{
  "abstract": "Introduction Preclinical models are critical in development and evaluation of novel mechanical thrombectomy (MT) devices for acute ischemic stroke (AIS); yet reported efficacy and safety outcomes in these models often do not translate clinically. While several studies have reviewed in vivo models evaluating thrombectomy, comparisons of predictive performance across in vivo models and their translational relevance remains limited. Aim: This is a first systematic literature review and meta-analysis to compare performance across models, assess how effectively outcomes from MT device testing in in-vivo models translate to clinical practice, and provide insights into experimental design.Methodology A systematic search was performed using Pubmed, Scopus and Ovid MEDLINE on Nested Knowledge to identify preclinical studies evaluating MT within in vivo models. Studies lacking extractable efficacy/safety outcomes or not in vivo models were excluded. Clinical benchmarks were extracted from trials included in the HERMES meta-analysis, representing standard MT clinical evidence base. The primary outcome assessed recanalization rates, mTICI≥2b. Data was extracted on vessel characteristics, clot properties, procedural efficacy (first-pass effect (FPE), time to recanalization), and safety outcomes (hemorrhagic, thromboembolic, vasospasm, mortality, vessel injury events). Results were compared using Means, with p<0.05 considered statistically significant, and analyses were performed using Graphpad based on normality.Results A total of 1160 articles were screened; 41 studies were included. Swine and canine models were the most used platforms. Canine models demonstrated slightly higher mTICI≥2b (98.6% vs 91.6%, p>0.05). Vessels used per model, diameters and their reported human occlusion correlates were aggregated across studies and summarised to streamline new model design. Average vessel diameters were 3.06 mm in swine and 2.23 mm in canine, with no significant differences in clot dimensions between models. FPE (77.8% vs 77.3%), mean number of attempts (1.83 vs 1.78) and time to recanalization (20.6 vs 26.1 minutes) were comparable (p>0.05) across both groups. As for safety, canine models demonstrated higher thromboembolic events (28.6% vs 4.6%, p<0.05). Swine models reported higher vasospasm (42.2% vs 31.3%) and vessel wall injury, though not statistically significant. Compared with clinical benchmarks, both models overestimated clinical mTICI≥2b (71.2% vs 91.6-98.6%, p<0.05). Similarly, hemorrhagic events (16.2%) and mortality (14.5%) were reported clinically but were almost negligible in both models. Vasospasm was more frequent in animal models (31-42% vs 9%, p<0.05). Swine models closely approximated clinical observations (4.6% vs 6.0%, p>0.05) for thromboembolic events unlike canine models. Thrombectomy was performed 5-180 minutes post-occlusion, shorter than the 6-8 hour treatment window accepted in clinical guidelines. Furthermore, no studies reported functional outcome measures comparable to the modified Rankin Scale, which remains a primary endpoint in clinical trials.Conclusion Preclinical thrombectomy models provide valuable testing platforms but demonstrate important translational limitations. Both swine and canine models overestimate recanalization efficacy while failing to capture key clinical complications such as hemorrhage and mortality. Qualitative assessment highlights limitations in comparable treatment time and absence of functional endpoints. These findings highlight the need for improved model design, standardized outcome reporting, and incorporation of clinically relevant endpoints to enhance the translational validity of preclinical thrombectomy testing.Disclosures N. Silva: None. N. Moreton: None. P. Brouwer: None. C. Ulfert: None.",
  "authors": [
    {
      "affiliations": [
        "Medical Affairs, Johnson&Johnson Neurovascular MedTech, Galway, Ireland"
      ],
      "name": "N Silva"
    },
    {
      "affiliations": [
        "Medical Affairs, Johnson&Johnson Neurovascular MedTech, Galway, Ireland"
      ],
      "name": "N Moreton"
    },
    {
      "affiliations": [
        "Neuro Thromboembolic Initiative, Cerenovus, Galway, Ireland"
      ],
      "name": "M Mirza"
    },
    {
      "affiliations": [
        "Medical Affairs, Johnson&Johnson Neurovascular MedTech, Galway, Ireland"
      ],
      "name": "P Brouwer"
    },
    {
      "affiliations": [
        "Medical Affairs, Johnson&Johnson Neurovascular MedTech, Galway, Ireland"
      ],
      "name": "C Ulfert"
    }
  ],
  "title": "E-033 Preclinical to clinical translation in mechanical thrombectomy: a systematic review and meta-analysis of in vivo stroke models compared with clinical outcomes",
  "uid": "75463359-7e2c-596d-99b4-1e620f39335b"
}