Serial monitoring of circulating tumor DNA in patients with primary breast cancer for detection of occult metastatic disease
Lao Saal, M.D, Ph.D., Head of the Translational Oncogenomics Unit
The potential for cell free DNA to be used as biomarkers for prognostic, predictive and diagnostic means are great and a fantastic recent publication by Saal et al., shows the potential use and clinical applications of circulating tumor DNA (ctDNA) which can be found here. The findings establish the rationale for larger validation studies in early breast cancer to evaluate ctDNA as a monitoring tool for early metastasis detection, therapy modification, and to aid in avoidance of overtreatment.
The headline figures are summarized below:
- ctDNA‐based detection preceded clinical detection of metastasis in 86% of patients with an average lead time of 11 months
- ctDNA quantity was predictive of poor survival.
- ctDNA analysis can discriminate patients with eventual metastasis from those with long‐term disease‐free survival with 93% sensitivity and 100% specificity
- No ctDNA could be detected at any time‐point after surgery for patients with long‐term disease‐free survival.
I caught up with Lao Saal, M.D, Ph.D., Head of the Translational Oncogenomics Unit and asked a few questions:
Who is the Translational Oncogenomics Unit and what motivated you to conduct your study?
I lead the Translational Oncogenomics Unit, a research group within the Division of Oncology and Pathology, Lund University Cancer Center. At present, there are 8 members in the group: Eleonor Olsson, Christof Winter, Yilun Chen, Christian Brueffer, Tony George, Robert Rigo, and Marina Villamor. On the more basic side, we focus on PTEN/PI3K signaling in cancer, and on the applied side, we focus on development of genomics-based diagnostics, prognostics, and treatment-predictive biomarkers for clinical use. You can read more on the group website: http://www.med.lu.se/saalgroup
The observation that tumor-derived DNA can be found in the blood circulation of cancer patients was made decades ago, however it has not been until recently that the technologies matured and it became feasible to study circulating tumor DNA (ctDNA) in a robust way. For example, using advances in next-generation sequencing and digital PCR, a number of excellent proof-of-concept papers came out some years ago from groups at Johns Hopkins and the Wellcome Trust Sanger Institute, showing the promise of ctDNA analysis on an individual basis. We also thought that ctDNA had a lot of potential as a biomarker, given that it is non-invasive and amenable to all the molecular methods we have now for working with DNA. We wanted to evaluate ctDNA as a biomarker for early metastasis detection in breast cancer patients, and through our collaboration with Helena Jernström in our department, we had access to an excellent collection of early-stage breast cancer patients with serially-collected blood plasma samples and long clinical follow-up.
We had many plasma samples, so direct sequencing of ctDNA would have been an expensive undertaking, and even with exome sequencing, breast cancer has only a few commonly mutated genes and then a long tail. Since chromosomal rearrangements are inherently tumor-specific and amenable to PCR-based assays, we focused on these rather than point mutations. My PhD student Eleonor (who just defended her thesis last month) and my bioinformatics postdoc Christof were the first authors of the study. We performed low-pass whole-genome sequencing of the primary tumors and utilized the chromosomal rearrangements as tumor-specific fingerprints that we monitored in the blood samples using personalized droplet digital PCR assays.
What were the key findings from your study
The main findings from our study are that
1) ctDNA monitoring is an accurate method for detection of breast cancer metastasis in patients diagnosed with primary breast cancer;
2) ctDNA monitoring detected the breast cancer metastases earlier than symptoms- and imaging-based diagnosis for most patients, by an average of 11 months (and up to 36 months); and
3) the amount of ctDNA was a quantitative predictor for metastasis and death with odds ratios of 2 and 1.3, respectively, for each doubling of the ctDNA level.
Any surprising findings?
We were rather limited in the volume of plasma available to us for this study (0.5 ml). Despite this, we were quite pleased that we could get robust signals. Nevertheless, I do think our sensitivity was affected by having only 0.5 ml plasma per timepoint: there likely is some ctDNA present at even earlier timepoints after surgery in patients who go on to have metastasis, but we were unable to detect it because we didn’t have sufficient input. One thing that was interesting (although our patient series is small so one cannot make any definitive conclusions) is that there didn’t appear to be any relationship between the number of chromosomal rearrangements present in the primary tumor and the patient’s outcome. Another interesting anecdote is that, for the one patient with bilateral breast cancer, it was actually the tumor with the more favorable prognostic factors that gave rise to the eventual metastastic disease. This illustrates the difficulty in predicting who will have recurrent disease, and why ctDNA monitoring could be useful.
Overall recommendations based on your conclusion
We have provided proof-of-concept that ctDNA monitoring may have potential uses in the early breast cancer setting. However, there isn’t enough evidence yet to make any type of medical recommendation. From a research perspective, I would recommend that large volumes of plasma be collected when possible since the sensitivity of any method will generally be related linearly to the amount of plasma that can be analyzed. Droplet digital PCR is an excellent method for quantification of ctDNA, particularly for personalized analysis of chromosomal rearrangements. Although we didn’t look at point mutations in our EMBO Molecular Medicine publication, we do have technological innovations for detection of hotspot mutations that drastically reduces the problem of false-positives (patent pending). So we definitely see that both chromosomal rearrangements and gene mutations can be excellent biomarkers to monitor non-invasively using blood samples.
What is next for you/your laboratory?
Clearly, additional research is necessary to further delineate the utility of ctDNA in monitoring cancer. We have started prospective and retrospective follow-up studies. We are also looking at other cancer forms, and are very interested in collaborating with other researchers and pharma and biotech companies to do studies for example evaluating ctDNA as a marker of response to therapy, as a companion diagnostic, and for detection of minimal residual disease and early detection of metastasis.