# A machine-learning clustering approach for reference interval estimation of liver enzymes from hospital laboratory big-data

## TL;DR

This paper uses machine learning and big data from hospital labs to determine accurate reference intervals for liver enzymes AST and ALT.

## Contribution

A novel combination of clustering and outlier detection methods is applied to estimate liver enzyme reference intervals from real-world data.

## Key findings

- DBSCAN with Tukey's fences or Local Outlier Factor performed best in covering validation data.
- Estimated AST and ALT reference intervals are 15-41 U/L and 11-46 U/L, respectively.

## Abstract

It is of interest to establish clinically valid reference intervals (RIs) for the liver enzymes aspartate transaminase (AST) and
Alanine aminotransferase (ALT) using a combination of unsupervised machine learning clustering and robust outlier detection applied to
real-world laboratory big data. Four outlier detection methods were each combined with four clustering algorithms to identify homogeneous
subgroups and the largest cluster from each combination was used to estimate RIs based on percentile cut-offs. Among the tested
combinations, DBSCAN with Tukey's fences or Local Outlier Factor achieved optimal performance, covering 100% of the validation data. The
widest intervals were derived using Local Outlier Factor, while Isolation Forest yielded the narrowest. Ultimately, the study estimated
the reference intervals for AST and ALT to be 15-41 U/L and 11-46 U/L, respectively.