Deep learning in sex estimation from a peripheral quantitative computed tomography scan of the fourth lumbar vertebra-a proof-of-concept study.

Sex estimation is a key element in the analysis of unknown skeletal remains. The vertebrae display clear sex discrepancy and have proven accurate in conventional morphometric sex estimation. This proof-of-concept study aimed to investigate the possibility to develop a deep learning algorithm for sex estimation even from a single peripheral quantitative computed tomography (pQCT) slice of the fourth lumbar vertebra (L4). The study utilized a total of 117 vertebrae from the Terry Anatomical Collection. There were 58 male and 59 female cadavers, all of the white ethnicity, with the average age at death 49 years and a range of 24 to 77 years. A coronal pQCT scan was taken from the midway of the L4 corpus. Sex estimation was performed in a total of 19 neural network architectures implemented in the AIDeveloper software. Of the explored architectures, a LeNet5-based algorithm reached the highest accuracy of 86.4% in the test set. Sex-specific classification rates were 90.9% among males and 81.8% among females. This preliminary finding advances the field by encouraging and directing future research on artificial intelligence-based methods in sex estimation from individual skeletal traits such as the vertebrae. Combining quickly obtained imaging data with automated deep learning algorithms may establish a valuable pipeline for forensic anthropology and provide aid when combined with traditional methods.

Altered IGF-I activity and accelerated bone elongation in growth plates precede excess weight gain in a mouse model of juvenile obesity.

Nearly one-third of children in the United States are overweight or obese by their preteens. Tall stature and accelerated bone elongation are characteristic features of childhood obesity, which cooccur with conditions such as limb bowing, slipped epiphyses, and fractures. Children with obesity paradoxically have normal circulating IGF-I, the major growth-stimulating hormone. Here, we describe and validate a mouse model of excess dietary fat to examine mechanisms of growth acceleration in obesity. We used in vivo multiphoton imaging and immunostaining to test the hypothesis that high-fat diet increases IGF-I activity and alters growth plate structure before the onset of obesity. We tracked bone and body growth in male and female C57BL/6 mice (n = 114) on high-fat (60% kcal fat) or control (10% kcal fat) diets from weaning (3 wk) to skeletal maturity (12 wk). Tibial and tail elongation rates increased after brief (1-2 wk) high-fat diet exposure without altering serum IGF-I. Femoral bone density and growth plate size were increased, but growth plates were disorganized in not-yet-obese high-fat diet mice. Multiphoton imaging revealed more IGF-I in the vasculature surrounding growth plates of high-fat diet mice and increased uptake when vascular levels peaked. High-fat diet growth plates had more activated IGF-I receptors and fewer inhibitory binding proteins, suggesting increased IGF-I bioavailability in growth plates. These results, which parallel pediatric growth patterns, highlight the fundamental role of diet in the earliest stages of developing obesity-related skeletal complications and validate the utility of the model for future studies aimed at determining mechanisms of diet-enhanced bone lengthening.NEW & NOTEWORTHY This paper validates a mouse model of linear growth acceleration in juvenile obesity. We demonstrate that high-fat diet induces rapid increases in bone elongation rate that precede excess weight gain and parallel pediatric growth. By imaging IGF-I delivery to growth plates in vivo, we reveal novel diet-induced changes in IGF-I uptake and activity. These results are important for understanding the sequelae of musculoskeletal complications that accompany advanced bone age and obesity in children.

Great apes and humans evolved from a long-backed ancestor.

There is current debate whether the Homo/Pan last common ancestor (LCA) had a short, stiff lumbar column like great apes or a longer, flexible column observed in generalized Miocene hominoids. Beyond having only four segments, three additional features contribute to lumbar stiffening: the position of the transitional vertebra (TV), orientation of the lumbar spinous processes, and entrapment of lumbar vertebrae between the iliac blades. For great apes, these features would be homologous if inherited from a short-backed LCA but likely functionally convergent through dissimilar phenotypes if evolved from a long-backed LCA. We quantitatively and qualitatively analyzed human, ape, and monkey thoracic and lumbar vertebrae using 3D surface scanning and osteological measurements to compare spinous process morphology and sacral depth. We also used a large sample of hominoid vertebral counts to assess variation in the position of the TV and lumbosacral boundary. All extant hominoids modally place the TV at the ultimate thoracic. However, humans and orangutans place the TV at the 19th postcranial vertebral segment, whereas other apes place the TV at the 20th. Furthermore, chimpanzees, gorillas, and orangutans each have distinct patterns of spinous process angulation and morphology associated with lumbar stiffening, while human spinous process morphology is similar to that of longer backed gibbons, monkeys, and Miocene hominoids Morotopithecus and Pierolapithecus. Finally, chimpanzees are unique compared with other hominoids with a greater sacral depth facilitating lumbar entrapment, and there are differences among African apes with respect to the mechanisms governing variation in the lumbosacral boundary. These differences suggest that lumbar stiffening is convergent among great apes and that human bipedalism evolved from a more generalized long-backed ancestor. Such a model is more consistent with evidence of TV placement in Australopithecus.

Developmental identity versus typology: Lucy has only four sacral segments.

OBJECTIVES: Both interspecific and intraspecific variation in vertebral counts reflect the action of patterning control mechanisms such as Hox. The preserved A.L. 288-1 ("Lucy") sacrum contains five fused elements. However, the transverse processes of the most caudal element do not contact those of the segment immediately craniad to it, leaving incomplete sacral foramina on both sides. This conforms to the traditional definition of four-segmented sacra, which are very rare in humans and African apes. It was recently suggested that fossilization damage precludes interpretation of this specimen and that additional sacral-like features of its last segment (e.g., the extent of the sacral hiatus) suggest a general Australopithecus pattern of five sacral vertebrae. METHODS: We provide updated descriptions of the original Lucy sacrum. We evaluate sacral/coccygeal variation in a large sample of extant hominoids and place it within the context of developmental variation in the mammalian vertebral column. RESULTS: We report that fossilization damage did not shorten the transverse processes of the fifth segment of Lucy's sacrum. In addition, we find that the extent of the sacral hiatus is too variable in apes and hominids to provide meaningful information on segment identity. Most importantly, a combination of sacral and coccygeal features is to be expected in vertebrae at regional boundaries. DISCUSSION: The sacral/caudal boundary appears to be displaced cranially in early hominids relative to extant African apes and humans, a condition consistent with the likely ancestral condition for Miocene hominoids. While not definitive in itself, a four-segmented sacrum accords well with the "long-back" model for the Pan/Homo last common ancestor. Am J Phys Anthropol 160:729-739, 2016. © 2016 Wiley Periodicals, Inc.

First steps of bipedality in hominids: evidence from the atelid and proconsulid pelvis.

Upright walking absent a bent-hip-bent-knee gait requires lumbar lordosis, a ubiquitous feature in all hominids for which it can be observed. Its first appearance is therefore a central problem in human evolution. Atelids, which use the tail during suspension, exhibit demonstrable lordosis and can achieve full extension of their hind limbs during terrestrial upright stance. Although obviously homoplastic with hominids, the pelvic mechanisms facilitating lordosis appear largely similar in both taxa with respect to abbreviation of upper iliac height coupled with broad sacral alae. Both provide spatial separation of the most caudal lumbar(s) from the iliac blades. A broad sacrum is therefore a likely facet of earliest hominid bipedality. All tailed monkeys have broad alae. By contrast all extant apes have very narrow sacra, which promote "trapping" of their most caudal lumbars to achieve lower trunk rigidity during suspension. The alae in the tailless proconsul Ekembo nyanzae appear to have been quite broad, a character state that may have been primitive in Miocene hominoids not yet adapted to suspension and, by extension, exaptive for earliest bipedality in the hominid/panid last common ancestor. This hypothesis receives strong support from other anatomical systems preserved in Ardipithecus ramidus.