High Pretreatment ctDNA Levels Correlate With Short Diagnosis-to-Treatment Intervals in DLBCL


Pretreatment circulating tumor DNA levels were found to more objectively measure disease burden compared with diagnosis-to-treatment intervals in patients with diffuse large B-cell lymphoma and can therefore be used to quantify and mitigate selection bias in prospective DLBCL clinical trials.

Stefan Alig, MD

Stefan Alig, MD

Pretreatment circulating tumor DNA (ctDNA) levels were found to more objectively measure disease burden compared with diagnosis-to-treatment intervals (DTIs) in patients with diffuse large B-cell lymphoma (DLBCL) and can therefore be used to quantify and mitigate selection bias in prospective DLBCL clinical trials, according to findings from a study published in the Journal of Clinical Oncology.

Both DTI and pretreatment ctDNA levels were found to correlate with tumor burden. Notably, short DTI was associated with higher levels of ctDNA. Patients with a median DTI of 14 days or less had a median ctDNA level of 2.9 log hGE/mL compared with 2.1 log hGE/mL in patients with a median DTI of more than 14 days (P < .001; RS = -.33).

Findings from a univariate logistic regression model showed that ctDNA and International Prognostic Index (IPI) individually predicted for short DTI (odds ratio [OR], 1.9; 95% CI, 1.5-2.5; P < .001). Notably, ctDNA levels remained statistically significant when evaluated in a multivariable analysis (OR, 1.7; 95% CI, 1.3-2.3; P < .001).

“Both DTI and ctDNA levels were found to be associated with conventional measures of tumor burden such as stage, [IPI], and total metabolic tumor volume [TMTV]. ctDNA levels predicted short DTI better than the IPI and were independently prognostic of event-free [survival (EFS)] and overall survival [OS] in multivariable models also including the IPI and DTI,” wrote lead study author Stefan Alig, MD, a postdoctoral research fellow at Stanford Medicine, and co-authors in the study publication.

DTI is measured as the time between pathologic diagnosis and initiation of chemoimmunotherapy. Prior studies have shown that DTI is associated with prognostic clinical factors and outcomes for patients with DLBCL, with shorter DTI appearing to confer more adverse clinical risk factors and worse EFS.

“Intriguingly, the prognostic impact of DTI was independent of the IPI, suggesting that widely applied prognostic scores do not adequately reflect disease aggressiveness and other factors considered for clinical decision making,” Alig and co-authors wrote.

Although still investigational, pretreatment ctDNA levels are associated with tumor burden and can potentially predict treatment outcomes.

As such, DTI, pretreatment ctDNA levels, and potentially other factors may be key to improving patient selection for clinical trials. Currently, patients with short DTI could be excluded from clinical trials on the basis that they cannot delay therapy, and therefore, cannot complete the screening or consent processes required to enroll.

Overall, 267 patients with initial diagnoses of DLBCL per the 2008 World Health Organization criteria, including T-cell/histiocyte rich large B-cell lymphoma and antecedent low-grade lymphoma with histologic transformation, were included on the study. Patients were a median age of 60 years (range, 17-84). The majority of patients (n = 171; 64%) had stage III or IV disease and a low IPI (n = 105; 39%).

Patients were accrued from within the PETAL trial in Germany (n = 126; 47%), and treatment centers in Stanford, California (n = 46; 17%), Novara, Italy (n = 36; 13%), Dijon, France (n = 23; 9%), The University of Texas MD Anderson Cancer Center in Texas (n = 19; 7%), and the National Cancer Institute in Maryland (n = 17; 6%).

All patients were previously treated with chemoimmunotherapy regimens, including rituximab (Rituxan) plus cyclophosphamide, doxorubicin, vincristine, and prednisone (R-CHOP; 75%), dose-adjusted etoposide plus prednisone, vincristine, cyclophosphamide, doxorubicin, and rituximab (19%), and others (6%).

The median DTI was 21 days in the patient cohort (range, 0-154 days) and the distribution of DTI varied significantly by treatment center (P = .002). Conversely, no significant difference in ctDNA levels was observed between treatment centers despite blood samples having been taken at varying timepoints relative to therapy initiation (P = .001). Notably, 80% of pretreatment blood samples evaluated were collected within 7 days of treatment initiation; however, no significant correlation between DTI and the sample-to-treatment interval was observed (RP = .012).

Additional findings revealed that patients with more advanced-stage DLBCL had shorter DTI vs patients with limited-stage disease (median DTI, 18 days vs 26 days, respectively; P < .001). Patients with higher IPI also had shorter DTI compared with patients with longer IPI (P < .001). The median DTI was 25 days in patients with low IPI, 22 days in patients with low-intermediate IPI, 21 days in patients with high-intermediate IPI, and 13 days in patients with high IPI.

Additionally, higher TMTV strongly correlated with shorter DTI (RS = -0.37; P < .001).

Total cell free DNA levels demonstrated less significant correlation with tumor burden (RS = .25; P = .002).

Patients with advanced-stage DLBCL had significantly higher median levels of ctDNA (2.5 log hGE/mL) compared with individuals with lower median levels of ctDNA (1.8 log hGE/mL; P < .001).

Additionally, IPI strongly correlated with ctDNA levels. The median ctDNA levels measured in log hGE/mL were 1.8 with low IPI, 2.2 with low-intermediate IPI, 2.5 with high-intermediate IPI, and 3.4 with high IPI (P < .001).

Notably, ctDNA levels significantly correlated with TMTV (RS = .06; P < .001).

No significant difference in median DTI was observed between Germinal center B-cell and non–Germinal center B-cell subtypes or mutations in individual genes. Similarly, no significant difference between ctDNA levels and cell-of-origin subgroups or individual gene mutations was observed.

Additional findings confirmed that short DTI reflects inferior EFS (P = .007) and OS (P = .04). High levels of pretreatment ctDNA also reflected shorter EFS (P < .001) and OS (P = .002). Moreover, a univariable Cox regression demonstrated that ctDNA levels (HR, 1.7; 95% CI, 1.4-2.1; P < .001), IPI (HR, 1.3; 95% CI, 1.1-1.6; P < .001), and DTI (HR, 1.2; 95% CI, 1.1-1.4; P = .004) were individually prognostic for EFS; however, a multivariable analysis showed that just ctDNA levels significantly correlated with EFS (HR, 1.5; 95% CI, 1.2-2.0; P < .001). Similar findings were revealed in analyses for OS.

Notably, the prognostic potential of ctDNA levels were reflected for EFS even when TMTV was added to the multivariable model in a subset of patients.

An additional analysis was conducted showing that DTI and ctDNA levels were associated with tumor burden in a subset of patients with de novo DLBCL who were treated with R-CHOP.

“Collectively, our data suggest that ctDNA more objectively measures disease burden than DTI and could therefore have immediate utility to quantify potential selection biases in prospective DLBCL clinical trials,” concluded Alig and co-authors.


  1. Alig S, Macaulay CW, Kurtz DM, et al. Short diagnosis-to-treatment interval is associated with higher circulating tumor DNA levels in diffuse large B-cell lymphoma. J Clin Oncol. Published online April 28, 2021. Accessed May 10, 2021. doi:10.1200/JCO.20.02573
Related Videos
Rahul Banerjee, MD, FACP, assistant professor, Clinical Research Division, Fred Hutchinson Cancer Center; assistant professor, Division of Hematology and Oncology, University of Washington
Julie M. Vose, MD, MBA
Lori A. Leslie, MD
David J. Andorsky, MD
Michael R. Cook, MD
Peter Riedell, MD
Paolo Strati, MD
Stephen M. Ansell, MD, PhD
Jason R. Westin, MD
Loretta J. Nastoupil, MD