Introduction: Clinical experience has shown that addressing variations in bone morphology is important in the development of successful hip implant designs. Numerous studies of femoral bone morphology have been published utilizing various techniques. This study has developed a method which consistently measures large quantities of 3-dimensional digital femura geometry segmented from computed tomography (CT) scans and can accurately make anatomical measurements from these images

Methods: CT images of left femora on five hundred fifty six left femura (57% male, 43% female), consisting of 69% Caucasian, 16% Asian and 14% unknown were analyzed. The average age was 66 years, ranging from 40 to 93 years. Segmentation of the outer cortical, inner cortical, and marrow boundaries were consistently performed over all CT scans. The positions identified on the reference bone are transformed to the equivalent position on the clinical bone images, from which the dimensional data is extracted and stored. The mediolateral width (MLW), medial offset (MO) and lateral offset (LO) were measured in 10mm increments, ranging from 20mm above the lesser trochanter (LT) to 130mm below the lesser trochanter. The canal flare index was defined as a ratio of the mediolateral width at a section 20mm above the lesser trochanter to the mediolateral width at the isthmus level.

Results: The mean mediolateral width at 20mm above the lesser trochanter was 47.0 ± 4.5 (35.1–61.8; n=556). Noble reported 45.4 ± 5.3 (31.0–60.0; n=200), Husmann reported in a neck oriented study 46.3 ± 6.9 (27.6–63.6; n=310) and Laine reported 47.1 ± 4.9 (n=50). The mean medial offset at a section 20mm above the lesser trochanter was 25.1 ± 2.9 (16.7–33.4). In the study by Husmann, a mean of 25.0 ± 5.2 (9.4–45.5) was reported. The mean canal flare index was 4.49 ±.8. Noble reported a mean canal flare index of 3.80 ±.074, Husmann 3.81 ±.83 and Laine 4.3 ±.93.

Discussion: In general, the study showed minor differences to published data of proximal bone morphology. However, this more powerful study has shown that there is a higher mean canal flare index than determined by Noble and a similar mean canal flare index as determined by Laine. As reported by Laine, the canal flare index varies significantly with the placement of measurements in the canal. In this study the measurements were performed in a plane oriented by the femoral neck as a hip stem would be placed. The CFI over the isthmus width showed a greater correlation than previously shown by Noble. The novel software tool allows for anatomical measurements that can be applied to an unlimited population size enabling further applications and studies.