A systematic review of the diagnostic accuracy of physical examination for the detection of cirrhosis

The Use of Deep Learning and Machine Learning on Longitudinal Electronic Health Records for the Early Detection and Prevention of Diseases: Scoping Review

BACKGROUND

Electronic health records (EHRs) contain patients' health information over time, including possible early indicators of disease. However, the increasing amount of data hinders clinicians from using them. There is accumulating evidence suggesting that machine learning (ML) and deep learning (DL) can assist clinicians in analyzing these large-scale EHRs, as algorithms thrive on high volumes of data. Although ML has become well developed, studies mainly focus on engineering but lack medical outcomes.

OBJECTIVE

This study aims for a scoping review of the evidence on how the use of ML on longitudinal EHRs can support the early detection and prevention of disease. The medical insights and clinical benefits that have been generated were investigated by reviewing applications in a variety of diseases.

METHODS

This study was conducted according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. A literature search was performed in 2022 in collaboration with a medical information specialist in the following databases: PubMed, Embase, Web of Science Core Collection (Clarivate Analytics), and IEEE Xplore Digital Library and computer science bibliography. Studies were eligible when longitudinal EHRs were used that aimed for the early detection of disease via ML in a prevention context. Studies with a technical focus or using imaging or hospital admission data were beyond the scope of this review. Study screening and selection and data extraction were performed independently by 2 researchers.

RESULTS

In total, 20 studies were included, mainly published between 2018 and 2022. They showed that a variety of diseases could be detected or predicted, particularly diabetes; kidney diseases; diseases of the circulatory system; and mental, behavioral, and neurodevelopmental disorders. Demographics, symptoms, procedures, laboratory test results, diagnoses, medications, and BMI were frequently used EHR data in basic recurrent neural network or long short-term memory techniques. By developing and comparing ML and DL models, medical insights such as a high diagnostic performance, an earlier detection, the most important predictors, and additional health indicators were obtained. A clinical benefit that has been evaluated positively was preliminary screening. If these models are applied in practice, patients might also benefit from personalized health care and prevention, with practical benefits such as workload reduction and policy insights.

CONCLUSIONS

Longitudinal EHRs proved to be helpful for support in health care. Current ML models on EHRs can support the detection of diseases in terms of accuracy and offer preliminary screening benefits. Regarding the prevention of diseases, ML and specifically DL models can accurately predict or detect diseases earlier than current clinical diagnoses. Adding personally responsible factors allows targeted prevention interventions. While ML models based on textual EHRs are still in the developmental stage, they have high potential to support clinicians and the health care system and improve patient outcomes.

Indications and methods for measuring portal hypertension in cirrhosis

Background and objectives: Over the last decade our understanding of the pathophysiology of portal hypertension has increased. Novel diagnostic technologies have facilitated and improved the diagnosis and treatment of hepatic fibrosis and cirrhosis. With this review we aim to provide an overview of contemporary diagnostic principles of portal hypertension and indications for measuring portal pressure in cirrhosis.Methods: By review of current literature, we assessed new and old principles of measuring portal hypertension and the diagnostic values of the methods.Results: Invasive measurement of the portal pressure is still the gold standard to quantitate portal hypertension and to assess response to vasoactive treatment. The size of the portal pressure is important to assess since it contains information on the course of the disease and risk of developing hepatic decompensation, hepatocellular carcinoma, and mortality. Reliable non-invasive Elastography techniques are emerging that adequately assess portal pressure, but the available methods are not yet sufficiently accurate.Conclusion: Although elastography techniques provide valuable information and are good monitoring tools, liver vein catheterization remains valuable in diagnosing and monitoring portal hypertension, especially in combination with a trans-jugular liver biopsy.

Missed diagnosis of cirrhosis in the inpatient setting

Cirrhosis accounts for a large number of deaths in the United States and worldwide, leading to an increasing burden on the healthcare system. Cirrhosis is, however, a progressive disease with different potential complications related to liver dysfunction and portal hypertension. Often, patients may present with complications of cirrhosis without having been diagnosed previously. It is pertinent that clinicians recognize these signs to place patients on an appropriate course of management to help delay or avoid further disease progression while avoiding deleterious outcomes and unnecessary utilization. We will discuss the epidemiology of liver disease, cirrhosis, and its complications (hepatic encephalopathy, ascites, and varices). In this study, we will discuss the rationale and impact of missing these diagnoses on the healthcare system and patient.

Indian College of Radiology and Imaging Evidence-Based Guidelines for Interventions in Portal Hypertension and Its Complications

Portal hypertension is a complication of chronic liver disease. Various radiological interventions are being done to aid in the diagnosis of portal hypertension; further, an interventional radiologist can offer various treatments for the complications of portal hypertension. Diagnosis of portal hypertension in its early stage may require hepatic venous pressure gradient measurement. Measurement of gradient also guides in diagnosing the type of portal hypertension, measuring response to treatment and prognostication. This article attempts to provide evidence-based guidelines on the management of portal hypertension and treatment of its complications.

[Diagnosis and Severity Assessment of Alcohol-Related Liver Disease]

A diagnosis of alcohol-related liver disease (ALD) requires information on the history of excessive alcohol consumption (average intake of 40 g or more in men and 20 g or more in women a day). Furthermore, blood tests, such as GGT, AST, ALT, and mean corpuscular volume, and imaging studies, including abdominal ultrasound or transient elastography, are also useful. A liver biopsy can be useful for confirming the diagnosis and has prognostic value. ALD includes alcoholic fatty liver, alcoholic hepatitis, and alcoholic cirrhosis, and in most cases, clinical manifestations can overlap. The prognostic scoring systems of ALD are limited mainly to alcoholic hepatitis, and the early mortality and treatment response can be predicted using various scoring systems. This review summarizes how to diagnose and evaluate the severity of ALD in clinical practice.

Early Multiplexed Detection of Cirrhosis using Giant Magnetoresistive Biosensors with Protein Biomarkers

Liver cirrhosis is one of the leading causes of death in adults worldwide. It is highly prevalent in developing countries and is growing in prevalence in developed countries mostly because of chronic liver diseases, such as chronic hepatitis B and C and alcoholic and nonalcoholic fatty liver disease. However, the prevalence of cirrhosis may be highly underestimated because early stages are asymptomatic and current early detection methods are inadequate. Here, we evaluate the potential of a set of novel cirrhotic protein biomarkers, including soluble intercellular adhesion molecule-1 and mac-2 binding protein glycosylation isomer, for early detection of cirrhosis in a multiplexed assay using our giant magnetoresistive (GMR) sensor arrays. We evaluated the diagnostic performance of the biomarkers, individually and in combination, using multivariate logistic regression and random forest in a blinded proof-of-concept retrospective case-controlled study. The biomarkers in combination exhibited high diagnostic performance in both logistic regression and random forest models, with an area under the curve of 0.98 (0.94-1.00). In addition, the combination of biomarkers resulted in a high sensitivity of 0.97 (0.95-1.00) and a high specificity of 1.00. We showed that the diagnostic performance of our novel set of cirrhotic protein biomarkers on our multiplexed GMR sensor arrays is higher than the performance of currently used clinical biomarkers and factors (i.e., age, sex, alanine aminotransferase, aspartate aminotransferase, etc.). With this combination of novel biomarkers and GMR technology, we could potentially boost the diagnostic power of early cirrhosis detection.

Candidate Biomarkers of Liver Fibrosis: A Concise, Pathophysiology-oriented Review

Repair of sustained liver injury results in fibrosis (i.e. the accumulation of extracellular matrix proteins), and ultimately the complete distortion of parenchymal architecture of the liver, which we call cirrhosis. Detecting and staging of fibrosis is thus a mainstay in the management of chronic liver diseases, since many clinically relevant decisions, such as starting treatment and/or monitoring for complications including hepatocellular carcinoma, may depend on it. The gold standard for fibrosis staging is liver biopsy, the role of which, however, is questioned nowadays because of cost, hazards and poor acceptance by patients. On the other hand, imaging techniques and/or measurement of direct and indirect serum markers have not proved to be completely satisfactory under all circumstances as alternatives to liver biopsy. Making progress in this field is now more crucial than ever, since treatments for established fibrosis appear on the horizon. Fine dissection of the pathways involved in the pathophysiology of liver diseases has put forward several novel candidate biomarkers of liver fibrosis, such as growth arrest-specific6, Mac-2-binding protein, osteopontin, placental growth factor, growth/differentiation factor 15 and hepatocyte growth factor. All molecules have been suggested to have potential to complement or substitute methods currently used to stage liver diseases. Here, we review the pros and cons for their use in this setting.

A Clinical Decision Rule Based on the AST-to-Platelet Ratio Index Improves Adherence to Published Guidelines on the Management of Acute Variceal Bleeding

BACKGROUND

Optimal management of acute upper gastrointestinal bleeding (UGIB) depends on identifying a variceal versus nonvariceal etiology. An objective measure predicting etiology could guide early management pending endoscopy. The AST-to-platelet ratio index (APRI) score has been studied as a marker of cirrhosis and portal hypertension, but has not been evaluated in the setting of acute UGIB.

METHODS

In this single-center retrospective cohort study, we reviewed endoscopy reports and other data for patients with acute UGIB, and classified episodes as variceal bleeds or other. We assessed the diagnostic utility of the APRI score relative to other objective measures by Area Under the Receiver Operating Characteristic (AUROC) curve analysis. We constructed a clinical decision rule based on the APRI score, and assessed how it would have changed management.

RESULTS

The APRI score performed well in predicting a variceal etiology of acute UGIB, with AUROC 0.89. We developed a clinical decision rule using an APRI score of 0.4 to guide early management of acute UGIB patients. Retroactively applying this to our cohort, adherence to published guidelines for administration of octreotide and antibiotics would have increased from 56% to 91%.

CONCLUSIONS

The APRI score is an objective metric that helps predict a variceal etiology of acute UGIB. Using our proposed decision rule could improve adherence to guidelines on management of acute variceal bleeding. Although we were unable to demonstrate a survival benefit, improved adherence to evidence-based guidelines serves as a metric related to this most important outcome measure. Prospective study to validate these findings is indicated.

Assessing portal hypertension in liver diseases

Portal hypertension is a common complication of chronic liver diseases and is responsible for most clinical consequences of cirrhosis, which represent the more frequent causes of death and liver transplantation in these patients. This review is aimed at clarifying the state-of-the art assessment of portal hypertension and at discussing recent developments in this field. Particular attention is paid to new noninvasive techniques that will be soon available for potential routine use.

Noninvasive evaluation of portal hypertension: emerging tools and techniques

Portal hypertension is the main cause of complications in patients with cirrhosis. However, evaluating the development and progression of portal hypertension represents a challenge for clinicians. There has been considerable focus on the potential role of noninvasive markers of portal hypertension that could be used to stratify patients with respect to the stage of portal hypertension and to monitor disease progression or treatment response in a longitudinal manner without having to undertake repeated invasive assessment. The pathogenesis of portal hypertension is increasingly understood and emerging knowledge of the vascular processes that underpin portal hypertension has paved the way for exploring novel biomarkers of vascular injury, angiogenesis, and endothelial dysfunction. In this paper we focus on the pathogenesis of portal hypertension and potential non-invasive biomarkers with particular emphasis on serum analytes.

Mechanisms and implications of age-related changes in the liver: nonalcoholic Fatty liver disease in the elderly

Nonalcoholic fatty liver disease (NAFLD) is hepatic steatosis associated with metabolic abnormalities such as overweight/central obesity, insulin resistance, type 2 diabetes (T2D), and dyslipidemia. NAFLD is becoming the most common liver disease in contemporary society, with the highest prevalence in those over 60 years. NAFLD pathology ranges from simple steatosis to a necroinflammatory fibrosing disorder called steatohepatitis (SH), the latter associated with high risk of developing cirrhosis, often occuring in the seventh to ninth decades of life. While the main health implications of NAFLD are increased risk of developing T2D, cardiovascular diseases, and common cancers, there is substantantially increased standardized mortality, and deaths from decompensated cirrhosis and hepatocellular carcinoma (HCC). Little is known about the interactive effects of ageing and NAFLD, with most studies focusing on the younger population. This paper summarises the epidemiology, pathogenesis, and clinical course of NAFLD, with particular attention to persons over age 60 years. An approach to the management of NASH and its complications in the elderly, will also be presented here.

Mechanisms and implications of age-related changes in the liver: nonalcoholic Fatty liver disease in the elderly

Nonalcoholic fatty liver disease (NAFLD) is hepatic steatosis associated with metabolic abnormalities such as overweight/central obesity, insulin resistance, type 2 diabetes (T2D), and dyslipidemia. NAFLD is becoming the most common liver disease in contemporary society, with the highest prevalence in those over 60 years. NAFLD pathology ranges from simple steatosis to a necroinflammatory fibrosing disorder called steatohepatitis (SH), the latter associated with high risk of developing cirrhosis, often occuring in the seventh to ninth decades of life. While the main health implications of NAFLD are increased risk of developing T2D, cardiovascular diseases, and common cancers, there is substantantially increased standardized mortality, and deaths from decompensated cirrhosis and hepatocellular carcinoma (HCC). Little is known about the interactive effects of ageing and NAFLD, with most studies focusing on the younger population. This paper summarises the epidemiology, pathogenesis, and clinical course of NAFLD, with particular attention to persons over age 60 years. An approach to the management of NASH and its complications in the elderly, will also be presented here.

Alcoholic liver disease

These recommendations provide a data-supported approach. They are based on the following: (i) a formal review and analysis of the recently published world literature on the topic (Medline search); (ii) American College of Physicians Manual for Assessing Health Practices and Designing Practice Guidelines (1); (iii) guideline policies, including the American Association for the Study of Liver Diseases (AASLD) Policy on the development and use of practice guidelines and the AGA Policy Statement on Guidelines (2); and (iv) the experience of the authors in the specified topic. Intended for use by physicians, these recommendations suggest preferred approaches to the diagnostic, therapeutic, and preventive aspects of care. They are intended to be flexible, in contrast to the standards of care, which are inflexible policies to be followed in every case. Specific recommendations are based on relevant published information. To more fully characterize the quality of evidence supporting the recommendations, the Practice Guideline Committee of the AASLD requires a Class (reflecting the benefit vs. risk) and Level (assessing the strength or certainty) of Evidence to be assigned and reported with each recommendation (Table 1, adapted from the American College of Cardiology and the American Heart Association Practice Guidelines) (3,4).

Non invasive diagnosis of portal hypertension in cirrhotic patients

The measure of disease progression in chronic liver disease represents a key challenge in any of the different stages of evolution. Indeed, a correct and reliable measure of the stage of the disease has relevant implications for assessing the effectiveness of the current therapeutic regimens and for predicting the occurrence of complication. Accordingly, a current major effort is directed at evaluating methodologies characterized by no or low invasiveness to be employed as clinical discriminators in patients populations potentially requiring invasive assessment. This appears particularly relevant in patients with compensated cirrhosis, where the only reference standard is the measurement of portal pressure by hepatic venous pressure gradient (HVPG). In this particular context, transient elastography (TE) appears to be promising and needs to be further investigated, possibly in combination with other non-invasive methodologies such as serum markers algorithms and/or imaging techniques. On the other hand, the application of non-invasive methods for monitoring the response to vasoactive treatment for the reduction of portal pressure and the prevention of related complications seems at the moment not realistic.

The management of portal hypertension: rational basis, available treatments and future options

Variceal bleeding is the last step in a chain of events initiated by an increase in portal pressure, followed by the development and progressive dilation of varices until these finally rupture and bleed. This sequence of events might be prevented - and reversed - by achieving a sufficient decrease in portal pressure. A different approach is the use of local endoscopic treatments at the varices. This article reviews the rationale for the management of patients with cirrhosis and portal hypertension, the current recommendations for the prevention and treatment of variceal bleeding, and outlines the unsolved issues and the perspectives for the future opened by new research developments.