Why Karyopharm's 5.3-month improvement still counts as a failed trial

Karyopharm reported topline results from its phase 3 XPORT-EC-042 trial on 30 July, and the two headline numbers look like a win. Median progression-free survival was 12.75 months for patients on selinexor versus 7.43 months on placebo. The trial failed. Shares fell about 69% after hours.

A result that reads like success and scores as a failure is the most instructive kind of readout there is. Two things are going on here, and they are worth separating: what the statistics actually said, and why the market took two-thirds of the company off the board rather than a slice of it.

A median is not a result

Median progression-free survival is the point at which half the patients in an arm had their disease progress or died. It is one coordinate on a curve. It tells you nothing about the shape of the curve on either side of it, and with a few hundred patients it can swing on a handful of events.

The number the trial was actually judged on is the hazard ratio, which was 0.76, with a one-sided p-value of 0.0791 in a modified intent-to-treat population of 236 patients. Translated: at any given moment in the trial, a patient on selinexor was roughly 24% less likely to progress or die than a patient on placebo. That is a real-looking effect. The p-value is the part that sank it. The conventional bar for a one-sided test is 0.025, and 0.0791 is about three times that. With this many patients and this many events, a separation this size would turn up by chance often enough that chance cannot be ruled out.

That is a narrower statement than "the drug does not work." The trial could not prove it works. Those are different claims, and the difference is exactly what determines whether a company runs the study again or walks away.

What the topline did not say

Three absences matter more than anything in the release:

• No overall survival. Progression-free survival was the primary endpoint. Whether patients on selinexor lived longer is not yet known, and follow-up continues.
• No discontinuation rate. "Consistent with selinexor's established profile, with no new safety signals" is a sentence, not a safety table. The share of patients who stopped treatment because of side effects is the cleanest read on real-world tolerability, and it was not in the topline.
• No subgroup split. The trial enrolled TP53 wild-type patients with both mismatch-repair-proficient and mismatch-repair-deficient tumors. The topline reported one number for everyone.

Karyopharm says it will present the full dataset at a future medical meeting. Until then, the honest position is that nobody outside the company knows the shape of those curves.

Why the move was 69% and not 20%

Phase 3 trials miss routinely. Of programs that reach phase 3, 57.8% go on to a regulatory filing — meaning something close to two in five do not. A miss on its own usually costs a company some fraction of the value the market had assigned to that one program.

A move of this size says the market was pricing something else. Karyopharm ended the first quarter of 2026 with $90.9 million in cash and said its existing liquidity would fund operating plans into late in the third quarter of 2026. That window is now. The same quarterly filing flagged going-concern risk, and the company has Centerview Partners advising it on financing and strategic alternatives.

This is the part that generalizes. A failed readout at a company with two years of cash is a setback to a program. The same readout at a company whose own disclosed runway ends this quarter is a financing event, because the trial was the thing that would have set the price of the next raise. The clinical result and the balance sheet do not add. They multiply.

It also cuts the other way, which is the reason to do this work before the date rather than after: the same binary can re-rate a company hard in either direction, and how far it travels depends as much on the cash line as on the curve.

What is actually left

The endometrial program is being wound down — the company said it will reduce investment there and prioritize myelofibrosis and multiple myeloma. The near-term catalyst is a supplemental new drug application, planned for August 2026, seeking accelerated approval of selinexor combined with ruxolitinib in myelofibrosis, based on the phase 3 SENTRY trial's spleen-volume response at week 24. Karyopharm intends to request priority review, which would put an FDA decision roughly six months out.

One caveat in the filing deserves translation. The FDA "requires further discussion on the data to be used to support the sNDA and convert potential accelerated approval to traditional approval." Accelerated approval is provisional: it is granted on a surrogate measure, and it has to be confirmed later by evidence the agency accepts. The company and the regulator have not settled what that confirming evidence would be. That is an open question sitting underneath the one remaining near-term catalyst, not a formality.

Chief executive Richard Paulson said the results "do not diminish our confidence in the broader potential of selinexor." That is a reasonable thing for a chief executive to say and not, on its own, evidence about anything.

What to watch

• Whether the myelofibrosis sNDA is actually filed in August, and whether priority review is granted.
• The outcome and the terms of any financing or strategic transaction. Terms are the story; a rescue on punishing terms is not the same as a rescue.
• The full XPORT-EC-042 dataset at a medical meeting: the overall survival trend, the shape of the curves, and the discontinuation numbers.
• Any public change in the FDA's position on what would confirm an accelerated approval.

The durable takeaway is smaller and more portable than this one company. When a readout produces a number that looks good and a verdict that is bad, the medians are usually the reason you were fooled and the hazard ratio is usually the reason the market was not. And when the move is far larger than the program was worth, stop reading the trial and start reading the cash.

This is analysis, not investment advice. It describes what was disclosed and what was not, and it does not recommend any position in any security.