‘Liquid biopsy’ predicts lymphoma therapy success within days

Changes in circulating tumor DNA levels quickly predict how patients with diffuse large B cell lymphoma are responding to therapy, according to a Stanford-led study. Currently, patients wait months for the results.

- By Krista Conger

Maximilian Diehn (left) and Ash Alizadeh led a study that found a blood test could show whether patients with a type of blood cancer were responding well to initial treatment or needed more agressive treatment.
Mark Tuschman

A blood test can predict which patients with a type of cancer called diffuse large B cell lymphoma are likely to respond positively to initial therapy and which are likely to need more aggressive treatment, according to a multicenter study led by researchers at the Stanford University School of Medicine.

The study validates the clinical usefulness of tracking the rise and fall of circulating tumor DNA, or ctDNA, in the blood of patients before and after therapy. It suggests that clinicians may soon be able to determine how a patient is responding to treatment within days or weeks of starting therapy rather than waiting until therapy is completed five to six months later.

“Although conventional therapy can cure the majority of patients with even advanced B cell lymphomas, some don’t respond to initial treatment,” said associate professor of medicine Ash Alizadeh, MD, PhD. “But we don’t know which ones until several months have passed. Now we can predict nonresponders within 21 days after the initiation of treatment by tracking the levels of ctDNA in a patient’s blood. We can look earlier and make a reliable prediction about outcome.”

The study was published online Aug. 20 in the Journal of Clinical Oncology. Alizadeh shares senior authorship with associate professor of radiation oncology Maximilian Diehn, MD, PhD. Instructor of medicine David Kurtz, MD, PhD, and postdoctoral scholar Florian Scherer, MD, are the lead authors.

Varying responses to treatment

Diffuse large B cell lymphoma, a blood cancer, is the most common type of non-Hodgkin lymphoma. Because it is highly biologically variable, patients vary widely in their response to treatment. Although most people are cured by conventional therapy, about one-third are not. Being able to predict early in the course of treatment those who will need additional or more aggressive therapies would be a significant boon to both clinicians and patients.

Circulating tumor DNA is released into the blood by dying cancer cells. Learning to pick out and read these DNA sequences among the thousands or even millions of other noncancerous sequences in the blood can provide valuable insight into the course of the disease and the effectiveness of therapy. Recently, Diehn and Alizadeh showed that ctDNA tracking can also predict lung cancer recurrence weeks or months before any clinical symptoms arise.

“Combined with our recent study on lung cancer, our new findings speak to the power and likely utility of using ctDNA to assess how well cancer treatments are working in an individual patient. We are very hopeful that the approach will ultimately be extensible to most if not all cancer types,” Diehn said.

If we can identify those people who are responding extremely well, we could spare them additional treatments.

In this study, the researchers tracked ctDNA levels in 217 people with diffuse large B cell lymphoma who were treated at six medical centers — three in the United States and three in Europe. For each patient, they compared levels of ctDNA before treatment began with the levels after the first and second rounds of conventional chemotherapy. They then correlated those changes with each patient’s outcome.

They found that ctDNA was detectable prior to the initiation of therapy in 98 percent of the people studied. And, as would be expected, the amount of ctDNA in the blood dropped in all patients once treatment began. But the precipitousness of the decline varied. Those people whose ctDNA levels dropped a hundredfold after the first round or three-hundredfold by the second round were much more likely to live 24 months or more without experiencing a recurrence of their disease than those whose ctDNA levels declined more slowly.

“We found that ctDNA levels serve as a very sensitive and specific biomarker of response to therapy within as few as 21 days,” Kurtz said. “Every year, about 30,000 people in the United States are diagnosed with diffuse large B cell lymphoma and, for the most part, they’re treated with six cycles of combination therapy. But we know that not all patients need six cycles. A large fraction could be cured with fewer cycles — maybe even just two. If we can identify those people who are responding extremely well, we could spare them additional treatments. Conversely, we could intensify the therapy or seek other options for those who are not responding as well as we would have hoped.”

Hopes for expansion

The researchers are encouraged that they saw a similar correlation between changes in ctDNA levels and outcomes in patients from each of the six participating medical centers, confirming the global usefulness of the analysis. They’re currently planning a clinical trial based on the results, and they’re eager to learn whether they can make similar predictions about the prognoses of patients other than those with diffuse large B cell lymphomas.

“These findings confirm the value of tracking cancer genetics in the blood in real time,” Alizadeh said. “We are thinking about how to use the tools to best benefit patients, and are very excited to test this approach in other types of cancers.”

The work is an example of Stanford Medicine’s focus on precision health, the goal of which is to anticipate and prevent disease in the healthy and precisely diagnose and treat disease in the ill.

Other Stanford authors of the study are research assistants Michael Jin and Alexander Craig; visiting student researcher Joanne Soo; postdoctoral scholars Mohammad Esfahani, PhD, and Jacob Chabon, PhD; bioinformatics scientist Henning Stehr, PhD; laboratory manager Chih Long Liu, PhD; professor of biomedical data science and of statistics Robert Tibshirani, PhD; assistant professor of medicine Lauren Maeda, MD; assistant professor of medicine Neel Gupta, MD; assistant professor of medicine Michael Khodadoust, MD, PhD; professor of medicine Ranjana Advani, MD;  professor of medicine Ronald Levy, MD; and assistant professor of biomedical data science Aaron Newman, PhD.

Authors from University Hospital Essen, Germany; Hopitaux Universitaires Henri Mondor, France; Centre Hospitalier Universitaire, France; the MD Anderson Cancer Center; the National Cancer Institute; the University of Eastern Piedmont, Italy; and the Oncology Institute of Southern Switzerland and the Institute of Oncology Research, Switzerland, also contributed to the study. 

Alizadeh is a member of the Stanford Child Health Research Institute, the Stanford Institute for Stem Cell Biology and Regenerative Medicine, the Stanford Cancer Institute and Stanford Bio-X.

The research was supported by the National Institutes of Health (grant 1DP2CA186569), the Damon Runyon Cancer Research Foundation, the American Society of Hematology, the V Foundation for Cancer Research, the German Research Foundation, the Conquer Cancer Foundation, the Emerson Collective Cancer Research Fund, a Stinehart/Reed award, a Stanford TRAM pilot grant and the Ludwig Institute for Cancer Research.

Stanford’s Department of Medicine also supported the work.

About Stanford Medicine

Stanford Medicine is an integrated academic health system comprising the Stanford School of Medicine and adult and pediatric health care delivery systems. Together, they harness the full potential of biomedicine through collaborative research, education and clinical care for patients. For more information, please visit med.stanford.edu.

2023 ISSUE 3

Exploring ways AI is applied to health care