1. |
- Mcteer, Matthew, et al.
(författare)
-
Machine learning approaches to enhance diagnosis and staging of patients with MASLD using routinely available clinical information
- 2024
-
Ingår i: PLOS ONE. - : PUBLIC LIBRARY SCIENCE. - 1932-6203. ; 19:2
-
Tidskriftsartikel (refereegranskat)abstract
- Aims Metabolic dysfunction Associated Steatotic Liver Disease (MASLD) outcomes such as MASH (metabolic dysfunction associated steatohepatitis), fibrosis and cirrhosis are ordinarily determined by resource-intensive and invasive biopsies. We aim to show that routine clinical tests offer sufficient information to predict these endpoints.Methods Using the LITMUS Metacohort derived from the European NAFLD Registry, the largest MASLD dataset in Europe, we create three combinations of features which vary in degree of procurement including a 19-variable feature set that are attained through a routine clinical appointment or blood test. This data was used to train predictive models using supervised machine learning (ML) algorithm XGBoost, alongside missing imputation technique MICE and class balancing algorithm SMOTE. Shapley Additive exPlanations (SHAP) were added to determine relative importance for each clinical variable.Results Analysing nine biopsy-derived MASLD outcomes of cohort size ranging between 5385 and 6673 subjects, we were able to predict individuals at training set AUCs ranging from 0.719-0.994, including classifying individuals who are At-Risk MASH at an AUC = 0.899. Using two further feature combinations of 26-variables and 35-variables, which included composite scores known to be good indicators for MASLD endpoints and advanced specialist tests, we found predictive performance did not sufficiently improve. We are also able to present local and global explanations for each ML model, offering clinicians interpretability without the expense of worsening predictive performance.Conclusions This study developed a series of ML models of accuracy ranging from 71.9-99.4% using only easily extractable and readily available information in predicting MASLD outcomes which are usually determined through highly invasive means.
|
|
2. |
- Lee, Jenny, et al.
(författare)
-
Machine learning algorithm improves the detection of NASH (NAS-based) and at-risk NASH: A development and validation study
- 2023
-
Ingår i: Hepatology. - : LIPPINCOTT WILLIAMS & WILKINS. - 0270-9139 .- 1527-3350. ; 78:1, s. 258-271
-
Tidskriftsartikel (refereegranskat)abstract
- Background and Aims: Detecting NASH remains challenging, while at-risk NASH (steatohepatitis and F >= 2) tends to progress and is of interest for drug development and clinical application. We developed prediction models by supervised machine learning techniques, with clinical data and biomarkers to stage and grade patients with NAFLD. Approach and Results: Learning data were collected in the Liver Investigation: Testing Marker Utility in Steatohepatitis metacohort (966 biopsy-proven NAFLD adults), staged and graded according to NASH CRN. Conditions of interest were the clinical trial definition of NASH (NAS >= 4;53%), at-risk NASH (NASH with F >= 2;35%), significant (F >= 2;47%), and advanced fibrosis (F >= 3;28%). Thirty-five predictors were included. Missing data were handled by multiple imputations. Data were randomly split into training/validation (75/25) sets. A gradient boosting machine was applied to develop 2 models for each condition: clinical versus extended (clinical and biomarkers). Two variants of the NASH and at-risk NASH models were constructed: direct and composite models.Clinical gradient boosting machine models for steatosis/inflammation/ballooning had AUCs of 0.94/0.79/0.72. There were no improvements when biomarkers were included. The direct NASH model produced AUCs (clinical/extended) of 0.61/0.65. The composite NASH model performed significantly better (0.71) for both variants. The composite at-risk NASH model had an AUC of 0.83 (clinical and extended), an improvement over the direct model. Significant fibrosis models had AUCs (clinical/extended) of 0.76/0.78. The extended advanced fibrosis model (0.86) performed significantly better than the clinical version (0.82). Conclusions: Detection of NASH and at-risk NASH can be improved by constructing independent machine learning models for each component, using only clinical predictors. Adding biomarkers only improved the accuracy of fibrosis.
|
|
3. |
- Vali, Yasaman, et al.
(författare)
-
Biomarkers for staging fibrosis and non-alcoholic steatohepatitis in non-alcoholic fatty liver disease (the LITMUS project) : a comparative diagnostic accuracy study
- 2023
-
Ingår i: The Lancet Gastroenterology & Hepatology. - : Elsevier Ltd. - 2468-1253. ; 8:8, s. 714-725
-
Tidskriftsartikel (refereegranskat)abstract
- Background: The reference standard for detecting non-alcoholic steatohepatitis (NASH) and staging fibrosis—liver biopsy—is invasive and resource intensive. Non-invasive biomarkers are urgently needed, but few studies have compared these biomarkers in a single cohort. As part of the Liver Investigation: Testing Marker Utility in Steatohepatitis (LITMUS) project, we aimed to evaluate the diagnostic accuracy of 17 biomarkers and multimarker scores in detecting NASH and clinically significant fibrosis in patients with non-alcoholic fatty liver disease (NAFLD) and identify their optimal cutoffs as screening tests in clinical trial recruitment. Methods: This was a comparative diagnostic accuracy study in people with biopsy-confirmed NAFLD from 13 countries across Europe, recruited between Jan 6, 2010, and Dec 29, 2017, from the LITMUS metacohort of the prospective European NAFLD Registry. Adults (aged ≥18 years) with paired liver biopsy and serum samples were eligible; those with excessive alcohol consumption or evidence of other chronic liver diseases were excluded. The diagnostic accuracy of the biomarkers was expressed as the area under the receiver operating characteristic curve (AUC) with liver histology as the reference standard and compared with the Fibrosis-4 index for liver fibrosis (FIB-4) in the same subgroup. Target conditions were the presence of NASH with clinically significant fibrosis (ie, at-risk NASH; NAFLD Activity Score ≥4 and F≥2) or the presence of advanced fibrosis (F≥3), analysed in all participants with complete data. We identified thres holds for each biomarker for reducing the number of biopsy-based screen failures when recruiting people with both NASH and clinically significant fibrosis for future trials. Findings: Of 1430 participants with NAFLD in the LITMUS metacohort with serum samples, 966 (403 women and 563 men) were included after all exclusion criteria had been applied. 335 (35%) of 966 participants had biopsy-confirmed NASH and clinically significant fibrosis and 271 (28%) had advanced fibrosis. For people with NASH and clinically significant fibrosis, no single biomarker or multimarker score significantly reached the predefined AUC 0·80 acceptability threshold (AUCs ranging from 0·61 [95% CI 0·54–0·67] for FibroScan controlled attenuation parameter to 0·81 [0·75–0·86] for SomaSignal), with accuracy mostly similar to FIB-4. Regarding detection of advanced fibrosis, SomaSignal (AUC 0·90 [95% CI 0·86–0·94]), ADAPT (0·85 [0·81–0·89]), and FibroScan liver stiffness measurement (0·83 [0·80–0·86]) reached acceptable accuracy. With 11 of 17 markers, histological screen failure rates could be reduced to 33% in trials if only people who were marker positive had a biopsy for evaluating eligibility. The best screening performance for NASH and clinically significant fibrosis was observed for SomaSignal (number needed to test [NNT] to find one true positive was four [95% CI 4–5]), then ADAPT (six [5–7]), MACK-3 (seven [6–8]), and PRO-C3 (nine [7–11]). Interpretation: None of the single markers or multimarker scores achieved the predefined acceptable AUC for replacing biopsy in detecting people with both NASH and clinically significant fibrosis. However, several biomarkers could be applied in a prescreening strategy in clinical trial recruitment. The performance of promising markers will be further evaluated in the ongoing prospective LITMUS study cohort. Funding: The Innovative Medicines Initiative 2 Joint Undertaking. © 2023 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license
|
|