Background: The frequency and arrangement of the subsegmental branches of the third portion of the vertebral artery (V3) have been assessed in small samples by autopsy, but they have not been assessed by computed tomography angiography (CTA). To determine the frequency and arrangement of subsegmental branches of V3 by CTA and to analyze the interrelationships among frequency, arrangement, sex, and side-to-side.Methods: First, the radiology records of 668 consecutive patients who underwent cervical or craniocervical CTA scans from October 1, 2017 to October 31, 2019 were retrospectively retrieved. Second, the four demarcation points were ascertained to define the three subsegments by reviewing the resource images, namely, the vertical portion of V3(V3v), horizontal portion of V3(V3h), and extradural portion of V3(V3e).Then, the numerical value and the location of the bifurcating branch in each subsegment of V3 were recorded. Third, the frequency and arrangement of the branches was assessed, and the relationships between sex, laterality and frequency and arrangement were analyzed as well as the interobserver performance.Results: On the left, 25%, 20.21%, and 6.59% were the frequencies of one branch in the V3v, V3h, and V3e, and 1.05%, 0.00%, and 0.45% were the frequencies of two branches, respectively. On the right, 25.04%, 17.07%, and 6.44% were the frequencies of one branch in the V3v, V3h, and V3e, and 1.04%, 0.3%, and 0.15% were the frequencies of two branches, respectively. There were no differences between the side-to-side and numerical values of the branches according to the t-test (P=0.4341, P=0.7968), and there were no differences between the side-to-side variable, sex variable and number of branches according to the t-test (P=0.4474, P=0.3593). There were no differences between the side-to-side and eight arrangements (permutation of 000; 100; 110; 111; 010; 011; 001; 101) by using the two-sample KS test (P=0.942), and there were no differences between sex and the eight arrangements according to the two-sample KS test (P=0.9973, P=0.8519). The interobserver reliability was excellent (Spearman's ranked correlation: 0.9927).
Conclusions:The frequency and arrangement of V3 subsegmental branches could be displayed by source imaging with CTA, and there were no significant differences according to the sex or side of the individual. It was imperative to acquaint the subsegmental branches before the operation involving V3 in the craniocervical junction to determine the surgical approach and reduce bleeding during the surgical procedures.