Ascelia's share price has fallen sharply since the full Complete Response Letter was published, extending the decline that began when the CRL was first disclosed on 3 July. At the current level, the market is starting to price in the most negative scenarios, namely a new clinical trial or no viable regulatory path, even though the less severe scenarios remain consistent with both the wording of the full CRL letter and the management commentary.

It should be acknowledged that, with uncertainty high until a Type A meeting brings more clarity, the market can be expected to apply a higher-than-normal discount to the range of possible outcomes

At yesterday's close, Ascelia was valued at a market capitalisation of around SEK 64m, which is close to the company's cash balance. Liquid assets stood at SEK 33.9m at the end of Q1 2026 (31 March), plus the SEK 20m direct share issue completed on 23 April. Cash burn was around SEK 16m in Q1 2026.

Looking at our valuation framework, the current share price corresponds to a market-implied probability of success of approximately 6% based on our base case. That base case rests on a royalty-based model, with an addressable market of around USD 800m, a 25% royalty rate and a 50% peak EBIT margin, discounted at 15%. Following the CRL, we have moved the assumed launch to 2028. 6% is measured against the de-risked value the model produces on that basis.

With the market starting to price in the most negative scenarios, it is worth looking at the different scenarios and what points to each, balanced between the full CRL letter and management's comments.

Scenario: Clarification and documentation

The FDA's concerns are resolved through clarification and the product-documentation fixes, without new analysis of the imaging data. This is the fastest path, with a resubmission inside the one-year window that runs to early July 2027 and approval potentially in 2027.

What points to this: the company describes the clinical concern as related to the image-reading process, states the product documentation matters should be resolved in the near term, and says its re-read approach was discussed with the FDA before submission. We would note, however, that a clarificatory resolution goes beyond clearing a general evidentiary bar. The FDA has already reached a negative conclusion on the existing analysis, finding that SPARKLE did not meet its pre-specified success criteria on the initial read and rejecting the subsequent re-evaluation on two specific grounds, the possible bias in the new readers' training and the scoring limited to T1-weighted images. Because the agency has rejected the method rather than asked a clarifying question, we see a resolution on clarification and documentation alone as the least likely of the faster outcomes, and the company's stated position that its re-read approach was discussed with the FDA before submission would need to carry substantial weight at the Type A meeting for this path to hold.

Scenario: A corrected re-read of the existing SPARKLE images

A new, prospectively defined and properly controlled re-evaluation of the images already acquired, scoring the complete set of sequences rather than T1 alone. A re-read of this kind, which we estimate at six to ten months, could still fit inside the resubmission window on a six-month review clock, pointing to approval around late 2027 or early 2028.

What points to this: Ascelia has confirmed the full sequence set was collected per protocol, and the two deficiencies, the FDA cites, reader training and the T1-only scoring, both concern how the images were read rather than the data itself. Importantly, the FDA recommends “a new adequate and well-controlled study,” and that phrase does not by its terms require a new Phase 3 trial. An adequate and well-controlled study is defined by its design characteristics, such as pre-specified endpoints, appropriate controls, blinding and reader training, rather than by whether the patient data is freshly collected. Because the imaging data already exists, a re-read that is prospectively designed to meet those characteristics, scoring the complete set of sequences under proper reader training, could in principle qualify as such a study without enrolling new patients. Whether the FDA would accept a re-read on that basis, rather than a new trial, is the central question the Type A meeting has to resolve.

Scenario: A new clinical trial

The most demanding reading of the letter would be an entirely new adequate and well-controlled trial with fresh patient data. This would push a realistic timeline to around 2029 to 2030 and require financing well beyond the current cash runway into 2027. We view it as the least likely of the scenarios to be pursued to completion, given the funding involved.

What points to this: several features of the letter pull towards the more demanding reading. First, a literal reading of “conduct of a new adequate and well-controlled study” can be taken to mean a fresh trial rather than a re-analysis of existing images. Second, the FDA rejected the earlier re-evaluation partly because it was post hoc, that is, conducted after the initial read had already failed, which raises the question of whether any further re-read of the same images can escape that objection, however well designed. Third, the agency has historically applied a stringent standard when judging whether a CRL response fully resolves the deficiencies cited, and here it reached a negative conclusion on effectiveness rather than requesting additional analysis. Taken together, these points mean a new trial cannot be ruled out, even if the specific fixes the FDA names point more towards the reading process than the underlying data.

Scenario: No viable regulatory path

No path forward is agreed at the Type A meeting, and Orviglance does not advance. In that case the valuation would rest largely on the cash position and any cost savings taken.

What makes us regard this as a tail scenario rather than a central expectation: the FDA's letter identifies specific, addressable deficiencies in the image reading and scoring rather than a fundamental problem with the drug's safety or mechanism, and it sets out a route back through resubmission within one year. The agency also recommends a concrete remedy rather than closing the file. On the company's side, management has stated it intends to pursue an expedited Type A meeting and remains confident in the dataset. A complete dead end would require the Type A dialogue to produce no workable path at all, which we see as less likely than one of the outcomes above, though it cannot be excluded if the funding required for the FDA's preferred route proves out of reach.

The company's press release states that manganese, the active substance in Orviglance, is a T1-enhancing compound, and that all planned MRI sequences, including T1, T2 and diffusion-weighted imaging, were acquired in line with the study protocol.

In our view, these two points together suggest the deficiency the FDA identifies is methodological, related to how the images were read and scored rather than in the conduct of the trial or the completeness of the data collected. Since the drug's contrast effect is expressed on T1 imaging by mechanism, and the full standard-of-care sequence set is already available in the dataset, a re-read designed to score lesions against the complete baseline could address the comparison the FDA is asking for without new patients. We would balance this by noting that the agency's underlying question, whether post-contrast imaging adds clinical value over non-T1 standard of care, is one the re-read design would still need to answer to the FDA's satisfaction.

Disclaimer: HC Andersen Capital receives payment from Ascelia Pharma for a Digital IR/Corporate Visibility subscription agreement. Michael Friis, 15/07-2026. kl 08:02