Measuring changes in levels of BCR-ABL in patients with chronic myeloid leukemia (CML) can help predict treatment outcomes and disease progression, according to a new study.
Measuring changes in levels of BCR-ABL in patients with chronic myeloid leukemia (CML) can help predict treatment outcomes and disease progression, according to a new study.
Response assessment at 3 months with imatinib has become an important tool to predict outcomes, according to study authors led by Benjamin Hanfstein, MD, of the University of Heidelberg in Germany. “In theory, the prognostic impact of 3-month response landmarks could be driven by individual differences in tumor load at diagnosis, that is, BCR-ABL transcript levels, or by the individual reduction of transcript levels within the first 3 months of treatment,” the authors wrote.
In the new study, they analyzed data from 301 patients who participated in the German CML-Study IV. All were treatment-naive at sample collection, and were then treated with imatinib. The BCR-ABL transcript levels at diagnosis ranged very widely, and no particularly cutoff at that point could be identified that predicted overall survival (OS) or progression-free survival (PFS).
At 3 months, however, predictive cutoffs did emerge. A 3-month reduction ratio of 0.35-fold was found to be the best cutoff point, which corresponded to a 0.46-log reduction in BCR-ABL transcript levels. This amount yielded a high-risk group of 48 patients (16%), and discriminated significantly with regard to OS and PFS (P = .001 for both). The 5-year OS in the high-risk group was 83%, compared with 98% in the low-risk group.
The median BCR-ABL transcript ratio was 33% at diagnosis, and dropped to 1.4% at 3 months. The researchers looked at four different landmarks, all of which turned out to be significantly predictive of OS. Using a 6% cutoff yielded 65 patients as high-risk (22%), and a 5-year OS of 85% vs 98% (hazard ratio [HR] = 6.1; P = .002). HRs were similar for 10%, 14%, and 21% cutoffs, though each yielded fewer high-risk patients. Thus, the 6% cutoff was selected as the best available option.
The authors noted that the sample size was too small to also provide a validation cohort for these results, so further study is still needed to confirm them.
“The definition of cutoff levels to be used as predictive response landmarks that imply clinical consequences, for example, a change of TKI treatment or the evaluation of stem cell transplantation, is difficult,” the authors wrote. These cutoff values, though, do seem promising for doing just that. “The lack of achieving a half-log reduction at 3 months is a negative prognostic indicator and may trigger treatment intervention.”