**4. Discussion**

This study investigated the anatomical level of the conus medullaris and analyzed factors associated with the conus medullaris levels in 629 healthy volunteers. In the present study, the majority (92.2%) of the participants had the conus medullaris at the caudal level of the T12-L1 disk, and the conus medullaris was located cranially to the T12 vertebral level in only 7.8% of the participants. Among them, the T12-L1 disk and L1 vertebral body were the most common conus medullaris levels, which were 32.4% and 45.8%, respectively.

Our study demonstrated that shorter height and smaller PI were significantly associated with a caudally placed conus medullaris. This result might indicate that the length of the spinal cord varies little among individuals and that the skeletal difference affects the conus medullaris level. In addition to height, PI was a key driver of the conus medullaris level. Individuals with a larger PI typically have greater LL and thoracic kyphosis, and the end of the spinal cord might be located more cranially in the twisted spinal canal. However, as far as we know, there is no paper showing the relationship between the PI and the conus level due to the lack of studies investigating the conus level by using both lumbar MRI and X-rays. For this reason, the current results will need to be verified in future studies.

The location of the conus medullaris varies by developmental stage [6,8]. At birth, the cord fills the vertebral canal and terminates at the lumbosacral junction [8]. The distal end of the spinal cord then moves toward the cranial direction with infant development [6,8], probably because of the differential growth between the spinal column and spinal cord. In adults, the tip usually terminates at the mid aspect of the L1 vertebra. However, its position varies between the lower 11th thoracic and upper third lumbar vertebrae [5]. In a cadaveric study, the spinal cord measured roughly 45 cm in the adult male and 42 cm in the adult female [10]. The current results might indicate that the variation in spinal cord length is limited, and the skeletal anatomy of height and spino-pelvic sagittal alignment varies among individuals.

PI is one of the most important radiographical parameters in the case of spinal sagittal alignment [11]. The PI increases during childhood as the spine adapts to bipedal walking and stabilizes after adulthood [12]. PI strongly correlates with LL through the sacral slope (SS), and the larger PI is associated with a larger LL. Despite its grea<sup>t</sup> importance, PI varies from 33◦ to 85◦ among adults [13] and largely affects spinal sagittal alignment. Recent retrospective studies suggested that distal LL (L4-S1) is comparable between low to moderate and high PI groups. Proximal LL (L1-L4), however, is significantly influenced by the PI value (greater PI, and greater proximal lumbar lordosis) [14,15]. Furthermore, not only does the LL magnitude increase in cases of a larger PI but also the LL apex and inflection point are located more toward the cranial side [14]. Thus, in cases with a large PI the local lordosis around L1, where the conus medullaris is often located [1–6], might be greater, and the conus medullaris might be located more toward the cranial side in the twisted spinal canal. The present study did not measure local sagittal alignment around L1, and so this discussion is only speculative. The relationship between PI and conus medullaris needs to be further investigated.

#### *Strengths and Limitations*

A strength of this study was that it was a relatively large-scale study including ≥50 individuals of each sex and decade of age (20s–70s). Furthermore, both MRI and Xp were obtained in all subjects. As a limitation of the current study, the participants were a single race of Japanese. This limitation might affect the size and place of the spinal column and spinal cord. An international large-scale multicenter study is warranted to validate our results. As a second limitation, cases with lumbosacral transitional anomalies were excluded in the current study, however, it is necessary to examine the level of conus in these cases of transitional vertebra in the future. Lastly, we could not compare spinal alignment and the level of conus medullaris by degrees of pain, although the degrees of pain might affect the results. Future detailed studies assessing the pain are needed.
