HER2-Positive Brain Metastases: Switching Systemic Therapy or Adding Radiosurgery
Three small, asymptomatic lesions on a surveillance scan force the same question twice over: is this a local problem to treat locally, or a signal that the systemic drug controlling everything else has quietly stopped reaching her brain.
Marisol K., a 52-year-old accountant, was diagnosed with HER2-positive metastatic breast cancer three years ago and has been through two prior lines of HER2-directed therapy, most recently switching to trastuzumab deruxtecan eleven months ago after her disease progressed on a taxane-based combination. That switch has gone well — her liver and bone lesions have shrunk on every scan since, she is back to working full time, and she reports feeling, in her own words, better than she has in two years. Her most recent surveillance brain MRI, ordered as routine monitoring rather than for any new symptom, found three new lesions, each under one centimeter, without surrounding edema and without any correlating neurologic complaint on exam. She is entirely asymptomatic and functioning at full capacity — the finding exists only on the scan.
The question the team keeps circling back to is what three small, silent lesions actually mean for a drug that is visibly working everywhere else in her body. Trastuzumab deruxtecan's own trials have shown real intracranial response activity, but not complete central-nervous-system control, so new small lesions appearing on it are consistent with a drug that reaches the brain imperfectly rather than one that has failed outright. Tucatinib, by contrast, was tested in HER2CLIMB specifically alongside trastuzumab and capecitabine in a population that included patients with active, untreated brain metastases, and it produced a real central-nervous-system-specific survival benefit in exactly that population — a genuinely different evidentiary claim than 'works in the brain too.' Whether that trial's population match justifies switching her off a regimen that is otherwise clearly working, on the strength of three asymptomatic millimeter-scale findings from one scan, is the actual disagreement, not whether tucatinib's central-nervous-system data are real.
There is also a simpler fact the team keeps returning to: this is her first surveillance scan to show any intracranial finding at all in eleven months on the current regimen. A single new observation, however real, is not yet a trend, and the team's working distinction — between new metastatic seeding, which a drug can produce even while working well elsewhere, and documented progression of a previously treated site, which would be a much stronger signal of central-nervous-system failure — turns out to depend entirely on which of those two categories these three lesions actually belong to, a question no single scan can answer on its own.
Case conference, incidental brain metastases
Three visible, measurable lesions on a scan is a local-control problem with an excellent, low-toxicity answer: stereotactic radiosurgery has control rates well above 90 percent for lesions this size, and treating them doesn't require touching a systemic regimen that is otherwise clearly working for her. I'd treat the lesions and continue trastuzumab deruxtecan unchanged.
I'd weigh this differently given what HER2CLIMB actually showed. That trial enrolled patients with active, including untreated, brain metastases and still found a real central-nervous-system-specific survival benefit from adding tucatinib — a much stronger CNS claim than anything shown for her current regimen. New lesions appearing on trastuzumab deruxtecan, even small and asymptomatic ones, are a real signal that this drug's brain penetration has a ceiling she's now testing.
I'm not arguing her current drug has failed everywhere — her body disease is clearly controlled. I'm arguing specifically about the brain, where the evidence for tucatinib's benefit is more direct than the evidence that her current regimen will keep intracranial disease fully suppressed going forward.
I want to name what's actually being compared here, because I think it's being conflated. New lesions on surveillance imaging are new metastatic seeding, not documented progression of a previously treated site — the literature doesn't treat those as the same event, and one surveillance scan with three silent, millimeter-scale findings isn't yet the kind of failure signal that would justify abandoning a regimen this clearly effective everywhere else.
You're right that HER2CLIMB's population included active brain disease and showed real benefit there — I'm not disputing the trial. What I'd dispute is treating 'a drug with strong CNS data exists' as sufficient reason to switch before this drug has actually been shown to be losing intracranial control in her, specifically. Radiosurgery to the visible lesions now, continue her current regimen, and reserve tucatinib for if she shows a genuine intracranial progression pattern on a future scan — that keeps both options live instead of spending one early.
There's also a cumulative-radiation argument worth naming: if this turns out to be the first of several rounds of new small lesions over what could be years of otherwise well-controlled disease, treating each round with focal radiosurgery carries its own accumulating radionecrosis risk. That's a real cost of choosing local therapy repeatedly rather than switching regimens, worth tracking explicitly if this isn't a one-time finding.
Agreed: stereotactic radiosurgery to all three lesions this month, trastuzumab deruxtecan continued unchanged, and a repeat brain MRI at three months rather than the standard six, given the new finding.
Not agreed, and carried forward explicitly: what would count as a genuine intracranial-progression signal significant enough to trigger switching to tucatinib. The medical oncologist would treat any new lesion at the three-month scan as that signal; the radiation oncologist would want to see growth of a treated lesion or a larger new lesion before conceding the current regimen's brain penetration has a real ceiling for her. Neither position was adopted as the standing threshold — the team agreed to decide it concretely once the three-month scan is actually in hand, rather than pre-commit to a rule now.