple severity stages, offering a comprehensive analysis of shoulder bone structure pathology. The pipeline comprised two cascaded CNN architectures: 3D CEL-UNet for segmentation and 3D Arthro-Net for threefold classification. A retrospective dataset of 571 CT scans featuring patients with various degrees of GH osteoarthritic-related pathologies was used to train, validate, and test the pipeline. Root mean squared error and Hausdorff distance median values for 3D reconstruction were 0.22 mm and 1.48 mm for the humerus and 0.24 mm and 1.48 mm for the scapula, outperforming state-of-the-art architectures and making it potentially suitable for a PSI-based shoulder arthroplasty preoperative plan context. The classification accuracy for OS, JS, and HSA consistently reached around 90% across all three categories. The computational time for the entire inference pipeline was less than 15 s, showcasing the framework's efficiency and compatibility with orthopedic radiology practice. The achieved reconstruction and classification accuracy, combined with the rapid processing time, represent a promising advancement towards the medical translation of artificial intelligence tools. This progress aims to streamline the preoperative planning pipeline, delivering high-quality bone surfaces and supporting surgeons in selecting the most suitable surgical approach according to the unique patient joint conditions.

Magnetic resonance imaging (MRI) is a widely used radiological modality renowned for its radiation-free, comprehensive insights into the human body, facilitating medical diagnoses. However, the drawback of prolonged scan times hinders its accessibility. The k-space undersampling offers a solution, yet the resultant artifacts necessitate meticulous removal during image reconstruction. Although deep learning (DL) has proven effective for fast MRI image reconstruction, its broader applicability across various imaging scenarios has been constrained. Challenges include the high cost and privacy restrictions associated with acquiring large-scale, diverse training data, coupled with the inherent difficulty of addressing mismatches between training and target data in existing DL methodologies. Here, we present a novel Physics-Informed Synthetic data learning Framework for fast MRI, called PISF. PISF marks a breakthrough by enabling generalizable DL for multi-scenario MRI reconstruction through a single trained model. Our approach separates the reconstruction of a 2D image into many 1D basic problems, commencing with 1D data synthesis to facilitate generalization. We demonstrate that training DL models on synthetic data, coupled with enhanced learning techniques, yields in vivo MRI reconstructions comparable to or surpassing those of models trained on matched realistic datasets, reducing the reliance on real-world MRI data by up to 96 %. With a single trained model, our PISF supports the high-quality reconstruction under 4 sampling patterns, 5 anatomies, 6 contrasts, 5 vendors, and 7 centers, exhibiting remarkable generalizability. Its adaptability to 2 neuro and 2 cardiovascular patient populations has been validated through evaluations by 10 experienced medical professionals. In summary, PISF presents a feasible and cost-effective way to significantly boost the widespread adoption of DL in various fast MRI applications.