The INTERCEPT study looks at ctDNA as a part of routine care after curative intent procedures for colorectal cancer from stage 1-4. What’s unique is that it’s a big proportion of stage 4 patients in there, where it’s almost half of the patients. Here we wanted to look at ctDNA biology and the behaviour of ctDNA in depth because ctDNA clearances are being used more and more as a surrogate endpoint in many clinical trials. We know that some of the patients have a spontaneous clearance so we wanted to look at that proportion and how common that is and also how durable those responses are.

This was done in the INTERCEPT study and, as I said, it’s real-world material from MD Anderson where we followed them after the surgery. 1,300 patients included in this current analysis, 47% stage 4 and the ctDNA was measured after the surgery and then roughly every three months during follow-up. Clearances were seen in 5% of all patients after the event and then if we only consider clearances after adjuvant therapy then it was in 3% of the patients. Those that had clearances either on one or more, we did a distinction of if it was just one sample or if it was two or more subsequent tests, and there we saw that both of those groups had a better DFS and in the stage 1-3 patients, especially the ones with clearances on two or more tests, had a much better DFS.

Then if we go on to looking at the spontaneous clearances, those were seen in 16 out of 403 patients, so roughly 4.2% of the patients. Half of those were durable, meaning two or more tests, and 1.7% out of these 403 didn’t have any recurrences during the follow-up. The median duration of the clearance was 11.2 months.

If we look at the mean tumour molecules per millilitre value among these patients, those were usually in the lower range with a median of 0.06 but a handful also had higher MTM values but generally at the lower end and then going to negative and then some of them coming up again.

The main conclusions would be that ctDNA clearance is connected with a better DFS and if we look specifically at adjuvant therapy, I didn’t mention that earlier but there it was a quarter of the patients that cleared their ctDNA with adjuvant therapy. DFS was better in the ones that had clearances and the clearance rate that are spontaneous, that’s quite low – only 4%. That’s a useful value to have as a null hypothesis when you design MRD trials to know how much higher you should get from that if you want to use ctDNA clearance as an endpoint.

Another conclusion would be that ctDNA clearance is a useful endpoint, especially as you can get a fast readout. Of course, you need to verify it as well with DFS or some other more established endpoint. But the readout, as I said, is going to be much quicker than with traditional endpoints.

What could be the implications of these findings?

The main one is that we have this spontaneous clearance proportion of around 4% and that needs to be kept in mind as a baseline if you want to use ctDNA clearance as an endpoint for a trial. You want to reach higher than that and show a meaningful difference to that 4% of spontaneous clearances.