New Software Program May Help Prevent Leukemia

Two University of Florida (UF) medical students are now playing a major role in the battle against acute myeloid leukemia (AML). They have found it may be possible to use personalized medicine to catch this type of leukemia before it starts. They reported at the American Society of Hematology’s 57^th Annual meeting in Orlando this month that it may be possible to treat myelodysplastic syndrome (MDS) and potentially prevent it from turning into leukemia.

Myelodysplastic syndrome is a rare disease in which bone marrow loses the ability to produce healthy blood cells, and in about 30% of the cases it can progress into acute myeloid leukemia. University of Florida medical student Cindy Medina found that a software program that interprets cancer gene mutations and tests for drug sensitivity was successful in predicting patient outcomes in an examination of three previously published clinical studies.

The new personalized medicine software program, iCare for Cancer Patients, reads the genes that make up an individual patient’s cancer. It also creates a unique map of the inside of the patient’s cancer cell using information from published medical literature and then screens for US Food and Drug Administration (FDA)-approved drugs with therapeutic potential.

In the previously published studies, MDS patients were treated with three different kinds of medications: azacitidine (Vidaza), decitabine (Dacogen), and lenalidomide (Revlimid). In the first study, there were 37 patients who responded well to lenalidomide, and the software predicted 35 patients would respond well. Of the nine patients that did not respond to lenalidomide, the software predicted four would not respond well.

In the second study, seven patients responded well to the chemotherapy drugs azacitidine or decitabine, and the software predicted all of them correctly. Of the eight patients who did not respond well to the medications, the software was correct in five cases. In the final study, the software correctly predicted clinical outcomes in all 10 MDS patients treated with the combination of azacitidine and lenalidomide.

The researchers found that the software did a good job at predicting who would respond to treatment because the software relies on published studies available in PubMed to make an accurate prediction. If there were more publications on genes and proteins linking to improved clinical response, then the software would be stronger in predicting who will respond well. By incorporating new information as soon as it is published in PubMed, the software continually learns about MDS and provides better modeling results.

“In the long run, we really want to be more efficient in treatment selection, and spare patients from lengthy or toxic treatments that maybe they were not going to benefit from in the first place,” said Medina in a UF Health news release.

University of Florida medical student Shannon Stockton discovered that 21 different gene mutations are recurrently found in patients with MDS, and that some of the gene mutations link with particular blood count abnormalities. Stockton found these associations by combining traditional blood cell counting and chromosome staining with next-generation sequencing.

Stockton said that although all MDS patients have blood count abnormalities, about half have chromosomes that appear normal. Stockton’s research showed that all MDS patients, no matter what their chromosomes looked like under the microscope, had one to four mutated genes out of the 21 genes examined.