Dataset Information

Machine learning-based diagnosis for disseminated intravascular coagulation (DIC): Development, external validation, and comparison to scoring systems.

ABSTRACT: The major challenge in the diagnosis of disseminated intravascular coagulation (DIC) comes from the lack of specific biomarkers, leading to developing composite scoring systems. DIC scores are simple and rapidly applicable. However, optimal fibrin-related markers and their cut-off values remain to be defined, requiring optimization for use. The aim of this study is to optimize the use of DIC-related parameters through machine learning (ML)-approach. Further, we evaluated whether this approach could provide a diagnostic value in DIC diagnosis. For this, 46 DIC-related parameters were investigated for both clinical findings and laboratory results. We retrospectively reviewed 656 DIC-suspected cases at an initial order for full DIC profile and labeled their evaluation results (Set 1; DIC, n = 228; non-DIC, n = 428). Several ML algorithms were tested, and an artificial neural network (ANN) model was established via independent training and testing using 32 selected parameters. This model was externally validated from a different hospital with 217 DIC-suspected cases (Set 2; DIC, n = 80; non-DIC, n = 137). The ANN model represented higher AUC values than the three scoring systems in both set 1 (ANN 0.981; ISTH 0.945; JMHW 0.943; and JAAM 0.928) and set 2 (AUC ANN 0.968; ISTH 0.946). Additionally, the relative importance of the 32 parameters was evaluated. Most parameters had contextual importance, however, their importance in ML-approach was different from the traditional scoring system. Our study demonstrates that ML could optimize the use of clinical parameters with robustness for DIC diagnosis. We believe that this approach could play a supportive role in physicians' medical decision by integrated into electrical health record system. Further prospective validation is required to assess the clinical consequence of ML-approach and their clinical benefit.

SUBMITTER: Yoon JG

PROVIDER: S-EPMC5931474 | biostudies-literature | 2018

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Machine learning-based diagnosis for disseminated intravascular coagulation (DIC): Development, external validation, and comparison to scoring systems.

Yoon Jihoon G JG Heo JoonNyung J Kim Minkyu M Park Yu Jin YJ Choi Min Hyuk MH Song Jaewoo J Wyi Kangsan K Kim Hakbeen H Duchenne Olivier O Eom Soowon S Tsoy Yury Y

PloS one 20180502 5

The major challenge in the diagnosis of disseminated intravascular coagulation (DIC) comes from the lack of specific biomarkers, leading to developing composite scoring systems. DIC scores are simple and rapidly applicable. However, optimal fibrin-related markers and their cut-off values remain to be defined, requiring optimization for use. The aim of this study is to optimize the use of DIC-related parameters through machine learning (ML)-approach. Further, we evaluated whether this approach co ...[more]

PMID: 29718941

Similar Datasets

Project description:BackgroundThere is no gold standard to diagnose septic shock-induced disseminated intravascular coagulation (DIC). The objective of our multicenter prospective study was to assess the performances of the different major scoring systems in terms of mortality prediction and DIC diagnosis. The JAAM-DIC 2016 score, the ISTH overt-DIC 2001 score, the associations of sepsis-induced coagulopathy (SIC) score with JAAM-DIC 2016 or ISTH overt-DIC scores were tested in patients within 12 h of their admission in ICU for septic shock (day 1) and at day 2.Results582 patients were enrolled in the study. 182/567 (32.1%) were diagnosed with DIC according to ISTH overt-DIC score, and 193/561 (34.4%) according to JAAM-DIC score; 486/577 patients (84.2%) were diagnosed with a coagulopathy according to SIC score. A moderate concordance was observed between ISTH overt-DIC and JAAM-DIC [κ = 0.67 (0.60, 0.73), p < 0.001]. The delay of positivity of the scores for early DIC patients was not different between JAAM-DIC and ISTH overt-DIC scores. Although it was positive earlier, SIC score had worse diagnosis specificity, as 84.2% of the patients with septic shock were diagnosed with "coagulopathy". The specificity of SIC score alone to predict mortality was very low [0.18 (0.15; 0.22)], compared to the ones of JAAM-DIC score [0.71 (0.67; 0.75)], and of ISTH overt-DIC score [0.76 (0.72; 0.80)], p < 0.001. The sensitivity of SIC score to predict mortality was 0.95 [0.89; 0.98], and the ones of JAAM-DIC score and ISTH overt-DIC score were 0.61 [0.50; 0.70] and 0.68 [0.58; 0.77], respectively. There was no benefit in sensitivity and specificity in combining SIC score to JAAM-DIC score or to ISTH overt-DIC score, compared to JAAM-DIC score or ISTH overt-DIC score alone.ConclusionsOur data suggest that the added value of SIC score alone or combined with other scores is limited, and that both JAAM-DIC score and ISTH overt-DIC score can be used in septic shock patients. Trial registration clinicaltrial; Trial registration number: NCT02391792; Date of registration: 18/03/2015; URL of trial registry record: https://clinicaltrials.gov/ct2/show/NCT02391792?term=meziani&draw=4&rank=1.

Project description:IntroductionSince disseminated intravascular coagulation (DIC) was first described, it has been considered a serious disease of the coagulation system and a major challenge to clinicians. Currently, several important knowledge gaps remain. The DANish Disseminated Intravascular Coagulation (DANDIC) Cohort Study will aim to answer questions regarding the incidence and mortality of patients with DIC including time trends. The study will also identify prognostic factors that may guide personalised prevention and treatment. Furthermore, the study will describe treatment patterns and the safety and effectiveness of various treatment modalities.Methods and analysisWe will establish the DANDIC Cohort using data collected in daily clinical practice from the Central Denmark Region, which covers approximately 1.3 million residents. The study period will encompass 1 January 2011 through 1 July 2021. Potential DIC cases will be identified from the hospital laboratory database, based on coagulation biomarkers, and diagnoses will be adjudicated by medical experts. The dataset will be enriched with detailed clinical data from electronic medical charts on aetiologies, bleeding, microthrombus formation, organ failure, thrombosis, treatments and comorbidities. The dataset will also take advantage of in-hospital data with longitudinal information on laboratory records, transfusions, microbiology and treatments. It will be possible to merge this dataset with other unique Danish health registries with more than 10 years of virtually complete follow-up. The project will use state-of-the-art epidemiological and biostatistical methods.Ethics and disseminationThe project has been approved by the Danish Patient Safety Authority (31-1521-452), the Central Denmark Region (1-45-70-83-21), the Danish Data Protection Agency (1-16-02-258-21) and all the hospital chairs. Register-based studies require no ethical approval in Denmark. The results will be disseminated in international peer-reviewed journals.

Dataset Information

Machine learning-based diagnosis for disseminated intravascular coagulation (DIC): Development, external validation, and comparison to scoring systems.

Publications

Machine learning-based diagnosis for disseminated intravascular coagulation (DIC): Development, external validation, and comparison to scoring systems.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets