New classification of spinal lipomas based on embryonic stage.

OBJECTIVE Spinal lipomas are generally thought to occur as a result of failed primary neurulation. However, some clinical features cannot be explained by this theory. The authors propose a novel classification of spinal lipomas based on embryonic changes seen during primary and secondary neurulation. METHODS A total of 677 patients with occult spinal dysraphism underwent 699 surgeries between August 2002 and May 2015 at the National Center for Child Health and Development and Tokyo Metropolitan Children's Medical Center. This group of patients had 378 spinal lipomas, including 119 conus spinal lipomas, 27 lipomyelomeningoceles, and 232 filum lipomas, which the authors classified into 4 types based on neural tube formation during embryonic development. Type 1 is defined as pure primary neurulation failure; Type 2 ranges from primary to secondary neurulation failure; Type 3 consists of secondary neurulation failure (early phase); and Type 4 is defined as secondary neurulation failure (late phase). The authors also review embryogenesis in secondary neurulation and analyze the clinical utility of the new classification. RESULTS There were 55 Type 1 spinal lipomas, 29 Type 2, 62 Type 3, and 232 Type 4. All filum lipomas fell into the Type 4 spinal lipoma category. Association with anorectal and/or sacral anomalies was seen in none of the Type 1 cases, 15 (52%) of Type 2, 35 (56%) of Type 3, and 31 (13%) of Type 4. Urogenital anomalies were observed in none of the Type 1 or Type 2 cases, 1 (2%) of Type 3, and 28 (12%) of Type 4. Anomaly syndromes were present in none of the Type 1 cases, 6 (21%) of Type 2, 3 (5%) of Type 3, and 16 (7%) of Type 4. Associated anomalies or anomaly syndromes were clearly observed only for Type 2-4 spinal lipomas encompassing failed secondary neurulation. Radical resection was feasible for Type 1 spinal lipomas. CONCLUSIONS Secondary neurulation of the spinal cord gives rise to the conus medullaris and filum terminale, which are often involved in spinal lipomas. Formation of spinal lipomas seems to be a continuous process overlapping primary and secondary neurulation in some cases. Association with other anomalies was higher in Type 2-4 spinal lipomas, which included failed secondary neurulation, than in Type 1 lipomas, with failed primary neurulation. On the other hand, radical resection was indicated for Type 1, but not for Type 2, spinal lipomas. The new classification of spinal lipomas based on embryonic stage has the potential for clinical use and agrees well with both clinical and surgical findings. The classification proposed here is still preliminary. Further studies and verification are necessary to establish its clinical utility.