Predicting Therapy Response in ALL Patients May Come Down to 9 Genes

To determine why some patients with cancer respond better to a particular treatment than others, researchers often retrospectively analyze data from multiple studies for clues. A team of investigators from University of Texas M D Anderson Cancer Center sought to determine whether adult patients with acute lymphoblastic leukemia (ALL) might have genetic biomarkers that predict treatment response. They identified a 9-gene signature that was highly accurate in predicting which patients with Philadelphia chromosome-positive (Ph+) ALL would have persistent disease or experience relapse and presented the results in a poster session.

The investigators looked at data for 672 ALL patients who participated in earlier genome-wide gene expression studies. They identified 46 genes in 7 pathways that they believed might be indicative of therapy response (P <.001). Using Applied Biosystems low-density reverse-transcription quantitative PCR arrays, they analyzed bone marrow samples from 27 patients who had undergone treatment with standard chemotherapy and a tyrosine kinase inhibitor. Patients were categorized as “optimal,” “persistent,” or “relapse,” based on measurable BCR/ABL1 levels in response to therapy. Researchers noted that the three groups were similar in “age, initial peripheral blood white cell and bone marrow blast counts, and initial normalized BCR/ABL1 levels.” The groups differed significantly in disease-free survival, however, which reached a median 12 months in the optimal group and 5 months in the relapse group (P = .002).

Analyzing gene expression in 5 patients from each group, researchers were able to narrow the gene signature down to 9 markers. Further analysis of the gene expression profile for 12 more patients found the 9-gene profile had 91.7% accuracy in predicting therapy response.

There are approximately 1000 new cases of adult ALL diagnosed in the United States annually. Adult ALL has a high rate of treatment resistance or relapse. The ability to determine which patients are likely to respond to standard treatments beforehand could allow oncologists to select other treatments more likely to be effective. It could also spur development of new treatments for patients with this particular 9-gene signature.

ASCO Abstract No. 7014.