Liver cancer imaging: role of CT, MRI, US and PET

Tackling Modality-Heterogeneous Client Drift Holistically for Heterogeneous Multimodal Federated Learning

Multimodal Federated Learning (MFL) has emerged as a collaborative paradigm for training models across decentralized devices, harnessing various data modalities to facilitate effective learning while respecting data ownership. In this realm, notably, a pivotal shift from homogeneous to heterogeneous MFL has taken place. While the former assumes uniformity in input modalities across clients, the latter accommodates modality-incongruous setups, which is often the case in practical situations. For example, while some advanced medical institutions have the luxury of utilizing both MRI and CT for disease diagnosis, remote hospitals often find themselves constrained to employ CT exclusively due to its cost-effectiveness. Although heterogeneous MFL can apply to a broader scenario, it introduces a new challenge: modality-heterogeneous client drift, arising from diverse modality-coupled local optimization. To address this, we introduce FedMM, a simple yet effective approach. During local optimization, FedMM employs modality dropout, randomly masking available modalities, and promoting weight alignment while preserving model expressivity on its original modality combination. To enhance the modality dropout process, FedMM incorporates a task-specific inter- and intra-modal regularizer, which acts as an additional constraint, forcing that weight distribution remains more consistent across diverse input modalities and therefore eases the optimization process with modality dropout enabled. By combining them, our approach holistically addresses client drift. It fosters convergence among client models while considering each client's unique input modalities, enhancing heterogeneous MFL performance. Comprehensive evaluations in three medical image segmentation datasets demonstrate FedMM's superiority over state-of-the-art heterogeneous MFL methods.

Solitary necrotic nodule of the liver: imaging features, differential diagnosis and management

Solitary necrotic nodule of the liver (SNNL) is a rare and benign liver lesion often discovered incidentally. Despite its occurrence, the exact cause of SNNL remains unknown, with various theories proposing traumatic, infectious, degenerative or transformative origins. The variable imaging characteristics of SNNLs frequently lead to misdiagnosis as malignant tumors, prompting patients to undergo unnecessary and high-risk procedures such as biopsies and surgeries. Moreover, biopsies often yield inconclusive results due to the presence of necrotic tissue within the lesion, posing challenges for accurate histologic diagnosis. This review aims to offer guidance on differentiating SNNLs from other liver lesions using multimodality imaging approaches. It will analyze essential imaging steps that should be performed and highlight those that should be avoided to enhance diagnostic accuracy and prevent unnecessary interventions.

Prognostic and diagnostic value of [18F]FDG, 11C-acetate, and [68Ga]Ga-FAPI-04 PET/CT for hepatocellular carcinoma

OBJECTIVES

To assess the prognostic value of Fluorine 18-labeled fluorodeoxyglucose [18F]FDG, gallium 68-labeled fibroblast-activation protein inhibitor-04 [68Ga]Ga-FAPI-04, 11C-acetate in hepatocellular carcinoma (HCC) and evaluate the potential usefulness and advantages of different combinations for accurate diagnosis.

MATERIALS AND METHODS

Thirty-six patients with suspected hepatic masses were prospectively enrolled from May 2021 to September 2022 and underwent [18F]FDG, [68Ga]Ga-FAPI-04, and 11C-acetate PET/CT scans before surgery. PET/CT results and histopathologic examinations were independently interpreted by two radiologists and pathologists, respectively. Kaplan-Meier overall survival curves were calculated and the sensitivity among [18F]FDG, 11C-acetate, [68Ga]Ga-FAPI-04, and different combinations were compared.

RESULTS

Of the 36 included patients (mean age, 59 years ± 10 (standard deviation)), 29 were diagnosed with HCC, four with non-HCC malignant tumors, and three with benign tumors. Patients with HCC lesions negative for 11C-acetate or [68Ga]Ga-FAPI-04 exhibited poorer overall survival. Out of 36 patients, 44 HCC lesions were detected. The dual-tracer [68Ga]Ga-FAPI-04/11C-acetate exhibited the highest sensitivity (39 of 44 lesions (88.6%)) among all schemes. HCC lesions with higher histological grade and microvascular invasion (MVI) showed higher maximum standardized uptake value (SUVmax) and tumor-to-background ratio (TBR) of [18F]FDG, but no evidence of significant differences was found in [68Ga]Ga-FAPI-04 and 11C-acetate PET/CT. Higher expression of fibroblast activation protein (FAP) showed higher uptake of [68Ga]Ga-FAPI-04 and [18F]FDG.

CONCLUSION

[68Ga]Ga-FAPI-04 and 11C-acetate PET/CT exhibited good predictive value for HCC patients, with their combination showing the highest sensitivity for HCC detection, suggesting potential for improved diagnostic protocols.

KEY POINTS

Question What are the prognostic and diagnostic values of PET/CT tracers, including [18F]FDG, [68Ga]FAPI-04, and 11C-acetate? Findings Hepatocellular carcinoma, with differing findings across [18F]FDG, [68Ga]GaFAPI-04, and 11C-acetate PET/CT, showed varied prognoses; [68Ga]GaFAPI-04 and 11C-acetate combined offered the highest detection sensitivity. Clinical relevance Evaluating the prognostic value and diagnostic efficacy of different tracer combinations in patients with hepatocellular carcinoma helps to guide the optimal selection of tracers in clinical practice.

Prognostic Modeling for Liver Cirrhosis Mortality Prediction and Real-Time Health Monitoring from Electronic Health Data

Liver cirrhosis stands as a prominent contributor to mortality, impacting millions across the United States. Enabling health care providers to predict early mortality among patients with cirrhosis holds the potential to enhance treatment efficacy significantly. Our hypothesis centers on the correlation between mortality and laboratory test results along with relevant diagnoses in this patient cohort. Additionally, we posit that a deep learning model could surpass the predictive capabilities of the existing Model for End-Stage Liver Disease score. This research seeks to advance prognostic accuracy and refine approaches to address the critical challenges posed by cirrhosis-related mortality. This study evaluates the performance of an artificial neural network model for liver disease classification using various training dataset sizes. Through meticulous experimentation, three distinct training proportions were analyzed: 70%, 80%, and 90%. The model's efficacy was assessed using precision, recall, F1-score, accuracy, and support metrics, alongside receiver operating characteristic (ROC) and precision-recall (PR) curves. The ROC curves were quantified using the area under the curve (AUC) metric. Results indicated that the model's performance improved with an increased size of the training dataset. Specifically, the 80% training data model achieved the highest AUC, suggesting superior classification ability over the models trained with 70% and 90% data. PR analysis revealed a steep trade-off between precision and recall across all datasets, with 80% training data again demonstrating a slightly better balance. This is indicative of the challenges faced in achieving high precision with a concurrently high recall, a common issue in imbalanced datasets such as those found in medical diagnostics.

Enhancing in situ cancer vaccines using delivery technologies

In situ cancer vaccination refers to any approach that exploits tumour antigens available at a tumour site to induce tumour-specific adaptive immune responses. These approaches hold great promise for the treatment of many solid tumours, with numerous candidate drugs under preclinical or clinical evaluation and several products already approved. However, there are challenges in the development of effective in situ cancer vaccines. For example, inadequate release of tumour antigens from tumour cells limits antigen uptake by immune cells; insufficient antigen processing by antigen-presenting cells restricts the generation of antigen-specific T cell responses; and the suppressive immune microenvironment of the tumour leads to exhaustion and death of effector cells. Rationally designed delivery technologies such as lipid nanoparticles, hydrogels, scaffolds and polymeric nanoparticles are uniquely suited to overcome these challenges through the targeted delivery of therapeutics to tumour cells, immune cells or the extracellular matrix. Here, we discuss delivery technologies that have the potential to reduce various clinical barriers for in situ cancer vaccines. We also provide our perspective on this emerging field that lies at the interface of cancer vaccine biology and delivery technologies.

Ultrasmall Mn-doped iron oxide nanoparticles with dual hepatobiliary and renal clearances for T1 MR liver imaging

Although magnetic nanoparticles demonstrate significant potential as magnetic resonance imaging (MRI) contrast agents, their negative contrasts, liver accumulation, and limited excretion hinder their application. Herein, we developed ultrasmall Mn-doped iron oxide nanoparticles (UMIOs) with distinct advantages as T1 MRI contrast agents. Exceptionally small particle sizes (ca. 2 nm) and magnetization values (5 emu gMn+Fe-1) of UMIOs provided optimal T1 contrast effects with an ideally low r2/r1 value of ∼1. Furthermore, the use of Mn as a dopant facilitated hepatocyte uptake of the particles, allowing liver imaging. In animal studies, UMIOs exhibited significantly enhanced contrasts for sequential T1 imaging of blood vessels and the liver, distinguishing them from conventional magnetic nanoparticles. UMIOs were systematically cleared via dual hepatobiliary and renal excretion pathways, highlighting their safety profile. These characteristics imply substantial potential of UMIOs as T1 contrast agents for the accurate diagnosis of liver diseases.

Advances in Magnetic Carbon Dots: A Theranostics Platform for Fluorescence/Magnetic Resonance Bimodal Imaging and Therapy for Tumors

Theranostics technology that combines tumor diagnosis or monitoring with therapy is an important direction for the future development of tumor treatment. It takes advantage of efficiently observing tumor tissues, monitoring tumor treatment in real time, and significantly improving the cure efficiency. Magnetic carbon dots (CDs) are of wide interest as molecular imaging probes, drug carriers, photosensitizers, and radiosensitizers in the integration of tumor fluorescence/magnetic resonance bimodal diagnosis and treatment because of their small size, good optical stability, magnetic relaxation rate, and biocompatibility. This review first analyzes and compares the synthesis methods and physicochemical properties of magnetic CDs in recent years and then concludes their mechanism in tumor fluorescence/magnetic resonance bimodal imaging and therapy in details. Subsequently, the research progress of their application in tumor theranostics are summarized. Finally, the problems and challenges of magnetic CDs for development at this stage are prospected. This review provides new ideas for their controlled synthesis and application in efficient and precise therapy for tumors.

Evaluation of health economic impact of initial diagnostic modality selection for colorectal cancer liver metastases in suspected patients in China, Japan and the USA

AIMS

To compare cost offsets and contributing factors (false-negative rates and confirmatory imaging requirements, potentially leading to longer waiting times for diagnosis) as well as long-term cost effectiveness associated with the diagnostic and treatment pathways for colorectal cancer liver metastases (CRCLM) in the US, Japan, and China according to initial imaging modality used. Gadoxetate disodium (ethoxylbenzyl-diethylenetriaminepentaacetic acid)-enhanced magnetic resonance imaging (EOB-MRI) was compared to multidetector computed tomography (MDCT), extracellular contrast media enhanced-MRI (ECCM-MRI) (the US and China only) and contrast-enhanced ultrasound (CEUS).

MATERIALS AND METHODS

Decision tree models were developed to simulate the clinical pathway, from first diagnostic test to initial treatment decision, based on local clinical guidelines and validated by experts. Input data were derived from the literature (up to 31st December 2020) as well as from interviews with local experts. A Markov model extension was built to evaluate the number of false-negative patients and associated costs, over a lifetime horizon.

RESULTS

The decision-tree models showed that, increasing proportionate use of initial EOB-MRI resulted in a cost-offset per patient (excluding false-negative patients) in all countries (USD 201 for the US, JPY 6,284 for Japan and CNY 446 for China) driven by reductions in follow-on diagnostic procedures and unnecessary treatment. The use of EOB-MRI was also associated with a shorter average waiting time to a final diagnosis and treatment decision compared to MDCT, ECCM-MRI and CEUS. The Markov model showed that with an increase in EOB-MRI use, there are fewer false-negative diagnoses over a lifetime horizon. In all three countries, the incremental cost-effectivenes ratio (ICER) was below standard willingness-to-pay thresholds.

CONCLUSION

The findings of these models demonstrate that use of EOB-MRI early in the diagnostic pathway for CRCLM results in short-term cost savings, as well as being cost effective in the long term.

A flexible three-dimensional heterophase computed tomography hepatocellular carcinoma detection algorithm for generalizable and practical screening

Hepatocellular carcinoma (HCC) can be potentially discovered from abdominal computed tomography (CT) studies under varied clinical scenarios (e.g., fully dynamic contrast-enhanced [DCE] studies, noncontrast [NC] plus venous phase [VP] abdominal studies, or NC-only studies). Each scenario presents its own clinical challenges that could benefit from computer-aided detection (CADe) tools. We investigate whether a single CADe model can be made flexible enough to handle different contrast protocols and whether this flexibility imparts performance gains. We developed a flexible three-dimensional deep algorithm, called heterophase volumetric detection (HPVD), that can accept any combination of contrast-phase inputs with adjustable sensitivity depending on the clinical purpose. We trained HPVD on 771 DCE CT scans to detect HCCs and evaluated it on 164 positives and 206 controls. We compared performance against six clinical readers, including two radiologists, two hepatopancreaticobiliary surgeons, and two hepatologists. The area under the curve of the localization receiver operating characteristic for NC-only, NC plus VP, and full DCE CT yielded 0.71 (95% confidence interval [CI], 0.64-0.77), 0.81 (95% CI, 0.75-0.87), and 0.89 (95% CI, 0.84-0.93), respectively. At a high-sensitivity operating point of 80% on DCE CT, HPVD achieved 97% specificity, which is comparable to measured physician performance. We also demonstrated performance improvements over more typical and less flexible nonheterophase detectors. Conclusion: A single deep-learning algorithm can be effectively applied to diverse HCC detection clinical scenarios, indicating that HPVD could serve as a useful clinical aid for at-risk and opportunistic HCC surveillance.

Automatic Detection of Liver Cancer Using Hybrid Pre-Trained Models

Liver cancer is a life-threatening illness and one of the fastest-growing cancer types in the world. Consequently, the early detection of liver cancer leads to lower mortality rates. This work aims to build a model that will help clinicians determine the type of tumor when it occurs within the liver region by analyzing images of tissue taken from a biopsy of this tumor. Working within this stage requires effort, time, and accumulated experience that must be possessed by a tissue expert to determine whether this tumor is malignant and needs treatment. Thus, a histology expert can make use of this model to obtain an initial diagnosis. This study aims to propose a deep learning model using convolutional neural networks (CNNs), which are able to transfer knowledge from pre-trained global models and decant this knowledge into a single model to help diagnose liver tumors from CT scans. Thus, we obtained a hybrid model capable of detecting CT images of a biopsy of a liver tumor. The best results that we obtained within this research reached an accuracy of 0.995, a precision value of 0.864, and a recall value of 0.979, which are higher than those obtained using other models. It is worth noting that this model was tested on a limited set of data and gave good detection results. This model can be used as an aid to support the decisions of specialists in this field and save their efforts. In addition, it saves the effort and time incurred by the treatment of this type of cancer by specialists, especially during periodic examination campaigns every year.

Development of an abdominal phantom for the validation of an oligometastatic disease diagnosis workflow

PURPOSE

The liver is a common site for metastatic disease, which is a challenging and life-threatening condition with a grim prognosis and outcome. We propose a standardized workflow for the diagnosis of oligometastatic disease (OMD), as a gold standard workflow has not been established yet. The envisioned workflow comprises the acquisition of a multimodal image dataset, novel image processing techniques, and cone beam computed tomography (CBCT)-guided biopsy for subsequent molecular subtyping. By combining morphological, molecular, and functional information about the tumor, a patient-specific treatment planning is possible. We designed and manufactured an abdominal liver phantom that we used to demonstrate multimodal image acquisition, image processing, and biopsy of the OMD diagnosis workflow.

METHODS

The anthropomorphic abdominal phantom contains a rib cage, a portal vein, lungs, a liver with six lesions, and a hepatic vessel tree. This phantom incorporates three different lesion types with varying visibility under computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography CT (PET-CT), which reflects clinical reality. The phantom is puncturable and the size of the corpus and the organs is comparable to those of a real human abdomen. By using several modern additive manufacturing techniques, the manufacturing process is reproducible and allows to incorporate patient-specific anatomies. As a first step of the OMD diagnosis workflow, a pre-interventional CT, MRI, and PET-CT dataset of the phantom was acquired. The image information was fused using image registration and organ information was extracted via image segmentation. A CBCT-guided needle puncture experiment was performed, where all six liver lesions were punctured with coaxial biopsy needles.

RESULTS

Qualitative observation of the image data and quantitative evaluation using contrast-to-noise ratio (CNR) confirms that one lesion type is visible only in MRI and not CT. The other two lesion types are visible in CT and MRI. The CBCT-guided needle placement was performed for all six lesions, including those visible only in MRI and not CBCT. This was possible by successfully merging multimodal pre-interventional image data. Lungs, bones, and liver vessels serve as realistic inhibitions during needle path planning.

CONCLUSIONS

We have developed a reusable abdominal phantom that has been used to validate a standardized OMD diagnosis workflow. Utilizing the phantom, we have been able to show that a multimodal imaging pipeline is advantageous for a comprehensive detection of liver lesions. In a CBCT-guided needle placement experiment we have punctured lesions that are invisible in CBCT using registered pre-interventional MRI scans for needle path planning. This article is protected by copyright. All rights reserved.

Non-Coding RNA-Based Biosensors for Early Detection of Liver Cancer

Primary liver cancer is an aggressive, lethal malignancy that ranks as the fourth leading cause of cancer-related death worldwide. Its 5-year mortality rate is estimated to be more than 95%. This significant low survival rate is due to poor diagnosis, which can be referred to as the lack of sufficient and early-stage detection methods. Many liver cancer-associated non-coding RNAs (ncRNAs) have been extensively examined to serve as promising biomarkers for precise diagnostics, prognostics, and the evaluation of the therapeutic progress. For the simple, rapid, and selective ncRNA detection, various nanomaterial-enhanced biosensors have been developed based on electrochemical, optical, and electromechanical detection methods. This review presents ncRNAs as the potential biomarkers for the early-stage diagnosis of liver cancer. Moreover, a comprehensive overview of recent developments in nanobiosensors for liver cancer-related ncRNA detection is provided.

68Ga-FAPI-04 Versus 18F-FDG PET/CT in the Detection of Hepatocellular Carcinoma

Background

Fibroblast activation protein (FAP) is commonly expressed in activated stromal fibroblasts in various epithelial tumours. Recently, ⁶⁸Ga-FAPI-04 has been used for tumour imaging in positron emission tomography/computed tomography (PET/CT). This study aimed to compare the diagnostic performances of ⁶⁸Ga-FAPI-04 PET/CT and ¹⁸F-FDG PET/CT in hepatocellular carcinoma (HCC), and to assess factors associated with ⁶⁸Ga-FAPI-04 uptake in HCC.

Materials and Methods

Twenty-nine patients with suspiciously HCC who received both ¹⁸F-FDG and ⁶⁸Ga-FAPI-04 PET/CT were included in this retrospective study. The results were interpreted by two experienced nuclear medicine physicians independently. The maximum and mean standardized uptake values (SUV_max and SUV_mean) were measured in the lesions and liver background, respectively. The tumour-to-background ratio (TBR) was then calculated as lesion's SUV_max divided by background SUV_mean.

Results

A total of 35 intrahepatic lesions in 25 patients with HCC were finally involved in the statistical analysis. ⁶⁸Ga-FAPI-04 PET/CT showed a higher sensitivity than ¹⁸F-FDG PET/CT in detecting intrahepatic HCC lesions (85.7% vs. 57.1%, P = 0.002), including in small (≤ 2 cm in diameter; 68.8% vs. 18.8%, P = 0.008) and well- or moderately-differentiated (83.3% vs. 33.3%, P = 0.031) tumors. SUV_max was comparable between ⁶⁸Ga-FAPI-04 and ¹⁸F-FDG (6.96 ± 5.01 vs. 5.89 ± 3.38, P > 0.05), but the TBR was significantly higher in the ⁶⁸Ga-FAPI-04 group compared with the ¹⁸F-FDG group (11.90 ± 8.35 vs. 3.14 ± 1.59, P < 0.001). SUV_max and the TBR in ⁶⁸Ga-FAPI-04 positive lesions were associated with tumour size (both P < 0.05), but not the remaining clinical and pathological features (all P > 0.05).

Conclusions

⁶⁸Ga-FAPI-04 PET/CT is more sensitive than ¹⁸F-FDG PET/CT in detecting HCC lesions, and ⁶⁸Ga-FAPI-04 uptake is correlated mainly with tumour size.

Endoscopic Ultrasound for Detection of Liver Metastasis: Hope or Hype?

Transabdominal ultrasonography, contrast-enhanced computed tomography, and magnetic resonance imaging (MRI) are the common diagnostic tests for the detection of hepatic lesions. Use of enhanced and advanced MRI technique, that is, diffusion weighted MRI and hepatocyte-specific contrast agents, has further improved the accuracy of detection of metastatic liver lesions ≤10 mm in diameter. However, even with these advanced imaging modalities sensitivity is low for lesions smaller than 10 mm when compared with standard intraoperative ultrasound. Endoscopic ultrasound (EUS) is an emerging imaging modality with resolution sufficient to detect and sample lesions as small as 5 mm in diameter. In this news and views, we have discussed the role of standard and enhanced EUS for the detection of metastatic liver lesions.

Assessment of SE-MRE-derived shear stiffness at 3.0 Tesla for solid liver tumors characterization

OBJECTIVES

To evaluate the feasibility and diagnostic value of using a 2D spin-echo MR elastography (SE-MRE) sequence at 3.0 Tesla for solid focal liver lesions (FLL) characterization.

METHODS

This prospective study included 55 patients with solid FLL (size > 20 mm), who underwent liver SE-MRE at 3 Tesla between 2016 and 2019. Stiffness measurements were performed by two independent readers blinded to the complete MRI exam or patient information. Histological confirmation or typical behavior on the complete MRI exam evaluated in consensus by expert abdominal radiologists was used as reference standard. FLLs were grouped and compared (malignant vs. benign) using the Mann-Whitney and Kruskal-Wallis tests. MRE diagnostic performance was assessed, and stiffness cutoffs were obtained by analysis of ROC curves from accuracy maximization. A linear regression plot was used to evaluate inter-rater agreement for FLLs stiffness measurements. p values < 0.05 were considered statistically significant.

RESULTS

The final study group comprised 57 FLLs (34 malignant, 23 benign). Stiffness measurements were technically successful in 91.23% of lesions. To both readers, the median stiffness of the lesions categorized as benign was 4.5 ± 1.5 kPa and in the malignant group 6.8 ± 1.7 and 7.5 ± 1.5 kPa depending on the reader. A cutoff of 5.8 kPa distinguished malignant and benign lesions with 88% specificity and 75-85% accuracy depending on the reader. The inter-rater agreement was 0.90 ± 0.04 with a correlation coefficient of 0.94.

CONCLUSION

2D-SE-MRE at 3.0 T provides high specificity and PPV to differentiate benign from malignant liver lesions. Trial registration 18FFUA-A02.

Silica nanoparticle assists determining liver cancer gene sequence on interdigitated electrode surface

A high-performance interdigitated electrode (IDE) biosensing surface was reported here by utilizing self-assembled silica nanoparticle (SiNP). The modified surface was used to evaluate the complementation of hairpin forming region from Mitoxantrone resistance gene 7 (MXR7; liver cancer-related short gene). The conjugated SiNPs on 3-aminopropyl triethoxysilane functionalization were captured with probe sequence on IDE biosensing surface. The physical and chemically modified surface was used to quantify MXR7 and an increment in the current response upon complementation was noticed. Limit of target DNA detection was calculated (1-10 fM) and this label-free detection is at the comparable level to the fluorescent-based sensing. A linear regression was calculated [y = 0.243x - 0.0773; R² = 0.9336] and the sensitivity was 1 fM on the linear range of 1 fM to 10 pM. With the strong attachment of capture DNA on IDE through SiNP, the surface clearly discriminates the specificity (complementary) versus nonspecificity (complete-, single-, and triple-mismatched sequences). This detection strategy helps to determine liver cancer progression and the similar strategy can be followed for other gene sequence complementation.

Ensemble classification for predicting the malignancy level of pulmonary nodules on chest computed tomography images

Early identification and classification of pulmonary nodules are essential for improving the survival rates of individuals with lung cancer and are considered to be key requirements for computer-assisted diagnosis. To address this topic, the present study proposed a method for predicting the malignant phenotype of pulmonary nodules based on weighted voting rules. This method used the pulmonary nodule regions of interest as the input data and extracted the features of the pulmonary nodules using the Denoising Auto Encoder, ResNet-18. Moreover, the software also modifies texture and shape features to assess the malignant phenotype of the pulmonary nodules. Based on their classification accuracy (Acc), the different classifiers were assigned to different weights. Finally, an integrated classifier was obtained to score the malignant phenotype of the pulmonary nodules. The present study included training and testing experiments conducted by extracting the corresponding lung nodule image data from the Lung Image Database Consortium-Image Database Resource Initiative. The results of the present study indicated a final classification Acc of 93.10±2.4%, demonstrating the feasibility and effectiveness of the proposed method. This method includes the powerful feature extraction ability of deep learning combined with the ability to use traditional features in image representation.

Mapping hepatocyte size in vivo using temporal diffusion spectroscopy MRI

PURPOSE

The goal of this study is to implement a noninvasive method for in vivo mapping of hepatocyte size. This method will have a broad range of clinical and preclinical applications, as pathological changes in hepatocyte sizes are relevant for the accurate diagnosis and assessments of treatment response of liver diseases.

METHODS

Building on the concepts of temporal diffusion spectroscopy in MRI, a clinically feasible imaging protocol named IMPULSED (Imaging Microstructural Parameters Using Limited Spectrally Edited Diffusion) has been developed, which is able to report measurements of cell sizes noninvasively. This protocol acquires a selected set of diffusion imaging data and fits them to a model of water compartments in tissues to derive robust estimates of the cellular structures that restrict free diffusion. Here, we adapt and further develop this approach to measure hepatocyte sizes in vivo. We validated IMPULSED in livers of mice and rats and implemented it to image healthy human subjects using a clinical 3T MRI scanner.

RESULTS

The IMPULSED-derived mean hepatocyte sizes for rats and mice are about 15-20 µm and agree well with histological findings. Maps of mean hepatocyte size for humans can be achieved in less than 15 minutes, a clinically feasible scan time.

CONCLUSION

Our results suggest that this method has potential to overcome major limitations of liver biopsy and provide noninvasive mapping of hepatocyte sizes in clinical applications.

Value of ultrasound combined with MRI in the diagnosis of primary and recurrent hepatocellular carcinoma

Ultrasound (US) combined with magnetic resonance imaging (MRI) in the diagnosis of primary hepatocellular carcinoma (PHCC) and recurrent hepatocellular carcinoma (RHCC) were compared. The clinical data of 329 patients with hepatocellular carcinoma (HCC) admitted to Qingdao Women and Children's Hospital from June 2015 to December 2017 were collected. One hundred and sixty patients with PHCC were regarded as the PHCC group, and the other 169 patients with RHCC were regarded as the RHCC group. US and MRI were used in the imaging diagnosis of both groups and the results of US combined with MRI, US, and MRI alone were compared. The lesion size in the PHCC group was significantly higher than that in the RHCC group (P<0.05). The MRI fast-in and fast-out rates of the two groups were significantly higher than those of the other three methods (P<0.05). The coincidence rate of MRI in the two groups was higher than that of computed tomography (CT), US, and US combined with MRI (P<0.05). The coincidence rates of CT, US, MRI, and US combined with MRI in PHCC group were significantly higher than those in RHCC group. In PHCC group, MRI was superior to the other methods in the detection of micro HCC (P<0.05). In RHCC group, MRI was significantly better than US in the detection of micro HCC (P<0.05). The sensitivity, specificity, positive predictive value and negative predictive value of MRI were significantly better than the other three methods (P<0.05). MRI alone has the best diagnostic efficacy for micro HCC-type lesions. The diagnostic efficacy of MRI, US, CT, and US combined with MRI in PHCC was better than those in RHCC. In addition to imaging examination, the diagnosis of RHCC should be combined with other indicators for comprehensive diagnosis.

Domain-Agnostic Learning with Anatomy-Consistent Embedding for Cross-Modality Liver Segmentation

Domain Adaptation (DA) has the potential to greatly help the generalization of deep learning models. However, the current literature usually assumes to transfer the knowledge from the source domain to a specific known target domain. Domain Agnostic Learning (DAL) proposes a new task of transferring knowledge from the source domain to data from multiple heterogeneous target domains. In this work, we propose the Domain-Agnostic Learning framework with Anatomy-Consistent Embedding (DALACE) that works on both domain-transfer and task-transfer to learn a disentangled representation, aiming to not only be invariant to different modalities but also preserve anatomical structures for the DA and DAL tasks in cross-modality liver segmentation. We validated and compared our model with state-of-the-art methods, including CycleGAN, Task Driven Generative Adversarial Network (TD-GAN), and Domain Adaptation via Disentangled Representations (DADR). For the DA task, our DALACE model outperformed CycleGAN, TD-GAN, and DADR with DSC of 0.847 compared to 0.721, 0.793 and 0.806. For the DAL task, our model improved the performance with DSC of 0.794 from 0.522, 0.719 and 0.742 by CycleGAN, TD-GAN, and DADR. Further, we visualized the success of disentanglement, which added human interpretability of the learned meaningful representations. Through ablation analysis, we specifically showed the concrete benefits of disentanglement for downstream tasks and the role of supervision for better disentangled representation with segmentation consistency to be invariant to domains with the proposed Domain-Agnostic Module (DAM) and to preserve anatomical information with the proposed Anatomy-Preserving Module (APM).

Unsupervised Domain Adaptation via Disentangled Representations: Application to Cross-Modality Liver Segmentation

A deep learning model trained on some labeled data from a certain source domain generally performs poorly on data from different target domains due to domain shifts. Unsupervised domain adaptation methods address this problem by alleviating the domain shift between the labeled source data and the unlabeled target data. In this work, we achieve cross-modality domain adaptation, i.e. between CT and MRI images, via disentangled representations. Compared to learning a one-to-one mapping as the state-of-art CycleGAN, our model recovers a manyto-many mapping between domains to capture the complex cross-domain relations. It preserves semantic feature-level information by finding a shared content space instead of a direct pixelwise style transfer. Domain adaptation is achieved in two steps. First, images from each domain are embedded into two spaces, a shared domain-invariant content space and a domain-specific style space. Next, the representation in the content space is extracted to perform a task. We validated our method on a cross-modality liver segmentation task, to train a liver segmentation model on CT images that also performs well on MRI. Our method achieved Dice Similarity Coefficient (DSC) of 0.81, outperforming a CycleGAN-based method of 0.72. Moreover, our model achieved good generalization to joint-domain learning, in which unpaired data from different modalities are jointly learned to improve the segmentation performance on each individual modality. Lastly, under a multi-modal target domain with significant diversity, our approach exhibited the potential for diverse image generation and remained effective with DSC of 0.74 on multi-phasic MRI while the CycleGAN-based method performed poorly with a DSC of only 0.52.

Characterization of high- and low-risk hepatocellular adenomas by magnetic resonance imaging in an animal model of glycogen storage disease type 1A

Hepatocellular adenomas (HCAs) are benign tumors, of which the most serious complications are hemorrhage and malignant transformation to hepatocellular carcinoma (HCC). Among the various subtypes of HCA, the β-catenin-activated subtype (bHCA) is associated with greatest risk of malignant transformation. Magnetic resonance imaging (MRI) is an important tool to differentiate benign and malignant hepatic lesions, and preclinical experimental approaches may help to develop a method to identify MRI features associated with bHCA. HCAs are associated with various pathologies, including glycogen storage disease 1a (GSD1a). Here, we utilized a mouse model for GSD1a that develops HCA and HCC, and analyzed the mice in order to distinguish low-risk from high-risk tumors. Animals were scanned by MRI using a hepato-specific contrast agent. The mice were sacrificed after MRI and their lesions were classified using immunohistochemistry. We observed that 45% of the animals developed focal lesions, and MRI identified four different patterns after contrast administration: isointense, hyperintense and hypointense lesions, and lesions with peripheral contrast enhancement. After contrast administration, only bHCA and HCC were hypointense in T1-weighted imaging and mildly hyperintense in T2-weighted imaging. Thus, high-risk adenomas display MRI features clearly distinguishable from those exhibited by low-risk adenomas, indicating that MRI is a reliable method for early diagnosis and classification of HCA, necessary for correct patient management.

Biomedical Imaging: Principles, Technologies, Clinical Aspects, Contrast Agents, Limitations and Future Trends in Nanomedicines

This review article presents the state-of-the-art in the major imaging modalities supplying relevant information on patient health by real-time monitoring to establish an accurate diagnosis and potential treatment plan. We draw a comprehensive comparison between all imagers and ultimately end with our focus on two main types of scanners: X-ray CT and MRI scanners. Numerous types of imaging probes for both imaging techniques are described, as well as reviewing their strengths and limitations, thereby showing the current need for the development of new diagnostic contrast agents (CAs). The role of nanoparticles in the design of CAs is then extensively detailed, reviewed and discussed. We show how nanoparticulate agents should be promising alternatives to molecular ones and how they are already paving new routes in the field of nanomedicine.

Endoscopic Ultrasound for the Diagnosis and Staging of Liver Tumors

Endoscopic ultrasound examination may provide complementary information to cross-sectional imaging in lesions of the liver, portal vein, and surrounding lymph nodes. With fine needle aspiration, endoscopic ultrasound examination is a powerful tool for the diagnosis of focal liver lesions and has usefulness in the evaluation of indeterminate liver lesions. Endoscopic ultrasound examination may influence hepatocellular cancer staging and Endoscopic ultrasound examination with fine needle aspiration of locoregional nodes and portal vein thromboses changes management. Contrast-enhanced endoscopic ultrasound examination and endoscopic ultrasound examination elastography are likely to expand the usefulness of endoscopic ultrasound examination in evaluating liver malignancy with technologic improvements.

Lean Body Weight-Tailored Iodinated Contrast Injection in Obese Patient: Boer versus James Formula

Purpose

To prospectively compare the performance of James and Boer formula in contrast media (CM) administration, in terms of image quality and parenchymal enhancement in obese patients undergoing CT of the abdomen.

Materials and Methods

Fifty-five patients with a body mass index (BMI) greater than 35 kg/m² were prospectively included in the study. All patients underwent 64-row CT examination and were randomly divided in two groups: 26 patients in Group A and 29 patients in Group B. The amount of injected CM was computed according to the patient's lean body weight (LBW), estimated using either Boer formula (Group A) or James formula (Group B). Patient's characteristics, CM volume, contrast-to-noise ratio (CNR) of liver, aorta and portal vein, and liver contrast enhancement index (CEI) were compared between the two groups. For subjective image analysis readers were asked to rate the enhancement of liver, kidneys, and pancreas based on a 5-point Likert scale.

Results

Liver CNR, aortic CNR, and portal vein CNR showed no significant difference between Group A and Group B (all P ≥ 0.177). Group A provided significantly higher CEI compared to Group B (P = 0.007). Group A and Group B returned comparable overall subjective enhancement values (3.54 and vs 3.20, all P ≥ 0.199).

Conclusions

Boer formula should be the method of choice for LBW estimation in obese patients, leading to an accurate CM amount calculation and an optimal liver contrast enhancement in CT.

The Combination of Glycolytic Inhibitor 2-Deoxyglucose and Microbubbles Increases the Effect of 5-Aminolevulinic Acid-Sonodynamic Therapy in Liver Cancer Cells

Sonodynamic therapy (SDT) overcomes the shortcoming of photodynamic therapy in the treatment of cancer. Previous studies indicated that the glycolysis inhibitor 2-deoxyglucose (2-DG) potentiated photodynamic therapy induced tumor cell death and microbubbles (MBs) improved the SDT performance. We hypothesized that the combination of 2-DG and MBs will increase the effect of 5-aminolevulinic acid (ALA)-SDT in HepG2 liver cancer cells. When cells were treated with 5-min ALA-SDT and 2-mmol/L 2-DG, the cell survival rate decreased to 73.0 ± 7.1% and 75.2 ± 7.9%, respectively. Furthermore, 2 mmol/L 2-DG increased 5-min ALA-SDT induced growth inhibition and augmented ALA-SDT induced cell apoptotic rate from 9.8 ± 0.7% to 17.4 ± 2.2%. In the combination group (2-DG and ALA-SDT group), HepG2 cells possessed typical apoptotic characters. 2-DG also increased ALA-SDT associated intracellular reactive oxygen species generation and loss of mitochondrial membrane potential. Moreover, SonoVue MBs had stimulatory function on cell viability inhibition, apoptosis, reactive oxygen species production and mitochondrial membrane potential loss for combination treatment. This study suggests a promising therapeutic strategy using a combination of 2-DG, MBs and ALA-SDT for treating liver cancer.

Imaging in Colorectal Cancer: Progress and Challenges for the Clinicians

The use of imaging in colorectal cancer (CRC) has significantly evolved over the last twenty years, establishing important roles in surveillance, diagnosis, staging, treatment selection and follow up. The range of modalities has broadened with the development of novel tracer and contrast agents, and the fusion of technologies such as positron emission tomography (PET) and computed tomography (CT). Traditionally, the most widely used modality for assessing treatment response in metastasised colon and rectal tumours is CT, combined with use of the RECIST guidelines. However, a growing body of evidence suggests that tumour size does not always adequately correlate with clinical outcomes. Magnetic resonance imaging (MRI) is a more versatile technique and dynamic contrast-enhanced (DCE)-MRI and diffusion-weighted (DW)-MRI may be used to evaluate biological and functional effects of treatment. Integrated fluorodeoxyglucose (FDG)-PET/CT combines metabolic and anatomical imaging to improve sensitivity and specificity of tumour detection, and a number of studies have demonstrated improved diagnostic accuracy of this modality in a variety of tumour types, including CRC. These developments have enabled the progression of treatment strategies in rectal cancer and improved the detection of hepatic metastatic disease, yet are not without their limitations. These include technical, economical and logistical challenges, along with a lack of robust evidence for standardisation and formal guidance. In order to successfully apply these novel imaging techniques and utilise their benefit to provide truly personalised cancer care, advances need to be clinically realised in a routine and robust manner.

Monitoring the Growth of an Orthotopic Tumour Xenograft Model: Multi-Modal Imaging Assessment with Benchtop MRI (1T), High-Field MRI (9.4T), Ultrasound and Bioluminescence

BACKGROUND

Research using orthotopic and transgenic models of cancer requires imaging methods to non-invasively quantify tumour burden. As the choice of appropriate imaging modality is wide-ranging, this study aimed to compare low-field (1T) magnetic resonance imaging (MRI), a novel and relatively low-cost system, against established preclinical techniques: bioluminescence imaging (BLI), ultrasound imaging (US), and high-field (9.4T) MRI.

METHODS

A model of colorectal metastasis to the liver was established in eight mice, which were imaged with each modality over four weeks post-implantation. Tumour burden was assessed from manually segmented regions.

RESULTS

All four imaging systems provided sufficient contrast to detect tumours in all of the mice after two weeks. No significant difference was detected between tumour doubling times estimated by low-field MRI, ultrasound imaging or high-field MRI. A strong correlation was measured between high-field MRI estimates of tumour burden and all the other modalities (p < 0.001, Pearson).

CONCLUSION

These results suggest that both low-field MRI and ultrasound imaging are accurate modalities for characterising the growth of preclinical tumour models.

Approach to the Solitary Liver Lesion: Imaging and When to Biopsy

The characterization and management of focal liver lesions is a commonly encountered problem in radiology. While the imaging findings will often be diagnostic, in equivocal cases the decision of how to proceed may be challenging. The primary modalities for liver lesion characterization are multiphase contrast-enhanced computed tomography and magnetic resonance imaging. Most lesions have typical imaging features, and when taken in conjunction with patient demographics and biochemistry the diagnosis can usually be made. Ancillary imaging modalities such as contrast-enhanced ultrasound and hepatobiliary specific contrast agents are also useful. Cirrhotic livers present a challenge due to the spectrum of benign, dysplastic, and malignant nodules that can occur. The report should include information necessary for accurate staging, and published standardized reporting guidelines should be taken into consideration. A decision to proceed to biopsy should be made only after multidisciplinary review of the case. If biopsy is required, fine needle aspiration is usually sufficient, though core needle biopsy may be required in certain circumstances.

The Research of Feature Extraction Method of Liver Pathological Image Based on Multispatial Mapping and Statistical Properties

We propose a new feature extraction method of liver pathological image based on multispatial mapping and statistical properties. For liver pathological images of Hematein Eosin staining, the image of R and B channels can reflect the sensitivity of liver pathological images better, while the entropy space and Local Binary Pattern (LBP) space can reflect the texture features of the image better. To obtain the more comprehensive information, we map liver pathological images to the entropy space, LBP space, R space, and B space. The traditional Higher Order Local Autocorrelation Coefficients (HLAC) cannot reflect the overall information of the image, so we propose an average correction HLAC feature. We calculate the statistical properties and the average gray value of pathological images and then update the current pixel value as the absolute value of the difference between the current pixel gray value and the average gray value, which can be more sensitive to the gray value changes of pathological images. Lastly the HLAC template is used to calculate the features of the updated image. The experiment results show that the improved features of the multispatial mapping have the better classification performance for the liver cancer.

Percutaneous real-time sonoelastography as a non-invasive tool for the characterization of solid focal liver lesions: A prospective study

BACKGROUND

Real-time sonoelastography is currently used for the characterization of superficial solid lesions such as thyroid and breast masses. This study evaluates the usefulness of percutaneous sonoelastography for the characterization of solid focal liver lesions.

METHODS

30 out of 43 patients with 38 known liver lesions were included in a prospective, diagnostic study. Qualitative analysis (pattern of deformation, elasticity type of liver tumour) and semi-quantitative measurements (strain ratio, hardness percentage, histogram) were evaluated. Sensitivity, specificity, positive and negative predictive values were calculated and the area under the receiver operating characteristics curve was constructed.

RESULTS

Patterns A and C-D are specific of benign lesions and metastases respectively. The patterns for haemangiomas, focal nodular hyperplasia and metastases were significantly different to each other in terms of strain ratio, hardness percentage and histogram (p<0.05). A statistically significant difference (p<0.001) was observed between the median values of the 3 measured parameters for benign (1.02; 12%; 47) and malignant lesions (1.66; 65%; 20.5) respectively. The area under the receiver operating characteristics curve values for strain ratio, hardness percentage and histogram were 0.88, 0.89, and 0.86 respectively for cut-off values of 1.2, 45, and 30.

CONCLUSIONS

By percutaneous sonoelastography it is possible to differentiate benign versus malignant focal liver lesions, metastases in particular, with good diagnostic performance.

Semiautomated hybrid algorithm for estimation of three-dimensional liver surface in CT using dynamic cellular automata and level-sets

Liver segmentation continues to remain a major challenge, largely due to its intense complexity with surrounding anatomical structures (stomach, kidney, and heart), high noise level and lack of contrast in pathological computed tomography (CT) data. We present an approach to reconstructing the liver surface in low contrast CT. The main contributions are: (1) a stochastic resonance-based methodology in discrete cosine transform domain is developed to enhance the contrast of pathological liver images, (2) a new formulation is proposed to prevent the object boundary, resulting from the cellular automata method, from leaking into the surrounding areas of similar intensity, and (3) a level-set method is suggested to generate intermediate segmentation contours from two segmented slices distantly located in a subject sequence. We have tested the algorithm on real datasets obtained from two sources, Hamad General Hospital and medical image computing and computer-assisted interventions grand challenge workshop. Various parameters in the algorithm, such as [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], [Formula: see text], and [Formula: see text], play imperative roles, thus their values are precisely selected. Both qualitative and quantitative evaluation performed on liver data show promising segmentation accuracy when compared with ground truth data reflecting the potential of the proposed method.

Evaluation of Selective Arterial Embolization Effect by Chitosan Micro-Hydrogels in Hindlimb Sarcoma Rodent Models Using Various Imaging Modalities

PURPOSE

Embolization is mainly used to reduce the size of locally advanced tumors. In this study, selective arterial catheterization with chitosan micro-hydrogels (CMH) into the femoral artery was performed and the therapeutic effect was validated using different imaging methods.

METHODS

Male SD rats (n = 18, 6 weeks old) were randomly assigned into three groups: Group 1 as control, Group 2 without any ligation of distal femoral artery, and Group 3 with temporary ligation of the distal femoral artery. RR1022 sarcoma cell lines were inoculated into thigh muscle. After 1 week, CMH was injected into the proximal femoral artery. Different imaging modalities were performed during a 3-week follow-up.

RESULTS

The tumor size was significantly (P < 0.001) decreased in both Group 2 and Group 3 (P < 0.001) after selective arterial embolization therapy. (18)F-FDG-PET/CT revealed decreased intensity of (18)F-FDG uptake in tumors. The accumulation status of (125)I-CMH near the tumor was verified by gamma camera.

CONCLUSIONS

Appropriate selective arterial embolization therapy with CMH was.

Cancer stem cells in primary liver cancers: pathological concepts and imaging findings

There is accumulating evidence that cancer stem cells (CSCs) play an integral role in the initiation of hepatocarcinogenesis and the maintaining of tumor growth. Liver CSCs derived from hepatic stem/progenitor cells have the potential to differentiate into either hepatocytes or cholangiocytes. Primary liver cancers originating from CSCs constitute a heterogeneous histopathologic spectrum, including hepatocellular carcinoma, combined hepatocellular-cholangiocarcinoma, and intrahepatic cholangiocarcinoma with various radiologic manifestations. In this article, we reviewed the recent concepts of CSCs in the development of primary liver cancers, focusing on their pathological and radiological findings. Awareness of the pathological concepts and imaging findings of primary liver cancers with features of CSCs is critical for accurate diagnosis, prediction of outcome, and appropriate treatment options for patients.

Upconversion improvement by the reduction of Na+-vacancies in Mn2+ doped hexagonal NaYbF4:Er3+ nanoparticles

A method of Mn²⁺ doping for the simultaneous control of lattice defects and luminescence output in β-NaYbF₄:Er³⁺ upconversion nanoparticles with a fixed composition of both host and dopants of Ln³⁺ is demonstrated.

Establishment of animal models with orthotopic hepatocellular carcinoma

Hepatocellular carcinoma (HCC) is one of the most serious health problems worldwide. Many researchers have investigated HCC at the level of genes, ribonucleic acid, proteins, cells, and animals. The resultant development of animal models and monitoring methods has improved the effectiveness of guidelines provided to researchers working with preclinical HCC models. HCC in animal models and clinical patients is monitored by various current imaging modalities such as ultrasound (US) imaging, computed tomography (CT), magnetic resonance imaging (MRI), single photon emission computed tomography (SPECT), positron emission tomography (PET) and bioluminescence imaging (BLI). These techniques are currently used for both preclinical and clinical assessment, and provide valuable diagnostic information. In this article, we have mainly reviewed the established animal models and the assessment of orthotopic HCC using imaging modalities. Additionally, we have introduced a method of orthotopic HCC rat model developed in our laboratory. We have furthermore evaluated the occurrence of tumor mass using molecular imaging techniques.