Tuberous Sclerosis with Three Everolimus-Responsive Findings: Whose Threshold Sets the Dose
A boy with tuberous sclerosis has a growing SEGA, refractory seizures, and a renal angiomyolipoma — all three genuinely responsive to the same drug, everolimus, but each with its own trial, its own target level, and its own monitoring plan. When they disagree on how aggressively to dose, whose threshold wins?
T.A. is nine years old, diagnosed with tuberous sclerosis complex in infancy after infantile spasms led to the brain MRI that found his cortical tubers, and has grown up with a care plan that until this year felt manageable to his parents even as it touched nearly every organ system. Three manifestations are now converging on the same conversation: a subependymal giant cell astrocytoma identified at age 4 with interval growth on his most recent surveillance MRI, drug-resistant focal epilepsy that has failed two anticonvulsants and still produces several seizures a week, and a right renal angiomyolipoma that has reached 3.2 centimeters — the size at which mTOR inhibition begins to be recommended for a growing asymptomatic lesion, though well below the four centimeters at which aneurysm and hemorrhage risk classically climbs. His creatinine and eGFR are normal for age and he is treatment-naive to mTOR inhibition.
All three are documented everolimus indications, and the team agrees on starting the drug. The dosing question is where the trials stop describing him. At nine, T.A. sits essentially at EXIST-1's median enrolled age of 9.5 years — that trial's population is his population, and it started everolimus at 4.5 mg/m²/day titrated to a whole-blood trough of 5–15 ng/mL. EXIST-3 also targeted trough directly, randomizing to a low range of 3–7 or a high range of 9–15, with the high arm outperforming the low (roughly 40 percent versus 28 percent achieving at least a halving of seizures, against about 15 percent on placebo). EXIST-2 describes him least of all: median age 31, range 18 to 61, on a flat 10 mg daily dose with no trough target whatsoever.
So the difficulty is not that three trials point to three incompatible numbers. EXIST-3's high arm falls inside EXIST-1's own 5–15 window, and a single trough can satisfy both. The difficulty is that they imply different places to aim within that shared band, and that one of his three findings has no pediatric exposure benchmark at all. His epileptologist wants the upper half, where EXIST-3's seizure benefit actually concentrated. His neuro-oncology team is watching the SEGA on EXIST-1's own monitoring cadence. His nephrologist is tracking a kidney lesion against no trial-derived target whatever, since the only pediatric angiomyolipoma data come from EXIST-1's post-hoc analysis of children treated for SEGA — where about three-quarters with a target lesion responded — rather than from EXIST-2 at all. No published protocol tells this team how to reconcile that.
At the multidisciplinary TSC clinic visit
I want to dose toward what actually controls his seizures. He's had drug-resistant epilepsy for years now, several seizures a week despite two prior agents, and every week of uncontrolled seizures in a nine-year-old is a week of real, cumulative neurodevelopmental cost that doesn't reverse the way a shrinking tumor's risk profile does. EXIST-3 showed real seizure-frequency reduction in exactly this population, treatment-resistant TSC-associated epilepsy, and it showed it dose-dependently — the 9–15 ng/mL arm did better than the 3–7 arm, roughly 40% versus 28% reaching a halving of seizures. I'm not asking to go outside EXIST-1's range; 9–15 sits inside it. I'm asking that we aim at the top half of that shared window rather than settle at the bottom of it and accept an inadequate seizure response as the cost.
I want to be clear I'm not arguing his SEGA doesn't matter — I'm arguing that if we have to pick a starting target when the three don't overlap, ongoing seizure burden is the manifestation actively harming him right now, today, in a way the other two currently are not.
You're right that his seizure burden is causing real, ongoing harm, and I'm not proposing we accept an undertreated seizure disorder as an acceptable tradeoff. But I think the SEGA has to anchor the floor of this conversation, because it's the one manifestation among the three that can become a true emergency without much warning — obstructive hydrocephalus from a growing SEGA can present acutely, and EXIST-1 specifically demonstrated that dosing within its target range reliably shrinks these tumors. I'd want us dosing at least to that floor before we push higher chasing seizure control, and watching his SEGA closely enough that we'd know quickly if higher dosing for seizure control were somehow allowing tumor growth to resume.
I don't think 'dose to the EXIST-3 seizure target and let the SEGA ride along' is actually safe by default just because everolimus addresses the same pathway — EXIST-3 wasn't designed to confirm SEGA response holds anywhere in particular within its range, only to compare seizure outcomes between two exposures. The reassuring thing you're relying on is that 9–15 overlaps EXIST-1's window, and I agree that it does — but overlap is not the same as EXIST-1 having demonstrated SEGA shrinkage specifically at the top of its own range. We'd be extrapolating within a range, which is a smaller leap than extrapolating past it, but still a leap.
I don't have a strong opinion about whether seizure control or SEGA shrinkage should set the starting dose between the two of you — both are real, structural arguments. What I want to make sure doesn't get lost is that his AML crossed 3.2 centimeters, which is the size at which guidelines start recommending mTOR inhibition for a growing asymptomatic lesion — the aneurysm and bleeding risk classically climbs above 4 centimeters, so he isn't there yet, but he is on that trajectory, and unlike a seizure or a symptomatic SEGA, an AML bleed typically doesn't give any warning beforehand. I also want to be honest that I have the weakest trial footing of the three of us: EXIST-2 enrolled adults only, median age 31, on a flat 10 milligram dose with no trough target, so I cannot tell you what exposure his kidney lesion needs. Whatever dose this ends at, I need his renal ultrasound on an explicit surveillance interval built into his plan — not assumed to be covered because he's on the right drug for a different reason.
If his dosing lands at the top of the shared window because of the seizure-control argument, that's not necessarily wrong — the closest thing we have to pediatric AML evidence is EXIST-1's post-hoc look at children treated for SEGA, where about three-quarters of those with a target angiomyolipoma responded at that trial's 5–15 target. I just want us watching to confirm his AML is actually responding, not assuming it must be because a different endpoint is being hit.
Agreed: start everolimus targeting the EXIST-3 seizure-control exposure range, given the epileptologist's argument that ongoing seizure burden is the actively accumulating harm today, with the explicit condition that SEGA and AML surveillance both continue on their own trial-derived schedules rather than being assumed to track along with seizure control.
Not agreed: what to do if the SEGA or AML fails to respond at the higher, seizure-driven dose — the neurologist wants a pre-specified plan to escalate SEGA-specific monitoring frequency if imaging shows any growth at the next interval, while the epileptologist is reluctant to commit in advance to lowering the dose for a structural finding that hasn't actually changed yet. That question was left for the next visit, once the first surveillance imaging is back.