GISTT RACE · Research Alliance for CurE

Is a cure realistic?

Patients with localized GIST can be cured by surgery. That is a fact. But many with localized GIST relapse despite surgery, and most patients with metastatic GIST cannot be cured as of today. Still, there is evidence that medical treatment may achieve one of the three forms of cure. We would like you to understand what our reference points are for the optimistic goal of cure, and why it is not unrealistic.

Surgery alone already cures many patients

60%still free of GIST 15 years after surgery alone

These are records of 2,560 patients who had their GIST removed, pooled from ten population-based series. About six in ten patients with the diagnosis GIST never had the disease come back. Risk scores can help assess in advance who is likely to be cured by the operation.

Joensuu H, et al. Lancet Oncol 2012. 15-year relapse-free survival 59.9%.

On the timeline: what surgery alone achieves

Many of these patients had small or low-risk tumors. For high-risk and spread disease, surgery alone is not enough. And certainly, even in this cohort, 40.1% relapse or metastatic disease is no reason for inaction.

Three years of treatment with imatinib after surgery: (slightly) more patients free of GIST, without any drug

In a randomized trial of 397 patients at high risk of relapse, half took imatinib for one year after surgery, half for three. Then treatment stopped. The charts show the efficacy population: patients whose tumor was confirmed as GIST, without the patients who had metastases removed at surgery.

Alive and relapse-free
0204060801000246810Years after randomization3 years1 yearenlarged below
The tail of the curves, enlarged
4045505560891011Years after randomization3 years1 yearfew patients left10 years52.4%44.2%8.2 points

At 10 years, 52.4% of patients in the three-year group were alive and had never relapsed, against 44.2% in the one-year group: a gap of 8.2 points. These patients had taken no drug for seven years.

What the gap means, if we read it as cure

Assume that one year of imatinib does not cure anyone who still has microscopic GIST cells after surgery; it only delays their relapse. Then, out of 100 patients:

44 were cured by surgery. They did not need imatinib at all; with better risk prediction, they could have been spared it. 8 were cured by the two extra years of imatinib. 48 relapsed despite three years of imatinib (or died).

Of the 56 patients who would have relapsed after one year, three years of imatinib cured about 8: one in seven (14.7%). Put differently, three years lowered the 10-year risk of relapse from 55.8% to 47.6%, a relative reduction of about 15%. About 12 patients take imatinib for two extra years for each additional patient cured.

The evidence of cure: Imatinib alone, a drug that does nothing but block KIT, cured a small subset of patients who still had GIST cells after surgery: a “cure-cure”. That is the good news, and it is the precedent: KIT inhibition alone can, in principle, cure.

Alive
0204060801000246810Years after randomization3 years1 year10 years81.6%66.8%14.8 points

The effect on survival is larger, and it lasts. At 10 years, 81.6% of patients in the three-year group were alive, against 66.8% in the one-year group. The relapse curves almost meet; the survival curves do not. Three years of imatinib clearly prolonged life for some patients who relapsed, and we do not know why. One possibility: in the one-year group, relapse came earlier, so these patients lived longer with visible disease, and with more time for resistance to develop. If follow-up after stopping were done at short intervals, and relapses were removed early by surgery, the difference might shrink. This is speculation. For patients old enough, it amounts to something like a “sorta-cure”.

Joensuu H, et al. JAMA Oncol 2020 (SSG XVIII, 10-year follow-up). Efficacy population: relapse-free survival HR 0.70 (95% CI 0.52–0.94), overall survival HR 0.50 (95% CI 0.32–0.80). All 397 patients (intention to treat): relapse-free survival at 10 years 52.5% vs 41.8%, overall survival 79.0% vs 65.3%. Curves redrawn from the published figure; their shape is approximate, the labeled values are the published ones.

On the timeline: ten years on, the benefit holds

Open questions: It would be interesting to understand WHICH patients were the ones that were cured. That is no easy question to answer, because we don’t know which patients had been destined to relapse. And while the trial was reasonably big, it is relatively small for questions like this.

GISTT-RACE projects: It could be helpful to collect real-world data for patients who have received 3 years of imatinib, and see how the distribution of genotypes differs between the pre-treatment cohort and the patients who have relapsed. That could generate hypotheses on whether certain genotypes have a greater benefit (and there are substudies of SSG XVIII that address this question).

Caveats: With a trial this size, there is a small chance that part of the 8.2-point gap reflects an uneven distribution of risk factors that the patient characteristics do not capture. At 5 years the gap was 18 points; most of the early benefit was delayed relapse, not cure. After 10 years few patients remain under observation, and in the last steps the curves move closer still (about 3 to 4 points apart at 11 years), so part of the gap may yet turn out to be a late relapse. Relapse-free survival also counts deaths from other causes, so the 48 include a few patients who died without relapse. And the trial included genotypes that respond poorly or not at all to imatinib, which dilutes the effect in sensitive tumors.

Even with spread disease, some patients live for decades

Of 424 patients with advanced GIST who started imatinib in a French trial in the early 2000s, followed for more than 18 years:

Patients whose tumors disappeared completely on the drug, or whose remaining disease could be removed completely by surgery, lived longest (median survival around 14 years in both groups).

Alive after the start of imatinib. Blay JY, et al. Ann Oncol 2025 (BFR14, 20-year follow-up).

On the timeline: should treatment stop?

Long survival is not the same as cure: many of these patients still take a drug. Who among them is free of disease is not known. That is exactly what a curative protocol has to find out.

Some patients with spread disease stop imatinib and never relapse

In the French BFR14 trial, patients doing well on imatinib were randomly assigned to stop or continue. For most who stopped, the disease came back. Five never progressed and never restarted the drug:

Years without treatment and without progression; ▸ = still progression-free at last follow-up. Four of the five had a complete response to imatinib alone, without surgery.

Blay JY, et al. Lancet Oncol 2024 (BFR14, long-term follow-up of the interruption groups).

On the timeline: stopping imatinib

Five patients is a small number, and "no progression on CT" cannot rule out a few remaining cells. For the group as a whole, stopping is not recommended. Finding out in advance who these patients are is one of the questions below.

No visible disease left: almost half stay free without treatment

A Norwegian-led trial stopped imatinib in 12 patients with only a few metastases, after at least five years of treatment, once no tumor was visible on scans. In 11 of them, the metastases had been removed by surgery or heat ablation.

5 still progression-free off imatinib, 3–5 years so far 7 relapsed, after a median of 10 months

All six patients who restarted imatinib after relapse responded again, and all patients were alive.

Hompland I, et al. Acta Oncol 2024. Open access (CC BY 4.0).

On the timeline: stopping after complete remission

Twelve carefully selected patients and three to five years of follow-up. Most relapses were in the abdominal lining, where small deposits are hard to see on scans. A test that can find remaining tumor cells is what would make stopping safe.

What a curative protocol has to solve

  1. Which cells survive the best response to treatment, and what keeps them alive?
  2. Can surgery or other local treatment, at the right moment, remove what the drugs leave behind?
  3. Which drug combinations hit the driver mutation and the surviving cells together, without overlapping side effects?
  4. Can a blood test show when no disease is left, and who can safely stop treatment?
  5. Does longer treatment cure, or only delay relapse?
See the projects working on these questions →

Charts are redrawn from the numbers published in each paper, not copied from the papers.