Imaging meets clinical judgment: lessons from the CT/MRI endpoints of the EMBOLISE trial

Subacute and chronic subdural hematomas (cSDH) represent some of the most common conditions managed by neurosurgeons worldwide, and its incidence is rising rapidly as populations age and antithrombotic use expands (1,2). Despite surgical evacuation remaining the standard of care for symptomatic disease, recurrence rates of up to 12% after burr-hole drainage or craniotomy have frustrated patients, families, and clinicians alike (3,4). This unmet need created fertile ground for middle meningeal artery embolization (MMAE), a minimally invasive procedure based on the theory that dural neovascularity sustains these hematomas, such that interrupting this supply breaks the cycle of microhemorrhage and membrane formation (5,6).

Over the past decade, MMAE has moved rapidly from isolated case reports to the subject of four large multicentre randomized controlled trials (RCTs): EMBOLISE, MAGIC-MT, STEM, and EMPROTECT (7-10). In November 2024, the EMBOLISE investigators published its primary results in the New England Journal of Medicine, reporting that adjunctive MMAE reduced the 90-day reoperation rate from 11.3% to 4.1% [relative risk (RR) 0.36; 95% confidence interval (CI): 0.11–0.80; P=0.008]. Now, Kniep et al. extend this work in Radiology, presenting the prespecified secondary imaging endpoints of EMBOLISE alongside an informative set of post hoc radiological analyses (11).

What the paper reports, and why the apparent paradox matters

The primary question posed by Kniep et al. is whether adjunctive MMAE produces measurable differences in hematoma volume, thickness, and midline shift compared with surgery alone, as assessed by quantitative imaging. At face value, the prespecified secondary imaging endpoint, capturing relative changes in these metrics between 24 hours and 90 days post-procedure, was not significantly different between groups. This contrasts with the clinical primary endpoint, which showed a dramatic reduction in reoperation rates.

The resolution to this apparent paradox lies in the details of the analysis and in the biological timing of MMAE’s mechanism of action. When Kniep et al. shift from examining relative change to comparing absolute hematoma volumes, a different picture emerges. At both 90 and 180 days, patients who received MMAE plus surgery had significantly lower absolute cSDH volumes—20.6 mL vs. 28.9 mL at 90 days (P=0.03) and 19.4 mL vs. 31.5 mL at 180 days (P=0.04) in the modified intention-to-treat cohort (11). Mixed-effects modelling confirmed an approximately 25–30% lower absolute volume at 90 days in the MMAE group. The clinical significance of these absolute differences, though individually modest in millilitres, is supported by lower midline shift and hematoma thickness observed at 180 days, aligning with reduced reoperation rates.

The prespecified endpoint measured change from 24 hours to 90 days—that is, the period after initial surgical decompression. However, MMAE likely exerts its principal radiological effect either intraoperatively, when performed before surgery, or during the immediate post-acute phase before the 24-hour imaging anchor point. In the parent EMBOLISE trial, embolization was performed after surgery in 78 of 185 treated patients (42.2%), not in the majority. Importantly, the treatment groups were already beginning to separate by the 24-hour scan: the MMAE-plus-surgery arm had lower midline shift and hematoma volume at that time point. The 24-hour scan is consequently an imperfect analytical anchor: it was obtained after surgical evacuation and variably in relation to MMAE, which was permitted within 48 hours of randomisation. The 24-hour measurements may therefore reflect a combination of baseline imbalance, surgical decompression, treatment timing, and potentially early treatment effects, and the existing data cannot determine which of these factors accounts for the observed early separation. A change-from-24-hour endpoint may therefore have attenuated a genuine between-group difference and biased the prespecified analysis toward the null. The post hoc analyses, by focusing on absolute volumes at fixed time points rather than change from an already-treated baseline, are more biologically revealing, and this methodological lesson should directly inform the design of future imaging endpoint frameworks in MMAE trials.

Clinical implications

The imaging findings reported by Kniep et al. serve an important function beyond confirming the primary endpoint. The EMBOLISE trial’s primary outcome, reoperation rate, has been criticised as subject to physician judgment bias, since treating neurosurgeons were not blinded to treatment assignment and held discretion over whether to reoperate. The authors themselves acknowledged this risk, implementing mitigation strategies including a requirement for documented clinical and radiological rationale at each follow-up visit. The imaging data in Radiology provide a complementary quantitative signal that is less dependent on retreatment thresholds, and they corroborate the direction and magnitude of clinical benefit, although they are not free from bias. Logistic regression also confirmed a strong association between both prespecified and post hoc imaging metrics and the requirement for reoperation, lending mechanistic plausibility to the clinical endpoint.

The corollary question—whether clinicians should order imaging surveillance routinely to guide treatment decisions after MMAE—is more nuanced. Rather than demonstrating that routine surveillance imaging adds little over clinical monitoring, this study mainly suggests that short-term change-from-baseline imaging may be an insensitive biomarker of embolization response. The benefit of MMAE in this trial appears best captured at fixed time horizons (90 and 180 days) using absolute volume rather than relative change metrics.

These imaging findings should be interpreted as evidence of biological activity rather than proof of patient-centred benefit. Hematoma volume is a surrogate endpoint, and the clinically decisive question is not whether MMAE shrinks the hematoma but how much volume reduction is sufficient to produce measurable improvement in function, symptoms, cognition, or quality of life—a threshold that has not been established. Across the major randomized trials, MMAE has not demonstrated superiority in modified Rankin Scale (mRS)-defined functional outcome or mortality, and reductions in length of stay, readmission, or quality-of-life burden have not been demonstrated. EMBOLISE did show fewer reoperations, an outcome that is clinically relevant but partly dependent on physician decision-making. The imaging analysis therefore supports anatomical-clinical concordance without establishing how much better patients feel or function. Two non-exclusive hypotheses may account for this dissociation: the achieved volume reduction may be insufficient to cross a clinically meaningful threshold, or global instruments such as the mRS may lack sensitivity to outcomes relevant to frail older adults. These remain hypotheses to be tested, not post hoc explanations that negate the null functional findings. Future trials should prespecify imaging thresholds linked to granular functional, cognitive, patient-reported, and health-resource outcomes.

Contextualising EMBOLISE within the broader RCT evidence

These findings must be read alongside the rapidly maturing landscape of MMAE RCT data. A recent systematic review and meta-analysis synthesising all four major trials (EMBOLISE, MAGIC-MT, STEM, and EMPROTECT) involving 1,774 patients confirms that MMAE significantly reduces treatment failure overall (RR 0.55; 95% CI: 0.41–0.74) (12). The strongest pooled signal is in non-surgically managed patients, where standalone MMAE produces a 64% RR reduction (RR 0.36; 95% CI: 0.22–0.60), with a corresponding number needed to treat (NNT) of approximately 4; however, this derives from subgroup rather than primary analyses. As a surgical adjunct, the pooled relative effect is more modest (RR 0.65; 95% CI: 0.48–0.89), reflecting heterogeneous treatment-failure composites that differ across trials in their inclusion of radiological recurrence, symptomatic progression, and reoperation. The NNT of 14 cited here is derived specifically from EMBOLISE’s 7.2% absolute reduction in 90-day reoperation (11.3% to 4.1%) and should not be interpreted as an average of trial-level NNTs, which would be inappropriate given the differing baseline risks and endpoint definitions across trials (7).

Notably, across all four trials, MMAE has not yet demonstrated a significant improvement in functional status as measured by the mRS or in all-cause mortality. The pooled analysis confirms no significant difference in the proportion of patients achieving favourable functional outcome (mRS 0–2; RR 1.01; 95% CI: 0.97–1.04) (12). The mRS may be insufficiently sensitive to some outcomes relevant to older patients with cSDH—including cognition, delirium, independence, and patient-reported wellbeing—and this insensitivity has been recognised in the geriatric literature, where cSDH is framed as a ‘sentinel health event’ whose trajectory is dominated by comorbid burden rather than the hematoma event alone (13,14). This possibility should be acknowledged as a hypothesis about under-detection, however, and must not be used to discount the null functional findings. None of the four trials reduced length of stay, readmission, or mortality, and most MMAE procedures were performed under general anaesthesia, with longer procedural times reported in STEM. At a population level, adding MMAE to the treatment pathway increases net procedural and anaesthetic burden rather than reducing it. For the individual patient in whom a reoperation is averted, that outcome carries intrinsic value; but whether the strategy produces net clinical benefit across the treated cohort, including effects on delirium, cognition, quality of life, and healthcare utilisation, remains unestablished.

The heterogeneity across trials regarding surgical subgroup benefit also warrants attention; for instance, MAGIC-MT showed nearly identical symptomatic recurrence rates in its surgical subgroup regardless of MMAE (RR 0.90; 95% CI: 0.50–1.64), while EMBOLISE demonstrated a reduction (12). Differences in MMAE timing likely contribute to this: in EMBOLISE, 42.2% of embolizations were performed post-operatively; in MAGIC-MT, 99.6% were performed pre-operatively (7,8). Whether this timing difference influences imaging-level outcomes, and whether a pre-operative strategy might produce more distinct 24-hour-to-90-day imaging changes, remains an open question that the current Radiology analysis flags but cannot resolve.

The guideline landscape reflects this evolving evidence base. National Institute for Health and Care Excellence (NICE) guidance (HTG706) currently recommends MMAE for cSDH only in research settings, citing uncertainty about which patients benefit given variable indications and techniques across the evidence base. European multidisciplinary consensus has endorsed MMAE as standalone therapy when surgery is precluded and as an adjunct in recurrent cSDH (15). More recently, the Society of Vascular and Interventional Neurology published updated guidance that moves toward endorsement of MMAE in selected patients, while still acknowledging residual uncertainty about optimal patient selection and follow-up (16). In this contested regulatory space, the Radiology imaging study strengthens the mechanistic plausibility that MMAE’s clinical benefit is accompanied by objectively measurable differences in hematoma resolution.

Safety: reassurance from imaging and clinical data

The safety profile of MMAE, as reflected in both the primary EMBOLISE publication and the imaging analysis, is generally reassuring. Serious adverse events directly attributable to MMAE occurred in 2.0% of treated patients at 30 days, including disabling stroke in 1.0% (7). No additional procedure-related events accrued between 30 and 180 days. The rate of stroke at 90 days did not differ significantly between groups (2.0% vs. 1.5%; P=0.72), and no ipsilateral vision loss was reported—a complication of particular concern given the anastomotic connections of the middle meningeal artery with the ophthalmic circulation. The pooled meta-analysis corroborates this favourable safety signal, with procedure-related complication rates of 0.6–3.0% across all four trials and no significant excess mortality attributable to MMAE (12).

The radiopacity of the Onyx embolic agent also introduces an unavoidable source of bias into imaging endpoint evaluation. Core laboratory readers assessing follow-up CT and MRI could not be blinded to treatment allocation, since the embolic material is distinctly visible on follow-up imaging. This limitation is acknowledged by both the original EMBOLISE investigators and by Kniep et al., and it is a structural issue that will affect any future trial using Onyx or similarly radiopaque agents (7,11). Development of embolic materials that are either isoattenuating on CT or selectively characterizable by MRI pulse sequences would represent a meaningful step toward better-blinded imaging studies.

Unresolved questions and future directions

Several important questions remain unanswered. The optimal timing of MMAE relative to surgical evacuation—pre- versus post-operative—remains undetermined from high-level prospective evidence. A recent multicentre retrospective study using propensity score matching found no significant difference in reoperation rates or technical success based on timing (17), but prospective data are lacking. The optimal embolic agent—liquid embolics such as Onyx or Squid versus particulate agents—has not been definitively established, though meta-analyses suggest a possible advantage for liquid embolics as upfront standalone therapy (18).

Radiological morphological subtypes of cSDH—separated, laminar, trabecular, or mixed density—have not been consistently incorporated into trial eligibility criteria or subgroup analyses, despite evidence that they may modify both recurrence risk and treatment response (19). Future trials should standardise the reporting of these features. Antithrombotic management represents another critical evidence gap; many cSDH patients require ongoing antiplatelet or anticoagulant therapy for cardiovascular indications, and optimal peri-procedural management remains contested in the retrospective literature (20,21).

On the therapeutic horizon, transarterial hematoma evacuation, combining MMAE with direct aspiration through a transarterial microchannel, may address MMAE’s principal limitation of slow radiological resolution and delayed mass effect relief (22). Preliminary feasibility data suggest that rapid and substantial hematoma reduction may be achievable in a single minimally invasive session, which, if confirmed in adequately powered trials, would substantially reframe the therapeutic algorithm. Further randomized evidence, including from MEMBRANE and OTEMACS, may help refine the role of MMAE across standalone and adjunctive pathways (23,24). Individual patient data meta-analyses will be important to resolve subgroup-level questions with adequate statistical power.

The cost-effectiveness of adjunctive MMAE remains unproven and should not be assumed. Because MMAE is added for every patient treated while reoperation occurs in only a minority, and because an MMAE procedure is not itself less costly than a conventional surgical evacuation, a treat-all adjunctive strategy adds upfront procedural, embolic-material, and anaesthetic expenditure to the pathway. Available randomised evidence has not demonstrated a reduction in length of stay or hospital readmission. The relevant question is therefore not whether avoiding a single reoperation saves money, but whether the savings from averted reoperations across a population exceed the cumulative cost of embolising all treated patients—a balance that is highly sensitive to baseline recurrence rates. Early economic modelling suggests this represents an institutional tipping point that is not universally favourable and depends critically on patient selection (25). Prospective registries embedding standardised volumetric imaging, explicit treatment decision rules, and health economic components will be essential to resolve these questions at scale.

An additional practical limitation of the current Radiology study is its analytic population. Of the 400 randomized patients in EMBOLISE, 352 were included in the imaging analysis; 11 unsuccessfully embolized patients were excluded from the modified intention-to-treat cohort, and post-retreatment 90- and 180-day radiologic outcomes were imputed from the latest available pretreatment imaging. These were reasonable analytic choices, but they should temper overinterpretation of the magnitude of the late imaging differences.

Conclusions

By demonstrating lower absolute hematoma volumes at 90 and 180 days in exploratory analyses, the CT/MRI endpoint study provides an anatomical signal consistent with an effect of adjunctive MMAE and directionally concordant with the lower reoperation rate observed in EMBOLISE. It does not establish that the observed volumetric differences improve function, quality of life, survival, hospital utilization, or cost-effectiveness. The imaging analysis therefore strengthens mechanistic plausibility but does not resolve whether the intervention produces a net benefit experienced directly by patients.

Taken alongside the broader randomized evidence, these findings support selective rather than universal use of MMAE. The strongest pooled efficacy signal is currently observed among non-surgically managed patients, whereas routine use as a surgical adjunct requires further stratification by recurrence risk and prospective evaluation of patient-centred and economic outcomes. Reported procedure-related complication rates have been low, although rare serious complications remain clinically important. Standardized imaging endpoints, outcome measures sensitive to this older and frequently frail population, and adequately powered comparisons of patient selection, timing, and technique will be necessary to determine when anatomical efficacy translates into meaningful clinical benefit.

Acknowledgments

None.

Footnote

Provenance and Peer Review: This article was commissioned by the editorial office, Journal of Neurointervention. The article has undergone external peer review.

Peer Review File: Available at https://jni.amegroups.com/article/view/10.21037/jni-26-0019/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jni.amegroups.com/article/view/10.21037/jni-26-0019/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Balser D, Farooq S, Mehmood T, et al. Actual and projected incidence rates for chronic subdural hematomas in United States Veterans Administration and civilian populations. J Neurosurg 2015;123:1209-15. [Crossref] [PubMed]
2. Stubbs DJ, Vivian ME, Davies BM, et al. Incidence of chronic subdural haematoma: a single-centre exploration of the effects of an ageing population with a review of the literature. Acta Neurochir (Wien) 2021;163:2629-37. [Crossref] [PubMed]
3. Almenawer SA, Farrokhyar F, Hong C, et al. Chronic subdural hematoma management: a systematic review and meta-analysis of 34,829 patients. Ann Surg 2014;259:449-57. [Crossref] [PubMed]
4. Maroufi SF, Farahbakhsh F, Macdonald RL, et al. Risk factors for recurrence of chronic subdural hematoma after surgical evacuation: a systematic review and meta-analysis. Neurosurg Rev 2023;46:270. [Crossref] [PubMed]
5. Edlmann E, Giorgi-Coll S, Whitfield PC, et al. Pathophysiology of chronic subdural haematoma: inflammation, angiogenesis and implications for pharmacotherapy. J Neuroinflammation 2017;14:108. [Crossref] [PubMed]
6. Fiorella D, Arthur AS. Middle meningeal artery embolization for the management of chronic subdural hematoma. J Neurointerv Surg 2019;11:912-5. [Crossref] [PubMed]
7. Davies JM, Knopman J, Mokin M, et al. Adjunctive Middle Meningeal Artery Embolization for Subdural Hematoma. N Engl J Med 2024;391:1890-900. [Crossref] [PubMed]
8. Liu J, Ni W, Zuo Q, et al. Middle Meningeal Artery Embolization for Nonacute Subdural Hematoma. N Engl J Med 2024;391:1901-12. [Crossref] [PubMed]
9. Fiorella D, Monteith SJ, Hanel R, et al. Embolization of the Middle Meningeal Artery for Chronic Subdural Hematoma. N Engl J Med 2025;392:855-64. [Crossref] [PubMed]
10. Shotar E, Mathon B, Salle H, et al. Meningeal Embolization for Preventing Chronic Subdural Hematoma Recurrence After Surgery: The EMPROTECT Randomized Clinical Trial. JAMA 2025;334:127-35. [Crossref] [PubMed]
11. Kniep H, Gellissen S, Bechstein M, et al. Middle Meningeal Artery Embolization for Subdural Hematoma: CT/MRI End Points of the EMBOLISE Trial. Radiology 2026;318:e251746. [Crossref] [PubMed]
12. Edward Liu J, Barras CD, Shaygi B, et al. A systematic review and Meta-Analysis of Middle meningeal artery embolization for Non-Acute Subdural Haematoma: A proposed treatment algorithm. J Clin Neurosci 2026;147:111917. [Crossref] [PubMed]
13. Rickard F, Gale J, Williams A, et al. New horizons in subdural haematoma. Age Ageing 2023;52:afad240. [Crossref] [PubMed]
14. Dumont TM, Rughani AI, Goeckes T, et al. Chronic subdural hematoma: a sentinel health event. World Neurosurg 2013;80:889-92. [Crossref] [PubMed]
15. Bartek J, Biondi A, Bonhomme V, et al. Multidisciplinary consensus-based statement on the current role of middle meningeal artery embolization (MMAE) in chronic SubDural hematoma (cSDH). Brain Spine 2024;4:104143. [Crossref] [PubMed]
16. Siddiq F, Shakir M, Nguyen TN, et al. Consensus Statement on Middle Meningeal Artery Embolization in Chronic Subdural Hematoma Treatment: A Guideline from the Society of Vascular and Interventional Neurology Guidelines and Practice Standards Committee. Stroke Vasc Interv Neurol 2025;5:e001814. [Crossref] [PubMed]
17. Sioutas GS, Salem MM, Kuybu O, et al. Order and Timing of Middle Meningeal Artery Embolization as a Perioperative Adjunct to Surgical Evacuation for Chronic Subdural Hematomas: A Multicenter Study. Radiology 2025;315:e241571. [Crossref] [PubMed]
18. Sioutas GS, Vivanco-Suarez J, Shekhtman O, et al. Liquid embolic agents for middle meningeal artery embolization in chronic subdural hematoma: Institutional experience with systematic review and meta-analysis. Interv Neuroradiol 2026;32:953-63. [Crossref] [PubMed]
19. Kan P, Fiorella D, Dabus G, et al. ARISE I Consensus Statement on the Management of Chronic Subdural Hematoma. Stroke 2024;55:1438-48. [Crossref] [PubMed]
20. Martinez-Gutierrez JC, Zeineddine HA, Nahhas MI, et al. Middle Meningeal Artery Embolization for Chronic Subdural Hematomas With Concurrent Antithrombotics. Neurosurgery 2023;92:258-62. [Crossref] [PubMed]
21. Zeineddine HA, Dawes BH, Wroe WW, et al. Middle Meningeal Artery Embolization Does Not Confer Protection against Symptomatic Recurrence in Patients with Early Antithrombotic Therapy Following Surgery for Chronic Subdural Hematomas. J Neuroendovasc Ther 2025;19:2025-0013. [Crossref] [PubMed]
22. Lylyk P, Lylyk P, Lylyk I, et al. O-019 A safety and feasibility clinical trial of middle meningeal artery embolization and transvascular drainage of non-acute subdural hematomas. J Neurointerv Surg 2025;17:A18-9.
23. Hoenning A, Lemcke J, Rot S, et al. Middle Meningeal Artery Embolization Minimizes Burdensome Recurrence Rates After Newly Diagnosed Chronic Subdural Hematoma Evacuation (MEMBRANE): study protocol for a randomized controlled trial. Trials 2022;23:703. [Crossref] [PubMed]
24. Ng S, Derraz I, Marnat G, et al. Embolization of the middle meningeal artery for chronic subdural hematoma: The OTEMACS multicenter, randomized, clinical trial protocol. Interv Neuroradiol 2025;15910199251356098. [Epub ahead of print]. [Crossref] [PubMed]
25. Debs LH, Rahimi SY, Vale FL. Tipping point in middle meningeal artery embolization: a cost-effectiveness and algorithm-based analysis. Neurosurg Focus 2025;59:E10. [Crossref] [PubMed]