Tim Forshew, Co-founder and Head of Science and Innovation at Inivata shares his thoughts on a recent publication from groups at Dana-Farber, Broad Institute and Boston University that demonstrate the value of personalized assays for detecting ctDNA with high sensitivity.
My thoughts on Parsons et al 2020 – a new study on the value of personalized ctDNA assays for residual disease and recurrence detection
There is a lot of compelling evidence that ctDNA can be used for detecting both residual disease and recurrence across a spectrum of cancers including breast, colon and lung. As ctDNA degrades very rapidly (~1h half-life) it enables an almost real-time assessment of tumor burden but also means that detecting the smallest tumors require exceptional sensitivity. There is a lot of debate on the best way to do this.
Should you track a single mutation (e.g. dPCR), a small number of mutations (e.g. with a fixed-panel NGS) or develop personalized assays to target large numbers of mutations?
A team across the Dana-Farber, Broad Institute and Boston University recently published an excellent article on the value of personalized assays (Parsons, H. et al, Clinical Cancer Research 2020). They combine duplex sequencing and personalized hybrid capture (from SNVs found through exome sequencing) to track large numbers of mutations. They showed with a dilution study that they could detect ctDNA levels well below that possible with dPCR (paper figure 1). When tracking 97 personalized mutations they could typically detect ctDNA at 0.01% VAF. When they looked at 488 variants they got down to 0.001% VAF (or 1/100k). These results mirror the findings of McDonald et al (Science Translational Medicine, 2019) who used a personalized platform called TARDIS. In the Parsons clinical cohort of early stage breast cancer patients, they achieved a median lead time of 18.9 months for recurrence detection (which beat previous dPCR studies).
One of the best examples of approaches that doesn’t require the development of personalized assays in my view is the Capp-Seq assay. It is a nice approach that currently uses a 188kb hybrid capture panel with molecular barcoding (Chaudhuri, A et al. Cancer Discovery 2017). In optimal settings, it can detect ctDNA at very low levels. That said, other than assay personalization they use almost identical methods to Parsons et al (molecular barcode ligation > hybrid capture > sequencing). Therefore ignoring technical optimizations which are happening all the time (ligation efficiency, error suppression etc) the difference between these methods is the number of mutant molecules they can track. Chaudhuri averaged 5 mutations detected pre-treatment (in lung) which they could track and in 7% of patients there were no pre-treatment variants. Parsons on the other hand averaged 57 mutations and never had none (in breast).
In my view, developing an assay that will detect cancer recurrence earlier and for more patients, will always be preferable both to clinicians and pharmaceutical companies as long as the approach is practical, rapid, cost effective and well validated. I think this paper further supports the argument for personalized assays. I am somewhat biased as Inivata has developed such an assay based on our eTAm-Seq platform and the personalized assay approach we first published in 2012 (Forshew et al, STM 2012). What are your thoughts?
- Parsons, H. A. et al. (2020) ‘Sensitive detection of minimal residual disease in patients treated for early-stage breast cancer’, Clinical Cancer Research, p. clincanres.3005.2019. doi: 10.1158/1078-0432.ccr-19-3005.
- McDonald, B. R. et al. (2019) ‘Personalized circulating tumor DNA analysis to detect residual disease after neoadjuvant therapy in breast cancer’, Science Translational Medicine, 11(504), pp. 1–14. doi: 10.1126/scitranslmed.aax7392.
- Chaudhuri, A. A. et al. (2017) ‘Early detection of molecular residual disease in localized lung cancer by circulating tumor DNA profiling’, Cancer Discovery, 7(12), pp. 1394–1403. doi: 10.1158/2159-8290.CD-17-0716.
- Forshew, T. et al. (2012) ‘Noninvasive Identification and Monitoring of Cancer Mutations by Targeted Deep Sequencing of Plasma DNA’, Science Translational Medicine, 4(136), pp. 136ra68–136ra68. doi: 10.1126/scitranslmed.3003726.