Circulating tumor DNA in lymphomas: Era of precision medicine
Abstract
Non-Hodgkin's lymphomas are heterogeneous lymphoid malignancies, usually with an excellent prognosis. Aggressive B-cell lymphomas, such as diffuse large B-cell lymphomas, are treated with a curative intent with first-line therapy achieving a cure rate in about 70% of patients. Indolent lymphomas, such as follicular lymphoma (FL), are considered less curable, with the goal of treatment directed toward inducting deep remissions. With treatment strategies evolving more toward precision medicine, there is a need for better diagnostic, prognostic biomarkers. Circulating tumor DNA (ct-DNA) shed into the blood by tumor cells can be used as a marker for identification of tumor, assessment of tumor response, and to predict long-term outcomes. Recent studies have demonstrated the potential of these techniques as a prognostic tool. We review the clinical data supporting the role of ct-DNA in non-Hodgkin's lymphomas.
1 INTRODUCTION
Non-Hodgkin's lymphomas are a heterogeneous group of tumors with excellent prognosis with the current therapy plans. Aggressive lymphomas, such as diffuse large B-cell lymphomas, have a cure rate of 70% with multi-agent chemotherapy.1 Indolent lymphomas, such as FL and marginal zone lymphomas, are considered incurable. However, with advances in therapy, the overall survival in these lymphomas has increased significantly.2
With an increasing understanding of the tumor biology and development of targeted agents, there is an opportunity to further improve these outcomes.
Precision medicine involves the development of targeted therapy, taking into account patient and tumor characteristics to achieve the best results. Dynamic monitoring of the tumor response to therapy, tailoring therapy accordingly, and accurately predicting the outcomes of disease are important aspects of precision medicine. Determining the molecular biology of the tumor requires tissue sampling. This step can often be limited due to the location of the tumor, heterogeneity of the tumor, and sampling error, among other factors.3 These drawbacks in the current traditional methods have led to an increased interest in developing techniques to identify circulating tumor DNA (ct-DNA) as a means to obtain information regarding molecular biology, response to therapy, and minimal residual disease (MRD).
We will review the present understanding of ct-DNA and their utility in clinical practice.
1.1 Circulating tumor DNA
Circulating tumor DNA is defined as tumor-specific DNA sequences from the cancer cells detected in the plasma or the serum.4 Ct-DNA is thought to arise from passive release from apoptotic and necrotic cells, or active secretion by lymphocytes. Ct-DNA was first described in 1977, where their presence was associated with invasive cancers.5 Since then, there has been a great interest in developing sensitive techniques for the detection of ct-DNA.
1.2 Methods to detect ct-DNA
Detection of tumor-specific DNA aberrations would be the ideal way to isolate and measure ct-DNA. However, due to the heterogeneity of lymphomas, there is variability of genetic mutations within the same tumor.3 This results in the requirement of a large panel of probes to establish sensitive DNA-based detection methods.6 Additionally, unlike in the case of solid tumors, lack of single tumor-specific mutations make ct-DNA detection in lymphomas challenging. The ct-DNA pool originating from lymphoma cells is often tiny, making it difficult to detect and isolate (Table 1).
| High-output VDJ sequencing | CAPP-Seq | Digital PCR | |
|---|---|---|---|
| Technique | Next-generation sequencing | Next-generation sequencing | Probe-based PCR assay |
| Pros |
|
|
|
| Cons |
|
|
|
The following 3 strategies are commonly employed for isolating ct-DNA in lymphomas.
1.2.1 High-output variable-diversity-joining (VDJ) rearrangement sequencing
VDJ rearrangement occurs during lymphocyte maturation and results in diverse immunoglobulins, T-cell, and B-cell receptors, thus playing an important role in adaptive immunity.7 This rearrangement results in clonal types, which are specific to the lymphocytes. Lymphomas can have unique clonal types, which may serve as markers for the detection of specific ct-DNA in the blood. Next-generation sequencing techniques can detect the ct-DNA that encrypts the VDJ junctions on the immunoglobulin chains.7 By analyzing the VDJ rearrangement patterns, clonal evolution patterns can be detected. Doing so may aid in the detection of resistant clones during the course of therapy. However, VDJ rearrangements are not detectable in all patients, particularly in patients with immunoglobulin-negative phenotype of primary mediastinal B-cell lymphoma.8
1.2.2 Cancer personalized profiling by deep sequencing (CAPP-seq)
CAPP-seq is a technique that measures ct-DNA through high-output sequencing and also allows the quantification of somatic variants.9 This next-generation sequencing method overcomes the limitations posed by tumor heterogeneity by detecting a large spectrum of genetic alterations. In a study evaluating 92 lymphoma patients, CAPP-seq was able to identify clonal evolution in transformed FL.10 CAPP-seq was also able to identify the cell of origin subtypes accurately in patients with diffuse large B cell lymphoma (DLBCL) by testing somatic mutations in the ct-DNA.11 In this study, the cell of origin subtypes detected in ct-DNA using CAPP-seq were highly concordant with the tissue biopsies.
1.2.3 Digital PCR
Studies in solid tumors have shown digital PCR (dPCR) to be more sensitive than quantitative PCR to detect ct-DNA.12 Digital PCR is a cost-effective, quick technique and requires less tissue sampling, making it ideal for ct-DNA analysis.13 Digital PCR assays are probe-based assays and are designed to detect mutations characteristic to certain lymphomas. For example, the MYD88 L265P mutation is seen in ct-DNA in patients with lymphoplasmacytic lymphomas.14 Similarly, immunoglobulin heavy chain (IGH)/CCND1 translocation, t(11;14), and the IGH/BCL2 translocation, t(14;18), are seen in patients with mantle cell lymphoma (MCL) and FL, respectively.
Additionally, newer ultra-deep next-generation sequencing methods can detect a large number of somatic mutations.15
1.3 ct-DNA in clinical practice
Circulating tumor DNA has been widely studied and is currently used in clinical practice in carcinomas as a diagnostic marker. However, its role in hematological malignancies, particularly in lymphomas, is less clear. This is largely due to limitations in the identification and quantification of ct-DNA owing to the tumor heterogeneity in lymphomas (Table 2).
| Role of ct-DNA |
|---|
|
At the time of diagnosis
|
|
Interim therapy
|
|
End of therapy
|
1.4 Ct-DNA in diagnosis and risk stratification
Imaging studies, such as PET/CT scans, help determine the tumor burden and risk stratification of lymphomas. Unlike imaging studies, ct-DNA are tumor-specific, can quantitatively measure the tumor burden, and provide important prognostic information.
Kornacker et al detected clone-specific DNA in the serum of a patient with Hodgkin's lymphoma, at diagnosis and relapse, as a proof of principle for the presence of ct-DNA.16 Frickhofen et al detected clonotypic DNA from rearranged IGH locus in 86% of plasma samples of patients with non hodgin lymphoma (NHL) at diagnosis.17
Hohaus et al used real-time quantitative PCR to detect the plasma ct-DNA in 142 patients with indolent and aggressive lymphomas and in 41 healthy controls.18 The levels of ct-DNA in patients with DLBCL, MCL, and Hodgkin lymphoma (HL) were significantly higher compared to the other forms.
Higher levels of ct-DNA were associated with an advanced stage of the disease, presence of B symptoms, elevated lactate dehydrogenase, and age resulting in a higher International Prognostic Index (IPI) score.18 Additionally, elevated ct-DNA levels were associated with an inferior failure-free survival rate in patients with DLBCL and HL.
Li. et al evaluated the levels of cell-free DNA in 170 patients with lymphomas and 80 healthy people. The mean concentration of cell-free DNA in patients with lymphoma was significantly higher when compared to healthy individuals (median 686 ng/mL vs 222 ng/mL). As noted in prior studies, patients with DLBCL and HL had higher levels of circulating cell-free DNA, when compared to indolent lymphomas. Higher stage, age, and IPI were associated with higher levels of ct-DNA.
This study also reported that elevated circulating cell-free DNA was associated with worse prognosis. Patients with a higher circulating cell-free DNA had a 2-year probability of progression free survival of 44% compared to 78% in patients with a lower circulating cell-free DNA (P = <.001).
Kurtz et al published one of the largest studies supporting the role of ct-DNA as a prognostic tool.19 In this study of 217 patients with DLBCL, pretreatment levels of ct-DNA were associated with both event-free and overall survival in patients receiving frontline or salvage therapy. Patients with levels greater than 2.5 log hGE/mL of ct-DNA had significantly worse event-free survival at 24 months compared to those with low levels in both frontline (hazard ratio [HR] = 2.6, P = .007) and salvage settings (HR = 2.9, P = .01). Higher pretreatment levels of ct-DNA were also predictive of worse overall survival in the salvage setting.19 In multivariate analysis, pretreatment ct-DNA remained prognostic for event-free survival in patients receiving frontline treatment when controlling for IPI, molecular subtype, and tumor burden (P = .018). Molecular responses were also evaluated in this study with superior outcomes noted among patients achieving an early molecular response or a major molecular response. In multivariable analyses including IPI and interim positron emission tomography/computed tomography scans, the molecular response was independently prognostic of outcomes, including event-free and overall survival.
These studies support the role of ct-DNA as a prognostic tool in lymphomas.
Ct-DNA can provide the genetic information of the tumor, which can complement or even replace the information from tissue biopsies. Scherer et al detected somatic mutations in 98% of the patients and 100% of BCL2/BCL6 rearrangements using the CAPP-seq technique.10 Biopsy-confirmed tumor mutations were detected with 99% specificity in 97% of the pretreatment samples. Rossi et al demonstrated that pretreatment ct-DNA sequenced via ultra-deep targeted next-generation detected mutations with over 90% sensitivity when compared to tissue biopsy.15 These studies show that ct-DNA genotyping of DLBCL is as accurate as tissue biopsy for the detection of somatic tumor mutations. New mutations that are acquired during disease progression can also be detected during the clinical course. Ct-DNA could become a noninvasive approach to track the emergence of treatment-resistant clones.19
On the basis of this study, Kurtz et.al developed a dynamic risk assessment model to predict outcomes for patients with DLBCL. The continuous risk assessment model, continuous individualized risk index (CIRI) included 6 factors, which comprised 3 well-established risk factors and 3 ct-DNA risk factors. The risk factors were IPI, molecular cell of origin, interim imaging studies, pretreatment ct-DNA levels, early molecular response, and major molecular response. The authors utilized a Bayesian methodology and considered both pretreatment and on-treatment risk factors to estimate outcomes. Using CIRI, the authors predicted event-free survival at 24 months. CIRI demonstrated improved outcome prediction when compared with conventional risk models such as IPI.20
1.5 Role of interim ct-DNA
PET/CT scans are usually performed after a few cycles of chemotherapy to identify patient's treatment effects and make decisions to either escalate therapy or to determine prognosis.
The role of interim PET/CT scan has been well established in HL.21 However, in NHL, interim PET/CT scans have a high rate of false positives and a positive predictive value (PPV) of 50%.22-24 Because of this, the role of interim PET/CT scans as a means to alter management plans in NHL is unclear.24
In DLBCL, patients with an undetectable quantitative ct-DNA mid-therapy have shown a superior progression-free survival, when compared with patients with a positive interim ct-DNA.
Studies show that a 2 log drop in ct-DNA levels following 2 cycles of chemotherapy is associated with a complete response. Alternately, a drop of fewer than 2 log in interim ct-DNA is associated with a higher rate of relapse. Studies show that ct-DNA, along with an interim PET/CT scan, can improve the accuracy of residual disease assessment compared to an interim PET/CT alone. In an NCI study evaluating the role of ct-DNA in patients with DLBCL, the 5-year time of progression was 41% (95% CI, 22%-69%) in patients with a positive interim ct-DNA compared to 80% in patients with negative interim Ct-DNA.25 Interim ct-DNA results after 2 cycles had a PPV of 62.5% (95% CI, 40.6%–81.2%) and a negative predictive value (NPV) of 79.8% (95% CI, 69.6%–87.8%) for prediction of treatment failures.
Interim ct-DNA could serve as an early marker of resistance to targeted therapy. A study of panobinostat in relapsed DLBCL and transformed lymphoma demonstrated that ct-DNA levels at day 15 were predictive of response to this novel targeted therapy.26 These results raise the possibility that ct-DNA could serve as a surrogate endpoint in assessing response to therapy.
In the setting of low tumor burden in DLBCL, methods to measure ct-DNA have limited sensitivities, which may restrict their applicability in a broader patient population. Nevertheless, these studies show that ct-DNA is a valuable tool and may enhance the accuracy of interim PET/CT scans.
1.6 Role of ct-DNA in surveillance in DLBCL
Ct-DNA could potentially play a role in the surveillance of patients with DLBCL. A patient who is in complete remission following therapy for DLBCL undergoes surveillance to detect relapses. Guidelines recommend surveillance with laboratory evaluation, clinical examination, and, occasionally, imaging studies. Role of imaging studies for surveillance is controversial, with increasing evidence recommending against the use of routine surveillance imaging.27 Although CT scans may identify relapses earlier in asymptomatic patients, studies have not been able to demonstrate a significant benefit in survival outcomes. Furthermore, the lack of specificity can lead to unnecessary biopsies. PET/CT scans are associated with a high rate of false positives and are not recommended for surveillance, with a study showing a PPV of only 21%.27, 28
Roschewski et al demonstrated that ct-DNA outperforms CT scans for detecting disease relapse during surveillance.25 Baseline tumor clonal sequences were identified in 126 patients with DLBCL. Following the completion of therapy, patients were monitored with periodic CT scans and ct-DNA levels. In patients who relapsed, ct-DNA was detectable in 88% of patients before relapse. Surveillance ct-DNA had a PPV and a NPV of 88% and 98%, respectively, and identified recurrence with a median range of 3.5 months (0-200) before evidence of clinical disease.
A similar study by Kurtz et al reported similar outcomes.29 Of 75 patients with DLBCL treated with combination chemotherapy, 25 patients had samples available during surveillance, and 5 patients eventually relapsed. The ct-DNA was detectable in the plasma in 3 (60%) patients, a median of 88 days (range 14-162) before clinical relapse, and 2 patients had detectable ct-DNA at the time of relapse, with a specificity of 100%. Imaging with PET scans in the same group of patients yielded a specificity of only 56%.
These studies support the role of ct-DNA as a tool for surveillance. It would be important to know if early detection of relapses translates into improved overall survival.
1.7 Circulating tumor DNA in indolent lymphomas
The utility of ct-DNA is being evaluated in indolent lymphomas such as FL. FL, despite being an indolent lymphoma with good survival outcomes, is an incurable lymphoma. FL has varied clinical presentations, with some patients experiencing prolonged periods of asymptomatic disease managed by observation, while others developing aggressive disease with short relapses. Prognostic scoring systems such as FLIPI (FL International Prognostic Index) use clinical factors to predict survival outcomes. However, tools to identify the biological mechanisms driving the disease process are lacking.
Baseline levels of ct-DNA correlate with prognosis in FL. Sarkozy et al reported high levels of ct-DNA, detected by next-generation sequencing, to be a negative independent predictor of progression-free survival.30 This was a retrospective study aimed to identify the tumor clonotypes of 34 patients. Next-generation sequencing was used to identify ct-DNA at the time of diagnosis. Ct-DNA was noted in 86% of patients, who had a clonotype which correlated to that of the tumor sample. Patients who had ct-DNA levels above the median had a significantly worse progression-free survival than patients with lower ct-DNA levels. Interestingly, the prognostic impact of elevated baseline ct-DNA levels declined in patients who received rituximab maintenance therapy. Thus, ct-DNA levels could potentially serve as a marker in identifying patients who could benefit from maintenance therapy.
A study by Galimberti et al reported baseline undetectable ct-DNA by qPCR was associated with a higher complete response rate and a longer progression-free survival.31 However, these studies have not shown a survival advantage. Further studies will be needed to determine whether ct-DNA levels could be used to guide treatment in FL.
In MCL, molecular remission following first-line therapy is a good prognostic factor. Suitable targets for MRD assessment include clonal IGH rearrangements as well as t(11;14). Detection of ct-DNA during and following induction therapy could help tailor postinduction consolidation therapy based on the risk of subsequent relapse.32, 33
2 LIMITATIONS OF ct-DNA
Ct-DNA is not without drawbacks. Next-generation sequencing techniques are limited by the quality of plasma samples, cross-contamination, and PCR sequencing errors.3 Inflammation and half-life can affect the size of ct-DNA and may impact its quantification. Although ct-DNA can be an effective tool in identifying the presence of residual disease, imaging studies will be needed to determine the location and extent of involvement of the disease. Thus, ct-DNA may supplement imaging studies but is unlikely to replace them.3 Although studies have shown early detection of relapses and treatment failure with ct-DNA, it remains to be seen if early interventions would translate into a survival advantage.
3 CONCLUSION
Ct-DNA is a promising new tool in the era of precision medicine. It is a convenient, noninvasive method for determining the prognosis and response to therapy in patients with lymphomas. Additionally, it may play an important role in the surveillance of both indolent and aggressive lymphomas. Interim ct-DNA measurement, along with imaging studies, has the potential to improve outcomes through response-adapted strategies. The ability to quantitatively assess the levels over time during therapy has the potential for early interventions to potentially improve outcomes in patients where there is suboptimal response to therapy. Ct-DNA has a potential to detect early relapses. Development of biomarkers using ct-DNA can help in therapy selection. DNA analysis can be used to better define mechanisms of resistance and to identify targeted therapies using a precision-based approach.
The challenge now remains as how best to integrate ct-DNA into clinical practice. Although dynamic risk assessments are capable of identifying high-risk patients, interventions in order to improve outcomes are unclear. Randomized prospective studies tailoring therapies according to patients’ risk as determined by the ct-DNA models are needed. Studies establishing the clinical role of ct-DNA in indolent lymphomas are ongoing. Results of these studies may help establish a role of ct-DNA in clinical practice.
With advances being made in methodologies to improve the detection of ct-DNA assays, it is expected that ct-DNA will evolve into an important decision-making tool in lymphomas.
4 CLINICAL IMPLICATIONS
Ct-DNA is a promising new tool in the era of precision medicine. Ct-DNA has shown to be a convenient, noninvasive method of determining the prognosis and response to therapy in patients with lymphomas. With advances being made in methodologies to improve detection of ct-DNA assays, it is expected that ct-DNA will evolve into an important decision-making tool in lymphomas.
ACKNOWLEDGMENT
None.
CONFLICT OF INTEREST
Kallam, A: None; Adusumalli, J: None; Fu, K: None; Armitage, JO: Consultant: Oncology Analytics 6/2014 – ongoing, Partner Therapeutics 9/2018 – ongoing, Samus Therapeutics 3/2017 – ongoing, Ascentage 12/2017 – ongoing, Board of Director's for Tesaro bio, Inc
ETHICS STATEMENT
The authors confirm that the ethical policies of the journal, as noted on the journal's author guidelines page, have been adhered to. No ethical approval was required as this is a review article with no original research data.
