Total Circulating Cell-Free DNA as a Prognostic Biomarker in Metastatic Colorectal Cancer Before First-Line Oxaliplatin-Based Chemotherapy

J. Hamfjord; T. K. Guren; O. Dajani; J. S. Johansen; B. Glimelius; H. Sorbye; P. Pfeiffer; O. C. Lingjærde; K. M. Tveit; E. H. Kure; N. Pallisgaard; K.-L. G. Spindler


Ann Oncol. 2019;30(7):1088-1095. 

In This Article

Abstract and Introduction


Background: Metastatic colorectal cancer (mCRC) is a heterogeneous disease where prognosis is dependent both on tumor biology and host factors. Total circulating cell-free DNA (cfDNA) has shown to harbor prognostic information in mCRC, although less is known about the biological correlates of cfDNA levels in this patient group. The primary objective was to evaluate the prognostic value of pretreatment cfDNA in patients receiving the first-line oxaliplatin-based chemotherapy for mCRC, by using a predefined upper limit of normal (ULN) from a cohort of presumed healthy individuals. The secondary objective was to model cfDNA levels as a function of predefined tumor and host factors.

Patients and methods: This was a retrospective post hoc study based on a prospective multicenter phase III trial, the NORDIC-VII study. DNA was purified from 547 plasma samples and cfDNA quantified by a droplet digital PCR assay (B2M, PPIA) with controls for lymphocyte contamination. Main clinical end point was overall survival (OS).

Results: cfDNA was quantified in 493 patients, 54 were excluded mainly due to lymphocyte contamination. Median cfDNA level was 7673 alleles/ml (1050–1 645 000) for B2M and 5959 alleles/ml (555–854 167) for PPIA. High cfDNA levels were associated with impaired outcome; median OS of 16.6 months for levels above ULN and 25.9 months for levels below ULN (hazard ratio = 1.83, 95% confidence interval 1.51–2.21, P < 0.001). The result was confirmed in multivariate OS analysis adjusting for established clinicopathological characteristics. A linear regression model predicted cfDNA levels from sum of longest tumor diameters by RECIST, the presence of liver metastases and systemic inflammatory response as measured by interleukin 6 (F(6, 357) = 62.7, P < 0.001).

Conclusion: cfDNA holds promise as a minimally invasive and clinically relevant prognostic biomarker in mCRC before initiating first-line oxaliplatin-based chemotherapy and may be a complex entity associated with tumor burden, liver metastases and systemic inflammatory response.


Colorectal cancer is the third most common cancer worldwide, with over 1.8 million new cases and 881 000 deaths every year.[1] Combination chemotherapy is the preferred first-line treatment of metastatic colorectal cancer (mCRC).[2] Adding anti-epidermal growth factor receptor therapy may provide further clinical benefit in RAS wild-type[3] and in particular left-sided RAS wild-type cancers.[4] mCRC is a heterogeneous disease where prognosis depends both on tumor biology and host factors.[5] There is a need for reliable biomarkers that can aid in clinical decision making throughout the patient's disease trajectory; selecting patients for optimal oncological and surgical strategies.

Small fragments of total circulating cell-free DNA (cfDNA) can be detected in the blood stream of humans in health and disease.[6,7] cfDNA originate primarily from cell turnover representing cells dying from apoptosis and necrosis[8] and can readily be detected in patients with advanced cancers[7,9] and diseases driven by inflammatory processes.[10,11] cfDNA has a short biological half-life and sampling is minimally invasive, making it an attractive biomarker at multiple decision points.

A negative prognostic significance of elevated cfDNA in patients with mCRC has been described.[12,13] The prognostic role has mainly been investigated in patients before second and subsequent lines of chemotherapy. It is uncertain if results are transferable to a first-line setting. Most studies lack external validation and no reference levels have been established.

Details regarding the release and possible biological correlates of cfDNA in mCRC still remain unclear. There is an association between tumor burden and cfDNA in human xenograft models.[14] In what way other tumor characteristics influence cfDNA levels is uncertain. Since cfDNA is released from both malignant and non-malignant cells, we hypothesized that additional host factors including systemic inflammatory response may further attenuate cfDNA levels.

The primary objective was to evaluate the prognostic value of cfDNA levels in plasma from mCRC patients before initiating first-line oxaliplatin-based chemotherapy, by using a predefined upper limit of normal (ULN) from a cohort of presumed healthy individuals. cfDNA was assessed alone and in combination with established prognostic clinicopathological and biochemical characteristics used in daily clinical practice.[15] The secondary objective was to model cfDNA levels in mCRC patients as a function of predefined tumor and host factors.