Grade 3 Colitis in Her Best Response Yet: Does Stopping Pembrolizumab Mean Losing It
The response is real and it is good. So is the colonic ulceration. The team has to decide whether continuing the drug that is working is worth the bowel it may be perforating.
Fourteen months after a routine skin check turned up more than her dermatologist expected, Denise A. — a 58-year-old retired court reporter diagnosed that day with metastatic melanoma — has had one of the best responses to pembrolizumab her oncologist says she's seen in a patient with this disease burden: her most recent CT scan showed a greater than 60 percent reduction in her largest measurable lesions, doing more for her disease than either of the two lines of therapy she needed before it. Eight days ago, after her sixth cycle, she developed watery diarrhea that progressed over 72 hours to more than ten bowel movements a day with visible blood, prompting admission and an urgent colonoscopy that neither she nor her oncologist saw coming given how well she had otherwise tolerated treatment through five prior cycles.
Biopsies confirmed immune-mediated colitis with diffuse mucosal ulceration extending through most of the visualized colon — a grade 3 severity by standard toxicity criteria, not the milder, more self-limited grade 1-2 pattern that resolves with a brief steroid course and a treatment pause. High-dose intravenous corticosteroids were started immediately; by day four her stool frequency had fallen to four to five per day without visible blood, a real, measurable improvement, though still short of the near-complete resolution that would typically precede any conversation about resuming therapy. A repeat scan obtained during this admission, done to rule out a bowel complication rather than to restage her cancer, incidentally showed her tumor response holding steady rather than progressing during the eight days pembrolizumab has already been held.
That last detail is the fact the team keeps returning to from different directions: it says nothing has been lost yet, but it does not by itself say what happens to a response this good if the drug that produced it is never given again.
Whether stopping means losing what’s already been gained
This is the best response she's had to any therapy, and there's real retrospective evidence in melanoma — the Memorial Sloan Kettering and Australian cohort analyses among others — linking the occurrence of immune-related toxicity with a stronger tumor response — her colitis, in a strange way, may be part of the same immune activation that's shrinking her tumor. Once this colitis is fully controlled, I'd want to resume pembrolizumab under close monitoring rather than walk away from a response this good.
I understand the oncologic argument, but I want to be precise about what her colon actually looks like, and equally precise about what the guidance does and doesn't say. Pembrolizumab's own labeling withholds the drug for grade 2 or grade 3 colitis and reserves permanent discontinuation for grade 4 — permanent discontinuation at grade 3 is the anti-CTLA-4 rule, not the PD-1 rule. So the label is on your side, and I'm going to argue against it anyway: what I'm looking at is diffuse ulceration through most of the visualized bowel, which is a tissue finding the grade doesn't fully carry, and I think it earns more caution than the grading table alone would give it.
And the irAE-response association you're describing is a population-level correlation, not a guarantee for her specifically. I'd rather not test whether it holds by rechallenging a bowel we've already seen ulcerate this extensively. I accept that's a judgment beyond what the label requires — I'd just want it recorded as my judgment rather than dressed up as a rule.
I think this is being framed as resuming treatment versus giving up her response, and I don't think that's actually the choice in front of us. Checkpoint-inhibitor-induced responses can persist for a meaningful period after the drug is stopped — response and ongoing drug exposure aren't the same thing, and her own interim scan, taken during eight days off the drug, already shows the response holding rather than reversing.
That doesn't resolve whether rechallenge is safe — that's still the Gastroenterologist's call to make on the bowel's own terms. But it does mean permanent discontinuation isn't automatically the sacrifice it's being described as, and I'd want a dedicated oncology conversation about surveillance-off-therapy before either resuming or writing off a response this strong.
Corticosteroids were continued toward full resolution rather than a partial improvement threshold, with a dedicated oncology surveillance plan started regardless of what is ultimately decided about pembrolizumab — the one piece of the plan every voice agreed on immediately.
Not agreed, and stated explicitly as unresolved rather than deferred quietly: whether pembrolizumab is ever restarted. The Oncologist wants that conversation reopened once colitis fully resolves, arguing the response is too strong to abandon without at least discussing rechallenge; the Gastroenterologist accepts that the label permits resumption once she recovers to grade 1 or less, but does not want rechallenge on the table regardless, on the strength of how extensively the bowel was ulcerated rather than on the grade assignment — a position he stated explicitly as exceeding what the guidance requires. Both agreed the surveillance data over the following weeks, not tonight's disagreement, should be what the final decision is actually built on.