Can pre-operative MRI estimate hamstring autograft diameter in anterior-cruciate ligament reconstruction? (110)

OBJECTIVES: Hamstring tendon autograft (HTA) is the most common graft source used worldwide for anterior cruciate ligament reconstruction (ACLR). The graft is comprised of a patient’s own semitendinosus tendon (ST) and gracilis tendon (GT), typically double stranded. Recent literature suggests that HTAs below 8mm in diameter are associated with higher failure rates and poorer outcome scores. Currently, surgeons do not have a reliable, user-friendly tool to estimate HTA diameter pre-operatively. The inability to do so leads to potentially harvesting an insufficient graft. There is a growing body of evidence that suggests a correlation between preoperative MRI measurements and the intraoperative measured HTA diameter. This could be used to identify patients at risk of having small HTA diameter, however, it is unknown if these patients also are at risk of smaller caliber alternative autograft tendon sources (i.e., quadriceps tendon and patellar tendon). Our hypotheses were: i) Intra-operative HTA diameter is strongly correlated with pre-operative MRI measurements of ST cross-sectional area (STCSA), GT cross-sectional area (GTCSA), and the sum of the ST and GT cross-sectional area (STGTCSA); and ii) patients with HTA diameters less than 8mm will also have smaller caliber patellar tendon and quadriceps tendon measurements on pre-operative MRI. METHODS: After appropriate IRB approval was obtained, patients undergoing ACLR with HTA between the period of 01/01/2013 to 05/31/2020 were retrospectively reviewed. Inclusion criteria included the following: MRI proven ACL tear, 3-Telsa MRI available for review, surgery performed within our institution using standard quadrupled hamstring technique, intra-operative HTA diameter recorded in the operative report, and age greater than 12 years old. The MRI measurements were performed by two physicians: one orthopedic sports medicine research fellow and one orthopedic surgery resident. Each physician was blinded to the intraoperative HTA diameter. CSA of the ST and GT was measured on axial MRI sequences using the axial slice that included the widest (medial-lateral (M-L) width) portion of the distal femur. This image was magnified 4 times and CSAs of the GT and ST were measured using the elliptical region of interest tool (Figure 1). In addition, the patellar tendon length (PTL), patellar tendon thickness (PTT), patellar tendon medial-lateral width (PTW), and quadriceps tendon thickness (QT) were measured. PTL was measured at the sagittal slice showing the most distal pole of the patella and tibial tubercle. PTT was also measured on this MRI slice at the tendon’s mid-point. PTW was measured in the sagittal view and cross referenced to an axial view as previously described. A point at the center of the tendon width (M-L width) was defined, and the width was then measured from this point to the medial and lateral borders separately in order to accommodate the tendon contour. The sum of the widths was regarded as the total tendon width. QT was measured in the anterior-posterior plane on a sagittal slice located 25mm proximal to the superior pole of the patella and measured at the mid-point of the tendon (M-L plane) orthogonal to the quadriceps tendon fibers. All measurements were taken using the universal viewer image analysis software. Pearson r values were calculated for MRI measurements from both readers and the average of their measurements against intra-operative HTA diameter. Receiver operator curves (ROC) were used to calculate sensitivity and specificity values for each MRI measurement. The measurement that best correlated with HTA diameter (e.g., GTCSA) was then compared to PTL, PTT, PTW, and QT among the patients with HTA less than 8mm. Intra-class correlation coefficients (ICC) were calculated for inter-rater reliability between reader 1 and reader 2 for all MRI measurements. RESULTS: Fifty-two patients (53 knees, 26 female and 26 male) met inclusion criteria, with a mean age of 23 years old. The mean intraoperative HTA diameter was 7.98mm, with 18 grafts (34%) measuring less than 8mm. Pearson r values for all MRI measurements and ICC values are shown in in table 1. HTA diameter was significantly correlated to all averaged MRI measurements with the exception of PTW and QT. The strongest correlation was seen with GTCSA (r=0.72, p<.01). By entering a patient’s GTCSA measurement as “x” into the line of best fit (y = 41.83x + 5.0846), the estimated HTA diameter “y” can be extrapolated (Figure 2). Using our dataset, we determined that a GTCSA cut off value of 0.0625mm can be used to identify patients who will have a HTA diameter of 8mm or greater with a sensitivity of 0.91. For our cohort of 53 knees, GTCSA significantly correlated with PTL (r=0.352, p<.01), QT (r=0.334, p<.05), and STCSA (r=0.531, p<.01). [AMT(S2] Of the 18 patients with HTA diameter less than 8mm, GTCSA showed a significant correlation with PTL (r=0.34, p<.05) and QT (r=0.33, p<.05) (Figure 3). No significant correlation was observed between GTCSA and ST, PTT, or QT. CONCLUSIONS: Pre-operative MRI measurements of STCSA and STGTCSA did not correlate with intra-operative HTA in our cohort. However, pre-operative MRI measurement of GTCSA did show a strong correlation with intra-operative HTA diameter in our cohort of patients. Among patients with HTA diameters less than 8mm, GTCSA on pre-operative MRI showed a significant correlation with PTL and PTW. GTCSA can help to estimate whether or not a patient will have a HTA greater than 8mm and may provide insight regarding alternative autograft characteristics. The methods described in this study are reproducible between observers at different levels of their orthopedic training. By knowing how likely a patient is to have a sufficient HTA, surgeons can better educate patients regarding the risks and benefits pre-operatively as well as plan for alternative graft sources as needed.