3 and 5 Parameter Model to estimate the severity of portal hypertension (HVPG) in patients with compensated advanced chronic liver disease (cACLD)
Based on machine-learning models (MLMs), we developed this online tool, that calculates:
- the probability of the presence of clinically significant portal hypertension (CSPH, i.e. HVPG ≥10 mmHg) – at which the risk for hepatic decompensating events starts to exist
- the probability of the presence of severe portal hypertension (i.e. HVPG ≥16 mmHg) – at which the risk for hepatic decompensating events is considerable
- the estimated numerical HVPG values in mmHg
Please refer to our study for more information (https://doi.org/10.1016/j.jhep.2022.09.012). The instructions and disclaimer provided below describe how to use it.
Disclaimer
This online tool was developed to aid in assessing the clinical risk for subsequent hepatic decompensation in patients with compensated advanced chronic liver disease (cACLD) or compensated cirrhosis.
The tool will indicate the likelihood for a given patient to suffer from CSPH or severe portal hypertension, and indicate an estimated HVPG value. However, it does not substitute for invasive HVPG measurement, which is the gold standard for assessing the severity of portal hypertension (Reiberger et al., 2020; de Franchis et al., 2021).
This tool should not be used for self-diagnosis or self-treatment. For more information on portal hypertension, please consult your physician.
When interpreting the result, keep in mind the limitations indicated in the study manuscript, as well as the specific clinical situation. In particular, the proposed machine learning model was trained exclusively on data of compensated patients with advanced chronic liver disease. Therefore, results from patients with decompensation or without advanced chronic liver disease should be interpreted cautiously. Due to the training dataset specifics, the model is expected to overestimate the predicted HVPG value for low (< 6 mmHg) or underestimate for high (> 20 mmHg) actual values.
The predicted probabilities and predicted HVPG values should only be interpreted with knowledge of the limitations discussed in the manuscript (Reiniš J, Petrenko O, et al., J Hepatol 2022).