NEW ORLEANS – A polygenic risk score designed to assess how well the body processes cytarabine could help doctors identify pediatric patients with acute myeloid leukemia who are likely to do better on a chemotherapy regimen augmented with clofarabine, according to data presented at the American Society of Hematology's annual meeting.
The test, called ACS10, was developed by Jatinder Lamba, a professor in the department of pharmacotherapy and translational research at Florida University's College of Pharmacy, and collaborators from St. Jude Hospital and Children's Oncology Group. ACS10 scores patients into high and low categories. And since African American pediatric AML patients are far more likely to have low ACS10 scores, the researchers are hopeful that their test could help doctors personalize therapy for this historically underserved group.
Standard therapy for AML since the 1970s has been the so-called 7+3 regimen, comprising seven days of cytarabine plus daunorubicin on days one through three. The therapy is usually given once, but may be repeated for a second cycle. In recent years, additional new agents have been added to the basic 7+3 regimen.
Cytarabine, for example, is a pro-drug that requires conversion to an active form, ara-CTP, through a multistep enzymatic process. High intracellular concentrations of ara-CTP are required to achieve an anti-leukemic effect. However, levels of intracellular ara-CTP vary considerably between patients. Genetic variation in key enzymes involved in the conversion of cytarabine to ara-CTP are responsible for this variation.
The ACS10 score that Lamba and colleagues built comprises 10 single nucleotide polymorphisms in the enzymes involved in cytarabine's metabolism. Patients with low ACS10 scores are not able to efficiently convert cytarabine into ara-CTP. Those with high ACS10 scores, conversely, are efficient activators of cytarabine.
In Lamba's initial report in the Journal of Clinical Oncology, a low ACS10 score identified patients who had unfavorable outcomes with 7+3 chemotherapy. When patients with low ACS10 scores were given a different chemotherapy regimen that did not rely on cytarabine, they experienced a 10 percent improvement in five-year event-free survival and overall survival compared to those receiving standard treatment.
Inspired by St. Jude's AML08 clinical trial of clofarabine and cytarabine compared to standard induction chemotherapy in AML patients, Lamba and her colleagues sought to evaluate whether ACS10 scores could predict outcomes in the 177 patients who received clofarabine and cytarabine (or augmented therapy) in the study. They looked for the same association between the polygenic risk score and the outcomes of 91 patients on standard cytarabine-containing chemotherapy induction in another St. Jude trial, AML02.
In both these studies, "when the patients who have low ACS10 score were given combination of clofarabine and [cytarabine], they did significantly better compared with the patients who received the low-dose [cytarabine]," said Lamba. In contrast, in the group of patients with high ACS10 score, those on low-dose cytarabine did better. "And this was quite a surprise, but in this group of patients, when you give them augmented therapy, actually their outcome was not as good."
Ranking the possible therapy options from AML02 and AML08, patients with low ACS10 scores responded best to the augmented therapy option, clofarabine with cytarabine. The second-best choice for that group was high-dose cytarabine, consistent with the theory that response is related to a reduced ability to activate the drug inside the cell. The patients with low ACS10 scores fared worst when given low-dose cytarabine.
In the AML02 trial, patients with low ACS10 scores had a 35 percent improvement in five-year event-free survival when they received high-dose cytarabine, as compared with low-dose cytarabine. In AML08, patients on clofarabine with low-dose cytarabine had 70 percent greater event-free survival than those on low-dose cytarabine.
For patients with high ACS10 scores, the therapy ranking went in the opposite direction. Patients did best on low-dose cytarabine and worst on clofarabine and cytarabine.
Lamba noted that ACS10 scores were not associated with initial risk group assignment, and using multivariate analysis, researchers found that ASC10 scores remained a significant predictor of therapy benefit after taking into account age, race, and FLT3 mutation status.
Next, Lamba and her colleagues are proposing to test ACS10 in a prospective clinical trial using a pharmacogenomics-based model. In such a trial, researchers would use ACS10 to stratify patients into ACS10 high and low score groups and give them treatments based on those scores. For the low ACS10 group, for example, treatment options would include either clofarabine with cytarabine or high-dose cytarabine.
Depending on patients' CD33 genotype, patients would receive low-dose cytarabine plus gemtuzumab ozogamicin but not standard low-dose cytarabine. For the high ACS10 group, meanwhile, treatment options would be low-dose cytarabine or cytarabine plus gemtuzumab ozogamicin.
Lamba highlighted that patients with African ancestry in her team's analysis were much more likely to have low ACS10 scores than patients with Caucasian ancestry. Seventy percent of patients of African ancestry had a low score, whereas only 30 percent of patients with Caucasian ancestry did. Black and Hispanic children with AML are known to have worse outcomes than white non-Hispanic children, with little known about any underlying genetic mechanism that may cause those differences.
"Can this score also contribute towards the observed racial disparities in outcomes that we've seen in different clinical trials?" said Lamba. "Our ongoing work is expanding this into the African ancestry patients by collating the data from multiple clinical trials."
Lamba's group is also planning to evaluate the score in adult AML patients and study its relationship to clinically relevant therapeutic toxicities. Further, they hope to gauge whether ACS10 can identify best responders to other agents, such as CPX-351 (cytarabine(araC) and daunorubicin) and AbbVie/Genentech's Venclexta (venetoclax).
For near-term clinical applicability of the polygenic risk score, Lamba said that the pathology lab at the University of Florida already offers CD33 testing and intends to integrate assessment of the 10 SNPs into its workflow.