MRI Tomoelastography to Assess the Combined Status of Vessels Encapsulating Tumor Clusters and Microvascular Invasion in Hepatocellular Carcinoma

BACKGROUND

Integrating vessels encapsulating tumor clusters (VETC) and microvascular invasion (MVI) (VM hereafter) is potentially useful in risk stratification of hepatocellular carcinoma (HCC). However, noninvasive assessment methods for VM are lacking.

PURPOSE

To investigate the diagnostic performance of tomoelastography in assessing the VM status in HCC.

STUDY TYPE

Retrospective.

POPULATION

One hundred sixty-eight patients with surgically confirmed HCC consisting of 115 training and 53 validation cohorts, divided into negative-VM and positive-VM groups with mild or severe-VMs. Of them, 127 patients completed the follow-up (median: 26.1 months).

FIELD STRENGTH/SEQUENCE

3D multifrequency tomoelastography with a single-shot spin-echo echo-planar imaging sequence, and liver MRI including T1-weighted in-phase and opposed-phase gradient echo (GRE), T2-weighted turbo spin echo, diffusion-weighted imaging and dynamic contrast-enhanced T1-weighted GRE sequences at 3.0 T.

ASSESSMENT

Shear wave speed (c) and phase angle of the shear modulus (φ) were calculated on tomoelastograms. Imaging features were visually analyzed and clinical features were collected. Conventional models used clinical and imaging features while nomograms combined tomoelastography, clinical and imaging features.

STATISTICAL TESTS

Univariable and multivariable logistic regression analyses, nomogram, area under the receiver operating characteristic curve (AUC), DeLong test, Kaplan-Meier analysis and log-rank test. P < 0.05 was considered statistically significant.

RESULTS

Tumor-to-liver parenchyma ratio of c (cr) and tumor c were independent risk factors for positive-VM and severe-VM, respectively. In validation cohort, the nomograms including cr and tumor c performed significantly better than the conventional models for diagnosing positive-VM (0.84 [95% CI: 0.72-0.93] vs. 0.77 [95% CI: 0.64-0.88]) and severe-VM (0.86 [95% CI: 0.68-0.96] vs. 0.75 [95% CI: 0.55-0.89]). Patients with estimated positive-VM (9.3 months)/severe-VM (9.2 months) based on nomograms had shorter median recurrence-free survival than those with estimated negative-VM (>20.0 months)/mild-VM (18.0 months) in validation cohort.

DATA CONCLUSION

Tomoelastography based-nomograms showed good performance for noninvasively assessing VM status in patients with HCC.

EVIDENCE LEVEL

3 TECHNICAL EFFICACY: Stage 2.

Magnetic Resonance Elastography Combined With PI-RADS v2.1 for the Identification of Clinically Significant Prostate Cancer

BACKGROUND

Multiparametric MRI may cause overdiagnosis of clinically significant prostate cancer (csPCa) with the Prostate Imaging Reporting and Data System version 2.1 (PI-RADS v2.1).

OBJECTIVES

To investigate the diagnostic performance of stiffness as a standalone and complementary marker to PI-RADS v2.1 for diagnosing csPCa.

STUDY TYPE

Prospective.

SUBJECTS

One hundred forty-seven participants with pathologically confirmed prostate lesions (≥1 cm), including 71 with csPCa.

FIELD STRENGTH/SEQUENCE

T1-weighted fast spin-echo, T2-weighted fast spin-echo, single-shot echo-planar diffusion-weighted imaging, fast 3D gradient-echo T1-weighted dynamic contrast-enhanced imaging, and 3D single-shot spin-echo based echo-planar MR elastography at 3.0 T.

ASSESSMENT

The PI-RADS v2.1 score was assessed by three radiologists independently. Lesion shear stiffness (SS) values at 60 Hz and 90 Hz were measured. A modified PI-RADS integrating stiffness with PI-RADS v2.1 was developed. Diagnostic performance for csPCa was compared between stiffness, PI-RADS v2.1 and the modified PI-RADS.

STATISTICAL TEST

Spearman's correlation, Fleiss κ and intraclass correlation, Pearson correlation, one-way analysis of variance, area under the receiver operating characteristic curve (AUC), and the Delong test. Significance level was P < 0.05.

RESULTS

In the peripheral zone, csPCa (N = 35) had significantly higher SS than non-csPCa at 60 Hz (3.22 ± 0.66 kPa vs. 2.56 ± 0.56 kPa) and at 90 Hz (5.64 ± 1.30 kPa vs. 4.48 ± 0.84 kPa). PI-RADS v2.1 showed 100% sensitivity, 58% specificity, and 0.79 AUC for detecting csPCa. SS achieved 97% sensitivity, 52% specificity, and 0.80 AUC at 60 Hz, while SS had 63% sensitivity, 87% specificity, and 0.78 AUC at 90 Hz. The modified PI-RADS, combing SS at 60 Hz with PI-RADS v2.1, resulted in a significantly increased AUC (0.86) compared to that of PI-RADS v2.1, with a sensitivity of 97% and specificity of 75%.

DATA CONCLUSION

Stiffness can help identifying csPCa in the peripheral zone. Combining stiffness with the PI-RADS v2.1 improved the diagnostic accuracy and specificity for csPCa.

EVIDENCE LEVEL

1 TECHNICAL EFFICACY: Stage 2.

MMR-Mamba: Multi-modal MRI reconstruction with Mamba and spatial-frequency information fusion

Multi-modal MRI offers valuable complementary information for diagnosis and treatment; however, its clinical utility is limited by prolonged scanning time. To accelerate the acquisition process, a practical approach is to reconstruct images of the target modality, which requires longer scanning time, from under-sampled k-space data using the fully-sampled reference modality with shorter scanning time as guidance. The primary challenge of this task lies in comprehensively and efficiently integrating complementary information from different modalities to achieve high-quality reconstruction. Existing methods struggle with this challenge: (1) convolution-based models fail to capture long-range dependencies; (2) transformer-based models, while excelling in global feature modeling, suffer from quadratic computational complexity. To address this dilemma, we propose MMR-Mamba, a novel framework that thoroughly and efficiently integrates multi-modal features for MRI reconstruction, leveraging Mamba's capability to capture long-range dependencies with linear computational complexity while exploiting global properties of the Fourier domain. Specifically, we first design a Target modality-guided Cross Mamba (TCM) module in the spatial domain, which maximally restores the target modality information by selectively incorporating relevant information from the reference modality. Then, we introduce a Selective Frequency Fusion (SFF) module to efficiently integrate global information in the Fourier domain and recover high-frequency signals for the reconstruction of structural details. Furthermore, we devise an Adaptive Spatial-Frequency Fusion (ASFF) module, which mutually enhances the spatial and frequency domains by supplementing less informative channels from one domain with corresponding channels from the other. Extensive experiments on the BraTS and fastMRI knee datasets demonstrate the superiority of our MMR-Mamba over state-of-the-art reconstruction methods. The code is publicly available at https://github.com/zoujing925/MMR-Mamba.

Impact of frailty on the long-term prognosis of the elderly with hepatocellular carcinoma treated with transarterial chemoembolization

Transcatheter arterial chemoembolization (TACE) is a standard treatment for unresectable, intermediate, or advanced-stage hepatocellular carcinoma (HCC), especially in elderly patients. The modified 5-item frailty index (mFI-5) is a concise, comorbidity-based risk stratification tool proven to effectively predict adverse outcomes. However, the prognostic capacity of the mFI-5 in elderly HCC patients after TACE is unclear. As such, we retrospectively analyzed clinical data from elderly HCC patients (age ≥ 65 years) who received their first TACE between November 2018 and November 2020 at the Department of Interventional Radiology, Affiliated Hospital of Xuzhou Medical University. The mFI-5 was calculated based on the presence of five co-morbidities: congestive heart failure within 30 days prior to surgery; insulin-dependent or noninsulin-dependent diabetes mellitus; chronic obstructive pulmonary disease (COPD) or pneumonia; partially dependent or totally dependent functional health status at time of surgery; and hypertension requiring medication. Patients were divided into two groups based on mFI-5 scores: mFI-5 ≥ 2 ('frail') and mFI-5 < 2 ('non-frail'). The primary outcomes were overall survival (OS) and progression-free survival (PFS).Among the 143 patients, 97 were in the mFI-5 < 2 group and 46 in the mFI-5 ≥ 2 group. The median OS was 40.0 months (95% confidence interval [CI]: 35.0-47.0) in the non-frail group vs. 24.0 months (95% CI: 22-NA) in the frail group (hazard ratio [HR] = 3.343, 95% CI: 1.802-6.201, p < .001). The median PFS was 7.0 months (95% CI: 4.0-11.0) vs. 3.0 months (95% CI: 2.0-9.0) (HR = 1.507, 95% CI: 0.996-2.280, p = .053). Cox regression identified mFI-5, alpha-fetoprotein (AFP) as independent predictors of OS. The mFI-5 is a useful predictor of long-term survival in elderly HCC patients treated with TACE, suggesting that incorporating frailty assessments can optimize treatment strategies and improve outcomes.

The state-of-the-art in cardiac MRI reconstruction: Results of the CMRxRecon challenge in MICCAI 2023

Cardiac magnetic resonance imaging (MRI) provides detailed and quantitative evaluation of the heart's structure, function, and tissue characteristics with high-resolution spatial-temporal imaging. However, its slow imaging speed and motion artifacts are notable limitations. Undersampling reconstruction, especially data-driven algorithms, has emerged as a promising solution to accelerate scans and enhance imaging performance using highly under-sampled data. Nevertheless, the scarcity of publicly available cardiac k-space datasets and evaluation platform hinder the development of data-driven reconstruction algorithms. To address this issue, we organized the Cardiac MRI Reconstruction Challenge (CMRxRecon) in 2023, in collaboration with the 26th International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI). CMRxRecon presented an extensive k-space dataset comprising cine and mapping raw data, accompanied by detailed annotations of cardiac anatomical structures. With overwhelming participation, the challenge attracted more than 285 teams and over 600 participants. Among them, 22 teams successfully submitted Docker containers for the testing phase, with 7 teams submitted for both cine and mapping tasks. All teams use deep learning based approaches, indicating that deep learning has predominately become a promising solution for the problem. The first-place winner of both tasks utilizes the E2E-VarNet architecture as backbones. In contrast, U-Net is still the most popular backbone for both multi-coil and single-coil reconstructions. This paper provides a comprehensive overview of the challenge design, presents a summary of the submitted results, reviews the employed methods, and offers an in-depth discussion that aims to inspire future advancements in cardiac MRI reconstruction models. The summary emphasizes the effective strategies observed in Cardiac MRI reconstruction, including backbone architecture, loss function, pre-processing techniques, physical modeling, and model complexity, thereby providing valuable insights for further developments in this field.

Histogram analysis of multiple mathematical diffusion-weighted imaging models for preoperative prediction of Ki-67 expression in hepatocellular carcinoma

Objective

To explore whether a combination of clinico-radiological factors and histogram parameters based on monoexponential, biexponential, and stretched exponential models derived from the whole-tumor volume on diffusion-weighted imaging (DWI) could predict Ki-67 expression in hepatocellular carcinoma(HCC).

Materials and Methods

Histogram parameters based on whole-tumor volumes were derived from monoexponential model, biexponential model, and stretched exponential model. Histogram parameters were compared between HCCs with high and low Ki-67 expression. Multivariate logistic regression and receiver operating characteristic curves were used to assess the ability to predict Ki-67 expression (expression index ≤ 20% vs. >20%).

Results

In the training and test set, the 5th percentile of distributed diffusion coefficient (DDC) yielded the area under the curve (AUC) value of 0.816 (95% CI 0.713 to 0.894) and 0.867 (95% CI 0.655 to 0.972), respectively. Multivariable analysis showed that alpha-fetoprotein (AFP) level, skewness of perfusion fraction(f), and 5th percentile of DDC were independent predictors of high Ki-67 expression in HCCs. In the training and test sets, the AUC of the combined model for predicting high Ki-67 expression in HCCs were 0.902 (95% CI 0.814 to 0.957) and 0.908 (95% CI 0.707 to 0.989), respectively.

Conclusion

Histogram parameters of multiple mathematical DWI models can be useful for predicting high Ki-67 expression in HCCs, and our combined model based on AFP level, skewness of f, and 5th percentile of DDC may be an effective approach for predicting Ki-67 expression in HCCs.

CMRxRecon2024: A Multimodality, Multiview k-Space Dataset Boosting Universal Machine Learning for Accelerated Cardiac MRI

The released CMRxRecon2024 dataset is currently the largest and most protocol-diverse publicly available k-space dataset including multimodality and multiview cardiac MRI data from 330 healthy volunteers, and each one covers standardized and commonly used clinical protocols.

Attention incorporated network for sharing low-rank, image and k-space information during MR image reconstruction to achieve single breath-hold cardiac Cine imaging

Cardiac Cine Magnetic Resonance Imaging (MRI) provides an accurate assessment of heart morphology and function in clinical practice. However, MRI requires long acquisition times, with recent deep learning-based methods showing great promise to accelerate imaging and enhance reconstruction quality. Existing networks exhibit some common limitations that constrain further acceleration possibilities, including single-domain learning, reliance on a single regularization term, and equal feature contribution. To address these limitations, we propose to embed information from multiple domains, including low-rank, image, and k-space, in a novel deep learning network for MRI reconstruction, which we denote as A-LIKNet. A-LIKNet adopts a parallel-branch structure, enabling independent learning in the k-space and image domain. Coupled information sharing layers realize the information exchange between domains. Furthermore, we introduce attention mechanisms into the network to assign greater weights to more critical coils or important temporal frames. Training and testing were conducted on an in-house dataset, including 91 cardiovascular patients and 38 healthy subjects scanned with 2D cardiac Cine using retrospective undersampling. Additionally, we evaluated A-LIKNet on the real-time prospectively undersampled data from the OCMR dataset. The results demonstrate that our proposed A-LIKNet outperforms existing methods and provides high-quality reconstructions. The network can effectively reconstruct highly retrospectively undersampled dynamic MR images up to 24× accelerations, indicating its potential for single breath-hold imaging.

Ultrasound-based artificial intelligence model for prediction of Ki-67 proliferation index in soft tissue tumors

RATIONALE AND OBJECTIVES

To investigate the value of deep learning (DL) combined with radiomics and clinical and imaging features in predicting the Ki-67 proliferation index of soft tissue tumors (STTs).

MATERIALS AND METHODS

In this retrospective study, a total of 394 patients with STTs admitted from January 2021 to December 2023 in two separate hospitals were collected. Hospital-1 was the training cohort (323 cases, of which 89 and 234 were high and low Ki-67, respectively) and Hospital-2 was the external validation cohort (71 cases, of which 23 and 48 were high and low Ki-67, respectively). Clinical and ultrasound characteristics including age, sex, tumor size, morphology, margins, internal echoes and blood flow were assessed. Risk factors with significant correlations were screened by univariate and multivariate logistic regression analyses. After extracting the radiomics and DL features, the feature fusion model is constructed by Support Vector Machine. The prediction results obtained from separate clinical features, radiomics features and DL features were combined to construct decision fusion models. Finally, the DeLong test was used to compare whether the AUCs between the models were significantly different.

RESULTS

The three feature fusion models and three decision fusion models constructed demonstrated excellent diagnostic performance in predicting Ki-67 expression levels in STTs. Among them, the feature fusion model based on clinical, radiomics, and DL performed the best with an AUC of 0.911 (95% CI: 0.886-0.935) in the training cohort and 0.923 (95% CI: 0.873-0.972) in the validation cohort, and proved to be well-calibrated and clinically useful. The DeLong test showed that the decision fusion models based on clinical, radiomics and DL performed significantly worse than the three feature fusion models on the validation set. There was no statistical difference in diagnostic performance between the other models.

CONCLUSION

The ultrasound-based fusion model of clinical, radiomics, and DL features showed good performance in predicting Ki-67 expression levels in STTs.

Systematic Review: AI Applications in Liver Imaging with a Focus on Segmentation and Detection

The liver is a frequent focus in radiology due to its diverse pathology, and artificial intelligence (AI) could improve diagnosis and management. This systematic review aimed to assess and categorize research studies on AI applications in liver radiology from 2018 to 2024, classifying them according to areas of interest (AOIs), AI task and imaging modality used. We excluded reviews and non-liver and non-radiology studies. Using the PRISMA guidelines, we identified 6680 articles from the PubMed/Medline, Scopus and Web of Science databases; 1232 were found to be eligible. A further analysis of a subgroup of 329 studies focused on detection and/or segmentation tasks was performed. Liver lesions were the main AOI and CT was the most popular modality, while classification was the predominant AI task. Most detection and/or segmentation studies (48.02%) used only public datasets, and 27.65% used only one public dataset. Code sharing was practiced by 10.94% of these articles. This review highlights the predominance of classification tasks, especially applied to liver lesion imaging, most often using CT imaging. Detection and/or segmentation tasks relied mostly on public datasets, while external testing and code sharing were lacking. Future research should explore multi-task models and improve dataset availability to enhance AI's clinical impact in liver imaging.

Imaging-Based Prediction of Ki-67 Expression in Hepatocellular Carcinoma: A Retrospective Study

AIM

This study aims to develop a non-invasive, preoperative predictive model for Ki-67 expression in HCC patients using enhanced computed tomography (CT) and clinical indicators to improve patient outcomes.

METHODS

This retrospective study analyzed 595 post-curative hepatectomy HCC patients. Patients were categorized into high (> 20%) and low (≤ 20%) Ki-67 expression groups based on cellular proliferation levels. Radiomic features were extracted from enhanced CT scans and combined with clinical parameters to develop a predictive model for Ki-67 expression.

RESULTS

Key clinical factors impacting Ki-67 expression in HCC included alpha-fetoprotein (AFP), non-smooth tumor margin, ill-defined pseudo-capsule, and peritumoral star node. From 1441 initially extracted radiomic features, 16 key features were selected using Lasso regression. These features were used to develop a radiomics model, which, when combined with clinical data, yielded an integrated predictive model with high accuracy. The combined model achieved an area under the curve (AUC) of 0.854 in the training group and 0.839 in the validation group. A nomogram based on this model was constructed, and its predictive accuracy was validated through calibration curves and decision curve analysis. A risk scorecard model was also constructed as a practical tool for clinicians to assess the risk level of high Ki-67 expression, facilitating personalized treatment planning. Survival analysis demonstrated significant differences in 3-year overall survival (OS) and progression-free survival (PFS) rates between patients with high and low Ki-67 expression, indicating the model's strong prognostic capability.

CONCLUSIONS

This study successfully developed a comprehensive model that integrates radiomic and clinical data for the preoperative prediction of Ki-67 expression in HCC patients.

Data- and Physics-Driven Deep Learning Based Reconstruction for Fast MRI: Fundamentals and Methodologies

Magnetic Resonance Imaging (MRI) is a pivotal clinical diagnostic tool, yet its extended scanning times often compromise patient comfort and image quality, especially in volumetric, temporal and quantitative scans. This review elucidates recent advances in MRI acceleration via data and physics-driven models, leveraging techniques from algorithm unrolling models, enhancement-based methods, and plug-and-play models to the emerging full spectrum of generative model-based methods. We also explore the synergistic integration of data models with physics-based insights, encompassing the advancements in multi-coil hardware accelerations like parallel imaging and simultaneous multi-slice imaging, and the optimization of sampling patterns. We then focus on domain-specific challenges and opportunities, including image redundancy exploitation, image integrity, evaluation metrics, data heterogeneity, and model generalization. This work also discusses potential solutions and future research directions, with an emphasis on the role of data harmonization and federated learning for further improving the general applicability and performance of these methods in MRI reconstruction.

Clinical Evaluation of 3D Motion-Correction Via Scout Accelerated Motion Estimation and Reduction Framework Versus Conventional T1-Weighted MRI at 1.5 T in Brain Imaging

OBJECTIVES

The aim of this study was to investigate the occurrence of motion artifacts and image quality of brain magnetic resonance imaging (MRI) T1-weighted imaging applying 3D motion correction via the Scout Accelerated Motion Estimation and Reduction (SAMER) framework compared with conventional T1-weighted imaging at 1.5 T.

MATERIALS AND METHODS

A preliminary study involving 14 healthy volunteers assessed the impact of the SAMER framework on induced motion during 3 T MRI scans. Participants performed 3 different motion patterns: (1) step up, (2) controlled breathing, and (3) free motion. The patient study included 82 patients who required clinically indicated MRI scans. 3D T1-weighted images (MPRAGE) were acquired at 1.5 T. The MRI data were reconstructed using either regular product reconstruction (non-Moco) or the 3D motion correction SAMER framework (SAMER Moco), resulting in 145 image sequences. For the preliminary and the patient study, 3 experienced radiologists evaluated the image data using a 5-point Likert scale, focusing on overall image quality, artifact presence, diagnostic confidence, delineation of pathology, and image sharpness. Interrater agreement was assessed using Gwet's AC2, and an exploratory analysis (non-Moco vs SAMER Moco) was performed.

RESULTS

Compared with non-Moco, the preliminary study demonstrated significant improvements across all imaging parameters and motion patterns with SAMER Moco (P < 0.001). Odds ratios favoring SAMER Moco were >999.999 for freedom of artifact and overall image quality (P < 0.0001). Excellent or good ratings for freedom of artifact were 52.4% with SAMER Moco, compared with 21.4% for non-Moco. Similarly, 66.7% of SAMER Moco images were rated excellent or good for overall image quality versus 21.4% for non-Moco. Multireader interrater agreement was excellent across all parameters.The patient study confirmed that SAMER Moco provided significantly superior image quality across all evaluated imaging parameters, particularly in the presence of motion (P < 0.001). Diagnostic confidence was rated as excellent or good in 95.1% of SAMER Moco cases, compared with 78.1% for non-Moco cases. Similarly, overall image quality was rated as excellent or good in 89.8% of SAMER Moco cases versus 65.9% for non-Moco cases. The odds ratios for diagnostic confidence and for overall image quality were 6.698 and 6.030, respectively, both favoring SAMER Moco (P < 0.0001). Multireader interrater agreement was excellent across all parameters.

CONCLUSIONS

The application of SAMER in T1-weighted imaging datasets is feasible in clinical routine and significantly increases image quality and diagnostic confidence in 1.5 T brain MRI by effectively reducing motion artifacts.

Cardiac MR image reconstruction using cascaded hybrid dual domain deep learning framework

Recovering diagnostic-quality cardiac MR images from highly under-sampled data is a current research focus, particularly in addressing cardiac and respiratory motion. Techniques such as Compressed Sensing (CS) and Parallel Imaging (pMRI) have been proposed to accelerate MRI data acquisition and improve image quality. However, these methods have limitations in high spatial-resolution applications, often resulting in blurring or residual artifacts. Recently, deep learning-based techniques have gained attention for their accuracy and efficiency in image reconstruction. Deep learning-based MR image reconstruction methods are divided into two categories: (a) single domain methods (image domain learning and k-space domain learning) and (b) cross/dual domain methods. Single domain methods, which typically use U-Net in either the image or k-space domain, fail to fully exploit the correlation between these domains. This paper introduces a dual-domain deep learning approach that incorporates multi-coil data consistency (MCDC) layers for reconstructing cardiac MR images from 1-D Variable Density (VD) random under-sampled data. The proposed hybrid dual-domain deep learning models integrate data from both the domains to improve image quality, reduce artifacts, and enhance overall robustness and accuracy of the reconstruction process. Experimental results demonstrate that the proposed methods outperform than conventional deep learning and CS techniques, as evidenced by higher Structural Similarity Index (SSIM), lower Root Mean Square Error (RMSE), and higher Peak Signal-to-Noise Ratio (PSNR).

Application of dynamic enhanced scanning with GD-EOB-DTPA MRI based on deep learning algorithm for lesion diagnosis in liver cancer patients

Background

Improvements in the clinical diagnostic use of magnetic resonance imaging (MRI) for the identification of liver disorders have been made possible by gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid (Gd-EOB-DTPA). Gd-EOB-DTPA-enhanced magnetic resonance imaging (MRI) technology is in high demand.

Objectives

The purpose of the study is to segment the liver using an enhanced multi-gradient deep convolution neural network (EMGDCNN) and to identify and categorize a localized liver lesion using a Gd-EOB-DTPA-enhanced MRI.

Methods

We provided the classifier images of the liver in five states (unenhanced, arterial, portal venous, equilibrium, and hepatobiliary) and labeled them with localized liver diseases (hepatocellular carcinoma, metastasis, hemangiomas, cysts, and scarring). The Shanghai Public Health Clinical Center ethics committee recruited 132 participants between August 2021 and February 2022. Fisher's exact test analyses liver lesion Gd-EOB-DTPA-enhanced MRI data.

Results

Our method could identify and classify liver lesions at the same time. On average, 25 false positives and 0.6 real positives were found in the test instances. The percentage of correct answers was 0.790. AUC, sensitivity, and specificity evaluate the procedure. Our technique outperforms others in extensive testing.

Conclusion

EMGDCNN may identify and categorize a localized hepatic lesion in Gd-EOB-DTPA-enhanced MRI. We found that one network can detect and classify. Radiologists need higher detection capability.

Radiomics predicting immunohistochemical markers in primary hepatic carcinoma: Current status and challenges

Primary hepatic carcinoma, comprising hepatocellular carcinoma (HCC), intrahepatic cholangiocarcinoma (ICC), and combined hepatocellular cholangiocarcinoma (cHCC-CCA), ranks among the most common malignancies worldwide. The heterogeneity of tumors is a primary factor impeding the efficacy of treatments for primary hepatic carcinoma. Immunohistochemical markers may play a potential role in characterizing this heterogeneity, providing significant guidance for prognostic analysis and the development of personalized treatment plans for the patients with primary hepatic carcinoma. Currently, primary hepatic carcinoma immunohistochemical analysis primarily relies on invasive techniques such as surgical pathology and tissue biopsy. Consequently, the non-invasive preoperative acquisition of primary hepatic carcinoma immunohistochemistry has emerged as a focal point of research. As an emerging non-invasive diagnostic technique, radiomics possesses the potential to extensively characterize tumor heterogeneity. It can predict immunohistochemical markers associated with hepatocellular carcinoma preoperatively, demonstrating significant auxiliary utility in clinical guidance. This article summarizes the progress in using radiomics to predict immunohistochemical markers in primary hepatic carcinoma, addresses the challenges faced in this field of study, and anticipates its future application prospects.

Three-Dimensional Multifrequency MR Elastography for Microvascular Invasion and Prognosis Assessment in Hepatocellular Carcinoma

BACKGROUND

Pretreatment identification of microvascular invasion (MVI) in hepatocellular carcinoma (HCC) is important when selecting treatment strategies.

PURPOSE

To improve models for predicting MVI and recurrence-free survival (RFS) by developing nomograms containing three-dimensional (3D) MR elastography (MRE).

STUDY TYPE

Prospective.

POPULATION

188 patients with HCC, divided into a training cohort (n = 150) and a validation cohort (n = 38). In the training cohort, 106/150 patients completed a 2-year follow-up.

FIELD STRENGTH/SEQUENCE

1.5T 3D multifrequency MRE with a single-shot spin-echo echo planar imaging sequence, and 3.0T multiparametric MRI (mp-MRI), consisting of diffusion-weighted echo planar imaging, T2-weighted fast spin echo, in-phase out-of-phase T1-weighted fast spoiled gradient-recalled dual-echo and dynamic contrast-enhanced gradient echo sequences.

ASSESSMENT

Multivariable analysis was used to identify the independent predictors for MVI and RFS. Nomograms were constructed for visualization. Models for predicting MVI and RFS were built using mp-MRI parameters and a combination of mp-MRI and 3D MRE predictors.

STATISTICAL TESTS

Student's t-test, Mann-Whitney U test, chi-squared or Fisher's exact tests, multivariable analysis, area under the receiver operating characteristic curve (AUC), DeLong test, Kaplan-Meier analysis and log rank tests. P < 0.05 was considered significant.

RESULTS

Tumor c and liver c were independent predictors of MVI and RFS, respectively. Adding tumor c significantly improved the diagnostic performance of mp-MRI (AUC increased from 0.70 to 0.87) for MVI detection. Of the 106 patients in the training cohort who completed the 2-year follow up, 34 experienced recurrence. RFS was shorter for patients with MVI-positive histology than MVI-negative histology (27.1 months vs. >40 months). The MVI predicted by the 3D MRE model yielded similar results (26.9 months vs. >40 months). The MVI and RFS nomograms of the histologic-MVI and model-predicted MVI-positive showed good predictive performance.

DATA CONCLUSION

Biomechanical properties of 3D MRE were biomarkers for MVI and RFS. MVI and RFS nomograms were established.

LEVEL OF EVIDENCE

2 TECHNICAL EFFICACY: Stage 2.

A Protocol for Body MRI/CT and Extraction of Imaging-Derived Phenotypes (IDPs) from the China Phenobank Project

Currently, standard protocols for body imaging and corresponding image processing pipelines in population-based cohort studies are unavailable, limiting the applications of body imaging. Based on the China Phenobank Project (CHPP), the present study described a body imaging protocol for multiple organs, including cardiac structures, liver, spleen, pancreas, kidneys, lung, prostate, and uterus, and the corresponding image processing pipelines promoted its development. Briefly, the body imaging protocol comprised a 40-min cardiac magnetic resonance imaging (MRI) scan, a 5-min computed tomography (CT) scan, a 20-min abdominal MRI scan, and a 10-min pelvic MRI scan. The recommended image processing pipeline utilized deep learning segmentation models to facilitate the analysis of large amount of data. This study aimed to provide a reference for planning studies based on the CHPP platform.

A Feasibility Study of AI-Assisted Compressed Sensing in Prostate T2-Weighted Imaging

RATIONALE AND OBJECTIVES

To evaluate the image quality and PI-RADS scoring performance of prostate T2-weighted imaging (T2WI) based on AI-assisted compressed sensing (ACS).

MATERIALS AND METHODS

In this prospective study, adult male urological outpatients or inpatients underwent prostate MRI, including T2WI, diffusion-weighted imaging and apparent diffusion coefficient maps. Three accelerated scanning protocols using parallel imaging (PI) and ACS: T2WIPI, T2WIACS1 and T2WIACS2 were evaluated through comparative analysis. Quantitative analysis included signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), slope profile, and edge rise distance (ERD). Image quality was qualitatively assessed using a five-point Likert scale (ranging from 1 = non-diagnostic to 5 = excellent). PI-RADS scores were determined for the largest or most suspicious lesions in each patient. The Friedman test and one-way ANOVA with post hoc tests were utilized for group comparisons, with statistical significance set at P < 0.05.

RESULTS

This study included 40 participants. Compared to PI, ACS reduced acquisition time by over 50%, significantly enhancing the CNR of sagittal and axial T2WI (P < 0.05), significantly improving the image quality of sagittal and axial T2WI (P < 0.05). No significant differences were observed in slope profile, ERD, and PI-RADS scores between groups (P > 0.05).

CONCLUSION

ACS reduced prostate T2WI acquisition time by half while improving image quality without affecting PI-RADS scores.

Early detection of myocardial iron overload in patients with β-thalassemia major using cardiac magnetic resonance T1 mapping

BACKGROUND

The T2* technique, used for quantifying myocardial iron content (MIC), has limitations in detecting early myocardial iron overload (MIO). The in vivo mapping of the myocardial T1 relaxation time is a promising alternative for the early detection and management of MIO.

METHODS

32 β-thalassemia major (βTM) patients aged 11.5 ± 4 years and 32 healthy controls were recruited and underwent thorough clinical and laboratory assessments. The mid-level septal iron overload was measured through T1 mapping using a modified Look-Locker inversion recovery sequence with a 3 (3 s) 3 (3 s) 5 scheme. Septum was divided at the mentioned level into 3 zones corresponding to segments 8 and 9 in the cardiac segmentation model.

RESULTS

21.9 % of βTM had clinical cardiac morbidity. The cut-off of T1 mapping of hepatic and myocardium to differentiate between the patients and control groups was ≤466 and ≥ 923 ms respectively. The T1 technique was able to detect 4 patients with high MIC, two of them were not detected by the T2* technique. There was a statistically significant correlation between the average T1 values of the studied zones in patients with βTM and the liver iron content (LIC), the T1 values within segment 8 of the liver, age of patients, the age at first transfusion, age of splenectomy and serum ferritin value.

CONCLUSION

The addition of the T1 mapping sequence to the conventional T2* technique was able to increase the efficacy of the MIC detection protocol by earlier detection of MIO. This would guide chelation therapy to decrease myocardial morbidity.

Deep-learning-based image reconstruction with limited data: generating synthetic raw data using deep learning

OBJECT

Deep learning has shown great promise for fast reconstruction of accelerated MRI acquisitions by learning from large amounts of raw data. However, raw data is not always available in sufficient quantities. This study investigates synthetic data generation to complement small datasets and improve reconstruction quality.

MATERIALS AND METHODS

An adversarial auto-encoder was trained to generate phase and coil sensitivity maps from magnitude images, which were combined into synthetic raw data. On a fourfold accelerated MR reconstruction task, deep-learning-based reconstruction networks were trained with varying amounts of training data (20 to 160 scans). Test set performance was compared between baseline experiments and experiments that incorporated synthetic training data.

RESULTS

Training with synthetic raw data showed decreasing reconstruction errors with increasing amounts of training data, but importantly this was magnitude-only data, rather than real raw data. For small training sets, training with synthetic data decreased the mean absolute error (MAE) by up to 7.5%, whereas for larger training sets the MAE increased by up to 2.6%.

DISCUSSION

Synthetic raw data generation improved reconstruction quality in scenarios with limited training data. A major advantage of synthetic data generation is that it allows for the reuse of magnitude-only datasets, which are more readily available than raw datasets.

Comparison of different acceleration factors of artificial intelligence-compressed sensing for brachial plexus MRI imaging: scanning time and image quality

BACKGROUND

3D brachial plexus MRI scanning is prone to examination failure due to the lengthy scan times, which can lead to patient discomfort and motion artifacts. Our purpose is to investigate the efficacy of artificial intelligence-assisted compressed sensing (ACS) in improving the acceleration efficiency and maintaining or enhancing the image quality of brachial plexus MR imaging.

METHODS

A total of 30 volunteers underwent 3D sampling perfection with application-optimized contrast using different flip angle evolution short time inversion recovery using a 3.0T MR scanner. The imaging protocol included parallel imaging (PI) and ACS employing acceleration factors of 4.37, 6.22, and 9.03. Radiologists evaluated the neural detail display, fat suppression effectiveness, presence of image artifacts, and overall image quality. Signal intensity and standard deviation of specific anatomical sites within the brachial plexus and background tissues were measured, with signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) subsequently calculated. Cohen's weighted kappa (κ), One-way ANOVA, Kruskal-Wallis and pairwise comparisons with Bonferroni-adjusted significance level. P < 0.05 was considered statistically significant.

RESULTS

ACS significantly reduced scanning times compared to PI. Evaluations revealed differences in subjective scores and SNR across the sequences (P < 0.05), with no marked differences in CNR (P > 0.05). For subjective scores, ACS 9.03 were lower than the other three sequences in neural details display, image artifacts and overall image quality. There was no significant difference in fat suppression. For objective quantitative evaluation, SNR of right C6 root in ACS 6.22 and ACS 9.03 was higher than that in PI; SNR of left C6 root in ACS 4.37, ACS 6.22 and ACS 9.03 was higher than that in PI; SNR of medial cord in ACS 6.22, ACS 9.03 was higher than that in PI.

CONCLUSION

Compared with PI, ACS can shorten scanning time while ensuring good image quality.

Joint suppression of cardiac bSSFP cine banding and flow artifacts using twofold phase-cycling and a dual-encoder neural network

BACKGROUND

Cardiac balanced steady state free precession (bSSFP) cine imaging suffers from banding and flow artifacts induced by off-resonance. The work aimed to develop a twofold phase cycling sequence with a neural network-based reconstruction (2P-SSFP+Network) for a joint suppression of banding and flow artifacts in cardiac cine imaging.

METHODS

A dual-encoder neural network was trained on 1620 pairs of phase-cycled left ventricular (LV) cine images collected from 18 healthy subjects. Twenty healthy subjects and 25 patients were prospectively scanned using the proposed 2P-SSFP sequence. bSSFP cine of a single RF phase increment (1P-SSFP), bSSFP cine of a single radiofrequency (RF) phase increment with a network-based artifact reduction (1P-SSFP+Network), the averaging of the two phase-cycled images (2P-SSFP+Average), and the proposed method were mutually compared, in terms of artifact suppression performance in the LV, generalizability over altered scan parameters and scanners, suppression of large-area banding artifacts in the left atrium (LA), and accuracy of downstream segmentation tasks.

RESULTS

In the healthy subjects, 2P-SSFP+Network showed robust suppressions of artifacts across a range of phase combinations. Compared with 1P-SSFP and 2P-SSFP+Average, 2P-SSFP+Network improved banding artifacts (3.85 ± 0.67 and 4.50 ± 0.45 vs 5.00 ± 0.00, P < 0.01 and P = 0.02, respectively), flow artifacts (3.35 ± 0.78 and 2.10 ± 0.77 vs 4.90 ± 0.20, both P < 0.01), and overall image quality (3.25 ± 0.51 and 2.30 ± 0.60 vs 4.75 ± 0.25, both P < 0.01). 1P-SSFP+Network and 2P-SSFP+Network achieved a similar artifact suppression performance, yet the latter had fewer hallucinations (two-chamber, 4.25 ± 0.51 vs 4.85 ± 0.45, P = 0.04; four-chamber, 3.45 ± 1.21 vs 4.65 ± 0.50, P = 0.03; and left atrium (LA), 3.35 ± 1.00 vs 4.65 ± 0.45, P < 0.01). Furthermore, in the pulmonary veins and LA, 1P-SSFP+Network could not eliminate banding artifacts since they occupied a large area, whereas 2P-SSFP+Network reliably suppressed the artifacts. In the downstream automated myocardial segmentation task, 2P-SSFP+Network achieved more accurate segmentations than 1P-SSFP with different phase increments.

CONCLUSIONS

2P-SSFP+Network jointly suppresses banding and flow artifacts while manifesting a good generalizability against variations of anatomy and scan parameters. It provides a feasible solution for robust suppression of the two types of artifacts in bSSFP cine imaging.

A review of deep learning-based reconstruction methods for accelerated MRI using spatiotemporal and multi-contrast redundancies

Accelerated magnetic resonance imaging (MRI) has played an essential role in reducing data acquisition time for MRI. Acceleration can be achieved by acquiring fewer data points in k-space, which results in various artifacts in the image domain. Conventional reconstruction methods have resolved the artifacts by utilizing multi-coil information, but with limited robustness. Recently, numerous deep learning-based reconstruction methods have been developed, enabling outstanding reconstruction performances with higher acceleration. Advances in hardware and developments of specialized network architectures have produced such achievements. Besides, MRI signals contain various redundant information including multi-coil redundancy, multi-contrast redundancy, and spatiotemporal redundancy. Utilization of the redundant information combined with deep learning approaches allow not only higher acceleration, but also well-preserved details in the reconstructed images. Consequently, this review paper introduces the basic concepts of deep learning and conventional accelerated MRI reconstruction methods, followed by review of recent deep learning-based reconstruction methods that exploit various redundancies. Lastly, the paper concludes by discussing the challenges, limitations, and potential directions of future developments.

An unrolled neural network for accelerated dynamic MRI based on second-order half-quadratic splitting model

The reconstruction of dynamic magnetic resonance images from incomplete k-space data has sparked significant research interest due to its potential to reduce scan time. However, traditional iterative optimization algorithms fail to faithfully reconstruct images at higher acceleration factors and incur long reconstruction time. Furthermore, end-to-end deep learning-based reconstruction algorithms suffer from large model parameters and lack robustness in the reconstruction results. Recently, unrolled deep learning models, have shown immense potential in algorithm stability and applicability flexibility. In this paper, we propose an unrolled deep learning network based on a second-order Half-Quadratic Splitting(HQS) algorithm, where the forward propagation process of this framework strictly follows the computational flow of the HQS algorithm. In particular, we propose a degradation-sense module by associating random sampling patterns with intermediate variables to guide the iterative process. We introduce the Information Fusion Transformer(IFT) to extract both local and non-local prior information from image sequences, thereby removing aliasing artifacts resulting from random undersampling. Finally, we impose low-rank constraints within the HQS algorithm to further enhance the reconstruction results. The experiments demonstrate that each component module of our proposed model contributes to the improvement of the reconstruction task. Our proposed method achieves comparably satisfying performance to the state-of-the-art methods and it exhibits excellent generalization capabilities across different sampling masks. At the low acceleration factor, there is a 0.7% enhancement in the PSNR. Furthermore, when the acceleration factor reached 8 and 12, the PSNR achieves an improvement of 3.4% and 5.8% respectively.

CRNN-Refined Spatiotemporal Transformer for Dynamic MRI reconstruction

Magnetic Resonance Imaging (MRI) plays a pivotal role in modern clinical practice, providing detailed anatomical visualization with exceptional spatial resolution and soft tissue contrast. Dynamic MRI, aiming to capture both spatial and temporal characteristics, faces challenges related to prolonged acquisition times and susceptibility to motion artifacts. Balancing spatial and temporal resolutions becomes crucial in real-world clinical scenarios. In the realm of dynamic MRI reconstruction, while Convolutional Recurrent Neural Networks (CRNNs) struggle with long-range dependencies, CRNNs require extensive iterations, impacting efficiency. Transformers, known for their effectiveness in high-dimensional imaging, are underexplored in dynamic MRI reconstruction. Additionally, prevailing algorithms fall short of achieving superior results in demanding generative reconstructions at high acceleration rates. This research proposes a novel approach for dynamic MRI reconstruction, named CRNN-Refined Spatiotemporal Transformer Network (CST-Net). The spatiotemporal Transformer initiates reconstruction, modeling temporal and spatial correlations, followed by refinement using the CRNN. This integration mitigates inaccuracies caused by damaged frames and reduces CRNN iterations, enhancing computational efficiency without compromising reconstruction quality. Our study compares the performance of the proposed CST-Net at 6 × and 12 × undersampling rates, showcasing its superiority over existing algorithms. Particularly, in challenging 25× generative reconstructions, the CST-Net outperforms current methods. The comparison includes experiments under both radial and Cartesian undersampling patterns. In conclusion, CST-Net successfully addresses the limitations inherent in existing generative reconstruction algorithms, thereby paving the way for further exploration and optimization of Transformer-based approaches in dynamic MRI reconstruction. Code and Datasets can be available: https://github.com/XWangBin/CST-Net.

Spatial attention-based CSR-Unet framework for subdural and epidural hemorrhage segmentation and classification using CT images

BACKGROUND

Automatic diagnosis and brain hemorrhage segmentation in Computed Tomography (CT) may be helpful in assisting the neurosurgeon in developing treatment plans that improve the patient's chances of survival. Because medical segmentation of images is important and performing operations manually is challenging, many automated algorithms have been developed for this purpose, primarily focusing on certain image modalities. Whenever a blood vessel bursts, a dangerous medical condition known as intracranial hemorrhage (ICH) occurs. For best results, quick action is required. That being said, identifying subdural (SDH) and epidural haemorrhages (EDH) is a difficult task in this field and calls for a new, more precise detection method.

METHODS

This work uses a head CT scan to detect cerebral bleeding and distinguish between two types of dural hemorrhages using deep learning techniques. This paper proposes a rich segmentation approach to segment both SDH and EDH by enhancing segmentation efficiency with a better feature extraction procedure. This method incorporates Spatial attention- based CSR (convolution-SE-residual) Unet, for rich segmentation and precise feature extraction.

RESULTS

According to the study's findings, the CSR based Spatial network performs better than the other models, exhibiting impressive metrics for all assessed parameters with a mean dice coefficient of 0.970 and mean IoU of 0.718, while EDH and SDH dice scores are 0.983 and 0.969 respectively.

CONCLUSIONS

The CSR Spatial network experiment results show that it can perform well regarding dice coefficient. Furthermore, Spatial Unet based on CSR may effectively model the complicated in segmentations and rich feature extraction and improve the representation learning compared to alternative deep learning techniques, of illness and medical treatment, to enhance the meticulousness in predicting the fatality.

Artificial intelligence techniques in liver cancer

Hepatocellular Carcinoma (HCC), the most common primary liver cancer, is a significant contributor to worldwide cancer-related deaths. Various medical imaging techniques, including computed tomography, magnetic resonance imaging, and ultrasound, play a crucial role in accurately evaluating HCC and formulating effective treatment plans. Artificial Intelligence (AI) technologies have demonstrated potential in supporting physicians by providing more accurate and consistent medical diagnoses. Recent advancements have led to the development of AI-based multi-modal prediction systems. These systems integrate medical imaging with other modalities, such as electronic health record reports and clinical parameters, to enhance the accuracy of predicting biological characteristics and prognosis, including those associated with HCC. These multi-modal prediction systems pave the way for predicting the response to transarterial chemoembolization and microvascular invasion treatments and can assist clinicians in identifying the optimal patients with HCC who could benefit from interventional therapy. This paper provides an overview of the latest AI-based medical imaging models developed for diagnosing and predicting HCC. It also explores the challenges and potential future directions related to the clinical application of AI techniques.

Considerations on Image Preprocessing Techniques Required by Deep Learning Models. The Case of the Knee MRIs

OBJECTIVES

This study aims to demonstrate the preprocessing steps for knee MRI images to detect meniscal lesions using deep learning models and highlight their practical implications in diagnosing knee conditions, especially meniscal injuries, often caused by degeneration or trauma. Magnetic resonance imaging (MRI) is key in this field, especially when combined with ligament evaluations, and our research underscores the relevance and applicability of these techniques in real-world scenarios. Importantly, our findings suggest a promising future for the diagnosis of knee conditions.

MATERIALS AND METHODS

We initially worked with DICOM-format images, the standard for medical imaging, utilizing the Python packages PyDicom and SimpleITK for preprocessing. We also addressed the NIfTI format commonly used in research. Our preprocessing methods, designed with efficiency in mind, encompassed modality-specific adjustments, orientation, spatial resampling, intensity normalization, standardization and conversion to algorithm input format. These steps ensure efficient data handling, accelerate training speeds, and reassure the audience about the effectiveness of our research.

RESULTS

Our study processed PD-sagittal images from 188 patients to create a test set for training a deep learning segmentation model. We successfully completed all preprocessing steps, including accessing DICOM header information using hexadecimal encoded identifiers and utilizing SimpleITK for efficient handling of both 2D and 3D DICOM data. Resampling was performed for all 188 sets. Additionally, manual segmentation was conducted on 188 MRI scans, focusing on regions of interest (ROIs), such as normal tissue and meniscus tears in both the medial and lateral menisci. This involved contrast adjustment and precise hand-tracing of the structures within the ROIs, demonstrating the effectiveness and potential of our research in diagnosing knee conditions, and offering hope for the future of knee MRI diagnosis.

CONCLUSIONS

Our study introduces innovative preprocessing methods that have the potential to advance the field. By enhancing researchers' understanding of the importance of preprocessing steps, we anticipate that our techniques will streamline the preparation of standardized formats for deep learning model training and significantly benefit radiologists and orthopedic surgeons. These techniques could reduce time and effort in tasks like meniscal tear segmentation or localization, inspiring hope for more efficient and effective achievements in the field.

Role of the Clinical Features and MRI Parameters on Ki-67 Expression in Hepatocellular Carcinoma Patients: Development of a Predictive Nomogram

PURPOSE

To develop a nomogram using clinical features and the MRI parameters for preoperatively predicting the expression of Ki-67 in patients with hepatocellular carcinoma (HCC).

METHODS

One hundred and forty patients (training cohorts: n = 108; validation cohorts: n = 32) with confirmed HCC were investigated. Mann-Whitney U test, independent sample t-test, and chi-squared test were used to analyze the continuous and categorical variables. Univariate and multivariate logistic regression analyses were performed to examine the clinical variables and parameters from MRI associated with Ki-67 expression. As a result, a nomogram was developed based on these associations in patients with HCC. The performance of the nomogram was evaluated using the area under the receiver operating characteristic curve (AUC) and calibration curves.

RESULTS

In the training set, multivariable logistic regression analysis revealed that lens culinaris agglutinin-reactive fraction of alpha-fetoprotein (AFP-L3) levels, protein induced by vitamin K absence or antagonist-II (PIVKA-II) levels, and tumor shape were independent predictors for Ki-67 expression (p < 0.05). These three variables and the apparent diffusion coefficient (ADC) value were used to establish a nomogram, while the ADC value was found to be a marginal significant predictor. The model demonstrated a strong ability to discriminate Ki-67 expression in both the training and validation cohorts (AUC = 0.862, 0.877).

CONCLUSION

A non-invasive preoperative prediction method, which incorporates MRI variables and clinical features was developed, and showed effectiveness in evaluating Ki-67 expression in HCC patients.

Model-based federated learning for accurate MR image reconstruction from undersampled k-space data

Deep learning-based methods have achieved encouraging performances in the field of Magnetic Resonance (MR) image reconstruction. Nevertheless, building powerful and robust deep learning models requires collecting large and diverse datasets from multiple centers. This raises concerns about ethics and data privacy. Recently, federated learning has emerged as a promising solution, enabling the utilization of multi-center data without the need for data transfer between institutions. Despite its potential, existing federated learning methods face challenges due to the high heterogeneity of data from different centers. Aggregation methods based on simple averaging, which are commonly used to combine the client's information, have shown limited reconstruction and generalization capabilities. In this paper, we propose a Model-based Federated learning framework (ModFed) to address these challenges. ModFed has three major contributions: (1) Different from existing data-driven federated learning methods, ModFed designs attention-assisted model-based neural networks that can alleviate the need for large amounts of data on each client; (2) To address the data heterogeneity issue, ModFed proposes an adaptive dynamic aggregation scheme, which can improve the generalization capability and robustness of the trained neural network models; (3) ModFed incorporates a spatial Laplacian attention mechanism and a personalized client-side loss regularization to capture the detailed information for accurate image reconstruction. The effectiveness of the proposed ModFed is evaluated on three in-vivo datasets. Experimental results show that when compared to six existing state-of-the-art federated learning approaches, ModFed achieves better MR image reconstruction performance with increased generalization capability. Codes will be made available at https://github.com/ternencewu123/ModFed.

Charged Gold Nanoparticles for Target Identification-Alignment and Automatic Segmentation of CT Image-Guided Adaptive Radiotherapy in Small Hepatocellular Carcinoma

Because of the challenges posed by anatomical uncertainties and the low resolution of plain computed tomography (CT) scans, implementing adaptive radiotherapy (ART) for small hepatocellular carcinoma (sHCC) using artificial intelligence (AI) faces obstacles in tumor identification-alignment and automatic segmentation. The current study aims to improve sHCC imaging for ART using a gold nanoparticle (Au NP)-based CT contrast agent to enhance AI-driven automated image processing. The synthesized charged Au NPs demonstrated notable in vitro aggregation, low cytotoxicity, and minimal organ toxicity. Over time, an in situ sHCC mouse model was established for in vivo CT imaging at multiple time points. The enhanced CT images processed using 3D U-Net and 3D Trans U-Net AI models demonstrated high geometric and dosimetric accuracy. Therefore, charged Au NPs enable accurate and automatic sHCC segmentation in CT images using classical AI models, potentially addressing the technical challenges related to tumor identification, alignment, and automatic segmentation in CT-guided online ART.

Knowledge-driven deep learning for fast MR imaging: Undersampled MR image reconstruction from supervised to un-supervised learning

Deep learning (DL) has emerged as a leading approach in accelerating MRI. It employs deep neural networks to extract knowledge from available datasets and then applies the trained networks to reconstruct accurate images from limited measurements. Unlike natural image restoration problems, MRI involves physics-based imaging processes, unique data properties, and diverse imaging tasks. This domain knowledge needs to be integrated with data-driven approaches. Our review will introduce the significant challenges faced by such knowledge-driven DL approaches in the context of fast MRI along with several notable solutions, which include learning neural networks and addressing different imaging application scenarios. The traits and trends of these techniques have also been given which have shifted from supervised learning to semi-supervised learning, and finally, to unsupervised learning methods. In addition, MR vendors' choices of DL reconstruction have been provided along with some discussions on open questions and future directions, which are critical for the reliable imaging systems.

MRI Radiomics Combined with Clinicopathological Factors for Predicting 3-Year Overall Survival of Hepatocellular Carcinoma After Hepatectomy

Background

A limited number of studies have examined the use of radiomics to predict 3-year overall survival (OS) after hepatectomy in patients with hepatocellular carcinoma (HCC). This study develops 3-year OS prediction models for HCC patients after liver resection using MRI radiomics and clinicopathological factors.

Materials and Methods

A retrospective analysis of 141 patients who underwent surgical resection of HCC was performed. Patients were randomized into two set: the training set (n=98) and the validation set (n=43) including the survival groups (n=111) and non-survival groups (n=30) based on 3-year survival after hepatectomy. Furthermore, x2 or Fisher's exact test, univariate and multivariate logistic regression analyses were conducted to determine independent clinicopathological risk factors associated with 3-year OS. 1688 quantitative imaging features were extracted from preoperative T2-weighted imaging (T2WI) and contrast-enhanced magnetic resonance imaging (CE-MRI) of arterial phase (AP), portal venous phases (PVP)and delay period (DP). The features were selected using the variance threshold method, the select K best method and the least absolute shrinkage and selection operator (LASSO) algorithm. By using Bernoulli Naive Bayes (BernoulliNB) and Multinomial Naive Bayes (MultinomialNB) classifiers, we constructed models based on the independent clinicopathological factors and Rad-scores. To determine the best model, receiver operating characteristics (ROC) and Delong's test were used. Moreover, calibration curves were used to determine the calibration ability of the model, while decision curve analysis (DCA) was implemented to evaluate its clinical benefit.

Results

The fusion model showed excellent prediction precision with AUC of 0.910 and 0.846 in training and validation set and revealed significant diagnostic accuracy and value in the calibration curve and DCA analysis.

Conclusion

Nomograms based on MRI radiomics and clinicopathological factors have significant predictive value for 3-year OS after hepatectomy and can be used for risk classification.

Dual-domain faster Fourier convolution based network for MR image reconstruction

Deep learning methods for fast MRI have shown promise in reconstructing high-quality images from undersampled multi-coil k-space data, leading to reduced scan duration. However, existing methods encounter challenges related to limited receptive fields in dual-domain (k-space and image domains) reconstruction networks, rigid data consistency operations, and suboptimal refinement structures, which collectively restrict overall reconstruction performance. This study introduces a comprehensive framework that addresses these challenges and enhances MR image reconstruction quality. Firstly, we propose Faster Inverse Fourier Convolution (FasterIFC), a frequency domain convolutional operator that significantly expands the receptive field of k-space domain reconstruction networks. Expanding the information extraction range to the entire frequency spectrum according to the spectral convolution theorem in Fourier theory enables the network to easily utilize richer redundant long-range information from adjacent, symmetrical, and diagonal locations of multi-coil k-space data. Secondly, we introduce a novel softer Data Consistency (softerDC) layer, which achieves an enhanced balance between data consistency and smoothness. This layer facilitates the implementation of diverse data consistency strategies across distinct frequency positions, addressing the inflexibility observed in current methods. Finally, we present the Dual-Domain Faster Fourier Convolution Based Network (D2F2), which features a centrosymmetric dual-domain parallel structure based on FasterIFC. This architecture optimally leverages dual-domain data characteristics while substantially expanding the receptive field in both domains. Coupled with the softerDC layer, D2F2 demonstrates superior performance on the NYU fastMRI dataset at multiple acceleration factors, surpassing state-of-the-art methods in both quantitative and qualitative evaluations.

CMRxRecon: A publicly available k-space dataset and benchmark to advance deep learning for cardiac MRI

Cardiac magnetic resonance imaging (CMR) has emerged as a valuable diagnostic tool for cardiac diseases. However, a significant drawback of CMR is its slow imaging speed, resulting in low patient throughput and compromised clinical diagnostic quality. The limited temporal resolution also causes patient discomfort and introduces artifacts in the images, further diminishing their overall quality and diagnostic value. There has been growing interest in deep learning-based CMR imaging algorithms that can reconstruct high-quality images from highly under-sampled k-space data. However, the development of deep learning methods requires large training datasets, which have so far not been made publicly available for CMR. To address this gap, we released a dataset that includes multi-contrast, multi-view, multi-slice and multi-coil CMR imaging data from 300 subjects. Imaging studies include cardiac cine and mapping sequences. The 'CMRxRecon' dataset contains raw k-space data and auto-calibration lines. Our aim is to facilitate the advancement of state-of-the-art CMR image reconstruction by introducing standardized evaluation criteria and making the dataset freely accessible to the research community.

Transfer Learning in Cancer Genetics, Mutation Detection, Gene Expression Analysis, and Syndrome Recognition

Artificial intelligence (AI), encompassing machine learning (ML) and deep learning (DL), has revolutionized medical research, facilitating advancements in drug discovery and cancer diagnosis. ML identifies patterns in data, while DL employs neural networks for intricate processing. Predictive modeling challenges, such as data labeling, are addressed by transfer learning (TL), leveraging pre-existing models for faster training. TL shows potential in genetic research, improving tasks like gene expression analysis, mutation detection, genetic syndrome recognition, and genotype-phenotype association. This review explores the role of TL in overcoming challenges in mutation detection, genetic syndrome detection, gene expression, or phenotype-genotype association. TL has shown effectiveness in various aspects of genetic research. TL enhances the accuracy and efficiency of mutation detection, aiding in the identification of genetic abnormalities. TL can improve the diagnostic accuracy of syndrome-related genetic patterns. Moreover, TL plays a crucial role in gene expression analysis in order to accurately predict gene expression levels and their interactions. Additionally, TL enhances phenotype-genotype association studies by leveraging pre-trained models. In conclusion, TL enhances AI efficiency by improving mutation prediction, gene expression analysis, and genetic syndrome detection. Future studies should focus on increasing domain similarities, expanding databases, and incorporating clinical data for better predictions.

Radial Undersampled MRI Reconstruction Using Deep Learning With Mutual Constraints Between Real and Imaginary Components of K-Space

The deep learning method is an efficient solution for improving the quality of undersampled magnetic resonance (MR) image reconstruction while reducing lengthy data acquisition. Most deep learning methods neglect the mutual constraints between the real and imaginary components of complex-valued k-space data. In this paper, a new complex-valued convolutional neural network, namely, Dense-U-Dense Net (DUD-Net), is proposed to interpolate the undersampled k-space data and reconstruct MR images. The proposed network comprises dense layers, U-Net, and other dense layers in sequence. The dense layers are used to simulate the mutual constraints between real and imaginary components, and U-Net performs feature sparsity and interpolation estimation for k-space data. Two MRI datasets were used to evaluate the proposed method: brain magnitude-only MR images and knee complex-valued k-space data. Several operations were conducted for data preprocessing. First, the complex-valued MR images were synthesized by phase modulation on magnitude-only images. Second, a radial trajectory based on the golden angle was used for k-space undersampling, whereby a reversible normalization method was proposed to balance the distribution of positive and negative values in k-space data. The optimal performance of DUD-Net was demonstrated based on a quantitative evaluation of inter-method and intra-method comparisons. When compared with other methods, significant improvements were achieved, PSNRs were increased by 10.78 and 5.74dB, whereas RMSEs were decreased by 71.53% and 30.31% for magnitude and phase image, respectively. It is concluded that DUD-Net significantly improves the performance of MR image reconstruction.

Predicting Ki-67 expression in hepatocellular carcinoma: nomogram based on clinical factors and contrast-enhanced ultrasound radiomics signatures

PURPOSE

To develop a contrast-enhanced ultrasound (CEUS) clinic-radiomics nomogram for individualized assessment of Ki-67 expression in hepatocellular carcinoma (HCC).

METHODS

A retrospective cohort comprising 310 HCC individuals who underwent preoperative CEUS (using SonoVue) at three different centers was partitioned into a training set, a validation set, and an external test set. Radiomics signatures indicating the phenotypes of the Ki-67 were extracted from multiphase CEUS images. The radiomics score (Rad-score) was calculated accordingly after feature selection and the radiomics model was constructed. A clinic-radiomics nomogram was established utilizing multiphase CEUS Rad-score and clinical risk factors. A clinical model only incorporated clinical factors was also developed for comparison. Regarding clinical utility, calibration, and discrimination, the predictive efficiency of the clinic-radiomics nomogram was evaluated.

RESULTS

Seven radiomics signatures from multiphase CEUS images were selected to calculate the Rad-score. The clinic-radiomics nomogram, comprising the Rad-score and clinical risk factors, indicated a good calibration and demonstrated a better discriminatory capacity compared to the clinical model (AUCs: 0.870 vs 0.797, 0.872 vs 0.755, 0.856 vs 0.749 in the training, validation, and external test set, respectively) and the radiomics model (AUCs: 0.870 vs 0.752, 0.872 vs 0.733, 0.856 vs 0.729 in the training, validation, and external test set, respectively). Furthermore, both the clinical impact curve and the decision curve analysis displayed good clinical application of the nomogram.

CONCLUSION

The clinic-radiomics nomogram constructed from multiphase CEUS images and clinical risk parameters can distinguish Ki-67 expression in HCC patients and offer useful insights to guide subsequent personalized treatment.

Insights into the Inherited Basis of Valvular Heart Disease

PURPOSE OF REVIEW: Increases in the availability of genetic data and advances in the tools and methods for their analyses have enabled well-powered genetic association studies that have significantly enhanced our understanding of the genetic factors underlying both rare and common valve diseases. Valvular heart diseases, such as congenital valve malformations and degenerative valve lesions, increase the risk of heart failure, arrhythmias, and sudden death. In this review, we provide an updated overview of our current understanding of the genetic mechanisms underlying valvular heart diseases. With a focus on discoveries from the past 5 years, we describe recent insights into genetic risk and underlying biological pathways. RECENT FINDINGS: Recently acquired knowledge around valvular heart disease genetics has provided important insights into novel mechanisms related to disease pathogenesis. Newly identified risk loci associated valvular heart disease mainly regulate the composition of the extracellular matrix, accelerate the endothelial-to-mesenchymal transition, contribute to cilia formation processes, and play roles in lipid metabolism. Large-scale genomic analyses have identified numerous risk loci, genes, and biological pathways associated with degenerative valve disease and congenital valve malformations. Shared risk genes suggest common mechanistic pathways for various valve pathologies. More recent studies have combined cardiac magnetic resonance imaging and machine learning to offer a novel approach for exploring genotype-phenotype relationships regarding valve disease. Progress in the field holds promise for targeted prevention, particularly through the application of polygenic risk scores, and innovative therapies based on the biological mechanisms for predominant forms of valvular heart diseases.

TransVFS: A spatio-temporal local-global transformer for vision-based force sensing during ultrasound-guided prostate biopsy

Robot-assisted prostate biopsy is a new technology to diagnose prostate cancer, but its safety is influenced by the inability of robots to sense the tool-tissue interaction force accurately during biopsy. Recently, vision based force sensing (VFS) provides a potential solution to this issue by utilizing image sequences to infer the interaction force. However, the existing mainstream VFS methods cannot realize the accurate force sensing due to the adoption of convolutional or recurrent neural network to learn deformation from the optical images and some of these methods are not efficient especially when the recurrent convolutional operations are involved. This paper has presented a Transformer based VFS (TransVFS) method by leveraging ultrasound volume sequences acquired during prostate biopsy. The TransVFS method uses a spatio-temporal local-global Transformer to capture the local image details and the global dependency simultaneously to learn prostate deformations for force estimation. Distinctively, our method explores both the spatial and temporal attention mechanisms for image feature learning, thereby addressing the influence of the low ultrasound image resolution and the unclear prostate boundary on the accurate force estimation. Meanwhile, the two efficient local-global attention modules are introduced to reduce 4D spatio-temporal computation burden by utilizing the factorized spatio-temporal processing strategy, thereby facilitating the fast force estimation. Experiments on prostate phantom and beagle dogs show that our method significantly outperforms existing VFS methods and other spatio-temporal Transformer models. The TransVFS method surpasses the most competitive compared method ResNet3dGRU by providing the mean absolute errors of force estimation, i.e., 70.4 ± 60.0 millinewton (mN) vs 123.7 ± 95.6 mN, on the transabdominal ultrasound dataset of dogs.

Ultrasound-based deep learning radiomics model for differentiating benign, borderline, and malignant ovarian tumours: a multi-class classification exploratory study

BACKGROUND

Accurate preoperative identification of ovarian tumour subtypes is imperative for patients as it enables physicians to custom-tailor precise and individualized management strategies. So, we have developed an ultrasound (US)-based multiclass prediction algorithm for differentiating between benign, borderline, and malignant ovarian tumours.

METHODS

We randomised data from 849 patients with ovarian tumours into training and testing sets in a ratio of 8:2. The regions of interest on the US images were segmented and handcrafted radiomics features were extracted and screened. We applied the one-versus-rest method in multiclass classification. We inputted the best features into machine learning (ML) models and constructed a radiomic signature (Rad_Sig). US images of the maximum trimmed ovarian tumour sections were inputted into a pre-trained convolutional neural network (CNN) model. After internal enhancement and complex algorithms, each sample's predicted probability, known as the deep transfer learning signature (DTL_Sig), was generated. Clinical baseline data were analysed. Statistically significant clinical parameters and US semantic features in the training set were used to construct clinical signatures (Clinic_Sig). The prediction results of Rad_Sig, DTL_Sig, and Clinic_Sig for each sample were fused as new feature sets, to build the combined model, namely, the deep learning radiomic signature (DLR_Sig). We used the receiver operating characteristic (ROC) curve and the area under the ROC curve (AUC) to estimate the performance of the multiclass classification model.

RESULTS

The training set included 440 benign, 44 borderline, and 196 malignant ovarian tumours. The testing set included 109 benign, 11 borderline, and 49 malignant ovarian tumours. DLR_Sig three-class prediction model had the best overall and class-specific classification performance, with micro- and macro-average AUC of 0.90 and 0.84, respectively, on the testing set. Categories of identification AUC were 0.84, 0.85, and 0.83 for benign, borderline, and malignant ovarian tumours, respectively. In the confusion matrix, the classifier models of Clinic_Sig and Rad_Sig could not recognise borderline ovarian tumours. However, the proportions of borderline and malignant ovarian tumours identified by DLR_Sig were the highest at 54.55% and 63.27%, respectively.

CONCLUSIONS

The three-class prediction model of US-based DLR_Sig can discriminate between benign, borderline, and malignant ovarian tumours. Therefore, it may guide clinicians in determining the differential management of patients with ovarian tumours.

A cardiac MRI motion artifact reduction method based on edge enhancement network

Cardiac magnetic resonance imaging (MRI) usually requires a long acquisition time. The movement of the patients during MRI acquisition will produce image artifacts. Previous studies have shown that clear MR image texture edges are of great significance for pathological diagnosis. In this paper, a motion artifact reduction method for cardiac MRI based on edge enhancement network is proposed. Firstly, the four-plane normal vector adaptive fractional differential mask is applied to extract the edge features of blurred images. The four-plane normal vector method can reduce the noise information in the edge feature maps. The adaptive fractional order is selected according to the normal mean gradient and the local Gaussian curvature entropy of the images. Secondly, the extracted edge feature maps and blurred images are input into the de-artifact network. In this network, the edge fusion feature extraction network and the edge fusion transformer network are specially designed. The former combines the edge feature maps with the fuzzy feature maps to extract the edge feature information. The latter combines the edge attention network and the fuzzy attention network, which can focus on the blurred image edges. Finally, extensive experiments show that the proposed method can obtain higher peak signal-to-noise ratio and structural similarity index measure compared to state-of-art methods. The de-artifact images have clear texture edges.

Role of radiomics in staging liver fibrosis: a meta-analysis

BACKGROUND

Fibrosis has important pathoetiological and prognostic roles in chronic liver disease. This study evaluates the role of radiomics in staging liver fibrosis.

METHOD

After literature search in electronic databases (Embase, Ovid, Science Direct, Springer, and Web of Science), studies were selected by following precise eligibility criteria. The quality of included studies was assessed, and meta-analyses were performed to achieve pooled estimates of area under receiver-operator curve (AUROC), accuracy, sensitivity, and specificity of radiomics in staging liver fibrosis compared to histopathology.

RESULTS

Fifteen studies (3718 patients; age 47 years [95% confidence interval (CI): 42, 53]; 69% [95% CI: 65, 73] males) were included. AUROC values of radiomics for detecting significant fibrosis (F2-4), advanced fibrosis (F3-4), and cirrhosis (F4) were 0.91 [95%CI: 0.89, 0.94], 0.92 [95%CI: 0.90, 0.95], and 0.94 [95%CI: 0.93, 0.96] in training cohorts and 0.89 [95%CI: 0.83, 0.91], 0.89 [95%CI: 0.83, 0.94], and 0.93 [95%CI: 0.91, 0.95] in validation cohorts, respectively. For diagnosing significant fibrosis, advanced fibrosis, and cirrhosis the sensitivity of radiomics was 84.0% [95%CI: 76.1, 91.9], 86.9% [95%CI: 76.8, 97.0], and 92.7% [95%CI: 89.7, 95.7] in training cohorts, and 75.6% [95%CI: 67.7, 83.5], 80.0% [95%CI: 70.7, 89.3], and 92.0% [95%CI: 87.8, 96.1] in validation cohorts, respectively. Respective specificity was 88.6% [95% CI: 83.0, 94.2], 88.4% [95% CI: 81.9, 94.8], and 91.1% [95% CI: 86.8, 95.5] in training cohorts, and 86.8% [95% CI: 83.3, 90.3], 94.0% [95% CI: 89.5, 98.4], and 88.3% [95% CI: 84.4, 92.2] in validation cohorts. Limitations included use of several methods for feature selection and classification, less availability of studies evaluating a particular radiological modality, lack of a direct comparison between radiology and radiomics, and lack of external validation.

CONCLUSION

Although radiomics offers good diagnostic accuracy in detecting liver fibrosis, its role in clinical practice is not as clear at present due to comparability and validation constraints.

Fat fraction quantification with MRI estimates tumor proliferation of hepatocellular carcinoma

Purpose

To assess the utility of fat fraction quantification using quantitative multi-echo Dixon for evaluating tumor proliferation and microvascular invasion (MVI) in hepatocellular carcinoma (HCC).

Methods

A total of 66 patients with resection and histopathologic confirmed HCC were enrolled. Preoperative MRI with proton density fat fraction and R2* mapping was analyzed. Intratumoral and peritumoral regions were delineated with manually placed regions of interest at the maximum level of intratumoral fat. Correlation analysis explored the relationship between fat fraction and Ki67. The fat fraction and R2* were compared between high Ki67(>30%) and low Ki67 nodules, and between MVI negative and positive groups. Receiver operating characteristic (ROC) analysis was used for further analysis if statistically different.

Results

The median fat fraction of tumor (tFF) was higher than peritumor liver (5.24% vs 3.51%, P=0.012). The tFF was negatively correlated with Ki67 (r=-0.306, P=0.012), and tFF of high Ki67 nodules was lower than that of low Ki67 nodules (2.10% vs 4.90%, P=0.001). The tFF was a good estimator for low proliferation nodules (AUC 0.747, cut-off 3.39%, sensitivity 0.778, specificity 0.692). There was no significant difference in tFF and R2* between MVI positive and negative nodules (3.00% vs 2.90%, P=0.784; 55.80s-1 vs 49.15s-1, P=0.227).

Conclusion

We infer that intratumor fat can be identified in HCC and fat fraction quantification using quantitative multi-echo Dixon can distinguish low proliferative HCCs.

IWNeXt: an image-wavelet domain ConvNeXt-based network for self-supervised multi-contrast MRI reconstruction

Objective.Multi-contrast magnetic resonance imaging (MC MRI) can obtain more comprehensive anatomical information of the same scanning object but requires a longer acquisition time than single-contrast MRI. To accelerate MC MRI speed, recent studies only collect partial k-space data of one modality (target contrast) to reconstruct the remaining non-sampled measurements using a deep learning-based model with the assistance of another fully sampled modality (reference contrast). However, MC MRI reconstruction mainly performs the image domain reconstruction with conventional CNN-based structures by full supervision. It ignores the prior information from reference contrast images in other sparse domains and requires fully sampled target contrast data. In addition, because of the limited receptive field, conventional CNN-based networks are difficult to build a high-quality non-local dependency.Approach.In the paper, we propose an Image-Wavelet domain ConvNeXt-based network (IWNeXt) for self-supervised MC MRI reconstruction. Firstly, INeXt and WNeXt based on ConvNeXt reconstruct undersampled target contrast data in the image domain and refine the initial reconstructed result in the wavelet domain respectively. To generate more tissue details in the refinement stage, reference contrast wavelet sub-bands are used as additional supplementary information for wavelet domain reconstruction. Then we design a novel attention ConvNeXt block for feature extraction, which can capture the non-local information of the MC image. Finally, the cross-domain consistency loss is designed for self-supervised learning. Especially, the frequency domain consistency loss deduces the non-sampled data, while the image and wavelet domain consistency loss retain more high-frequency information in the final reconstruction.Main results.Numerous experiments are conducted on the HCP dataset and the M4Raw dataset with different sampling trajectories. Compared with DuDoRNet, our model improves by 1.651 dB in the peak signal-to-noise ratio.Significance.IWNeXt is a potential cross-domain method that can enhance the accuracy of MC MRI reconstruction and reduce reliance on fully sampled target contrast images.

Radiomic Analysis Based on Gd-EOB-DTPA Enhanced MRI for the Preoperative Prediction of Ki-67 Expression in Hepatocellular Carcinoma

RATIONALE AND OBJECTIVES

To develop and validate a random forest model based on radiomic features in Gd-EOB-DTPA enhanced MRI for predicting the Ki-67 expression in solitary HCC.

MATERIALS AND METHODS

This retrospective study analyzed 258 patients with solitary HCC. Significant clinicoradiological factors were identified through univariate and multivariate analyses for distinguishing HCC with high (>20%) and low (≤20%) Ki-67 expression. Radiomic features were extracted at Gd-EOB-DTPA enhanced MRI. The recursive feature elimination (RFE) strategy was employed to screen robust radiomic features, and the Random Forest (RF) algorithm was utilized to rank radiomic features and construct prediction models. The AUC, accuracy, precision, recall, and f1-score were used to evaluate the performance of RF models.

RESULTS

Multivariate analysis identified serum AFP level, tumor size, growth type, and peritumoral enhancement as independent predictors for HCC with high Ki-67 expression. The clinicoradiological-radiomic model that incorporated the clinicoradiological predictors and the top ten radiomic features outperformed the clinicoradiological model in the training set (AUCs 0.876 vs. 0.780; p < 0.001), though the test set did not have a statistical significance (AUCs 0.809 vs. 0.723; p = 0.123). The addition of clinicoradiological predictors did not yield a significant improvement in the performance of radiomic features in both sets (training, p = 0.692; test, p = 0.229). Decision curve analysis further confirmed the clinical utility of the RF models.

CONCLUSION

The RF models based on radiomic features of Gd-EOB-DTPA enhanced MRI achieved satisfactory performance in preoperatively predicting Ki-67 expression in HCC.

MRI-Based Radiomics Methods for Predicting Ki-67 Expression in Breast Cancer: A Systematic Review and Meta-analysis

RATIONALE AND OBJECTIVES

The purpose of this systematic review and meta-analysis was to assess the quality and diagnostic accuracy of MRI-based radiomics for predicting Ki-67 expression in breast cancer.

MATERIALS AND METHODS

A systematic literature search was performed to find relevant studies published in different databases, including PubMed, Web of Science, and Embase up until March 10, 2023. All papers were independently evaluated for eligibility by two reviewers. Studies that matched research questions and provided sufficient data for quantitative synthesis were included in the systematic review and meta-analysis, respectively. The quality of the articles was assessed using Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) and Radiomics Quality Score (RQS) tools. The predictive value of MRI-based radiomics for Ki-67 antigen in patients with breast cancer was assessed using pooled sensitivity (SEN), specificity, and area under the curve (AUC). Meta-regression was performed to explore the cause of heterogeneity. Different covariates were used for subgroup analysis.

RESULTS

31 studies were included in the systematic review; among them, 21 reported sufficient data for meta-analysis. 20 training cohorts and five validation cohorts were pooled separately. The pooled sensitivity, specificity, and AUC of MRI-based radiomics for predicting Ki-67 expression in training cohorts were 0.80 [95% CI, 0.73-0.86], 0.82 [95% CI, 0.78-0.86], and 0.88 [95%CI, 0.85-0.91], respectively. The corresponding values for validation cohorts were 0.81 [95% CI, 0.72-0.87], 0.73 [95% CI, 0.62-0.82], and 0.84 [95%CI, 0.80-0.87], respectively. Based on QUADAS-2, some risks of bias were detected for reference standard and flow and timing domains. However, the quality of the included article was acceptable. The mean RQS score of the included articles was close to 6, corresponding to 16.6% of the maximum possible score. Significant heterogeneity was observed in pooled sensitivity and specificity of training cohorts (I2 > 75%). We found that using deep learning radiomic methods, magnetic field strength (3 T vs. 1.5 T), scanner manufacturer, region of interest structure (2D vs. 3D), route of tissue sampling, Ki-67 cut-off, logistic regression for model construction, and LASSO for feature reduction as well as PyRadiomics software for feature extraction had a great impact on heterogeneity according to our joint model analysis. Diagnostic performance in studies that used deep learning-based radiomics and multiple MRI sequences (e.g., DWI+DCE) was slightly higher. In addition, radiomic features derived from DWI sequences performed better than contrast-enhanced sequences in terms of specificity and sensitivity. No publication bias was found based on Deeks' funnel plot. Sensitivity analysis showed that eliminating every study one by one does not impact overall results.

CONCLUSION

This meta-analysis showed that MRI-based radiomics has a good diagnostic accuracy in differentiating breast cancer patients with high Ki-67 expression from low-expressing groups. However, the sensitivity and specificity of these methods still do not surpass 90%, restricting them from being used as a supplement to current pathological assessments (e.g., biopsy or surgery) to predict Ki-67 expression accurately.

MRI-Based Radiomics and Deep Learning in Biological Characteristics and Prognosis of Hepatocellular Carcinoma: Opportunities and Challenges

Hepatocellular carcinoma (HCC) is the fifth most common malignancy and the third leading cause of cancer-related death worldwide. HCC exhibits strong inter-tumor heterogeneity, with different biological characteristics closely associated with prognosis. In addition, patients with HCC often distribute at different stages and require diverse treatment options at each stage. Due to the variability in tumor sensitivity to different therapies, determining the optimal treatment approach can be challenging for clinicians prior to treatment. Artificial intelligence (AI) technology, including radiomics and deep learning approaches, has emerged as a unique opportunity to improve the spectrum of HCC clinical care by predicting biological characteristics and prognosis in the medical imaging field. The radiomics approach utilizes handcrafted features derived from specific mathematical formulas to construct various machine-learning models for medical applications. In terms of the deep learning approach, convolutional neural network models are developed to achieve high classification performance based on automatic feature extraction from images. Magnetic resonance imaging offers the advantage of superior tissue resolution and functional information. This comprehensive evaluation plays a vital role in the accurate assessment and effective treatment planning for HCC patients. Recent studies have applied radiomics and deep learning approaches to develop AI-enabled models to improve accuracy in predicting biological characteristics and prognosis, such as microvascular invasion and tumor recurrence. Although AI-enabled models have demonstrated promising potential in HCC with biological characteristics and prognosis prediction with high performance, one of the biggest challenges, interpretability, has hindered their implementation in clinical practice. In the future, continued research is needed to improve the interpretability of AI-enabled models, including aspects such as domain knowledge, novel algorithms, and multi-dimension data sources. Overcoming these challenges would allow AI-enabled models to significantly impact the care provided to HCC patients, ultimately leading to their deployment for clinical use. LEVEL OF EVIDENCE: 5 TECHNICAL EFFICACY: Stage 2.