CT-based assessment of lower limb surface area, volume, and tissue composition: Implications for ecogeographic rules of thermoregulation





Journal Title

Journal ISSN

Volume Title



Purpose: Anthropological research into human climatic adaptation has shown that global variation in skin surface-area to body volume (SA/V) ratio is generally consistent with theoretical predictions of ecogeographic rules. However, most studies have ignored the fact that internal body volume is comprised of various tissues (e.g., bone, muscle, fat) which exhibit different thermoregulatory properties. To address this issue, NSF has recently funded a research project in which human subjects will undergo full-body computed tomography (CT) scanning, and physiological testing during exposure to climatic extremes in an environmental chamber, to permit direct evaluation of associations between morphological variation and thermoregulatory physiology. Accordingly, the goal of the current project is to develop analytical methods for the NSF project, which will be used to 1) compare CT-derived SA/V ratios to traditional estimation methods, and 2) to evaluate potential volume differences in internal tissue compositions (e.g., bone, muscle, fat). Methods: 20 full body CT scans were selected from the New Mexico Decedent Image Database, a repository of CT scans made available to researchers by the New Mexico Office of the Medical Examiner. CT scan processing was performed in the Avizo software program. Appropriate anatomical landmarks were identified for segmentation of the hip, knee, and foot. The boundary between torso and lower limb is defined by a transverse plane between left and right greater trochanters. Separation of thigh and leg is defined by a similar plane between the medial and lateral midpoints of the knee on the tibial plateau. A final plane at the ankle passes through the medial and lateral malleoli. Identification of tissue types (bone, muscle, adipose) from the CT data began with generally accepted Hounsfield unit (i.e., density) ranges. These threshold ranges include +400 HU and higher for bone, -29 to +150 HU for muscle, and -190 to -30 HU for adipose. Results: Landmarks employed to orient planes of anatomical division were found to be readily identifiable across all specimens, permitting accurate collection of surface area and volume data for each of the 4 regions of interest (lower limb, thigh, leg, foot). In contrast, the use of basic Hounsfield unit threshold ranges was not found to permit reliable differentiation between bones, muscle, and adipose tissues. In particular, thresholding rages for bone had to be substantially altered to account for density-related differences between cortical bone, trabecular bones, and bone marrow. Hounsfield unit ranges for muscle and adipose did not need as much adjustment, but selections required more manual segmentation for contouring and removal of scanning artifacts. Conclusion: The methods developed by this project allow differences in tissue composition to be accurately accounted for during assessments of body volume. Thus, these methods provide advantages over traditional volumetric methods which assume tissue uniformity, and consequently, are well positioned to facilitate future experimental research into the relationships between anatomical variation and thermoregulatory physiology.