Algorithm Predicts Severe GVHD After Bone Marrow Transplant

Veronica Hackethal, MD

February 09, 2017

A new algorithm that uses a simple blood test performed 7 days after bone marrow transplant may predict serious, life-threatening graft-vs-host disease (GVHD), according to a study published online February 9 in the Journal of Clinical Investigation Insight.

The test could predict GVHD before clinical symptoms develop, the researchers emphasize, which could allow for early intervention.

Most (90%) clinical symptoms of GVHD develop 14 or more days after transplant, while this blood test is taken 7 days after the transplant.

"This test will help early identification of those patients who need aggressive treatment, and may be able to prevent the disease altogether. It is designed to work in any patient in any hospital and was tested in hundreds of patients from three continents," commented lead author, James Ferrara, MD, from the Icahn School of Medicine at Mount Sinai, New York, New York. Dr Ferrara is also codirector of the Mount Sinai Acute GVHD International Consortium (MAGIC).

The MAGIC algorithm is basically an equation. Plugging in the concentrations of two plasma proteins, ST2 and REG3α, gives the predicted probability of 6-month mortality from GVHD.

To develop the algorithm, researchers collected blood samples from 1287 patients 7 days after they underwent bone marrow transplant at 11 centers in the MAGIC consortium. The group then measured blood concentrations of four biomarkers of GVHD (ST2, REG3α, TNFR1, and IL-2Rα) and used these markers to model 6-month mortality from GVHD.

Results showed that the most accurate algorithm included the concentrations of just ST2 and REG3α and that the algorithm could differentiate high- vs low-risk groups. With use of this algorithm, the cumulative incidence of 6-month GVHD mortality was 28% in the high-risk group and 7% in the low-risk group (P < .001). Testing the algorithm in another group of patients produced similar results (33% for high risk vs 7% for low risk; P < .001). Testing a third time in a validation set of patients from multiple centers also yielded similar results (26% for high risk vs 10% for low risk; P < .001). Combining data from all three patient sets showed that high-risk patients were three times more likely to die of GVHD than low-risk patients (19% vs. 6%, respectively; P < .001).

A key part of the MAGIC algorithm was the identification of REG3α and ST2, Dr Ferrara told Medscape Medical News.

That's because, even though GVHD can affect other parts of the body, the GVHD that eventually kills people usually occurs in the gastrointestinal (GI) tract, he explained.

Both REG3α and ST2 are released into the blood when the GI tract becomes damaged. REG3α helps protect the surface of intestinal cells from infection. ST2 is a soluble form of the interleukin (IL)-33 receptor, which is released very early from damaged intestinal cells and acts as a powerful activator of T cells that cause GVHD.

"What's so interesting about this test is that these proteins increase in the bloodstream well before we see any evidence of the clinical signs of GVHD, so it's very sensitive to the damage," Dr Ferrara said.

That means that the algorithm could identify patients who need early intervention, he suggested. The only proven therapy for GVHD is high-dose systemic steroids, which greatly increase the risk for serious opportunistic infections. However, two new classes of drugs hold promise for the treatment of GVHD, Dr Ferrara commented.

The first is monoclonal antibodies, which prevent the trafficking of T cells to the GI tract. Examples include vedolizumab, natalizumab, and etrolizumab, which are currently used to treat inflammatory bowel disease. Two clinical trials of these agents in GVHD prophylaxis are underway.

The second option is using interleukins, such as IL-22, to strengthen the natural immune response in the GI tracts and protect it. Recombinant IL-22 is being tested in clinical trials for treating GVHD.

For now, though, the MAGIC algorithm remains a research tool. While Dr Ferrara stressed that it is a "major step toward precision medicine" for patients undergoing bone marrow transplant, he cautioned about rushing its use into clinical practice. Before it becomes standard of care, clinical trials are needed to prove that acting on the test can prevent GVHD, he said.

"I think basing treatment on the algorithm right now is premature. All of the monoclonal antibodies and cytokines have some risk. These are very sick, vulnerable patients, and we need to be careful that they don't have unintended, damaging side effects," he said.

However, he is hopeful that the test will be ready for clinical practice soon and that its use can make bone marrow transplant a safer option for many patients who could benefit from this potentially curative therapy.

"Our goal with the new trials is to cut the mortality rate from GVHD in half in the next 5 years. I'm hoping that by either 2020 or 2021 the algorithm should be ready for clinical practice," he concluded.

The study was supported by the National Cancer Institute, American Cancer Society, and Doris Duke Charitable Foundation. Authors and codirectors of the MAGIC consortium John Levine and James Ferrara are coinventors of a patent for GVHD biomarkers.

JCI Insight. 2017;2:e89798. Full text

Follow Medscape Oncology on Twitter: @MedscapeOnc


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.