Current clinical characteristics and demographics are not sufficient to capture aggressive disease in clinical trials of newly diagnosed DLBCL. Novel tools, such as measurement of tumor burden via ctDNA, are needed.

TRANSCRIPT

This JCO Podcast provides observations and commentary on the JCO article “Short Diagnosis-to-Treatment Interval is Associated with Higher Circulating Tumor DNA Levels in Diffuse Large B-Cell Lymphoma” by Alig et al. My name is Matthew Maurer, and I am a statistician at the Mayo Clinic in Rochester, MN. My oncologic specialty is lymphoid malignancies. I have no relevant conflicts to disclose.

The impact of any clinical research critically depends upon participating subjects being representative of the study population afflicted with the disease of interest and research efficiency is markedly enhanced when cohorts can be compared across studies. In newly diagnosed diffuse large B-cell lymphoma (DLBCL), there is a standard group of clinical variables that is typically captured and reported across studies. The International Prognostic Index, or IPI has been utilized as a prognostic model in aggressive lymphoma for nearly 30 years and remains relevant today. The IPI and its components, which consist of age, performance status, LDH, stage, and number of extranodal sites, provide the clinical characteristic backbone for trial eligibility and defining high risk disease in frontline DLBCL trials. Assessment of genomic features of the tumor using pathological techniques such as IHC and FISH can further identify high risk subsets of patients. However, despite these well-tested clinical tools to measure the aggressiveness of the disease, there remains significant heterogeneity in patient presentation and outcomes.

Clinically, some patients will present with a real or perceived clinical urgency to initiate therapy as soon as possible. This clinical urgency often precludes these patients from enrolling on frontline trials. My colleagues from the University of Iowa / Mayo Clinic Lymphoma SPORE and I explored this phenomenon in our observational lymphoma cohort study by evaluating the time between a patient’s diagnostic biopsy and their initiation of chemotherapy. We found that this simple measure of diagnosis to treatment interval, or DTI, was highly informative in the setting of newly diagnosed DLBCL. Patients with a shorter DTI were more likely to be symptomatic, have advanced stage disease, poor performance status, and elevated LDH. Further, patients with a short DTI had significantly inferior outcomes, even after accounting for the standard clinical details of the IPI. These results were validated in a cohort of clinical trial patients from the French Lymphoma Study Association as well as subsequent studies. These data suggest that clinicians are managing patients with aggressive disease more urgently, and our current set of clinical variables is not sufficient to describe and compare patients across studies.

Despite its simplicity and retrospective prognostic ability, DTI lacks specificity as a clinical characteristic for future studies. It can be easily influenced in an individual patient by numerous aspects unrelated to disease biology, such as physician preference or a patient’s available access to health care resources. In additional, the typical DTI can vary widely across health systems or institutions. In the paper accompanying this podcast, Alig and colleagues examine the relationship between DTI, conventional risk factors, circulating tumor DNA, and clinical outcomes. A strong association was observed between shorter DTI and increasing tumor burden as measured by baseline metabolic tumor volume and circulating tumor DNA. This is a key biologic confirmation that treatment urgency is directly related to disease biology. Importantly, the authors also showed that ctDNA was a highly informative variable in regards to prognosis in their dataset. ctDNA was an independent prognostic variable after adjusting for the IPI in Cox models for event-free and overall survival, and the prognostic ability of ctDNA was far superior to DTI in univariate models.

The association between DTI and disease biology has direct implications for clinical trials in frontline DLBCL. Clinical trials that did not adequately enroll patients in need of urgent therapy were likely biased towards enrollment of patients with lower tumor burden. In particular, single arm trials of novel therapies are particularly at risk, as these results are often evaluated in light of conventional clinical characteristics and compared to previous studies and/or clinical experience. This may also have contributed to the better than expected outcomes on the control arms in recent randomized trials of newly diagnosed DLBCL. Evaluation of tumor burden using ctDNA or metabolic tumor volume should allow us to better understand the impact of a study’s design on patient enrollment.

The amount of tumor is a long-standing prognostic feature for newly diagnosed DLBCL. This has standardly been measured by broad clinical features such as stage, number of extranodal sites, and bulky disease. As the retrospective studies on DTI have shown, however, these features are not sufficient. Novel tumor burden measures like circulating tumor DNA are needed. ctDNA is not without its drawbacks in frontline DLBCL. It is not clinically available for real-time assessment and significant work remains to be done to make it a routinely available and standardized biomarker, which includes independent validation of the results reported by Alig and colleagues. However, as we continue to develop and test new treatment strategies for DLBCL, we must also develop and test novel prognostic and predictive tools. Alig and colleagues have shown us that ctDNA has the potential to identify features of aggressive disease that our current tools do not. Capturing these features more precisely is vital for us to understand and interpret clinical trial results, as well as ensure future trial designs are enrolling the study’s intended patient population. As part of these efforts, clinical trials in newly diagnosed DLBCL should collect and store the necessary biospecimens to evaluate ctDNA in anticipation of a new generation of standard clinical characteristics.

This concludes this JCO Podcast. Thank you for listening.

ctDNA in DLBCL: A Blueprint for Better Clinical Trials? 07:00 Share

Journal of Clinical Oncology (JCO) Podcast

Shownotes Transcript

ctDNA in DLBCL: A Blueprint for Better Clinical Trials?