Study examines whether MRI findings predict AAI surgery outcomes: full analysis

A new retrospective study in Animals asks a practical question for clinicians managing atlantoaxial instability in toy and small-breed dogs: can preoperative MRI abnormalities help predict which dogs will do well after surgery? The answer, based on this 20-dog cohort, is mixed. MRI findings tracked with neurologic severity before surgery, but most did not reliably predict postoperative improvement. The strongest signal was around perioperative risk, with dogs that died around the time of surgery showing higher ventral compression index values than survivors. (preprints.org)

That matters because AAI remains one of the more challenging congenital cervical disorders in small dogs. It typically affects toy breeds at a young age, can range from cervical pain to nonambulatory tetraparesis, and often requires advanced imaging to define the lesion and guide stabilization planning. Review literature and specialty guidance have long supported surgery for dogs with neurologic deficits or persistent pain, but outcomes vary, and complications, including respiratory arrest and implant-related failure, remain part of the risk discussion. (pubmed.ncbi.nlm.nih.gov)

In the new study, investigators reviewed 28 cases and analyzed 20 dogs after excluding five postoperative deaths, two dogs lost to follow-up, and one dog with incomplete records. They assessed MRI-derived variables including ventral compression index, sagittal and transverse spinal cord compression, atlantoaxial bands, and ventricular enlargement. Greater ventral compression correlated with more severe neurologic deficits before surgery, but not with postoperative neurologic grade or degree of improvement among the dogs included in the final analysis. Sagittal spinal cord compression did show a significant negative correlation with neurologic grade at eight weeks post-op, suggesting some MRI findings may still carry limited prognostic value depending on the endpoint measured. (preprints.org)

The study also looked beyond the primary compression site. Atlantoaxial bands were identified in five dogs, but dogs with dorsal compression related to those bands did not show statistically different neurologic improvement. Lateral ventricular enlargement was common, seen in 15 dogs, yet ventricular measures were not associated with preoperative severity, postoperative status, or degree of improvement. That finding is notable given the broader literature on craniocervical junction abnormalities in small dogs, where concurrent malformations are common, but their direct clinical significance is often uncertain and may not independently change outcome. (preprints.org)

There does not appear to be a press release or formal outside commentary tied to this paper, but the findings line up with the direction of prior work. Earlier retrospective data in 46 surgical AAS cases found that age at onset, duration of clinical signs, and preoperative neurologic status affected the odds of success, while some technical and radiographic variables did not. Specialty references also continue to emphasize that MRI is especially useful for evaluating spinal cord parenchymal injury and concurrent abnormalities, rather than serving as a standalone crystal ball for recovery. Taken together, the emerging message is that imaging helps define risk and case complexity, but it should be interpreted alongside neurologic exam findings and clinical course. (researchgate.net)

Why it matters: For veterinarians, this study may help refine preoperative conversations with pet parents. MRI abnormalities can support diagnosis and surgical planning, and higher ventral compression may flag dogs presenting with more severe disease or possibly greater perioperative risk. But the paper does not support overpromising prognostic precision from MRI alone, especially given the small sample and the exclusion of postoperative deaths from the main outcome analysis. In practice, that reinforces a multimodal approach to prognosis: neurologic grade, chronicity, respiratory and anesthetic risk, concurrent craniocervical abnormalities, and surgical expertise likely remain as important as any single imaging metric. (preprints.org)

The study also highlights a familiar issue in referral neurology and surgery: the dogs most at risk may be the hardest to capture cleanly in retrospective outcome datasets. Because perioperative deaths were excluded from the final improvement analysis, the apparent disconnect between MRI findings and postoperative recovery may understate the clinical importance of severe compression in the full surgical population. That’s an inference, not a direct conclusion of the paper, but it is supported by the authors’ finding that nonsurvivors had higher ventral compression index values. (preprints.org)

What to watch: The next step is prospective, multicenter work with larger case numbers, standardized MRI positioning, and inclusion of perioperative mortality in primary outcome models, which could clarify whether MRI can move from descriptive tool to validated prognostic aid in dogs with AAI. (preprints.org)