# Improved Detection of Minimal Residual Disease in AML: Validation of IDH1/2 ddPCR Assays in the Perspective of Treatment with Target Inhibitors

## TL;DR

This study validates highly sensitive ddPCR assays for detecting IDH1/2 mutations in AML, which could improve monitoring of minimal residual disease during treatment.

## Contribution

The study introduces and validates optimized ddPCR assays for IDH1/2 mutations with high sensitivity and reproducibility for MRD monitoring in AML.

## Key findings

- Four ddPCR assays achieved detection limits as low as 0.07% for IDH1 R132H and 0.1% for IDH2 mutations.
- IDH fractional abundance showed a statistically significant positive correlation with NPM1 RQ-Ratio during follow-up.
- The assays demonstrated excellent intra-run reproducibility, making them suitable for patient follow-up.

## Abstract

Mutations in IDH1/2 are frequent in Acute Myeloid Leukemia (AML), defining a molecularly distinct subgroup with therapeutic implications due to the availability of specific inhibitors. Accurate monitoring of treatment response is crucial and Droplet Digital PCR (ddPCR) offers a sensitive approach for quantifying mutational burden in IDH-mutated AML. This study aimed to optimize and validate ddPCR assays specific for IDH1 R132 and IDH2 R172/R140 mutations for future use in Minimal Residual Disease (MRD) monitoring. Four ddPCR assays were set to evaluate the trend of IDH1/2 mutations in 191 diagnostic and follow-up samples. Each validation procedure included determining the limit of blank (LOB) and limit of detection (LOD) using titration series. Moreover, in AML harboring both IDH and NPM1 mutations, we performed generalized estimating equations (GEE) to assess the association between IDH fractional abundance and NPM1 RQ-Ratio across time points. Four IDH1/2 ddPCR assays were validated, demonstrating high sensitivity with limits of detection of 0.07% for IDH1 R132H, 0.1% for IDH2 R140Q and R172K, and 0.2% for IDH1 R132C. The method also exhibited excellent intra-run reproducibility, providing consistent results for patient follow-up. Comparison of IDH and NPM1 trends during follow-up revealed a statistically significant positive correlation, both in raw (β = 0.079, p = 0.001) and ranked data (β = 0.99, p = 0.004), suggesting a co-dynamic pattern potentially useful for surrogate monitoring. While our study cannot yet define the clinical role of IDH mutation assessment by ddPCR due to the lack of comparative follow-up studies, it establishes a solid methodological foundation for standardizing minimal residual disease evaluation via ddPCR, paving the way for future prospective validation.