A Diagnostic Test Meta-Analysis Evaluating Imaging-Based and Blood Biomarker-Based Assessment Tools for Fibrosis After Liver Transplantation

A deep learning framework for personalised dynamic diagnosis of graft fibrosis after liver transplantation: a retrospective, single Canadian centre, longitudinal study

BACKGROUND

Recurrent graft fibrosis after liver transplantation can threaten both graft and patient survival. Therefore, early detection of fibrosis is essential to avoid disease progression and the need for retransplantation. Non-invasive blood-based biomarkers of fibrosis are limited by moderate accuracy and high cost. We aimed to evaluate the accuracy of machine learning algorithms in detecting graft fibrosis using longitudinal clinical and laboratory data.

METHODS

In this retrospective, longitudinal study, we trained machine learning algorithms, including our novel weighted long short-term memory (LSTM) model, to predict the risk of significant fibrosis using follow-up data from 1893 adults who had a liver transplantation between Feb 1, 1987, and Dec 30, 2019, with at least one liver biopsy post transplantation. Liver biopsy samples with indefinitive fibrosis stage and those from patients with multiple transplantations were excluded. Longitudinal clinical variables were collected from transplantation to the date of last available liver biopsy. Deep learning models were trained on 70% of the patients as the training set and 30% of the patients as the test set. The algorithms were also separately tested on longitudinal data from patients in a subgroup of patients (n=149) who had transient elastography within 1 year before or after the date of liver biopsy. Weighted LSTM model performance for diagnosing significant fibrosis was compared against LSTM, other deep learning models (recurrent neural network and temporal convolutional network), and machine learning models (Random Forest, Support vector machines, Logistic regression, Lasso regression, and Ridge regression) and aspartate aminotransferase-to-platelet ratio index (APRI), fibrosis-4 index (FIB-4), and transient elastography.

FINDINGS

1893 people who had a liver transplantation (1261 [67%] men and 632 [33%] women) with at least one liver biopsy between Jan 1, 1992, and June 30, 2020, were included in the study (591 [31%] cases and 1302 [69%] controls). The median age at liver transplantation was 53·7 years (IQR 47·3-59·0) for cases and 55·3 years (48·0 to 61·2) for controls. The median time interval between transplant and liver biopsy was 21 months (5 to 71). The weighted LSTM model (area under the curve 0·798 [95% CI 0·790 to 0·810]) consistently outperformed other methods, including unweighted LSTM (0·761 [0·750 to 0·769]; p=0·031) Recurrent Neural Network (0·736 [0·721 to 0·744]), Temporal Convolutional Networks (0·700 [0·662 to 0·747], and Random Forest 0·679 [0·652 to 0·707]), FIB-4 (0·650 [0·636 to 0·663]) and APRI (0·682 [0·671 to 0·694]) when diagnosing F2 or worse stage fibrosis. In a subgroup of patients with transient elastography results, weighted LSTM was not significantly better at detecting fibrosis (≥F2; 0·705 [0·687 to 0·724]) than transient elastography (0·685 [0·662 to 0·704]). The top ten variables predictive for significant fibrosis were recipient age, primary indication for transplantation, donor age, and longitudinal data for creatinine, alanine aminotransferase, aspartate aminotransferase, total bilirubin, platelets, white blood cell count, and weight.

INTERPRETATION

Deep learning algorithms, particularly weighted LSTM, outperform other routinely used non-invasive modalities and could help with the earlier diagnosis of graft fibrosis using longitudinal clinical and laboratory variables. The list of most important predictive variables for the development of fibrosis will enable clinicians to modify their management accordingly to prevent onset of graft cirrhosis.

FUNDING

Canadian Institute of Health Research, American Society of Transplantation, Toronto General and Western Hospital Foundation, and Paladin Labs.

Liver transplantation immunology: Immunosuppression, rejection, and immunomodulation

Outcomes after liver transplantation have continuously improved over the past decades, but long-term survival rates are still lower than in the general population. The liver has distinct immunological functions linked to its unique anatomical configuration and to its harbouring of a large number of cells with fundamental immunological roles. The transplanted liver can modulate the immunological system of the recipient to promote tolerance, thus offering the potential for less aggressive immunosuppression. The selection and adjustment of immunosuppressive drugs should be individualised to optimally control alloreactivity while mitigating toxicities. Routine laboratory tests are not accurate enough to make a confident diagnosis of allograft rejection. Although several promising biomarkers are being investigated, none of them is sufficiently validated for routine use; hence, liver biopsy remains necessary to guide clinical decisions. Recently, there has been an exponential increase in the use of immune checkpoint inhibitors due to the unquestionable oncological benefits they provide for many patients with advanced-stage tumours. It is expected that their use will also increase in liver transplant recipients and that this might affect the incidence of allograft rejection. Currently, the evidence regarding the efficacy and safety of immune checkpoint inhibitors in liver transplant recipients is limited and cases of severe allograft rejection have been reported. In this review, we discuss the clinical relevance of alloimmune disease, the role of minimisation/withdrawal of immunosuppression, and provide practical guidance for using checkpoint inhibitors in liver transplant recipients.

Prediction of long-term morbidity and mortality after liver transplantation using two-dimensional shear wave elastography compared with liver biopsy

The role of noninvasive liver disease assessment by two-dimensional shear wave elastography (2D-SWE) to diagnose fibrosis is well described in patients with chronic liver disease. However, its role in prognosis, especially after liver transplantation (LT) has not been adequately examined. We hypothesized that elevated liver stiffness measurement (LSM) as measured by 2D-SWE after LT predicts future morbidity and mortality independent of fibrosis by liver biopsy. In a prospective cohort study, consecutive LT recipients underwent concomitant protocol 2D-SWE and protocol liver biopsy (2012-2014), with the assessor blinded to biopsy findings. We examined the baseline correlation of LSM with fibrosis stage and the association between elevated LSM and the development of subsequent clinical outcomes and all-cause mortality. A total of 187 LT recipients (median age 58 years, 38.5% women, median body mass index 26.5 kg/m² , 55.1% hepatitis C virus, 17.6% nonalcoholic steatohepatitis/cryptogenic) were examined. Median time between LT and biopsy/2D-SWE assessment was 4.0 years, and the median follow-up time after LSM determination was 3.5 years. Median LSM was 9 kPa (8 kPa [F0/F1], 11.5 kPa [F2], 12 kPa [F3/F4]). There was a positive correlation between LSM and fibrosis stage (r_s  = 0.41; p < 0.001). LSM ≥11 kPa was associated with lower survival within 3 years (84.8 vs. 93.7%; p = 0.04). After adjusting for age, sex, and fibrosis stage, LSM ≥11 kPa was independently associated with mortality (hazard ratio, 2.45; 95% confidence interval, 1.08-5.60). Elevated LSM by 2D-SWE is associated with increased mortality after LT independent of hepatic fibrosis. Given the overall decrease in the use of liver biopsy in the current era, 2D-SWE may serve as a novel noninvasive prognostic tool to predict relevant outcomes late after LT.

Non-alcoholic fatty liver disease increases risk of carotid atherosclerosis and ischemic stroke: An updated meta-analysis with 135,602 individuals

BACKGROUND/AIMS

Non-alcoholic fatty liver disease (NAFLD) is associated with the development of cardiovascular disease. While existing studies have examined cardiac remodeling in NAFLD, there has been less emphasis on the development of carotid atherosclerosis and stroke. We sought to conduct a meta-analysis to quantify the prevalence, risk factors, and degree of risk increment of carotid atherosclerosis and stroke in NAFLD.

METHODS

Embase and Medline were searched for articles relating to NAFLD, carotid atherosclerosis, and stroke. Proportional data was analysed using a generalized linear mixed model. Pairwise meta-analysis was conducted to obtain odds ratio or weighted mean difference for comparison between patients with and without NAFLD.

RESULTS

From pooled analysis of 30 studies involving 7,951 patients with NAFLD, 35.02% (95% confidence interval [CI], 27.36-43.53%) had carotid atherosclerosis with an odds ratio of 3.20 (95% CI, 2.37-4.32; P<0.0001). Pooled analysis of 25,839 patients with NAFLD found the prevalence of stroke to be 5.04% (95% CI, 2.74-9.09%) with an odds ratio of 1.88 (95% CI, 1.23-2.88; P=0.02) compared to non-NAFLD. The degree of steatosis assessed by ultrasonography in NAFLD was closely associated with risk of carotid atherosclerosis and stroke. Older age significantly increased the risk of developing carotid atherosclerosis, but not stroke in NAFLD.

CONCLUSION

This meta-analysis shows that a stepwise increment of steatosis of NAFLD can significantly increase the risk of carotid atherosclerosis and stroke development in NAFLD. Patients more than a third sufferred from carotid atherosclerosis and routine assessment of carotid atherosclerosis is quintessential in NAFLD.