For many years, conventional MRI of the ankle was the traditional imaging study used to assess for ligamentous injuries; however, there are many situations preclude the utility of MRI as in case of presence of a cardiac pacemaker, in certain types of aneurysm clips and in cochlear implant; in such cases, CT arthrography would be the excellent solution.
Multi-detector CTA can directly visualize a lot of ankle ligaments, which develop anatomically as intra-capsular structures and are normally highlighted by joint fluid. Depending on this fact, most of ankle ligaments could be readily outlined by intra-articular contrast during arthrographic studies, and this highlighting can be easily observed in CTA due to its high spatial resolution. In CTA, the high attenuation of injected contrast medium, in adequately distended joint, could surround the following ligaments: ATFL, PTFL, DDL, and syndesmotic ligaments; therefore, they will appear as hypo-attenuated structures, highlighted by high density of injected contrast, and this will not only occur along their inner contours, and also along their outer surfaces. The other ligaments (SDL and CFL) are delineated through their inner surfaces only as they are tightly applied to the joint capsule; their injuries are diagnosed by demonstrating contrast leakage beyond them into adjacent tendon sheaths (i.e., into the tibialis posterior tendon sheath in case of SDL and into the peroneal tendon sheath in case of CFL). Furthermore, CTA is a well-established imaging technique for detection of chondral defects and OCL especially in small joints. CTA of the ankle is considered superior to MR arthrography in hyaline cartilage imaging being very thin in the ankle. MR arthrography has low spatial resolution, and it is unable to separate the closely apposed tibial and talar cartilages and could not detect partial thickness chondral defect. On the other hand, the inherent high resolution of CTA permits easy depiction of the articular cartilage along each articular surface across the joint, and the articular cartilage is clearly seen by virtue of five thin alternating layers of high and low attenuations formed by hyperdense subchondral bone plate, hypodense hyaline cartilage on each side of the joint, and intervening hyperdense contrast material [14,15,16,17].
The purpose of the current study was to evaluate the diagnostic value of CTA in post-sprained ankle not only for detection of chondral defects and osteochondral lesions, which are the traditional indications for CTA, but also for detection of concurrent ligamentous tears.
Most of cases in the current study were presented with chronic or subacute ankle sprains with persistent pain, among the study cases; there were 34 (68%) which had ligamentous injury detected on CTA, with some had multiple ligament tears. To the best of authors’ knowledge, there is lack of studies concerned with the prevalence of ligamentous injuries in sprain-related ankle pain. The high percentage of ligamentous tear found in the current study could be expected particularly in the context of chronic or recurrent sprains, and in such cases, suspicion of ligamentous injury should be raised and kept in mind, as unhealed tear will be a source of persistent ankle pain and/or leads to abnormal joint motion which in turn leads to ankle instability and development of a vicious circle of recurrent sprains.
Regarding the ligamentous dimensions in the current study, there was no significant difference between partially torn ATFL and intact ones regarding their ligament thickness value (both had mean of 2-mm thickness) although the lower range value was thinner in partially torn ATFL (0.3 mm) than in intact ligament (1.1 mm). The mean thickness of normal CFL was 1.9 mm. These data are in concordance with Dimmick et al. 2009, who studied the thickness of ATFL and CFL in normal and abnormal ankles; they utilized MRI in their study and found that the mean thickness of the ATFL was 2.19 ± 0.6 mm and the CFL measured 2.13 ± 0.5 mm, with statistically insignificant difference between normal and abnormal ankles. This statistically insignificant difference between partially torn ATFL and normal ones could signify the utility of the morphologic abnormalities found in torn ligament (such as uneven thickness, bowing, or laxity) as indicators for ligament injury rather than relying on its numerical value of ligament thickness [18].
In this study, there were no cases with high ankle or syndesmotic sprain, all cases demonstrated intact anterior inferior tibio-fibular and posterior inferior tibio-fibular ligaments, and the mean syndesmotic recess length was 13.66 ± 3.9 mm (range 3.5–23 mm). Multiple reports studied the syndesmotic ligaments, however with no thorough documentation or clear cutoff values between normal measurement and abnormal ones. Brown et al. 2004 studied ankle syndesmotic recess by MRI; they found an average recess height of 5 mm in normal syndesmosis and of 14 mm and of 12 mm in chronic and acute syndesmosis injury respectively; these values are seemingly discordant with our results but, this would be attributed to large standard deviation of their means. Moreover, Kim et al. 2007 studied syndesmotic ankle injuries using contrast-enhanced MRI and measured the length of enhancing syndesmotic recess, although they measured the enhanced tissue about the recess, and although they extended the height measurement distally to the talar dome, their results are partially concordant with the current study, they found that the mean height of enhancing tissue delineating the syndesmotic recess was 12.6 mm ± 5.1 in non-injured ankles whereas it was higher in injured ankle group (16.2 mm ± 3.3) with statistically significant difference. Bartonicek 2003 studied the anatomy of the tibiofibular syndesmosis and found that the expansion of contrast solution as far as 1.2 cm above the joint could not be considered an indicator of tibiofibular syndesmosis disruption or injury of the interosseous tibiofibular ligament; furthermore, he did not give cutoff value between normal and abnormal syndesmotic height measures [19,20,21].
Regarding frequency of chondral and OCL in the current study, there were 36 cases (72%) which had either chondral defect or OCL; some of which had more than one osteochondral lesions or multi-focal cartilage defect. The occurrence of such lesions in study cases could be expected especially when poor recovery from ankle sprain occurs; however, such high percentage (72%) of chondral and OCLs might be a little bit surprising, as the indication of CTA in our study was centered about post-sprained ankle pain especially the chronic one which showed a positive correlation with the presence of chondral defect/OCL. The clinical suspicion of chondral/OCL was not an indication for CTA in our study. Kirschke et al. 2016 studied the diagnostic value of CTA for evaluation of osteochondral lesions at the ankle, although the primary indication for CTA in all their cases was to visualize chondral or osteochondral defects; their results are almost in agreement with our study; and they found that 51/79 (64%) of patients had talar cartilage defects/OCL and 38/79 (48%) of patients had tibial cartilage defects/OCL. Furthermore, Kirschke 2016 considered CTA as a reference standard technique in order to compare their findings with those of MRI, and they found that MRI was able to visualize only 83.1% of cartilage defects and 63.6% of defects of the subchondral bone. In contrast, the current study did not combine CTA with MRI or MR arthrography due to some logistic reasons; it only utilized CTA, in part because of a very thin normal ankle hyaline cartilage and in part due to inherent high resolution of CT images in examining osseous injuries. The high percentage of chondral defect/OCL discovered in the study, together with their positive correlation with chronic ankle pain and with recurrent sprain, could highlight their importance; there should be a high level of expectance of their presence when chronic post-sprained ankle pain workup is initiated [7, 9, 22].
Study limitation
Although surgical interventions of some patients in our study allowed for assessment of positive cases, negative case assessment was not available, since operative intervention was primarily targeted to the CTA detected ligamentous tear, and the relatively little number of surgically treated cases is also a limitation. Moreover, although CTA might be considered as a reference for OCL, there was lack of ankle arthroscopy as a reference standard method, as it is not frequently performed in our institution.