Paraneoplastic Cerebellar Degeneration: When Immunotherapy Isn't Working
A month of escalating immunotherapy for anti-Yo paraneoplastic cerebellar degeneration hasn't produced any real change — and the mechanism behind that antibody's own poor track record is exactly what the team now has to reckon with honestly.
E.V., 63, has spent every Saturday morning for a decade running the plant sale table at her neighborhood community garden, trading cuttings and giving unsolicited advice about tomato blight. Ten weeks ago her friends noticed she was unsteady climbing the porch steps, something everyone assumed was her arthritis flaring until it kept getting worse instead of better. Over six weeks it progressed from occasional stumbling to a gait she could not manage without a walker, slurred speech that made phone calls difficult, and hands unsteady enough that she stopped pouring her own tea. She has treated hypertension, well-controlled for years on a single agent, and no other real medical history. Anti-Yo antibody testing, sent once the picture stopped looking like anything ordinary, came back positive in serum and CSF; a subsequent workup found a 2.1cm invasive ductal carcinoma, ER-positive, treated with lumpectomy and now beginning adjuvant chemotherapy.
She has been on methylprednisolone pulses and IVIG for six weeks and rituximab for the last four, and none of it has meaningfully changed her exam — she remains unable to walk unassisted, her speech is no clearer than it was at diagnosis. Anti-Yo’s mechanism is part of why: Albert and colleagues’ analysis of cerebrospinal fluid in this disease found activated cytotoxic T cells in patients with active, progressive disease, and argued those cells are what destroys Purkinje neurons — a mechanism quite unlike a circulating antibody sitting on the surface waiting to be washed out by whatever agent is tried next. How much of the injury is T-cell-driven rather than antibody-driven is still genuinely contested, but the T-cell account is the one the field has moved toward. That distinction matters for what the team does now, because it is not simply a question of which drug is left to try — it is a question of how much of what has already happened to her cerebellum was ever going to be reversible by any immunotherapy at all.
Six weeks in, deciding whether to try again
She’s had exactly one second-line agent — rituximab — for one month. That’s not the same as having exhausted what immunotherapy can offer her. Dou and colleagues published a case in 2024 of confirmed anti-Yo PCD that had failed glucocorticoids and still achieved partial symptomatic relief after two doses of ofatumumab, a different anti-CD20 agent than the one she’s on now. Anti-Yo response rates are poor on average, but poor on average isn’t never, and we haven’t actually tried the next reasonable step.
You’re right that a single agent for a month isn’t exhaustive — but the mechanism matters here more than it does in most of what we treat with immunotherapy.
“The next reasonable step” assumes each additional agent carries roughly the same odds as the last one; it doesn’t. Albert and colleagues found activated cytotoxic T cells in the spinal fluid of these patients and made the case that those cells are what destroys Purkinje neurons — not an antibody effect sitting on the surface waiting to be washed out. That is still an argument rather than a closed question, but it’s where the mechanism data points, and by the time symptoms have been present this long, a meaningful share of that damage is neurons that are already gone, not neurons waiting to be rescued. I’d rather put our remaining leverage into the tumor, where the literature at least shows tumor control changes her overall trajectory even when it doesn’t reverse what’s already happened neurologically.
Neither of you is actually wrong, and that’s the problem — you could both be right and she still doesn’t know what happens if the next thing doesn’t work either. I’m not going to referee the immunology. What I’d ask for is one agreed reassessment point — six weeks — with an explicit answer, decided now, for what we do if there’s no change by then, told to her in those terms rather than as an open-ended ‘let’s see.’ Whichever of you is right about the mechanism, she deserves to know now what ‘this isn’t working’ would actually mean for her, instead of finding out by default three agents from now.
Agreed: rituximab continues to the six-week mark as already planned, adjuvant chemotherapy proceeds on its own oncologic timeline, and the team commits now — not at six weeks — to what happens if her exam is unchanged then: a switch to ofatumumab, explained to her in those terms, with a second explicit reassessment point set at that time.
Not agreed, and left stated rather than resolved: the neurologist and oncologist still disagree about how much further immunotherapy is actually worth pursuing given the mechanism, a disagreement neither the six-week plan nor the palliative care specialist’s framing was meant to settle. What the team could agree on was only that she should not be the one left holding that uncertainty without a plan attached to it.