MRI is considered gold standard for the diagnosis of presumptive acute hydrated non-compressive nucleus pulposus extrusions (AHNCNPE). This retrospective study describes the myelographic findings in dogs with AHNCNPE diagnosed by low-field MRI and their association with neurological grade, need of surgical decompression and outcome. Forty-two myelographies (21 dogs with presumptive AHNCNPE, 21 dogs with Hansen type I disc disease herniation) were blindly evaluated. Site of herniation, compression pattern, ratio of length of the lesion to length of the second lumbar vertebra (LL:L2) and degree of spinal cord compression (SCC) were measured on the myelographies of dogs with presumptive AHNCNPE and were compared with the corresponding MRI features. Percentage of extruded volume of nucleus pulposus (VNP) was calculated on MR images. Myelographic interobserver agreement for presumptive diagnosis of AHNCNPE was almost perfect (Îº=0.8). Accuracy of myelography to detect site of herniation was 80.9 per cent and to identify extradural compression was 57.1 per cent. Mean SCC was 5.8Â±2.6 per cent for myelography and 6.6Â±3 per cent for MRI. Mean LL:L2 ratio was 1.7Â±0.9 for myelography and 1.2Â±0.8 for MRI. Mean percentage of extruded VNP was 40Â±14 per cent, and it was positively associated with neurological grade.