Influence of the image levels of distal femur on the measurement of tibial tubercle-trochlear groove distance—a comparative study

Based on reliable measurements of the TT-TG distance across multiple image levels of the distal femur, we found that the levels had a significant influence on the results when cartilaginous landmarks were used. The influence was not significant when the values were based on osseous landmarks, with both MRI and CT measurements. However, large deviations of the TT-TG distance between levels could still be observed in some individual subjects. These findings supported our hypothesis that the image levels of the distal femur might affect the results of the TT-TG measurements.

The examination as well as the measurement techniques of the TT-TG distance has been evolving over the past decades. The TT-TG distance was initially described by Goutallier et al., who suggested the use of an axial radiograph taken at 30° of knee flexion with neutral axial rotation for the measurement [4]. With the development of CT, which was able to provide high-quality axial images, CT measurements prevailed in the following years. As suggested by Dejour et al., the CT scans were conducted with the knee in extension instead of 30° of flexion [2]. Early CT measurements were performed upon the so-called maximum intensity projection, which was reformatted from the original cross-sectional images in the way that all anatomical structures were superimposed [2, 7]. As this method had been criticized for sometimes difficultly detectable anatomical structures and low intra- and inter-rater reliability [20], new measurement approaches had been proposed. These new methods were commonly featured with the “two-slice measurement,” in which the bottom of the trochlear groove and the posterior tangent of the femoral condyles were identified based on one image slice selected from the range of the distal femur; the tip of the tibial tubercle was similarly identified on another image slice from the proximal tibia; measurements were conducted based on the two separate images rather than the all-in-one superimposed geometries [21]. The use of MRI has been prevailing in TT-TG measurement in recent years, owing to its additional benefits such as low radiation and superior soft tissue visualization. The two-slice measurement was still maintained as the major technique in the MRI measurements, as shown in previous studies [5, 14, 16, 19, 22]. Nevertheless, the disadvantage of the two-slice measurement has been noted by previous author. Lustig et al. inferred that the axial views of the knee might be a contributor to the measurement bias observed in their study, since choosing two identical axial views of exactly the same level at two separate measurements was difficult to achieve. This issue not only existed in the processes of measurement but also existed in the CT or MRI scanning when the axial images were generated [23].

As currently documented in the literature, various criterions have been held for the selection of image levels. These differing criterions may induce errors to the measurements, leading to less comparable results between studies. For example, according to our data, the most proximal and the most distal slices showing a complete cartilaginous trochlea had an average vertical distance of 16 mm; the average TT-TG distances measured based on cartilaginous landmarks at these two levels were 13.95 and 11.95 mm, respectively, which were significantly different (P??0.001). Although no significant difference in the measurements across levels was found when osseous landmarks were used, notable deviations between values as high as 5 mm could still be observed in some individual subjects. Therefore, a standardized and consistent criterion for the selection of image levels would be beneficial for the TT-TG measurements. According to our results, the interval of 0 through 4.5 mm below the trochlear entrance (approximating the proximal levels) had the greatest deviation to the mean TT-TG distance; in contrast, the interval of 6 through 9 mm below the trochlear entrance (approximating the mid levels) was the least variable. From this point of view, the latter may be a favorable range for the TT-TG measurements; yet, cautions should be taken when the former is employed.

We also noted an essential mismatch in the TT-TG distances of the cartilaginous-MRI group and the osseous-MRI group. From the proximal end to the distal end, the osseous landmarks lead to values closely around the average, yet the cartilaginous landmarks lead to a stable decline in the values. The cartilaginous-MRI group came with significantly higher values than the osseous-MRI group at most levels. The inconsistency of the cartilaginous and osseous measurements has also been reported by previous authors [5, 13, 14]. These findings may be explained by the mismatch in the shapes of the cartilaginous and osseous trochlea. As revealed in previous studies, the thickness of the cartilage was unevenly distributed in the trochlear groove, which resulted in an offset of the deepest points of the cartilaginous and the osseous trochlea, further leading to deviations in the dependent TT-TG distances [24, 25].

We acknowledge the limitations in the current study. The CT and MRI groups involved different subjects; with the same subjects involved, the results could have been more confident to show the potential difference between imaging modalities. Besides, the subjects involved were all healthy, with normal patellofemoral development. However, the TT-TG measurements were frequently conducted in patients with dysplastic trochlea. One can expect that the results observed in our study might be different when it comes to dysplastic trochlea. Moreover, the subjects were all drawn from Asian population; future investigations based on Caucasian population are still in favor to provide more complete data.