Six Clean Cycles of Temozolomide — And a Patient Who Doesn’t Want to Stop
A patient with a strong response to standard six-cycle temozolomide wants to keep going. The trials that tested exactly that question say his instinct, however understandable, isn’t backed by the evidence.
Daniel R., a 52-year-old high school chemistry teacher, was back in his classroom seven months ago, eight weeks out from his craniotomy and still wearing a soft cap over the healing incision, determined to finish the year with the AP students he’d already committed to before any of this started. Nine months ago an MRI ordered for two weeks of word-finding trouble and a single witnessed staring spell found a right frontal mass; resection was near-total, pathology confirmed glioblastoma, IDH-wildtype, with a methylated MGMT promoter. He completed six weeks of concurrent radiation and daily temozolomide without missing a session, then started the standard six cycles of adjuvant temozolomide, escalating from 150 to 200 mg/m². Today’s visit closes out cycle six. His counts have held — grade 1 lymphopenia, nothing dose-limiting — his Karnofsky score is 90, and the MRI performed for today’s visit shows the small rim of residual enhancement along the resection margin, present since cycle two, smaller again rather than larger, with no new lesion anywhere else. He wants to know when he starts cycle seven.
That the tumor is still visibly, if faintly, present at cycle six matters directly to how the two answers to his question apply to him. The Spanish GEINO 14-01 trial (Balana et al., 2020) enrolled patients exactly like him — no progression after six cycles — and randomized them to stop or continue for six more, stratifying deliberately by MGMT status and by whether measurable residual disease was present, which just over half its patients had. Extending changed nothing: no difference in six-month progression-free survival, none in overall survival, and more lymphopenia, thrombocytopenia, and nausea in the arm that carried on. Blumenthal et al. (2017) pooled individual patient data from four randomized trials and landed in the same place on survival, though its patients continued by their physicians’ choice rather than by randomization. Daniel therefore sits inside a population GEINO deliberately sorted for and still found no benefit in. What that trial did not sort is trajectory — it recorded whether residual enhancement was present, not whether it was still shrinking, and his has shrunk at every scan since cycle two. That is the distinction left for the room, and it is thinner than it first looks: the stratification that did exist placed him in the group with the worse prognosis, not the one with an unexamined reason to keep going.
Closing out cycle six
Six cycles is the floor these trials tested, not a ceiling anyone proved. Daniel’s tumor is MGMT-methylated, which is exactly the group most likely to still be responding to an alkylating agent at this point, and his residual enhancement has gotten smaller at every scan since cycle two — not stalled, not grown. Blumenthal’s pooled analysis found that continuing beyond six cycles was associated with a real, if modest, reduction in the risk of progression — hazard ratio 0.80 — and in the MGMT-methylated subgroup, which is his, 0.65. Neither figure moved overall survival. I’d rather offer him six more cycles of a drug he’s tolerating without difficulty than stop a treatment that’s still visibly working.
If his residual enhancement had plateaued rather than kept shrinking, I wouldn’t be making this argument — a stable-but-present lesion after six cycles is a genuinely different picture than one still trending down.
You’re right that his trajectory looks better than a generic trial average — I’m not disputing his scan. But the trial that actually answers this — GEINO 14-01 — didn’t enroll a generic glioblastoma population and extrapolate. It enrolled patients who’d finished six cycles without progression, randomized them to six more or none, and found no overall-survival difference between the arms, with more hematologic toxicity in the group that continued. That’s not an average diluting a real subgroup effect; that’s the specific decision in front of us, tested directly, in patients selected the same way Daniel was just selected — by getting through six cycles intact.
And I’d handle that hazard ratio carefully. Blumenthal’s patients weren’t randomized to continue — they and their oncologists chose it, and the patients who get offered six more cycles are the ones who look well enough to take them. GEINO took that choice away, and the progression signal went with it: no difference at six months, none overall, and more lymphopenia and thrombocytopenia in the arm that kept going.
Both of you are arguing about the wrong lever. If the real goal is squeezing more out of this than the standard Stupp regimen already gives him, Tumor Treating Fields is the one thing here with a randomized overall-survival result behind it — EF-14 (Stupp et al., 2017) reported 20.9 months against 16.0. Let me be honest about the fit, since you’ve both been honest about yours: progression-free survival was EF-14’s primary endpoint and survival its powered secondary, and it randomized patients within seven weeks of finishing chemoradiation, to wear the arrays alongside maintenance temozolomide. Daniel is six months past that door. Offering it now is an extrapolation — but it is an extrapolation from a positive survival trial, which is more than either of you can say for the drug you’re arguing over. I’d stop temozolomide at six per GEINO, and spend the conversation we’re about to have with Daniel on whether he’s willing to wear a device most of the day instead of swallowing six more months of pills — that’s the trade actually backed by an overall-survival number.
Agreed after some real back-and-forth: temozolomide stops at cycle six, matching the population GEINO 14-01 actually tested and found no benefit in extending. Daniel will be counseled today about Tumor Treating Fields as the alternative escalation with genuine randomized survival evidence behind it, rather than simply told to stop and come back in three months.
Not agreed, and left open for the visit itself rather than smoothed over: how hard to push if he declines Tumor Treating Fields once he hears what daily wear time actually involves — a shaved scalp, a battery pack, roughly eighteen hours a day of the arrays in place. The neuro-oncologist believes offering it once and respecting his answer is enough; the radiation oncologist, who has walked more patients through the device itself, wants at least one follow-up conversation before treating a first no as final, given how much of the early refusal rate in her own experience turns out to be about the described burden rather than the lived one. Both agree on the one thing they won’t do: let “he wants to keep going” become a reason to hand him six more cycles of a drug the trial already asked and answered for patients exactly like him.