Isolated Brain Progression on Crizotinib: Switch ROS1 Inhibitors or Treat Locally
Fourteen months into a good response to his first ROS1 inhibitor, two new small brain lesions raise a real question with no single right answer: abandon a systemic regimen that is still working, or treat the brain directly and leave it in place.
David M., 39, plans his long-haul freight routes months in advance around one fixed rule: he is never on the road for his daughter’s Saturday soccer games. Fourteen months ago that rule nearly broke when a persistent shoulder ache turned out to be a right lower lobe mass, biopsied to adenocarcinoma with a ROS1 fusion on tissue testing — an alteration more common in younger, lighter smokers, which fit him; he had quit ten years earlier after a fifteen pack-year history. He started crizotinib, the first ROS1 inhibitor to reach approval, and by three months had a partial response that has held steady since, with no new disease anywhere and no growth at the primary. His only real side effect has been mild, intermittent visual disturbances at night, a known crizotinib class effect he has adapted to by simply not driving after dark, and he was back on his daytime routes within six weeks of starting.
His most recent surveillance brain MRI, ordered on the same six-month schedule since diagnosis even though he had never had CNS disease, found two new lesions — 8mm and 5mm, both frontal, both asymptomatic — while the same day’s chest and abdomen scans showed nothing new anywhere else. That split is the actual finding worth reading closely: crizotinib is known to under-penetrate the CNS relative to newer ROS1-directed drugs, so a brain-only relapse on an otherwise fully responding regimen looks less like his cancer broadly outrunning treatment than like a known limitation of this specific drug reasserting itself. TRIDENT-1, the trial behind repotrectinib’s approval, showed real intracranial responses even in ROS1 patients who had already progressed on crizotinib, including some carrying the ROS1 G2032R resistance mutation the drug was purpose-built to reach — but whether David’s own tumor carries that mutation, or any resistance mutation at all, is something nobody has tested yet. His extracranial disease, notably, looks nothing like a tumor that has broadly outgrown crizotinib’s control: the primary mass measured 2.1cm at his last scan and 2.0cm today, still inching down rather than plateauing, and the mediastinal nodes that were the second site of disease at diagnosis remain below measurable threshold. There is no serum marker that would add anything here — ROS1-rearranged adenocarcinoma has no validated blood-based surveillance test, which is precisely why the resistance-testing question below is the one that carries information.
Two small spots on an otherwise-working drug
Switch him to repotrectinib. TRIDENT-1 showed meaningful intracranial responses in ROS1 patients who had already progressed on crizotinib, including a subset carrying the G2032R resistance mutation the drug was purpose-built to overcome. Crizotinib is well known to under-penetrate the CNS relative to newer ROS1-directed agents — a brain-only relapse on this specific drug is exactly the failure pattern its own pharmacology predicts. Waiting to see whether this is “just” two small spots risks missing that the biology producing them may already be building resistance we can’t see yet on a chest CT.
You’re right that crizotinib’s CNS penetration is its known weak point — that part isn’t in dispute.
But nothing about his extracranial disease looks like it’s failing. Two asymptomatic lesions under a centimeter, with no growth anywhere else, is close to the pattern Weickhardt and colleagues described when they showed local ablative therapy for oligoprogressive disease can prolong benefit from an otherwise-effective TKI substantially, sometimes by a year or more, without ever switching the systemic drug. Treat both lesions with stereotactic radiosurgery, continue crizotinib, and keep a regimen he has tolerated for over a year instead of trading it for one whose long-term tolerability in him is still completely unknown.
Both of you are arguing from what this pattern usually means, and neither of you actually knows what his tumor is doing yet. Get ctDNA or repeat tissue sampling for ROS1 resistance mutations before committing to either path — it’s a low-cost step that turns a guess into an answer.
And whichever way that comes back, David needs to hear plainly what repotrectinib would actually cost him: its approval came with a mandated low-dose titration schedule specifically because dizziness and gait disturbance are common early on — not a side effect he can treat as background noise behind the wheel of an eighty-thousand-pound truck. If resistance testing comes back negative and local therapy controls both lesions, we may never need to have that conversation with him at all.
Agreed: send ctDNA-based ROS1 resistance testing today, and refer to radiation oncology for stereotactic radiosurgery to both frontal lesions while continuing crizotinib in the interim, rather than switching systemic therapy pending that result.
Switch to repotrectinib with the mandated dose-titration schedule and explicit driving-safety counseling before he returns to routes.
Continue crizotinib indefinitely with brain MRI surveillance shortened to every three months rather than six.
Not agreed: the Medical Oncologist accepted the interim SRS-plus-continue plan but remained on record that a mutation-negative result would not fully resolve the concern that CNS under-penetration alone, apart from resistance, is reason enough to move proactively — a broader question the group left for the next real relapse rather than this one.