BRCA1-Mutant Ovarian Cancer: Stopping Maintenance Olaparib at Two Years
A durable responder to maintenance olaparib approaches the two-year mark SOLO1 built into its protocol — not because activity was shown to end there, but because the trial was designed to stop there. The disagreement is about what to do with a drug that is still working when the study that approved it never watched what happens next.
Diane R., a 58-year-old retired high school chemistry teacher who now spends most weekday mornings restoring stained-glass panels for a co-op that sells them at the county fair, was diagnosed two years ago with Stage IIIC high-grade serous ovarian carcinoma after a germline panel ordered at diagnosis returned a pathogenic BRCA1 mutation. She had debulking surgery followed by six cycles of carboplatin and paclitaxel, reaching a complete radiographic and biochemical response — CA-125 undetectable, no measurable disease on CT — and started maintenance olaparib the same month SOLO1's own frontline-maintenance regimen would have enrolled her.
She is now twenty-three months into that maintenance course, one month short of the fixed two-year point at which SOLO1's protocol required every patient to stop the drug regardless of how they were doing, and her most recent scan is unchanged from the one before it. That detail is the whole disagreement, and it turns on which SOLO1 patient she resembles. The protocol did allow some patients to continue past two years — but only those still carrying evidence of disease at that mark. Anyone in complete response, with nothing radiographically left to treat, stopped, and Diane's undetectable CA-125 and unchanged CT put her squarely in that second group. So what a twenty-fifth month buys a woman already in complete response is the one thing the trial deliberately declined to generate data on, in either direction. What it did generate is durable: 48.3% of the olaparib arm progression-free at five years against 20.5% on placebo, and at seven years 45.3% versus 20.6% still alive and not yet needing a second line of therapy. Cumulative myelosuppression and a small but real MDS/acute leukemia risk rise with prolonged PARP exposure in a genotype already carrying some baseline predisposition to hematologic malignancy, which is a cost that accrues whether or not the drug is still contributing anything.
Diane has said plainly that she would rather stop now, on the study's own terms, than spend years wondering whether a drug she can't be sure is still helping is also quietly raising a risk nobody has measured in her exact position. That preference doesn't settle the clinical question, but it does mean whichever plan the team lands on has to include a clear, monitored answer for what happens if her CA-125 moves in the first year afterward.
Twenty-three months in, deciding what month twenty-four means
I'd continue her. She is twenty-three months into a complete response with a drug she has tolerated well, and stopping it on a date SOLO1 chose for trial design reasons — not because her disease told us anything at month 24 — is a hard thing to explain to a patient who has been counting down to this exact appointment with dread rather than relief.
If she had progressed, or had significant cumulative cytopenias, I wouldn't be raising this at all — this is specifically about a patient who is doing better than the trial's own median.
I understand the instinct, but SOLO1's numbers are fixed-duration results — 48.3% progression-free at five years against 20.5%, and 45.3% versus 20.6% still off second-line therapy at seven — and every complete responder who generated them stopped at 24 months. The trial only continued patients who still had disease to treat, which she doesn't, so there's no arm anywhere in it that looks like her at month 25. Extending her therapy isn't a conservative choice, it's an experiment we'd be running on one patient without the safety data the fixed-duration trial actually generated.
"She's still responding" describes her disease at month 23, not what happens to her marrow at month 30 or 40 — PARP-associated MDS and acute leukemia are late, cumulative-dose phenomena, and BRCA1 carriers already start from a higher baseline hematologic-malignancy risk than the general population.
You're both arguing from real data — hers just isn't the data either trial actually collected. I'd stop at 24 months, because that's the regimen with seven years of follow-up behind it, not because I'm confident continuing would hurt her. But I wouldn't let 'stop' mean 'discharge her to routine surveillance.' Shorten her CA-125 and imaging interval for the first year off therapy specifically, since that's the window where an undertreated relapse would show itself fastest if the protocol duration turns out to have been too short for her.
Agreed: olaparib discontinued at the 24-month mark per SOLO1's protocol duration, with CA-125 and cross-sectional imaging shortened from the standard surveillance interval to every three months for the first year off therapy.
Not agreed, and left explicit rather than smoothed over: whether the fixed 24-month duration is actually the right stopping point for a patient still in complete response, or simply the point the pivotal trial happened to stop looking. The gynecologic oncologist's discomfort with the date wasn't resolved by the pharmacologist's data — it was acknowledged as a real gap the trial design leaves open, addressed practically through closer surveillance rather than settled on the merits.