Choosing Second-Line Therapy at a Borderline Platinum-Free Interval in Relapsed SCLC
Her small cell lung cancer relapsed 75 days after finishing first-line chemotherapy, close enough to the conventional 90-day line separating platinum-sensitive from platinum-resistant relapse that no single option is obviously correct. The disagreement is how much weight a somewhat arbitrary cutoff should carry when it falls this close to her actual number — sharpened by the fact that the field’s gentlest-looking second-line option came off the table this summer.
Louise P., 67, finished her first round of chemotherapy for extensive-stage small cell lung cancer and, ten days later, flew across the country to meet her first grandchild, born the week before — a trip she and her oncologist had deliberately timed around the end of treatment. Her initial four cycles of carboplatin, etoposide, and durvalumab produced a strong partial response, with her liver lesions shrinking by more than half and her original hilar mass barely visible on the last scan before maintenance began. Seventy-five days after that last cycle, new lower back pain sent her back for imaging that found two new lesions in her liver and one in a lumbar vertebra, consistent with relapsed disease; she remains fully independent and reports only mild fatigue alongside the pain.
Seventy-five days sits close enough to the conventional 90-day line separating platinum-sensitive from platinum-resistant relapse that her case falls into genuine gray territory rather than a clean call either way: platinum-sensitive relapse typically favors rechallenging with the original doublet, given a meaningfully higher chance of response; platinum-resistant relapse typically steers toward topotecan instead. The strength of her initial response is itself part of the ambiguity — a deep, durable partial response is generally read as a favorable sign for platinum sensitivity even at a borderline interval, though it is not part of the formal cutoff definition. Her new vertebral lesion adds a separate, time-sensitive consideration on top of the regimen debate: it has not caused any neurologic symptoms yet, but its location makes cord compression a real possibility if it grows before treatment starts. The third option on the list a year ago has since come off it. Lurbinectedin’s 2020 accelerated approval rested on a single-arm phase 2 trial reporting response rates in platinum-resistant disease; its confirmatory randomized trial, LAGOON, read out in June 2026 and did not meet its primary endpoint. Median overall survival was 8.7 months with lurbinectedin monotherapy against 10.7 months for investigator’s choice of topotecan or irinotecan — a hazard ratio of 1.19, pointing the wrong way — and the manufacturer has since announced it will submit a labeling supplement withdrawing the second-line indication. Louise’s own tolerance of her first regimen was, until this year, the kind of fact that argued for the gentler-looking of two non-platinum options; what it argues for now is a rechallenge question and a topotecan question, because the gentler option is no longer a second-line option at all.
Seventy-five days, and what a cutoff actually means
Seventy-five days is under ninety — by the convention we all use to sort platinum-sensitive from platinum-resistant relapse, that makes her platinum-resistant, and platinum rechallenge in truly platinum-resistant disease has a low enough response rate that I would not put her through it again. Topotecan is where that leaves us. She tolerated her first regimen without a dose reduction, which is the best predictor I have that she can absorb the myelosuppression topotecan reliably produces, and I would rather manage a known toxicity in a patient with demonstrated marrow tolerance than stretch a convention to justify a platinum she may not respond to.
The ninety-day convention is real and it is the number every major reference uses — I am not disputing that it exists.
But it is a historical convention derived from population-level response-rate data, not a hard biological threshold that flips at day ninety and not a day before. Seventy-five days is close enough to that line that treating her as confidently platinum-resistant, rather than as someone in a genuinely ambiguous zone, overstates what the cutoff actually tells us about her individually. She tolerated her original doublet well; a platinum rechallenge attempt remains a reasonable option here, not one the calendar has definitively ruled out.
Before the platinum question gets resolved either way, lurbinectedin has to come off the table, and not because it is a marginal call. LAGOON — the confirmatory trial required to validate its accelerated approval — read out this June and missed its primary endpoint outright: 8.7 months median overall survival on lurbinectedin monotherapy against 10.7 months for investigator’s choice, hazard ratio 1.19. That is not an unproven advantage, it is a point estimate favoring the comparator.
Its toxicity profile really is milder — grade 3-or-worse treatment-related events ran about 35 percent against roughly 64 percent in the control arm — and that is exactly the trap. A tolerability advantage is not a reason to give a drug that did not clear its own survival endpoint, and the second-line indication is being withdrawn in alignment with the FDA. So the real choice in front of Louise is two options, not three: rechallenge her original doublet, or treat her with topotecan.
Agreed: obtain updated bone imaging to assess the lumbar lesion for cord-compression risk, which independently changes urgency regardless of the systemic therapy chosen, and bring both remaining options — platinum rechallenge or topotecan — to Louise directly with the genuine uncertainty in the evidence stated plainly, rather than the team selecting one before she weighs in.
Not agreed: the Medical Oncologist and Clinical Pharmacologist remained split on whether her platinum-free interval genuinely supports a rechallenge attempt or effectively rules it out; the Oncology Pharmacist’s removal of lurbinectedin was accepted by both without argument as a settled regulatory and evidentiary fact rather than a judgment call, but narrowing the list from three to two did not resolve which of the remaining two either of them would actually recommend if forced to choose without Louise’s own input.