[Quantitative 3D anatomy of the lumbar spine].

PURPOSE OF THE STUDY: There are many anatomic descriptions of the spine, but most concentrate on qualitative information. Quantitative data is however important to achieve a better clinical approach, to adapt implant size and to construct geometric models of spine mechanics.

MATERIAL AND METHODS: We examined 32 dry spines (160 lumbar vertebrae) obtained from the Orfila Museum anatomy laboratory at the Saints-Pères School of Medicine in Paris. We drew 178 landmarks on the surface of each vertebra and recorded the 3D spatial coordinates of each point using a Fastrack electromagnetic system operating with +/- 0.2 mm precision. The coordinates of the digitalized points were expressed in a local x-y-z axis field (x=posteroanterior axis, y=right-left axis, z=caudocranial axis). The origin O was half way between the "centers" of the vertebral plates. After calculating 112 linear, angular and surface parameters, results were analyzed with the Statview statistics system.

RESULTS: All parameters exhibited gaussian distribution. The transpedicular vertebral depth, corresponding to the maximal penetration of a pedicular screw before touching the anterior wall, was nearly constant: 48 mm (mean). The mean height of the pedicle was approximately 16 mm for L1 to L4 and 21 mm for L5. Pedicle width was 7 mm for L1 and L2 then rapidly widened to reach 10 mm for L5. It was noteworthy that the narrowest pedicle (4 mm) was found in 10% of the L1 vertebrae. There was an exponential rise in the sagittal tilt of the pedicles from L1 to L5, measuring approximately 8 degrees for L1 and rising to 24 degrees for L5.

DISCUSSION: Our sample of human spines with unknown clinical characteristics (age, sex) is representative of anatomy laboratory populations, generally composed of subjects over 40 years of age, and is thus adapted for studies of the degenerative spine. Our findings are generally in agreement with data reported in the literature and also provided complementary quantitative data concerning the transpedicular vertebral depth that was found to be a rather constant feature of the lumbar spine. It measured between 40 and 56 mm for 95% of the study population. The dimensions of the pedicle is particularly important: the width must be known to determine the size of pedicle screws; it measured between 7 and 12 mm in 95% of the population. The largest mean cross section of the pedicle was found for L5 (82mm(2)), but measured less than 60 mm(2) in 10% of the vertebrae, suggesting predisposition to spondylolysis.