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

Clinical Deployment of Machine Learning Tools in Transplant Medicine: What Does the Future Hold?

Medical applications of machine learning (ML) have shown promise in analyzing patient data to support clinical decision-making and provide patient-specific outcomes. In transplantation, several applications of ML exist which include pretransplant: patient prioritization, donor-recipient matching, organ allocation, and posttransplant outcomes. Numerous studies have shown the development and utility of ML models, which have the potential to augment transplant medicine. Despite increasing efforts to develop robust ML models for clinical use, very few of these tools are deployed in the healthcare setting. Here, we summarize the current applications of ML in transplant and discuss a potential clinical deployment framework using examples in organ transplantation. We identified that creating an interdisciplinary team, curating a reliable dataset, addressing the barriers to implementation, and understanding current clinical evaluation models could help in deploying ML models into the transplant clinic setting.

Bayesian and deep-learning models applied to the early detection of ovarian cancer using multiple longitudinal biomarkers

BACKGROUND

Ovarian cancer is the most lethal of all gynecological cancers. Cancer Antigen 125 (CA125) is the best-performing ovarian cancer biomarker which however is still not effective as a screening test in the general population. Recent literature reports additional biomarkers with the potential to improve on CA125 for early detection when using longitudinal multimarker models.

METHODS

Our data comprised 180 controls and 44 cases with serum samples sourced from the multimodal arm of UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS). Our models were based on Bayesian change-point detection and recurrent neural networks.

RESULTS

We obtained a significantly higher performance for CA125-HE4 model using both methodologies (AUC 0.971, sensitivity 96.7% and AUC 0.987, sensitivity 96.7%) with respect to CA125 (AUC 0.949, sensitivity 90.8% and AUC 0.953, sensitivity 92.1%) for Bayesian change-point model (BCP) and recurrent neural networks (RNN) approaches, respectively. One year before diagnosis, the CA125-HE4 model also ranked as the best, whereas at 2 years before diagnosis no multimarker model outperformed CA125.

CONCLUSIONS

Our study identified and tested different combination of biomarkers using longitudinal multivariable models that outperformed CA125 alone. We showed the potential of multivariable models and candidate biomarkers to increase the detection rate of ovarian cancer.

The transformative potential of artificial intelligence in solid organ transplantation

Solid organ transplantation confronts numerous challenges ranging from donor organ shortage to post-transplant complications. Here, we provide an overview of the latest attempts to address some of these challenges using artificial intelligence (AI). We delve into the application of machine learning in pretransplant evaluation, predicting transplant rejection, and post-operative patient outcomes. By providing a comprehensive overview of AI's current impact, this review aims to inform clinicians, researchers, and policy-makers about the transformative power of AI in enhancing solid organ transplantation and facilitating personalized medicine in transplant care.

Assessment of hepatic fibrosis with non-invasive indices in subjects with diabetes before and after liver transplantation

Introduction

One of the most common complications of cirrhosis is diabetes, which prevalence is strictly related to severity of hepatopathy. Actually, there are no data on the persistence of post-transplant glucose abnormalities and on a potential impact of diabetes on development of fibrosis in the transplanted liver. To this aim, we evaluated liver fibrosis in cirrhotic subjects before and after being transplanted.

Methods

The study included 111 individuals who had liver transplantation. The assessment was performed before and two years after surgery to investigate a potential impact of the persistence of diabetes on developing de novo fibrosis in the transplanted liver. The degree of fibrosis was assessed using the Fibrosis Index Based on 4 Factors (FIB-4) and the Aspartate to Platelet Ratio Index (APRI).

Results

At pre-transplant evaluation, 63 out of 111 (56.8%) subjects were diabetic. Diabetic subjects had higher FIB-4 (Geometric mean, 95% confidence interval: 9.74, 8.32-11.41 vs 5.93, 4.71-7.46, P<0.001) and APRI (2.04, 1.69-2.47 vs 1.18, 0.90-1.55, P<0.001) compared to non-diabetic subjects. Two years after transplantation, 39 out of 111 (35.1%) subjects remained with diabetes and continued to show significantly higher FIB-4 (3.14, 2.57-3.82 vs 1.87, 1.55-2.27, P<0.001) and APRI (0.52, 0.39-0.69 vs 0.26, 0.21-0.32, P<0.001) compared to subjects without diabetes.

Discussion

Thus, persistence of diabetes after surgery is a possible risk factor for an evolution to fibrosis in the transplanted liver, potentially leading to worsened long-term outcomes in this population.

A Comprehensive Review of Liver Allograft Fibrosis and Steatosis: From Cause to Diagnosis

Despite advances in posttransplant care, long-term outcomes for liver transplant recipients remain unchanged. Approximately 25% of recipients will advance to graft cirrhosis and require retransplantation. Graft fibrosis progresses in the context of de novo or recurrent disease. Recurrent hepatitis C virus infection was previously the most important cause of graft failure but is now curable in the majority of patients. However, with an increasing prevalence of obesity and diabetes and nonalcoholic fatty liver disease as the most rapidly increasing indication for liver transplantation, metabolic dysfunction-associated liver injury is anticipated to become an important cause of graft fibrosis alongside alloimmune hepatitis and alcoholic liver disease. To better understand the landscape of the graft fibrosis literature, we summarize the associated epidemiology, cause, potential mechanisms, diagnosis, and complications. We additionally highlight the need for better noninvasive methods to ameliorate the management of graft fibrosis. Some examples include leveraging the microbiome, genetic, and machine learning methods to address these limitations. Overall, graft fibrosis is routinely seen by transplant clinicians, but it requires a better understanding of its underlying biology and contributors that can help inform diagnostic and therapeutic practices.

Development of a Radiomics-Based Model to Predict Graft Fibrosis in Liver Transplant Recipients: A Pilot Study

Liver Transplantation is complicated by recurrent fibrosis in 40% of recipients. We evaluated the ability of clinical and radiomic features to flag patients at risk of developing future graft fibrosis. CT scans of 254 patients at 3-6 months post-liver transplant were retrospectively analyzed. Volumetric radiomic features were extracted from the portal phase using an Artificial Intelligence-based tool (PyRadiomics). The primary endpoint was clinically significant (≥F2) graft fibrosis. A 10-fold cross-validated LASSO model using clinical and radiomic features was developed. In total, 75 patients (29.5%) developed ≥F2 fibrosis by a median of 19 (4.3-121.8) months. The maximum liver attenuation at the venous phase (a radiomic feature reflecting venous perfusion), primary etiology, donor/recipient age, recurrence of disease, brain-dead donor, tacrolimus use at 3 months, and APRI score at 3 months were predictive of ≥F2 fibrosis. The combination of radiomics and the clinical features increased the AUC to 0.811 from 0.793 for the clinical-only model (p = 0.008) and from 0.664 for the radiomics-only model (p < 0.001) to predict future ≥F2 fibrosis. This pilot study exploring the role of radiomics demonstrates that the addition of radiomic features in a clinical model increased the model's performance. Further studies are required to investigate the generalizability of this experimental tool.