A Single Silent Brain Lesion and the Choice Between Two ALK Inhibitors
A newly diagnosed ALK-positive lung cancer with one asymptomatic brain lesion too small to treat directly. The disagreement is which first-line ALK inhibitor’s intracranial protection is actually worth its own distinct cost to her daily life.
Renee K., 46, has spent nineteen years working nights in the same emergency department, the kind of job that runs on being able to trust your own read of a room in under a minute. She has never smoked, has no other medical problems, and had chalked up three weeks of a cough to a cold that wouldn’t quit until a single streak of blood in her sputum sent her for a chest X-ray on her way home from a shift. A 3.8cm right upper lobe mass with mediastinal lymphadenopathy followed, then a biopsy, then a molecular panel that came back positive for an EML4-ALK fusion — a driver mutation that, unlike most lung cancer diagnoses at her age, has almost nothing to do with the fact that she never smoked. PET-CT confirmed disease confined to the chest and one CNS site, with normal liver and adrenal glands, and her only abnormal vital at diagnosis was a resting heart rate in the low 90s, attributed to anxiety about the diagnosis itself rather than to anything cardiopulmonary.
Staging brain MRI, ordered routinely regardless of neurologic symptoms because ALK-positive disease has an outsized tendency to spread there, found one lesion: 4mm, right frontal, no surrounding edema, nothing on exam to suggest it exists. It is small enough that she would never have known about it outside a protocol built to look for exactly this. That fact is what actually organizes the decision in front of the team — not whether she has CNS disease, but how much of the treatment choice should be built around a lesion this size specifically, rather than around the two trials’ separate, larger-scale answers to a similar but not identical question. CROWN’s headline intracranial figure — a 96 percent probability of never developing brain metastases at five years — was measured in the patients who had no CNS disease at baseline, which is precisely what Renee is not. In the subgroup she actually belongs to, patients with brain metastases already present at enrollment, the five-year probability of remaining free of intracranial progression was 83 percent; ALEX showed alectinib cut CNS progression risk sharply against the same comparator. Neither trial enrolled for a lesion this small specifically, and neither was run against the other.
Lorlatinib is a third-generation inhibitor built deliberately for CNS penetration, and that broader reach comes with broader off-target activity — TRK and ROS1 inhibition alongside ALK, the mechanistic background to its mood, cognitive, and metabolic effects. Alectinib was designed for ALK selectivity, with a correspondingly narrower off-target footprint. Neither drug’s design was aimed at a patient with a single sub-5mm lesion; both were built for a heavier CNS burden than Renee currently carries.
Before the first dose, weighing two kinds of CNS protection
Start her on lorlatinib. CROWN’s three-year update reported a 92 percent reduction in the rate of intracranial progression against crizotinib, and at five years 83 percent of the patients who came in with brain metastases were still free of intracranial progression — the best numbers any ALK inhibitor has produced in a randomized dataset — and CNS progression, not systemic progression, is what actually ends most ALK-positive patients’ first remissions early. She already has a lesion. Waiting to see whether it grows before reaching for the more CNS-active drug concedes ground we don’t have to concede.
You’re right that CROWN’s intracranial numbers are the best any ALK inhibitor has produced — I’m not contesting that arithmetic.
But CROWN compared lorlatinib to crizotinib, not to alectinib, and ALEX independently showed alectinib itself already cuts CNS progression risk sharply against that same comparator. Treating two separately run trials as though they answer a head-to-head question is doing more work in this decision than the evidence actually supports.
And lorlatinib’s own label carries CNS-specific adverse effects — mood change, cognitive effects, dysarthria — in a meaningful share of patients, plus hypercholesterolemia in about seven in ten and hypertriglyceridemia in about two-thirds, at rates alectinib’s longer real-world track record simply doesn’t generate. She has told us directly that staying sharp at work is not incidental to her. A drug-attributable cognitive risk in a patient whose job depends on cognitive sharpness isn’t a side detail to note and move past.
Neither trial was built around a patient who looks like her — a single lesion under half a centimeter, found before it caused anything, in a population CROWN and ALEX would have called low CNS burden if they’d stratified for it that finely. That gap, between a population-level statistic and this specific 4mm finding, is the actual thing we’re deciding around, not which trial has the bigger number.
So use both trials for what they can honestly tell us: how much is really at stake in watching a lesion this small for a few weeks versus treating it immediately. Start alectinib — established, better tolerated, with a real independent CNS benefit of its own — get a brain MRI at eight weeks instead of the usual interval, and commit in writing now to switching to lorlatinib the day that lesion moves. If we’re wrong, we’re wrong for eight weeks, not for however long it takes a routine scan to catch it.
Agreed within the same visit: start alectinib today, obtain a brain MRI at eight weeks rather than the standard interval, and treat any growth in the frontal lesion — not merely new disease elsewhere — as an automatic trigger to switch to lorlatinib rather than a finding to re-discuss from scratch.
Not agreed: the Medical Oncologist remained on record believing lorlatinib should have been first-line purely on CROWN’s CNS numbers, and accepted the eight-week compromise as a reasonable plan without being persuaded the population-based argument was wrong — only that it lost, this time, to what Renee herself said mattered most to her.