A Second Attempt at Stopping CML Therapy, After the First One Failed
A single patient in sustained deep molecular response, asking to stop CML therapy for a second time after her first attempt failed. The disagreement isn't about whether TFR is ever appropriate for her — it's about how much weight her own prior failure should carry against a new, concrete reason to try again.
M.T., a 52-year-old woman who has run a small bakery out of her home kitchen for the better part of two decades, was diagnosed with chronic-phase CML seven years ago and started on imatinib. She reached a sustained MR4.5 by year four and, wanting to see whether she could come off a drug she had taken every day for most of a decade, stopped under monitoring — only to lose major molecular response by month four, restart imatinib, and regain MR4.5 within eight months. That eight-month recapture is worth pausing on: it is the reassuring half of her history, evidence that her disease came back sensitive rather than resistant, and it is the reason a second attempt is being discussed at all rather than dismissed. She has now held that response for another twenty-two months, and is back in clinic asking about stopping again, this time for a reason that wasn't in play the first time: she and her husband are trying to have a child, and imatinib's own labeling lists it as contraindicated in pregnancy.
Whether a second discontinuation attempt is reasonable after a first one already failed is a genuinely different question than the one EURO-SKI and similar trials answered for treatment-naive-to-stopping patients. The RE-STIM registry, built specifically to look at second attempts, found their success rate running below first-attempt rates in the same population — her own disease has already shown, once, what it does when the drug pressure comes off. What has changed since the first failure is duration. She is now twenty-two months past regaining MR4.5 — close to the two-year floor of deep response that the discontinuation literature associates with better first-attempt outcomes, though she has assembled that time across a relapse and a restart rather than in one uninterrupted run, and no trial has told anyone whether those two routes to the same number behave the same way afterward. Her monitoring access is the one variable that is unambiguously in her favour: monthly PCR, confirmed available at her regional lab without delay, and a documented history of actually attending for it through her first attempt.
A second conversation about stopping
Her request isn't really about wanting to try TFR again in the abstract — it's about wanting to conceive a child on a drug that isn't safe for that. Imatinib's teratogenicity is real and specific, not a theoretical caution, and that turns this from a probability question about molecular relapse into a question about how to get her off the drug as safely as anyone can, whether or not this attempt holds.
I'd weigh that seriously if her biology hadn't already answered part of this question once. RE-STIM's own second-attempt cohort relapsed molecularly at a higher rate than first-attempt populations, and she's now a second-attempt patient by definition. I'm not arguing she should never try — I'm arguing we should walk in expecting a real chance this doesn't hold, and have a restart plan already agreed, not improvised at the first rising PCR.
The pregnancy indication changes what she's willing to accept, not what the biology actually does once the drug comes off — those are two different conversations, and I don't want the first one to quietly answer the second.
Both of your numbers assume the same thing: monthly PCR, and restarting the moment major molecular response is lost rather than waiting for any symptom. That's confirmed available to her, which is why I think this attempt is reasonable — but it's worth saying plainly that if that monitoring access were in question, neither EURO-SKI's numbers nor RE-STIM's would describe what would actually happen to her.
Agreed: attempt a second discontinuation now, with restart triggered at loss of major molecular response exactly as the first attempt was, and explicit pre-conception counseling on the real chance this attempt doesn't hold.
Not agreed: whether, if this attempt also fails within the next year, a third attempt should ever be offered, or whether two failures should be read as this particular disease's answer. The transplant physician's reading — that RE-STIM's own numbers should set real expectations, not just get mentioned once and set aside — was accepted as the honest framing for today's decision without resolving what a second failure would mean going forward.