Diffuse large B-cell lymphoma (DLBCL) is an aggressive hematologic malignancy that is the most common form of non-Hodgkin lymphoma in adults. In the U.S., this disease occurs at a frequency of approximately eight cases per 100,000 people. Although several prognostic tools have been developed to stratify patients so that their treatment may be improved, these methodologies do not consistently predict DLBCL outcomes.
As a result, an international study was undertaken by a group of investigators, including Ash Alizadeh, MD, PhD, Associate Professor of Medicine at Stanford University, to evaluate the use of circulating tumor DNA (ctDNA) as a prognostic biomarker before and during DLBCL treatment (J Clin Oncol 2018;36(28):2845-2853). He worked with co-authors David M. Kurtz, MD, an instructor in medical oncology, and Maximilian Diehn, MD, PhD, Associate Professor of Radiation Oncology, both at Stanford.
"Our data suggest that both pretreatment and dynamic ctDNA assessments are feasible and can add to established risk factors. These approaches may allow novel clinical trial designs having wide applicability to patients with DLBCL and potentially other lymphomas," Alizadeh noted.
Background
DLBCL patient outcomes have been improved by the addition of rituximab to the standard therapy of cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP); however, a significant number of these patients still experience disease relapse or death. Consequently, a number of prognostic tools have been developed and used to stratify these patients into different risk groups, but these methods have not had a major impact on therapeutic outcomes.
"Two such tools are the International Prognostic Index (IPI) and interim positron emission tomography (PET), which have been used to select patients for intensified therapy," Alizadeh said. "These approaches have failed to improve survival and are confounded, in part, by imperfect risk stratification, including the variable specificity of interim PET/computed tomography (CT)."
The use of ctDNA as a biomarker has been explored for a number of different malignancies, including lymphomas. Prior studies have shown the potential for using ctDNA as a noninvasive means of molecular subtyping and tumor-specific mutation detection (Blood 2015;125:3679-3687).
"Because of the ease of sample collection, ctDNA offers unique opportunities for repeated tumor assessment before, during, and after therapy," Alizadeh noted.
"In our study, we sought to accomplish a number of goals; first, we wanted to apply personalized profiling by deep sequencing (CAPP-Seq) to examine the performance of ctDNA for both mutational genotyping and disease burden measurement in large B-cell lymphomas," he explained.
"Next, we tried to explore the utility of ctDNA quantification before and during therapy in order to predict overall survival (OS) and event-free survival (EFS) at 24 months, which is an important disease milestone in DLBCL. In a training and validation context, we wanted to define thresholds for molecular response capable of predicting outcomes after as few as one therapy cycle. Finally, we attempted to assess the utility of ctDNA in the context of established prognostic tools, demonstrating independent value for prediction of outcomes."
Methodology
For this study, patients seeking treatment for pathologically confirmed DLBCL or primary mediastinal large B-cell lymphoma (2008 WHO criteria) were enrolled at six different treatment centers across the U.S. and Europe. "Those participants that had preceding low-grade lymphoma with histologic transformation were deemed eligible, as were patients with MYC and BCL2/BCL6 rearrangements," Alizadeh stated.
Patients were divided into two cohorts based on their enrollment site. Patients in cohort 1 (n=144) were enrolled at Stanford Cancer Center, MD Anderson Cancer Center, and University of Eastern Piedmont; of these, 14 patients comprised a discovery set, while the remaining 130 comprised validation set 1. Patients from cohort 2 (n=73), which also served as validation set 2 for molecular response thresholds, were enrolled at the NCI, Centre Hospitalier Universitaire Dijon, and Essen University Hospital.
A discovery set of 14 patients from cohort 1 were utilized to determine the optimal timing and thresholds for molecular response; their samples were profiled throughout the first two cycles of therapy.
"After the optimal timing and threshold levels were identified, we profiled samples before the first, second, and third cycles of therapy from an additional 203 patients across all six institutions," Alizadeh explained.
"Targeted gene sequencing was performed using CAPP-Seq; somatic mutations were identified by paired analysis of either tumor or pretreatment plasma/serum and germline DNA." Assessment of blood ctDNA was accomplished by tracking somatic alterations in pretreatment and serial samples. Regarding ctDNA quantification, he noted, "Levels were measured in haploid genome equivalents per milliliter (hGE/mL) and expressed in log scale (log hGE/mL)."
Results
Somatic alterations for ctDNA quantitation and disease monitoring were identified by sequencing tumor or pretreatment cell-free DNA from all patients.
"Of the 217 patients included in our study, all but two patients (215, or 99%) had at least one tumor-specific alteration identified for tumor monitoring; 95 percent of patients had more than five such mutations," Alizadeh stated. "ctDNA was detected in 212 patients' (98%) cell-free DNA pre-therapy samples. Pretreatment ctDNA was significantly associated with both IPI and total metabolic tumor volume (TMTV) in patients receiving frontline therapy, which suggests ctDNA could serve as both a prognostic factor and a quantitative proxy for disease burden, another which is a documented prognostic factor for lymphomas." Regarding TMTV, Alizadeh noted, "This is an emerging biomarker that has shown strong correlation with ctDNA levels.
"In cohort 1, levels of ctDNA were continuously associated with both EFS and OS in those receiving either frontline or salvage therapy," he noted. "Using the ctDNA threshold value of 2.5 log hGE/mL, patients with high levels had significantly inferior rates of EFS at 24 months than those with low levels. This association was significant for EFS in both frontline and salvage settings (frontline: HR=2.6; p=0.007; salvage: HR=2.9; p=0.01)."
Additionally, high ctDNA levels predicted significantly worse OS in the salvage setting. "In multivariable analysis, pretreatment ctDNA remained prognostic for EFS in frontline treatment patients when controlling for IPI, molecular subtype, and TMTV," Alizadeh explained.
"Importantly, changes in ctDNA were prognostic of complete response. By the midpoint of the first cycle (6-16 days), patients could be perfectly discriminated as responders and nonresponders."
By the start of the second therapy cycle (i.e., 21 days after therapy initiation), a clear demarcation between groups emerged; a 2-log (100-fold) drop predicted an eventual complete response (CR). Similarly, a 2.5-log (~316-fold) drop in ctDNA by the start of cycle 3 also separated responding from nonresponding patients. The 2-log drop in ctDNA by the start of cycle 2 threshold was thus defined as an early molecular response (EMR), while the 2.5-log drop by the start of cycle 3 was defined as a major molecular response (MMR). These thresholds, which were initially found in the discovery set, were subsequently confirmed to be the optimum thresholds for determining EFS using bootstrap resampling.
"Notably, EMR and MMR were concordant in 92 percent of patients (57 of 62) in whom both were evaluable, demonstrating the robust performance of molecular response," Alizadeh commented.
The performance of EMR and MMR thresholds was then assessed for predicting survival in validation set 1 (cohort 1). In these patients, EMR and MMR were both prognostic for EFS and OS in patients receiving frontline therapy (EMR, p=0.0015 and p< 0.001; MMR, p<.0001 and p=0.0047, respectively). For those patients receiving salvage therapy, EMR was also prognostic for both EFS and OS (p=0.011 and p=0.011, respectively); however, an insufficient number of patients in this subgroup that received salvage therapy had data available for evaluating MMR.
Discussion
In summarizing what he thought were the main findings of their study, Alizadeh stated, "The fact that 98 percent of our DLBCL patients had quantifiable ctDNA was significant, showing that this procedure has the potential for broad applicability to DLBCL patients."
When queried about the utility of the molecular responses observed in their study, he replied, "We envision early milestones such as EMR and MMR will be useful in many areas; EMR may be used in drug development as an early surrogate endpoint in trials.
"Alternatively, ctDNA quantitation could be used in clinical practice as a prognostic factor for individual patients, and finally, these biomarkers could guide personalized approaches in novel clinical trial designs."
One potential first approach might involve intensifying therapy for patients not achieving a molecular response (EMR or MMR) and have a positive interim PET/CT scan. "Additional studies which could be useful are those that explore approaches for integrating ctDNA with traditional risk-assessment tools," Alizadeh commented.
"In that regard, the combination of molecular response and interim PET/CT response was able to robustly stratify both EFS and OS. Identification of those patients who are at exceptionally high risk (i.e., interim PET/CT positive and did not achieve EMR or MMR) could provide an opportunity for early intervention with alternative treatments, including autologous bone marrow transplantation or chimeric antigen receptor T cells (CAR-T)," he explained. It is hoped that identifying this highest-risk group may improve risk-adapted approaches and ultimately result in improved treatment outcomes.
When asked about the next steps for this promising tool, Alizadeh replied, "First, we need to try and replicate these findings in a larger and more diverse patient population. For those patients who did not respond to their therapies, we can use the genetic markers identified with the deep gene sequencing from the ctDNA to perhaps find directed therapies more appropriate for their specific disease.
"For example, more than 80 percent of the DLBCL patients in the study had mutations for which there were targeted therapies available. For those patients responding to their therapies, we can try to shorten dosing schedules so that they are not taking their medication longer than is necessary," Alizadeh concluded.
Richard Simoneaux is a contributing writer.
HemOnc Times Available Now!
Stay current on the latest hematology/oncology advances with HemOnc Times, your go-to source for insights and tools geared to enhancing your practice and research. Visit bit.ly/HemOncTimes for articles, podcasts, and more on innovations in the field.