MSI-High Metastatic Colon Cancer: Immunotherapy Against a Ticking Clock
The trial that made pembrolizumab standard first-line therapy here also showed some patients progress early on it — and this patient's liver may not have months to spare if that's how her case goes.
J.K., a 52-year-old woman, manages the produce section of a regional grocery chain and has spent the last month attributing her early fullness and fifteen-pound weight loss to stress over her daughter's upcoming wedding — until abdominal pain sent her to an emergency department, where CT found a large right colon mass and innumerable liver metastases, the largest 8cm and compressing the porta hepatis enough to nudge her bilirubin to 1.8. Molecular testing confirmed microsatellite instability-high, deficient mismatch repair disease, and germline testing — sent given her mother's own colon cancer at 46 — confirmed a pathogenic MLH1 mutation: Lynch syndrome, newly diagnosed alongside stage IV disease. She has two teenagers at home and, by her own account, "no interest in spending the next six months finding out if the fancy drug is going to work."
KEYNOTE-177 established pembrolizumab as preferred first-line therapy for MSI-H/dMMR metastatic colorectal cancer, with a dramatic improvement in median progression-free survival and response durability compared with chemotherapy. What the trial's own Kaplan-Meier curves also show, though, is real and specific to this decision: the two arms' progression-free survival curves cross early, in the first several months, because a meaningful subset of patients on pembrolizumab progress before ever responding. The overall response rate in fact favored pembrolizumab (43.8% versus 33.1%), so this is not equivalent response with different timing: chemotherapy produces fewer responses, but the ones it produces tend to arrive faster. For a patient whose tumor is already compressing her bile ducts, an early-progression outcome isn't an abstract statistical possibility; it's a plausible path to hepatic decompensation before pembrolizumab, if it is going to work for her, has had time to. Her bilirubin has already moved from 1.1 to 1.8 in two weeks, which puts a number on how much of that waiting period she can actually afford.
Multidisciplinary clinic, before first treatment
I'd start pembrolizumab. KEYNOTE-177 (André et al., NEJM 2020) showed median progression-free survival more than doubled compared to chemotherapy in MSI-H metastatic colorectal cancer, and the responses that do occur are far more durable — some patients years out with ongoing disease control chemotherapy essentially never produces in this biology. Starting with the historically weaker option because of a subset risk means potentially never getting her onto the drug that actually changes her long-term trajectory.
The long-term data are genuinely impressive, and I'm not arguing she should never get pembrolizumab. But her bilirubin has climbed in two weeks, and that same trial's own curves show an early crossing point — a real subset of patients progress on pembrolizumab before responding, while chemotherapy's responses, when they happen, tend to happen faster. If her biliary compression worsens before we know which kind of patient she is, we're managing a decompensating liver instead of a treatment decision.
This isn't a claim that chemotherapy is the better drug for her disease overall — it's a claim about what her liver can tolerate waiting to find out.
I looked at her imaging before this meeting — the compression is focal enough that percutaneous biliary drainage is technically straightforward here. If we decompress the biliary tree directly, the actual six-to-eight-week vulnerability window changes shape: a drained system can tolerate watching for a pembrolizumab response in a way an actively obstructing one can't. That doesn't settle which drug is better — it just means the timing argument doesn't have to be the reason we pick one.
Agreed: percutaneous biliary drainage this week, followed by pembrolizumab monotherapy once bilirubin trends down, with imaging reassessment at nine weeks — earlier than the usual twelve-week interval — specifically to catch an early-progression pattern promptly if it occurs.
Not fully agreed: the hepatologist's underlying discomfort with accepting any early-progression risk at all in a patient with this disease burden was not fully resolved by the drainage plan, and was recorded explicitly — chemotherapy remains the stated fallback if the nine-week scan shows progression rather than response, a threshold set now rather than negotiated under pressure later.