Injury of the anterior cruciate ligament (ACL) is a common sports injury of the young, physically active adult population. The poor healing capacity, as well as the need of an anterior tibial restrain in order to return to preinjury sporting activity level, made the ACL the most frequently reconstructed ligament in the human body1 . The techniques for reconstruction have constantly evolved over the last two decades.

Trans tibial tunnel drilling and the double bundle approach have not proved superior to the anatomic single bundle using hamstrings or bone-patellar tendonbone autografts (B-PT-B)2 .

Current trends aim, not only to restore the position and footprint of the native ACL, but also the shape and biomechanical function. Recent anatomical research3 found the shape of the ACL mid-substance to be flat, with a ribbon like appearance and broad, fan like expansions on both tibial and femoral insertions. Another study4 showed that conventionally prepared quadrupled hamstring autografts had a mid-portion cross-sectional area more than 20% larger than the native ACL.

These are cadaveric studies that do not account for the potential remodeling that may occur during healing. We therefore questioned whether the in vivo shape of the healed graft will also significantly differ from the native ACL, as determined by cross-sectional area and width-thickness ratio on postoperative MRIs. Patients and Methods We performed bilateral MRI examinations on 12 asymptomatic patients (3 females) with unilateral ACL reconstruction. Six patients had quadrupled hamstrings neoligaments and 6 had reconstructions using B-PT-B ipsilateral autografts for an average of 2.66 (range 1-5) and 3.33 (range 1-5) years respectively. All patients were operated by the same surgeon in the same location using an anatomical single bundle technique1 .

After initial arthroscopic exploration, the ipsilateral autografts were harvested and prepared by a resident on the back table. The bonepatellar tendon-bone was harvested from the mid portion using a longitudinal incision. The semitendinosus and gracilis were harvested through a small oblique incision over the pes anserine bursitis using closed stripper. Five patients had a partial meniscectomy: four of the medial and one of the lateral meniscus while 6 patients had grade 1 or 2 Outerbridge chondral lesions. The femoral tunnel was drilled first through the anteromedial portal with the aperture centered in the native footprint. The tibial tunnel intra articular aperture was centered on the remnant stump and the bony attachment of the anterior horn of the lateral meniscus5. Femoral fixation used either absorbable interference screws (2 knees) or cortical button-suspensory loop (4 knees).

Tibial fixation was done by absorbable interference screws augmented in one case by a staple. The imaging acquisitions were performed on a General Electric machine with 1.5T field strength using T2 FSE sequence at 2 mm spacing. We exported 6 DICOM images corresponding to the body of the ACL from each axial oblique sequence using RadiaAnt (Medixant, Poznan, Poland) DICOM Viewer 1.9.16 (64-bit Windows).

We then averaged the ratio between the maximum width and thickness as well as the surface area in pixels using ImageJ (64bit for Windows, National Institutes for Health, Bethesda, MD, USA) and compared it with the native ACLs using the paired ttest (GraphPad/QuickCalcs).

RESULTS: For both quadrupled hamstrings and B-PT-B neoligaments, the mid-portion area was significantly higher (p < 0.001) than the native contralateral ACL: 41.82/31.39 mm2 and 37.05/32.08 mm2, respectively. The surface area of the neoligaments mid-portion was on average 33.23% higher than the native ACL for the quadrupled hamstrings and 15.49% for the B-PT-B, respectively. The native contralateral ACL was also significantly thinner throughout the mid-portion (p < 0.001) than both B-PT-B and quadrupled hamstrings neoligaments, with a width-thickness ratio of 2.57 vs 1.97 and 2.57 vs 1.39, respectively.

CONCLUSIONS: Our study showed that, several years after successful anatomic single bundle ACL reconstruction using an ipsilateral autograft, the mid-portion shape and cross-sectional area are not restored compared to the uninjured contralateral knee. This effect was more prevalent with hamstrings and less prominent when B-PT-B