Minimal Disease Tracking by Next Generation Sequencing in AML

Acute myeloid leukemia (AML) is a type of cancer characterized by a myriad of molecular abnormalities, present in various combinations. Residual disease after therapy is instrumental in guiding treatment. AML can occur de novo, evolve from Myelodysplastic Syndromes, or by certain chemotherapeutic treatments. This patient-specific landscape of residual disease markers paints a picture of all the factors and mechanisms and drives recurrence of disease or complete remission.

The sensitivity of detection for minimal residual disease in AML depends on the testing methods used. In morphological examination alone, MRD is measured in terms of % blasts. Molecular methods encompass % blasts and the ability to detect whether they harbor clonal abnormalities and/or pre-leukemic clones. With error-corrected NGS, molecular detection of mutations is possible at 0.01%.

The central goal of this project was to determine if repeated analysis, utilizing molecular barcodes, laboratory-specific noise profiling and error-corrected NGS, could be instrumental in identifying the specific variants present in a patient to monitor and target for MRD, for the purpose of downstream treatment for individualized risk assessment.

After library optimization, a set of 68 primers were designed to target specific drivers in AML. Choosing a subset of these primers (n=<10) per patient, specific driver mutations were interrogated for their presence in samples that had previously shown clearance of these mutations.  The original patient samples were used to create both patient-specific and laboratory-specific noise profiles to filter out surrounding variants, resulting in the ability to discern variants well below the current limit of detection. Each target was assigned its own limit of detection based on the lowest variant detected and the limit of blank established from the normal profile for accurate MRD analysis. We used these results were used to develop error-corrected sequencing and subsequently develop a highly sensitive MRD assay for our patients with AML.