Deep Learning-Based Surgical Treatment Recommendation and Nonsurgical Prognosis Status Classification for Scaphoid Fractures by Automated X-ray Image Recognition.

Biomedical information retrieval for diagnosis, treatment and prognosis has been studied for a long time. In particular, image recognition using deep learning has been shown to be very effective for cancers and diseases. In these fields, scaphoid fracture recognition is a hot topic because the appearance of scaphoid fractures is not easy to detect. Although there have been a number of recent studies on this topic, no studies focused their attention on surgical treatment recommendations and nonsurgical prognosis status classification. Indeed, a successful treatment recommendation will assist the doctor in selecting an effective treatment, and the prognosis status classification will help a radiologist recognize the image more efficiently. For these purposes, in this paper, we propose potential solutions through a comprehensive empirical study assessing the effectiveness of recent deep learning techniques on surgical treatment recommendation and nonsurgical prognosis status classification. In the proposed system, the scaphoid is firstly segmented from an unknown X-ray image. Next, for surgical treatment recommendation, the fractures are further filtered and recognized. According to the recognition result, the surgical treatment recommendation is generated. Finally, even without sufficient fracture information, the doctor can still make an effective decision to opt for surgery or not. Moreover, for nonsurgical patients, the current prognosis status of avascular necrosis, non-union and union can be classified. The related experimental results made using a real dataset reveal that the surgical treatment recommendation reached 80% and 86% in accuracy and AUC (Area Under the Curve), respectively, while the nonsurgical prognosis status classification reached 91% and 96%, respectively. Further, the methods using transfer learning and data augmentation can bring out obvious improvements, which, on average, reached 21.9%, 28.9% and 5.6%, 7.8% for surgical treatment recommendations and nonsurgical prognosis image classification, respectively. Based on the experimental results, the recommended methods in this paper are DenseNet169 and ResNet50 for surgical treatment recommendation and nonsurgical prognosis status classification, respectively. We believe that this paper can provide an important reference for future research on surgical treatment recommendation and nonsurgical prognosis classification for scaphoid fractures.

eDRAM: Effective early disease risk assessment with matrix factorization on a large-scale medical database: A case study on rheumatoid arthritis.

Recently, a number of analytical approaches for probing medical databases have been developed to assist in disease risk assessment and to determine the association of a clinical condition with others, so that better and intelligent healthcare can be provided. The early assessment of disease risk is an emerging topic in medical informatics. If diseases are detected at an early stage, prognosis can be improved and medical resources can be used more efficiently. For example, if rheumatoid arthritis (RA) is detected at an early stage, appropriate medications can be used to prevent bone deterioration. In early disease risk assessment, finding important risk factors from large-scale medical databases and performing individual disease risk assessment have been challenging tasks. A number of recent studies have considered risk factor analysis approaches, such as association rule mining, sequential rule mining, regression, and expert advice. In this study, to improve disease risk assessment, machine learning and matrix factorization techniques were integrated to discover important and implicit risk factors. A novel framework is proposed that can effectively assess early disease risks, and RA is used as a case study. This framework comprises three main stages: data preprocessing, risk factor optimization, and early disease risk assessment. This is the first study integrating matrix factorization and machine learning for disease risk assessment that is applied to a nation-wide and longitudinal medical diagnostic database. In the experimental evaluations, a cohort established from a large-scale medical database was used that included 1007 RA-diagnosed patients and 921,192 control patients examined over a nine-year follow-up period (2000-2008). The evaluation results demonstrate that the proposed approach is more efficient and stable for disease risk assessment than state-of-the-art methods.

Mining disease risk patterns from nationwide clinical databases for the assessment of early rheumatoid arthritis risk.

Rheumatoid arthritis (RA) is a chronic autoimmune rheumatic disease that can cause painful swelling in the joint lining, morning stiffness, and joint deformation/destruction. These symptoms decrease both quality of life and life expectancy. However, if RA can be diagnosed in the early stages, it can be controlled with pharmacotherapy. Although many studies have examined the possibility of early assessment and diagnosis, few have considered the relationship between significant risk factors and the early assessment of RA. In this paper, we present a novel framework for early RA assessment that utilizes data preprocessing, risk pattern mining, validation, and analysis. Under our proposed framework, two risk patterns can be discovered. Type I refers to well-known risk patterns that have been identified by existing studies, whereas Type II denotes unknown relationship risk patterns that have rarely or never been reported in the literature. These Type II patterns are very valuable in supporting novel hypotheses in clinical trials of RA, and constitute the main contribution of this work. To ensure the robustness of our experimental evaluation, we use a nationwide clinical database containing information on 1,314 RA-diagnosed patients over a 12-year follow-up period (1997-2008) and 965,279 non-RA patients. Our proposed framework is employed on this large-scale population-based dataset, and is shown to effectively discover rich RA risk patterns. These patterns may assist physicians in patient assessment, and enhance opportunities for early detection of RA. The proposed framework is broadly applicable to the mining of risk patterns for major disease assessments. This enables the identification of early risk patterns that are significantly associated with a target disease.

Analysis of precore/core covariances associated with viral kinetics and genotypes in hepatitis B e antigen-positive chronic hepatitis B patients.

Hepatitis B virus (HBV) is one of the most common DNA viruses that can cause aggressive hepatitis, cirrhosis and hepatocellular carcinoma. Although many people are persistently infected with HBV, the kinetics in serum levels of viral loads and the host immune responses vary from person to person. HBV precore/core open reading frame (ORF) encoding proteins, hepatitis B e antigen (HBeAg) and core antigen (HBcAg), are two indicators of active viral replication. The aim of this study was to discover a variety of amino acid covariances in responses to viral kinetics, seroconversion and genotypes during the course of HBV infection. A one year follow-up study was conducted with a total number of 1,694 clones from 23 HBeAg-positive chronic hepatitis B patients. Serum alanine aminotransferase, HBV DNA and HBeAg levels were measured monthly as criteria for clustering patients into several different subgroups. Monthly derived multiple precore/core ORFs were directly sequenced and translated into amino acid sequences. For each subgroup, time-dependent covariances were identified from their time-varying sequences over the entire follow-up period. The fluctuating, wavering, HBeAg-nonseroconversion and genotype C subgroups showed greater degrees of covariances than the stationary, declining, HBeAg-seroconversion and genotype B. Referring to literature, mutation hotspots within our identified covariances were associated with the infection process. Remarkably, hotspots were predominant in genotype C. Moreover, covariances were also identified at early stage (spanning from baseline to a peak of serum HBV DNA) in order to determine the intersections with aforementioned time-dependent covariances. Preserved covariances, namely representative covariances, of each subgroup are visually presented using a tree-based structure. Our results suggested that identified covariances were strongly associated with viral kinetics, seroconversion and genotypes. Moreover, representative covariances may benefit clinicians to prescribe a suitable treatment for patients even if they have no obvious symptoms at the early stage of HBV infection.