Repeatability and reproducibility of deep-learning-based liver volume and Couinaud segment volume measurement tool

Assessing the performance of AI-assisted technicians in liver segmentation, Couinaud division, and lesion detection: a pilot study

BACKGROUND

In patients with primary and secondary liver cancer, the number and sizes of lesions, their locations within the Couinaud segments, and the volume and health status of the future liver remnant are key for informing treatment planning. Currently this is performed manually, generally by trained radiologists, who are seeing an inexorable growth in their workload. Integrating artificial intelligence (AI) and non-radiologist personnel into the workflow potentially addresses the increasing workload without sacrificing accuracy. This study evaluated the accuracy of non-radiologist technicians in liver cancer imaging compared with radiologists, both assisted by AI.

METHODS

Non-contrast T1-weighted MRI data from 18 colorectal liver metastasis patients were analyzed using an AI-enabled decision support tool that enables non-radiology trained technicians to perform key liver measurements. Three non-radiologist, experienced operators and three radiologists performed whole liver segmentation, Couinaud segment segmentation, and the detection and measurements of lesions aided by AI-generated delineations. Agreement between radiologists and non-radiologists was assessed using the intraclass correlation coefficient (ICC). Two additional radiologists adjudicated any lesion detection discrepancies.

RESULTS

Whole liver volume showed high levels of agreement between the non-radiologist and radiologist groups (ICC = 0.99). The Couinaud segment volumetry ICC range was 0.77-0.96. Both groups identified the same 41 lesions. As well, the non-radiologist group identified seven more structures which were also confirmed as lesions by the adjudicators. Lesion diameter categorization agreement was 90%, Couinaud localization 91.9%. Within-group variability was comparable for lesion measurements.

CONCLUSION

With AI assistance, non-radiologist experienced operators showed good agreement with radiologists for quantifying whole liver volume, Couinaud segment volume, and the detection and measurement of lesions in patients with known liver cancer. This AI-assisted non-radiologist approach has potential to reduce the stress on radiologists without compromising accuracy.

Liver magnetic resonance imaging, non-alcoholic fatty liver disease and metabolic syndrome risk in pre-pubertal Mexican boys

Rising global pediatric obesity rates, increase non-alcoholic fatty liver disease (NAFLD) and metabolic syndrome (MetS) prevalence, with MetS being a NAFLD risk factor. NAFLD can be asymptomatic, with liver function tests insensitive to mild disease, and liver biopsy, risking complications. Thus, we investigated multiparametric MRI (mpMRI) metrics of liver fat (proton density fat fraction, PDFF) and disease activity (fibro-inflammation; iron-corrected T1, cT1), in a Hispanic pre-pubertal pediatric cohort, with increased risk of NAFLD. Pre-pubertal boys (n = 81) of varying Body-Mass Index (BMI) were recruited in Mexico City. Most children (81%) had normal liver transaminase levels, 38% had high BMI, and 14% had ≥ 3 MetS risk factors. Applying mpMRI thresholds, 12%, 7% and 4% of the cohort had NAFLD, NASH and high-risk NASH respectively. Participants with ≥ 3 MetS risk factors had higher cT1 (834 ms vs. 737 ms, p = 0.004) and PDFF (8.7% vs. 2.2%, p < 0.001) compared to those without risk factors. Those with elevated cT1 tended to have high BMI and high insulin (p = 0.005), HOMA-IR (p = 0.005) and leptin (p < 0.001). The significant association of increased risk of MetS with abnormal mpMRI, particularly cT1, proposes the potential of using mpMRI for routine pediatric NAFLD screening of high-risk (high BMI, high MetS risk score) populations.

Precision and Robust Models on Healthcare Institution Federated Learning for Predicting HCC on Portal Venous CT Images

Hepatocellular carcinoma (HCC), the most common type of liver cancer, poses significant challenges in detection and diagnosis. Medical imaging, especially computed tomography (CT), is pivotal in non-invasively identifying this disease, requiring substantial expertise for interpretation. This research introduces an innovative strategy that integrates two-dimensional (2D) and three-dimensional (3D) deep learning models within a federated learning (FL) framework for precise segmentation of liver and tumor regions in medical images. The study utilized 131 CT scans from the Liver Tumor Segmentation (LiTS) challenge and demonstrated the superior efficiency and accuracy of the proposed Hybrid-ResUNet model with a Dice score of 0.9433 and an AUC of 0.9965 compared to ResNet and EfficientNet models. This FL approach is beneficial for conducting large-scale clinical trials while safeguarding patient privacy across healthcare settings. It facilitates active engagement in problem-solving, data collection, model development, and refinement. The study also addresses data imbalances in the FL context, showing resilience and highlighting local models' robust performance. Future research will concentrate on refining federated learning algorithms and their incorporation into the continuous implementation and deployment (CI/CD) processes in AI system operations, emphasizing the dynamic involvement of clients. We recommend a collaborative human-AI endeavor to enhance feature extraction and knowledge transfer. These improvements are intended to boost equitable and efficient data collaboration across various sectors in practical scenarios, offering a crucial guide for forthcoming research in medical AI.

Novel multiparametric MRI detects improved future liver remnant quality post-dual vein embolization

BACKGROUND

Optimisation of the future liver remnant (FLR) is crucial to outcomes of extended liver resections. This study aimed to assess the quality of the FLR before and after dual vein embolization (DVE) by quantitative multiparametric MRI.

METHODS

Of 100 patients with liver metastases recruited in a clinical trial (Precision1:NCT04597710), ten consecutive patients with insufficient FLR underwent quantitative multiparametric MRI pre- and post-DVE (right portal and hepatic vein). FLR volume, liver fibro-inflammation (corrected T1) scores and fat percentage (proton density fat fraction, PDFF) were determined. Patient metrics were compared by Wilcoxon signed-rank test and statistical analysis done using R software.

RESULTS

All patients underwent uncomplicated DVE with improvement in liver remnant health, median 37 days after DVE: cT1 scores reduced from median (interquartile range) 790 ms (753-833 ms) to 741 ms (708-760 ms) p = 0.014 [healthy range <795 ms], as did PDFF from 11% (4-21%), to 3% (2-12%) p = 0.017 [healthy range <5.6%]. There was a significant increase in median (interquartile range) FLR volume from 33% (30-37%)% to 49% (44-52%), p = 0.002.

CONCLUSION

This non-invasive and reproducible MRI technique showed improvement in volume and quality of the FLR after DVE. This is a significant advance in our understanding of how to prevent liver failure in patients undergoing major liver surgery.

Evaluation of liver regeneration after hemi-hepatectomy by combining computed tomography and post-operative liver function

Background

Accurate evaluation of postoperative liver regeneration is essential to prevent postoperative liver failure.

Aims

To analyze the predictors that affect liver regeneration after hemi-hepatectomy.

Method

Patients who underwent hemi-hepatectomy in Hangzhou First People's Hospital and Hangzhou Shulan Hospital from January 2016 to December 2021 were enrolled in this study. The regeneration index (RI) was calculated by the following equation: RI = [(postoperative total liver volume {TLVpost} - future liver remnant volume {FLRV}/FLRV] × 100 %. Hepatic dysfunction was defined according to the "TBilpeak>7" standard, which was interpreted as (peak) total bilirubin (TBil) >7.0 mg/dL. Good liver regeneration was defined solely when the RI surpassed the median with hepatic dysfunction. Logistic regression analyses were performed to estimate prognostic factors affecting liver regeneration.

Result

A total of 153 patients were enrolled, with 33 in the benign group and 120 patients in the malignant group. In the entire study population, FLRV% [OR 4.087 (1.405-11.889), P = 0.010], international normalized ratio (INR) [OR 2.763 (95%CI, 1.008-7.577), P = 0.048] and TBil [OR 2.592 (95%CI, 1.177-5.710), P = 0.018] were independent prognostic factors associated with liver regeneration. In the benign group, only the computed tomography (CT) parameter FLRV% [OR, 11.700 (95%CI, 1.265-108.200), P = 0.030] predicted regeneration. In the malignant group, parenchymal hepatic resection rate (PHRR%) [OR 0.141 (95%CI, 0.040-0.499), P = 0.002] and TBil [OR 3.384 (95%CI, 1.377-8.319), P = 0.008] were independent prognostic factors.

Conclusion

FLRV%, PHRR%, TBil and INR were predictive factors associated with liver regeneration.

Can Hisense computer-assisted surgery system (Hisense CAS) improve anatomy teaching in pediatric liver surgery?

PURPOSE

This study aimed to investigate the effectiveness of the Hisense computer-assisted surgery system (CAS) in teaching pediatric liver surgical anatomy.

METHODS

The research subjects were residents who underwent standardized training at the Department of Pediatric Surgery at Yijishan Hospital of Wannan Medical College from May 2022 to May 2023.

RESULTS

The study recruited a total of 62 students, with 31 students assigned to the Hisense CAS group (12 males and 19 females) and the remaining 31 students serving as controls (Control group, 15 males and 16 females). There were no significant differences in baseline characteristics observed between the two groups. This study found that the average scores of the Hisense CAS teaching group in the liver surgery evaluations were higher than those of the control group. Specifically, the Hisense CAS group had an average score of 84.25 ± 5.70 points in the liver surgery knowledge test, 77.10 ± 8.12 points in the image reading test, and 70.58 ± 8.79 points in the surgical simulation test, while the traditional teaching group had average scores of 73.45 ± 6.12 points, 69.81 ± 6.05 points, and 66.42 ± 6.61 points, respectively; the differences between the two groups were statistically significant (P < 0.05). Furthermore, this study also found that the Hisense CAS teaching model resulted in significantly better teaching satisfaction on the part of the residents in terms of standardized teaching for physicians in pediatric liver surgical anatomy.

CONCLUSION

In conclusion, this study demonstrated greater satisfaction of the residents with the use of 3D reconstruction added to traditional teaching sessions and better performance during the posttraining evaluation.

A deep-learning approach for segmentation of liver tumors in magnetic resonance imaging using UNet+

OBJECTIVE

Radiomic and deep learning studies based on magnetic resonance imaging (MRI) of liver tumor are gradually increasing. Manual segmentation of normal hepatic tissue and tumor exhibits limitations.

METHODS

105 patients diagnosed with hepatocellular carcinoma were retrospectively studied between Jan 2015 and Dec 2020. The patients were divided into three sets: training (n = 83), validation (n = 11), and internal testing (n = 11). Additionally, 9 cases were included from the Cancer Imaging Archive as the external test set. Using the arterial phase and T2WI sequences, expert radiologists manually delineated all images. Using deep learning, liver tumors and liver segments were automatically segmented. A preliminary liver segmentation was performed using the UNet + + network, and the segmented liver mask was re-input as the input end into the UNet + + network to segment liver tumors. The false positivity rate was reduced using a threshold value in the liver tumor segmentation. To evaluate the segmentation results, we calculated the Dice similarity coefficient (DSC), average false positivity rate (AFPR), and delineation time.

RESULTS

The average DSC of the liver in the validation and internal testing sets was 0.91 and 0.92, respectively. In the validation set, manual and automatic delineation took 182.9 and 2.2 s, respectively. On an average, manual and automatic delineation took 169.8 and 1.7 s, respectively. The average DSC of liver tumors was 0.612 and 0.687 in the validation and internal testing sets, respectively. The average time for manual and automatic delineation and AFPR in the internal testing set were 47.4 s, 2.9 s, and 1.4, respectively, and those in the external test set were 29.5 s, 4.2 s, and 1.6, respectively.

CONCLUSION

UNet + + can automatically segment normal hepatic tissue and liver tumors based on MR images. It provides a methodological basis for the automated segmentation of liver tumors, improves the delineation efficiency, and meets the requirement of extraction set analysis of further radiomics and deep learning.

The Role of the Multiparametric MRI LiverMultiScanTM in the Quantitative Assessment of the Liver and Its Predicted Clinical Applications in Patients Undergoing Major Hepatic Resection for Colorectal Liver Metastasis

Liver biopsy remains the gold standard for the histological assessment of the liver. With clear disadvantages and the rise in the incidences of liver disease, the role of neoadjuvant chemotherapy in colorectal liver metastasis (CRLM) and an explosion of surgical management options available, non-invasive serological and imaging markers of liver histopathology have never been more pertinent in order to assess liver health and stratify patients considered for surgical intervention. Liver MRI is a leading modality in the assessment of hepatic malignancy. Recent technological advancements in multiparametric MRI software such as the LiverMultiScanTM offers an attractive non-invasive assay of anatomy and histopathology in the pre-operative setting, especially in the context of CRLM. This narrative review examines the evidence for the LiverMultiScanTM in the assessment of hepatic fibrosis, steatosis/steatohepatitis, and potential applications for chemotherapy-associated hepatic changes. We postulate its future role and the hurdles it must surpass in order to be implemented in the pre-operative management of patients undergoing hepatic resection for colorectal liver metastasis. Such a role likely extends to other hepatic malignancies planned for resection.

Magnetic resonance imaging and ultrasound elastography in the context of preclinical pharmacological research: significance for the 3R principles

The 3Rs principles-reduction, refinement, replacement-are at the core of preclinical research within drug discovery, which still relies to a great extent on the availability of models of disease in animals. Minimizing their distress, reducing their number as well as searching for means to replace them in experimental studies are constant objectives in this area. Due to its non-invasive character in vivo imaging supports these efforts by enabling repeated longitudinal assessments in each animal which serves as its own control, thereby enabling to reduce considerably the animal utilization in the experiments. The repetitive monitoring of pathology progression and the effects of therapy becomes feasible by assessment of quantitative biomarkers. Moreover, imaging has translational prospects by facilitating the comparison of studies performed in small rodents and humans. Also, learnings from the clinic may be potentially back-translated to preclinical settings and therefore contribute to refining animal investigations. By concentrating on activities around the application of magnetic resonance imaging (MRI) and ultrasound elastography to small rodent models of disease, we aim to illustrate how in vivo imaging contributes primarily to reduction and refinement in the context of pharmacological research.

Update on the Use of Artificial Intelligence in Hepatobiliary MR Imaging

The application of machine learning (ML) and deep learning (DL) in radiology has expanded exponentially. In recent years, an extremely large number of studies have reported about the hepatobiliary domain. Its applications range from differential diagnosis to the diagnosis of tumor invasion and prediction of treatment response and prognosis. Moreover, it has been utilized to improve the image quality of DL reconstruction. However, most clinicians are not familiar with ML and DL, and previous studies about these concepts are relatively challenging to understand. In this review article, we aimed to explain the concepts behind ML and DL and to summarize recent achievements in their use in the hepatobiliary region.

Liver fat analysis using optimized support vector machine with support vector regression

BACKGROUND

Fatty liver disease is a common condition caused by excess fat in the liver. It consists of two types: Alcoholic Fatty Liver Disease, also called alcoholic steatohepatitis, and Non-Alcoholic Fatty Liver Disease (NAFLD). As per epidemiological studies, fatty liver encompasses 9% to 32% of the general population in India and affects overweight people.

OBJECTIVE

An Optimized Support Vector Machine with Support Vector Regression model is proposed to evaluate the volume of liver fat by image analysis (LFA-OSVM-SVR).

METHOD

The input computed tomography (CT) liver images are collected from the Chennai liver foundation and Liver Segmentation (LiTS) datasets. Here, input datasets are pre-processed using Gaussian smoothing filter and bypass filter to reduce noise and improve image intensity. The proposed U-Net method is used to perform the liver segmentation. The Optimized Support Vector Machine is used to classify the liver images as fatty liver image and normal images. The support vector regression (SVR) is utilized for analyzing the fat in percentage.

RESULTS

The LFA-OSVM-SVR model effectively analyzed the liver fat from CT scan images. The proposed approach is activated in python and its efficiency is analyzed under certain performance metrics.

CONCLUSION

The proposed LFA-OSVM-SVR method attains 33.4%, 28.3%, 25.7% improved accuracy with 55%, 47.7%, 32.6% lower error rate for fatty image classification and 30%, 21%, 19.5% improved accuracy with 57.9%, 46.5%, 31.76% lower error rate for normal image classificationthan compared to existing methods such as Convolutional Neural Network (CNN) with Fractional Differential Enhancement (FDE) (CNN-FDE), Fully Convolutional Networks (FCN) and Non-negative Matrix Factorization (NMF) (FCN-NMF), and Deep Learning with Fully Convolutional Networks (FCN) (DL-FCN).

Prediction of Decompensation and Death in Advanced Chronic Liver Disease Using Deep Learning Analysis of Gadoxetic Acid-Enhanced MRI

OBJECTIVE

This study aimed to evaluate the usefulness of quantitative indices obtained from deep learning analysis of gadoxetic acid-enhanced hepatobiliary phase (HBP) MRI and their longitudinal changes in predicting decompensation and death in patients with advanced chronic liver disease (ACLD).

MATERIALS AND METHODS

We included patients who underwent baseline and 1-year follow-up MRI from a prospective cohort that underwent gadoxetic acid-enhanced MRI for hepatocellular carcinoma surveillance between November 2011 and August 2012 at a tertiary medical center. Baseline liver condition was categorized as non-ACLD, compensated ACLD, and decompensated ACLD. The liver-to-spleen signal intensity ratio (LS-SIR) and liver-to-spleen volume ratio (LS-VR) were automatically measured on the HBP images using a deep learning algorithm, and their percentage changes at the 1-year follow-up (ΔLS-SIR and ΔLS-VR) were calculated. The associations of the MRI indices with hepatic decompensation and a composite endpoint of liver-related death or transplantation were evaluated using a competing risk analysis with multivariable Fine and Gray regression models, including baseline parameters alone and both baseline and follow-up parameters.

RESULTS

Our study included 280 patients (153 male; mean age ± standard deviation, 57 ± 7.95 years) with non-ACLD, compensated ACLD, and decompensated ACLD in 32, 186, and 62 patients, respectively. Patients were followed for 11-117 months (median, 104 months). In patients with compensated ACLD, baseline LS-SIR (sub-distribution hazard ratio [sHR], 0.81; p = 0.034) and LS-VR (sHR, 0.71; p = 0.01) were independently associated with hepatic decompensation. The ΔLS-VR (sHR, 0.54; p = 0.002) was predictive of hepatic decompensation after adjusting for baseline variables. ΔLS-VR was an independent predictor of liver-related death or transplantation in patients with compensated ACLD (sHR, 0.46; p = 0.026) and decompensated ACLD (sHR, 0.61; p = 0.023).

CONCLUSION

MRI indices automatically derived from the deep learning analysis of gadoxetic acid-enhanced HBP MRI can be used as prognostic markers in patients with ACLD.

Improving the repeatability of deep learning models with Monte Carlo dropout

The integration of artificial intelligence into clinical workflows requires reliable and robust models. Repeatability is a key attribute of model robustness. Ideal repeatable models output predictions without variation during independent tests carried out under similar conditions. However, slight variations, though not ideal, may be unavoidable and acceptable in practice. During model development and evaluation, much attention is given to classification performance while model repeatability is rarely assessed, leading to the development of models that are unusable in clinical practice. In this work, we evaluate the repeatability of four model types (binary classification, multi-class classification, ordinal classification, and regression) on images that were acquired from the same patient during the same visit. We study the each model's performance on four medical image classification tasks from public and private datasets: knee osteoarthritis, cervical cancer screening, breast density estimation, and retinopathy of prematurity. Repeatability is measured and compared on ResNet and DenseNet architectures. Moreover, we assess the impact of sampling Monte Carlo dropout predictions at test time on classification performance and repeatability. Leveraging Monte Carlo predictions significantly increases repeatability, in particular at the class boundaries, for all tasks on the binary, multi-class, and ordinal models leading to an average reduction of the 95% limits of agreement by 16% points and of the class disagreement rate by 7% points. The classification accuracy improves in most settings along with the repeatability. Our results suggest that beyond about 20 Monte Carlo iterations, there is no further gain in repeatability. In addition to the higher test-retest agreement, Monte Carlo predictions are better calibrated which leads to output probabilities reflecting more accurately the true likelihood of being correctly classified.

Precision medicine for liver tumours with quantitative MRI and whole genome sequencing (Precision1 trial): study protocol for observational cohort study

INTRODUCTION

Radiogenomic analysis of patients being considered for liver resection is seldom performed in the clinic despite recent evidence indicating that quantitative MRI could improve posthepatectomy outcomes. Meanwhile, the increasingly accessible results from whole genome sequencing reporting on clinically actionable genetic biomarkers are yet to be fully integrated into the clinical care pathway.

METHODS AND ANALYSIS

A prospective observational cohort study of up to 200 participants is planned, recruiting adults with primary or secondary liver cancer being considered for liver resection at Hampshire Hospitals NHS Foundation Trust. The data will be evaluated to address the primary endpoint to calculate the proportion of participants in which the results from whole genome sequencing would have resulted in a change in clinical management. Participants will be offered an additional non-invasive quantitative MRI scan prior to the operation and the impact of the imaging results on treatment decision-making will be evaluated.

ETHICS AND DISSEMINATION

This study was reviewed by the NHS Health Research Authority and given favourable opinion by the Brighton and Sussex Research Ethics Committee (REC reference: 20/PR/0222). Research findings will be discussed with a patient and public involvement and engagement group, presented at relevant scientific conferences and published in open access journals.

TRIAL REGISTRATION NUMBER

NCT04597710.

Automated segmentation of liver segment on portal venous phase MR images using a 3D convolutional neural network

OBJECTIVE

We aim to develop and validate a three-dimensional convolutional neural network (3D-CNN) model for automatic liver segment segmentation on MRI images.

METHODS

This retrospective study evaluated an automated method using a deep neural network that was trained, validated, and tested with 367, 157, and 158 portal venous phase MR images, respectively. The Dice similarity coefficient (DSC), mean surface distance (MSD), Hausdorff distance (HD), and volume ratio (RV) were used to quantitatively measure the accuracy of segmentation. The time consumed for model and manual segmentation was also compared. In addition, the model was applied to 100 consecutive cases from real clinical scenario for a qualitative evaluation and indirect evaluation.

RESULTS

In quantitative evaluation, the model achieved high accuracy for DSC, MSD, HD and RV (0.920, 3.34, 3.61 and 1.01, respectively). Compared to manual segmentation, the automated method reduced the segmentation time from 26 min to 8 s. In qualitative evaluation, the segmentation quality was rated as good in 79% of the cases, moderate in 15% and poor in 6%. In indirect evaluation, 93.4% (99/106) of lesions could be assigned to the correct segment by only referring to the results from automated segmentation.

CONCLUSION

The proposed model may serve as an effective tool for automated anatomical region annotation of the liver on MRI images.