As you may know, 3-6 months of oxaliplatin-based adjuvant chemotherapy is standard of care for stage 3 colon cancer patients, however, the absolute benefit for an individual patient is uncertain and they are exposed to toxicity, particularly oxaliplatin-related neuropathy that can be long-lasting. Post-operative circulating tumour DNA has been an emerging powerful biomarker which offers the potential for us to risk adapt patients’ chemo intensity after surgery.

So DYNAMIC-III is really designed based on this concept where we try to test if adjuvant chemotherapy intensity can be tailored according to post-operative ctDNA status. At ESMO this year we presented the data, the primary analysis of the ctDNA negative cohort, really asking the question can treatment de-escalation, so giving less intensive chemotherapy, is it non-inferior to standard of care.

What was the study design?

DYNAMIC-III is a randomised phase II/III study. It enrolled patients with resected stage III colon cancer who are fit for adjuvant chemotherapy. Patients were then randomised 1:1 to ctDNA-informed management or standard of care. Clinicians were to nominate their preferred chemotherapy regimen at baseline and ctDNA testing was done centrally at Johns Hopkins lab using a tumour-informed approach incorporating a targeted colorectal panel and a SaferSeqS MRD assay.

In the ctDNA-informed arm, those with a negative test received treatment de-escalation which involved omitting oxaliplatin or shortening the total duration of treatment. In the standard management arm treatment choices as per clinician’s discretion and they are blinded to the ctDNA results.

What were the results?

We randomised 1002 patients across 66 institutions from Australia, New Zealand and Canada. 73% of the patients tested negative and so when they were randomised to ctDNA-informed or standard management we found the compliance rate was very high. The key difference was the chemo intensity delivered as expected with significant reduction in oxaliplatin use from 89% for standard management down to 35% with ctDNA guidance. Along with this de-escalation we saw a better safety profile with a significant reduction in treatment-related hospitalisation as well as severe grade 3/4 treatment-related toxicity of interest, such as diarrhoea, nausea and mucositis.

At a median follow-up of 47 months the 3-year recurrence free survival was 85.3% with ctDNA guidance and 88.1% with standard of care with an absolute difference of 2.8%. Given the lower bound of the confidence interval of this difference, which is -8%, slightly exceeded our prespecified non-inferiority margin of 7.5%, the criteria for non-inferiority was not formally met in this study. Nonetheless, the difference is small and when we look at clinical risk, for clinical low-risk patients the overall outcomes were very good across both arms, both above 90%, so 91% for ctDNA-guided patients and 93.2% for standard of care with an absolute difference of 2.2%.

In contrast, though, for patients with clinical high risk features, so these are your patients with either T4 or N2 tumour, their outcome was less favourable and we saw a larger difference between the study arms with an absolute difference of 5.8%.

What are the next steps?

Perhaps before I talk about the next steps I maybe just want to mention about potential clinical implications, what does this mean for the patient and for clinical practice. I think it’s important to note that ctDNA negative patients do really well, overall their 3-year recurrence free survival was 87% across both the entire group of patients and that it is a feasible approach with high adherence. We reduce oxaliplatin use and along with that a better safety profile.

The outcomes are very similar although we haven’t quite reached non-inferiority. So it could potentially inform in practice currently the risk/benefit discussion with the patient. So I think this lays the foundation for future work in terms of reducing further chemotherapy using a ctDNA-guided approach.

In terms of future work we are working on increasing the sensitivity of the ctDNA assays. So there are some false negatives which likely explain the small difference that we currently detect in the study. So if we were to improve the sensitivity, and we have shown in our previous DYNAMIC phase II study that by tracking more mutations in the blood we can increase the sensitivity substantially. So if we can do that we can certainly close the gap in the difference that we see in the study. So there is also work ongoing looking at health economic analysis – whether using a ctDNA-guided approach can save costs which would be an important step for regulators.

We’re also looking at digital pathology and other tissue markers to try and see whether we can improve on prognostication for our patients along to complement ctDNA analysis.

There are some plans to take the de-escalation approach further to other disease stages, like resected stage 4 colon cancer and potentially a new trial with a newer assay, looking at really completely omitting chemotherapy. So in this study the majority of patients still receive single agent chemotherapy, reduced oxaliplatin, but they still receive single agent rather than no chemo. So we would love to have one day where I can tell patients they don’t need chemotherapy after surgery.