Molecular data may refine outcome prediction in elderly AML patients
Including molecular data, such as transcriptomic and mutational data, may help improve the prediction of outcomes in elderly patients with acute myeloid leukaemia (AML), a recent study has shown.
The study included 182 elderly AML patients (median age, 74 years; 50 percent female) who were intensively and palliatively being treated. Of the patients undergoing intensive chemotherapy, 59 percent reached complete remission, with median overall survival of 8.2 months. Each patient harboured a median of three somatic mutations.
The most commonly mutated gene in the cohort was the NPM1 gene, which saw mutations in 32 percent of the participants. This was followed by the TET2 (28 percent) and DNMT3A (25 percent) genes. The NPM1 gene (37 percent) was likewise the most commonly mutated gene in participant with de novo AML.
Multivariable Cox regression analysis found that mutations in the NPM1 gene were significantly correlated with better overall survival (OS; hazard ratio, 0.15, 95 percent confidence interval [CI], 0.06–0.35). The same was true for mutations in the IDH2 gene, while alterations in TP53 emerged as negative risk factors.
Similarly, NPM1 mutations correlated with better odds of achieving complete remission (CR; odds ratio, 0.46, 95 percent CI, 0.20–0.99), while TP53 mutations worsened the chances of CR.
In addition, transcriptomic data further improved outcome prediction. “When integrating the transcriptomic and mutational data with the aim to create an algorithm for CR prediction, two genes, ZBTB7A and EEPD1, added the most value to the molecular stratification,” the researchers said.
“Patients with high expression of both genes but without TP53 mutations formed a good-risk group with a CR rate of 97 percent.”