RESUMO
BACKGROUND: At present, increasing the accuracy of identification of knee ligament insertions is fundamental in developing accurate patient-specific three-dimensional (3D) models for preoperative planning surgeries, designing patient-specific instrumentation or implants, and conducting biomechanical analyses. The accuracy and reproducibility of magnetic resonance imaging (MRI) scans in identifying cruciate and collateral ligament insertions have not been investigated thus far, despite their wide use. This study aimed to define and validate a method for this purpose. METHODS: First, the femur and tibia bones were oriented. Then, the sites of cruciate and collateral ligament insertions were identified. The inter- and intra-class correlation coefficients (ICCs) were calculated after multiple operators were applied to the images of different patients. The effect of the quality of the available MRI scans on the accuracy of identifying ligament insertions was also investigated. Moreover, non-expert operators were also found to be capable of performing the procedure after a video tutorial, and its efficacy was tested. RESULTS: The mean intra-observer variability was always <1.5mm for all landmarks, whereas the mean inter-observer variability was always <2.5mm except for the medial collateral ligament localized on the tibia (6.7mm). The ICCs showed good results, up to 0.99. A lower image quality had no noticeable effect on the procedure for identifying cruciate ligaments. Video tutorials were found to improve the ICCs up to 45%. CONCLUSIONS: This procedure is suitable for the univocal and accurate identification of cruciate and lateral collateral ligaments on MRI scans. CLINICAL RELEVANCE: This study demonstrates that MRI scans are highly suited to identifying cruciate ligament insertions.