Circulating tumour DNA can guide adjuvant immunotherapy for mismatch repair-deficient tumours
Dr Yelena Yuriy Janjigian - Memorial Sloan Kettering Cancer Center, New York, USA
At AACR this year we’re presenting circulating tumour DNA status to direct adjuvant immunotherapy for mismatch repair deficient tumours. This is an investigator-initiated study that we conducted at Memorial Sloan Kettering across multiple disease management teams and multidisciplinary involvement was very important because it’s a study to elucidate the role of ctDNA to help guide adjuvant therapy.
We know that ctDNA is a real-time biomarker for minimal residual disease and cleared within hours after a complete resection and a definitive therapy and that persistent or recurrent post-op ctDNA strongly predicts a risk of cancer recurrence.
In mismatch repair deficiency tumours the role of adjuvant therapy with immunotherapy has not been well characterised because those tumours are cured with surgery alone and adjuvant chemotherapy offers limited benefit. We know that immunotherapy is highly effective in MMRd tumours in the metastatic and neoadjuvant setting so there have been no studies that evaluated ctDNA-guided adjuvant therapy in MMRd tumours.
Our hypothesis was that ctDNA detection post-surgery signals residual disease in MMRd tumours and may identify patients who require adjuvant immunotherapy. So our study looked at patients irrespective of tumour origin who had mismatch repair deficient tumour and based on tumour normal DNA sequencing after surgery and completion of standard chemotherapy, after the patient was told by their doctor that they’re finished and they’re cured we had a tumour-informed ctDNA assay performed using MSK-ACCESS. If the patient was ctDNA positive, if they were eligible for the study, we then obtained a CT scan to make sure the tumour has not recurred and if the patient was without evidence of disease then adjuvant pembrolizumab was given for ctDNA positive for no M1 disease on imaging. If they were ctDNA positive but they had already radiographic recurrence they weren’t eligible to receive adjuvant pembrolizumab. If they were ctDNA negative they were followed for recurrence and serial ctDNA in the observational cohort. So every patient across these cohorts had surveillance every three months with imaging and clinical follow-up and ctDNA.
We actually first designed this study as a placebo-controlled study but after having the study open for over a year only one patient consented and this was felt to be not feasible so we closed the placebo arm and enrolment was completed for December 2023 in a single-arm study.
For ctDNA detection we used MSK-ACCESS which is a custom hybrid capture ctDNA assay that targets 129 genes and incorporates tumour-informed analysis and de novo mutation calling so everybody had MSK-ACCESS but also MSK-IMPACT which is the tumour analysis. The mutations across the tumour and the ctDNA were compared and if they had at least one positive match they were defined as ctDNA positive. This was the match between blood and tumour and ruling out germline alterations.
So we screened 303 patients, ctDNA tested 184 and 174 were eligible ctDNA tested. Then of ctDNA positive patients that went on to receive adjuvant pembrolizumab, there were 13; ctDNA positive patients that had M1 disease before adjuvant PD-1 was six and then ctDNA negative, no adjuvant pembrolizumab, 152 patients.
These were a relatively high-risk population with node positivity seen in 40% at pathologic surgery. As I mentioned, 84% of patients had ctDNA done by MSK-ACCESS, we also had 16% who had outside CLEAR-approved assays. These were non-germline patients, mostly 72% did not have Lynch and the median tumour mutational burden was 50.
Looking across the different diseases, as I mentioned this was an irrespective of tumour origin study, the prominent population of patients we screened were colorectal, endometrial and gastrointestinal upper GI tumours, gastroesophageal cancers, and 9% of colorectal tumours that were screened were ctDNA positive while 35% of gastroesophageal cancers were ctDNA positive, again highlighting the difference in the tumour presentation and stage.
Overall, 11% of MSI high patients were ctDNA positive after completion of standard surgery and chemo treatment. All patients had tumour sequencing by MSK-IMPACT with matched ctDNA tumour match of six mutations with a range of 1-12.
So, looking at recurrence-free survival, as expected ctDNA negative patients, most of them are cured with no additional therapy and 94% recurrence-free survival at two years. ctDNA positive patients that recurred before anti-PD-1 therapy, all of them recurred within a year and there was a cohort of ctDNA positive adjuvant anti-PD-1 therapy, which is our primary therapeutic cohort, where recurrence-free survival at two years was 64%. So 94% recurrence-free survival for ctDNA negative patients, 0% for ctDNA positive patients at two years that did not receive anti-PD-1 adjuvant therapy, and 64% recurrence-free survival at two years for ctDNA positive treated patients with adjuvant PD-1 therapy.
Our primary endpoint was rate of ctDNA clearance at six months and looking at six months ctDNA clearance 85% of patients cleared ctDNA at six months which exceeded our threshold for success for this study, we defined it as 40% patients cleared ctDNA at six months. The median time to clearance was 3.9 months.
Despite clearing ctDNA at six months and 100% cleared at nine months, radiographic recurrence was still seen in 5 out of 13 patients with 38% of patients recurring overall. So ctDNA patients treated with pembrolizumab had a numerically similar time to recurrence as ctDNA negative patients who did not receive adjuvant therapy, suggesting the immunotherapy may suppress or delay outgrowth of residual clones. The time to ctDNA positivity to recurrence, it was approximately 20 months.
What is the significance of these results and what is next for this study?
In conclusion, our study showed that 94% of MSI-high ctDNA-negative patients remain disease free at two years. 11% of MSI-high patients had detectable ctDNA after standard therapy and 85% of ctDNA-positive patients cleared ctDNA within six months of adjuvant therapy. What’s important to note is that 38% of patients recurred despite ctDNA clearance at six months and so additional interventions are urgently needed for this high-risk group. Bespoke ctDNA assays may offer modest lead-time gains in MSI tumours but really what we need to focus on is more effective treatment strategies.
This study provides the first prospective clinical evidence that immunotherapy can clear residual disease before clinical recurrence, highlighting ctDNA’s dynamic tool to guide treatment in early stage cancers.
