Predicting MDS and AML Before it Begins: Targeted Sequencing Identifies Patients at High Risk for Disease Development

The diagnosis of myelodysplastic syndromes (MDS) remains difficult due to the subjective nature of morphologic assessment and the heterogeneity of the disease itself, but a new report published in Blood finds that, in patients who don’t yet have morphologic evidence of MDS or AML, certain mutations can predict those patients at high risk for developing the disease.

In this retrospective study, Catherine Cargo, MB, from the Hematological Malignancy Diagnostic Service at St. James’ University Hospital in the United Kingdom, and colleagues sought to molecularly characterize those patients with suspected MDS who have failed to meet strict diagnostic criteria using conventional techniques.

While cytogenetics can provide objective evidence of disease, this approach has been complicated by reports of frequent somatic mutations and large chromosomal abnormalities in healthy patients. Therefore, the authors stated, “[Patients] with clinically significant mutations are currently indistinguishable from those who will not progress.” Dr. Cargo and co-authors hypothesized that analyzing the molecular abnormalities in these patients would provide potential criteria to distinguish pre-clinical MDS from healthy individuals and detect those patients at high risk for developing the disease.

The researchers identified 82 patients from the Hematological Malignancy Diagnostic Service between 2004 and 2012 who, despite having an initial bone marrow with non-diagnostic features, went on to develop progressive dysplasia or acute myeloid leukemia (AML). They were compared with a cohort of healthy patients from several published studies reporting on clonal hematopoiesis.

Sixty-nine patients had adequate molecular material at both pre-diagnostic and diagnostic timepoints, and 59 patients had survival data available.

To investigate the impact of genetic abnormalities on overall survival (OS) and the association between genetic mutations and disease development, the researchers performed targeted sequencing for 26 commonly mutated genes in myeloid malignancies and array-based analyses on patient samples.

Most patients (63 patients; 91%) had a driver mutation and/or structural variant on pre-diagnostic samples, including 133 mutations among 62 patients, most commonly involving epigenetic regulators or spliceosome genes. At this point, the researchers noted, the spectrum of mutations mirrored the mutations reported in large MDS populations, with TET2 mutated in 39 percent of patients, SRSF2 in 26 percent, and ASXL1 in 20 percent.

DNMT3A was the most frequently mutated gene in the healthy population, though this was only seen in 10 percent of this patient cohort. A median of two mutations were detected per patient (range = 1-5), and the frequency of multiple mutations was greater in the AML/MDS patient population compared with the healthy cohort (64% vs. 0.8%; p value not reported). Of the 59 patients with OS data available, only 10 remained alive at the point of analysis. “Median OS from the pre-diagnostic sample was 43.6 months (95% CI 33.8-55.8) and, as expected, much shorter from diagnosis (13 months; 95% CI 9.9-24.6),” Dr. Vaughn and colleagues reported.

Patients diagnosed with AML had a median OS of 1.28 months (95% CI 0.789-12.625). Thirty-nine patients progressed to refractory anemia with excess blasts (RAEB) or AML and, notably, this occurred in a significantly shorter time than progression to refractory cytopenia with multilineage dysplasia (median = 403 vs. 606 days; HR=3.7 95% CI 2.1-6.6; p<0.001). Analysis of the most frequently mutated genes revealed that IDH2 and TP53 mutations were associated with a more rapid time to disease progression (hazard ratio [HR] = 4.2; 95% CI 1.3-13.8; p=0.017 for IDH2 and HR=5.5; 95% CI 1.1-27.7; p=0.038 for TP53).

Certain pre-diagnostic mutations were also associated with significantly worse OS, particularly for TP53 (HR=21.68; 95% CI 4.72-99.64; p<0.001) and U2AF1 (HR=2.63; 95% CI 1.0-6.4; p=0.049).

Thirty patients (43%) acquired a mutation between pre-diagnostic and diagnostic samples, most commonly involving transcription factors and cell-signaling genes. These mutations also correlated strongly with progression to RAEB and AML.

“The mutational profile in our cohort differs significantly from that in the healthy population and has the potential to identify patients with clonal hematopoiesis of indeterminate potential or clonal cytopenias of undetermined significance who are at greater risk of progression, even in the absence of morphological disease,” Dr. Vaughn and co-authors concluded. “Early detection would provide an increased window for therapeutic intervention in those with very poor prognosis.”

However, the diagnostic utility of these findings is limited by the lack of a control group of patients who did not progress to AML/MDS. “Our study design included only those patients re-investigated for cytopenia and both the mutation and disease frequency is likely to be greater,” the authors wrote, suggesting that the optimal way to explore this and refine molecular criteria for diagnosis is to prospectively study an unselected cytopenic patient cohort. A study like this is currently underway, they added.


Reference

Cargo CA, Rowbotham N, Evans PA, et al. Targeted sequencing identifies patients with pre-clinical MDS at high risk of disease progression. Blood. 2015 September 17. [Epub ahead of print]

SHARE