Quantitative MRI Biomarkers to Predict Risk of Reinjury Within 2 Years After Bridge-Enhanced ACL Restoration

Abstract

BACKGROUND:

Quantitative magnetic resonance imaging (qMRI) methods were developed to establish the integrity of healing ACLs and grafts. Whether qMRI variables predict risk of re-injury is unknown.

PURPOSE:

To determine if qMRI at 6 to 9 months after bridge-enhanced ACL restoration (BEAR) can predict the risk of revision surgery within 2 years of the index procedure.

STUDY DESIGN:

Cohort Study

METHODS:

124 patients underwent ACL restoration as part of the BEAR I, BEAR II and BEAR III prospective trials and consented to undergo an MRI of the surgical knee 6 to 9 months after surgery. Only one subject was lost to follow-up and 4 could not undergo MRI (n=119). qMRI techniques were used to determine the average cross-sectional area (CSA), normalized signal intensity (SI), and a qMRI-based predicted failure load, which was calculated using a pre-specified equation based on CSA and normalized SI. Patient-reported outcomes (International Knee Documented Committee (IKDC) Subjective Score), clinical measures (hamstring strength, quadricep strength, and side-to-side knee laxity), and functional outcomes (single-leg hop) were also measured at 6 to 9 months after surgery. Univariate and multivariable analyses were performed to determine the odds ratios (OR) for revision surgery based on the qMRI and non-imaging variables. Patient age and posterior tibial slope values were included as covariates.

RESULTS:

119 patients (97%), with a median age 17.6 years, underwent MRI between 6 to 9 months postoperatively. Sixteen of 119 patients (13%) required revision ACL surgery. In univariate analyses, higher IKDC subjective score at 6 to 9 months postoperatively (OR=1.66 per 10-point increase, p=.035) and lower qMRI-based predicted failure load (OR=0.66 per 100N increase, p=.014) were associated with increased risk of revision surgery. In the multivariable model, which adjusted for age and posterior tibial slope, the qMRI-based predicted failure load was the only significant predictor of revision surgery (OR= 0.71 per 100N, p=.044).

CONCLUSION:

qMRI-based predicted failure load of the healing ACL was a significant predictor of the risk of revision within 2 years after BEAR surgery. The current findings highlight the potential utility of early qMRI in postoperative management of patients undergoing the BEAR procedure.

INTRODUCTION

Anterior cruciate ligament (ACL) injuries are common in the young active population. For these patients, the desire to regain function and to return to sport is high, thus, and clinical outcomes with non-operative management are uniformly poor, thus, surgical treatments such as ACL reconstruction (ACLR) are commonly performed.²⁹ However, only about half of the patients following ACLR are able to return to their pre-injury activity level with an estimated 1 in 4 having a second ACL injury within the first-year after surgery.^22,29 Numerous factors, such as age, activity level, anatomical features, and graft type are associated with increased risk of reinjury following ACLR.²⁵ Given the shortcomings of ACLR (i.e., co-morbidities due to graft harvest, risk of posttraumatic osteoarthritis), new surgical methods to repair or restore the native ACL have been introduced.^17,21,33,48 One such approach, bridge-enhanced ACL restoration (BEAR), is a surgical procedure that stimulates ACL healing by placing an extracellular matrix-based implant within the injury site and using it to hold the patient’s blood in the space between the torn ends of the ligament.^32–34,51 The blood and collagen scaffold subsequently creates an environment that is biologically conducive to ligament healing. As the BEAR technique is relatively new, postoperative outcome measures that can predict risk factors for re-injury and subsequent revision surgery remain unknown in patients who undergo the BEAR procedure.

Younger age²⁵ and steeper posterior tibial slope^9,44,52 have been identified as risk factors for an ACL graft retear. Moreover, low postoperative patient-reported outcomes (e.g., International Knee Documentation Committee (IKDC) subjective score), ^24,36,43 clinical outcomes (e.g., arthrometer-based knee laxity testing),⁵⁰ and asymmetry in joint function (e.g., single-leg hop test),⁴³ have also been used to track ACLR recovery and to approximate reinjury risk. However, studies have shown that patient-reported outcome measurements (e.g., IKDC) introduce bias and inconsistency when compared to more objective methods.^1,2 Furthermore, these discrepancies become more apparent when comparing various validated knee outcome scores.^1,23 Most importantly, none of these commonly used outcome measures directly assess the structural integrity of the healing ACL or ACL graft.^1,2

To supplement the existing patient-reported, clinical, and functional assessments, non-invasive imaging methods have been developed to evaluate the graft maturity and to estimate the structural properties of the surgically treated ACL using quantitative magnetic resonance imaging (qMRI).^{6,8,26,27,46,49} Studies have determined that changes in qMRI parameters, such as signal intensity (SI), volume, and cross-sectional area (CSA) of the ligament or graft can be used to determine the integrity of the healing structures as these parameters have been shown to reflect the biomechanical and histological properties of the healing tissues.^4,6,8,54 Despite promising preclinical and clinical evidence on the utility of qMRI for the non-invasive assessment of the healing ACL structural properties and postoperative remodeling, these techniques have not yet become mainstream. Possible reasons for this include technical challenges in standardizing image acquisition parameters or post-hoc harmonization for consistent qMRI results,^16,49 and the lack of high-quality evidence on the relative performance of qMRI metrics in predicting ACL surgery outcomes and reinjury risk.

The purpose of this preliminary study was to analyze the prospectively collected data from the BEAR clinical trials to determine if early qMRI parameters (e.g., those obtained at 6 months or 9 months post-ACL surgery) were associated with the risk of revision surgery during the 2 years following the BEAR procedure. The 6-to-9-month window was selected as it corresponds to the time that patients are typically cleared to go back to sport. We also assessed the ability of common patient-reported (e.g., IKDC), clinical (e.g., anteroposterior knee laxity), and functional (e.g., single-leg hop ratio for distance, isometric quadriceps, and hamstring strength) outcomes, all collected at 6 to 9 months postoperatively, to predict the 2-year revision risk in the same cohort. We hypothesized that: 1) qMRI-based measures of the healing ACL structural properties (e.g., average CSA, normalized ligament SI, and predicted failure load) at 6 to 9 months would predict the risk of ipsilateral revision within 2 years of the BEAR procedure, and 2) qMRI-based measures could improve the ability to predict the risk of ipsilateral revision after considering baseline variables (e.g., age, posterior tibial slope),^9,13,44,52 and patient-reported, clinical, and functional outcomes at 6 to 9 months in the multivariable regression analyses.^24,36,43,50

METHODS

Subjects

Data were acquired from patients enrolled in the BEAR I (NCT02292004, IRB-P00012985),³⁴ BEAR II (NCT02664545, IRB-P00021470),³³ and BEAR III (NCT03348995, IRB-P00026162) clinical trials between February 2015 and January 2019. The three trials were approved by the Institutional Review Boards, and all subjects granted their informed consent prior to participating. BEAR I was a non-randomized controlled cohort study with 10 patients in the BEAR arm,³⁴ BEAR II was a randomized controlled trial with 65 patients in the BEAR arm,³³ and BEAR III was a prospective dual-center cohort study with 49 BEAR patients (Figure 1). Patients were excluded from these trials if they had a history of prior knee surgery, knee infection or potentially adverse risk factors including a history of nicotine use, corticosteroid use, chemotherapy, diabetes, inflammatory arthritis, sickle cell anemia, or anaphylaxis. Patients with concomitant injury to the posterolateral corner, grade III medial collateral ligament injury, or complete patellar dislocation were also excluded. Of the 124 patients who underwent the BEAR procedure, 119 had imaging data to carry out the analysis (97%; Figure 1) and revision surgery data were available for the 119 patients at two years (100%). A complete description of the inclusion/exclusion criteria have been previously published.^33,34

Figure 1.

STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) diagram detailing the flow of patients through the analysis. BEAR, bridge-enhanced ACL restoration; SSDKT, side to side difference in KT knee laxity.

Surgical Procedure

The BEAR procedure utilizes a resorbable implant to provide a platform on which the patient’s blood can stimulate the torn ACL to heal.³⁹ The implant is a scaffold composed of bovine extracellular matrix proteins, primarily collagen, which acts to hold autologous blood between the torn ligament ends. The scaffolds were manufactured at Boston Children’s Hospital.⁴⁰ Investigational Device Exemptions for use of the implant in the three trials were granted by the FDA.

Details of the BEAR procedure have been previously described.^33,34 In brief, a whipstitch (Vicryl; Ethicon) was placed in the tibial stump. A cortical button combined with a polyester suture stent (Ethibond; Ethicon) was passed through the femoral tunnel and secured to the proximal femoral cortex. The polyester sutures were threaded through the BEAR scaffold and a tibial tunnel and secured in place with an extracortical button. The scaffold was then saturated with 5 to 10 mL of the patient’s blood, and the tibial stump was pulled into the saturated scaffold to repair the ACL.^33,34

MR Imaging

MR imaging was performed 6 to 9 months post-surgery on a 3T scanner (Tim Trio or Prisma; Siemens, Erlangen, Germany) using a 15-channel transmit/receive knee coil (Siemens). For the BEAR I and II clinical trials, the Constructive Interference in Steady State (CISS) sequence was acquired for the surgical limb on the Tim Trio (Table 1). For the BEAR III trial, both the Tim Trio and Prisma were used (Table 1). Harmonization to standardize the images between the two scanners was performed as previously described.¹⁶ The Prisma scans were scaled to the Tim Trio scans due to hardware-necessitated differences in the sequence acquisition parameters (Table 1). The ACL was then segmented from the MR image stack by an observer with more than 8 years of experience in ACL segmentation (AMK) using commercial imaging software (Mimics; Materialise, Leuven, Belgium).

Table 1.

MRI acquisition parameters for constructive interference in steady state sequence on Tim Trio and Prisma scanners.

	BEAR I, II, & III	BEAR III
Scanner	Tim Trio (n=109)	Prisma (n=10)
Flip Angle (°)	35	35
Repetition time (ms)	12.78	12.92
Echo Time (ms)	6.39	6.46
Field of view (mm)	140 x 140	130 x 160
Acquisition matrix (−)	384 x 384	512 x 270
Voxel size (mm)	0.36 x 0.36 x 1.50	0.31 x 0.31 x 1.00

qMRI Outcome Measures

Postoperative MR images obtained at 6 to 9 months were used to determine the normalized SI and average CSA of the ACL. The normalized SI was calculated from the values of each voxel within the segmented ACL, normalized to the SI of the anterior cortex of the femur, and log base-2 transformed as the data were not normally distributed (Eq. 1).⁵ The anterior cortex of the femur was selected because it is consistently near the noise floor. The difference in background noise texture between scanners was used as a scaling term to harmonize the scanners (ΔNoise).⁵ The mean CSA was calculated by dividing the ACL volume by the ACL length. The qMRI prediction model of ACL failure load was previously determined using a porcine model of ACL repair, from which qMRI and tensile failure data were available.⁵ The models were scaled to humans by exchanging the ligament volume parameter of the previous study with the average CSA to account for ligament size differences between pigs and humans. The updated model used normalized SI and CSA as independent variables to predict the failure load of the ACL (F_max) (Eq. 2). All measurements were completed within one month of imaging and values stored until the two-year outcomes were measured. The imaging examiner was blinded to the postoperative outcomes.

$$S{I}_{\mathit{scaled}}={\log }_2\left(\frac{S{I}_{\mathit{ACL}}+\text{Δ}\mathit{Noise}}{S{I}_{\mathit{Bone}}+\text{Δ}\mathit{Noise}}\right)$$ (1)

$$\mathit{Predicted}{F}_{\mathit{max}}={\beta }_0+{\beta }_{1}SI+{\beta }_{2}CSA$$ (2)

Preoperative MR images were used to measure the posterior slope of the tibial plateau in the medial compartments as previously described.^26,27 The posterior slope of the medial tibial plateau was measured in a sagittal slice at the center of the medial plateau as the angle between a line that joined the peak points on the anterior and posterior rims of the plateau and as a line perpendicular to the longitudinal axis of the tibia. Measurements were performed based on established techniques by an experienced member (AMK) using a commercially available image viewer (Osirix Viewer v8.5, Pixmeo SARL).^26,27

Patient Follow-up and Non-Imaging Outcome Measures

Six patient-reported, clinical, and functional outcomes were assessed between 6 to 9 months after surgery: 1) the International Knee Documentation Committee Subjective Score (IKDC),^24,36 2) the side-to-side difference in KT knee laxity (SSDKT),¹² 3) the quadriceps strength ratio (%surgical:contralateral),³ 4) the hamstring strength ratio (%surgical:contralateral),³ 5) quadricep-hamstring strength ratio of the surgical leg,³ and 6) single-leg hop ratio (%surgical:contralateral).⁴³ The IKDC Subjective Score was calculated based on patients’ responses to the IKDC Questionnaire.^24,36 The arthrometer measurements (KT-1000; MEDMetric, San Diego, CA) were performed by an experienced independent examiner, who was blinded to surgical laterality and treatment using knee sleeves. Hamstring and quadriceps isometric muscle strength were measured using a handheld dynamometer (Microfet 2; Hoggan Scientific, LLC, Salt Lake City, UT). All measures were performed on each knee in duplicate and the duplicates were averaged.³ During the single-leg hop test, patients wore a brace on the surgically treated knee.

Patients were assessed for initial eligibility (n=124). Patients missing a 6- or 9-month MRI scan were excluded (n=4), and one patient was lost to follow-up at 2-years (n=1). The final cohort of BEAR subjects who returned 2 years after surgery for clinical assessment and to verify repair status included 119 patients (Figure 1).

Statistical Analyses

Logistic regression analyses were performed (GJB) to identify outcomes obtained at 6 to 9 months postoperatively, including qMRI, patient-reported, clinical, and functional outcomes, and baseline variables that were significant predictors of revision surgery prior to the 2-year assessment. First, unadjusted odds ratios were estimated for each variable using univariate logistic regression (i.e., one-predictor-at-a-time). Next, a multivariable logistic regression model utilizing a forward stepwise procedure was implemented to determine the best set of predictors of subsequent revision and their associated adjusted odds-ratios. To limit the number of potential predictors, only variables with p < .20 for their bivariate relationship with 2-year revision were considered as candidates for the multivariable model. Two baseline variables, age at surgery and posterior slope of the tibial plateau in the medial compartments (MTS), were forced into the multivariable stepwise model (i.e., step 0) as they have been previously shown to be predictive of revision.⁴⁵ Because listwise deletion of cases due to incomplete data on candidate clinical variables (Figure 1) adversely affects the sample size used for the stepwise multivariable regression model, a multiple imputation procedure was used (n=5 iterations) to allow the use of all cases in the multivariable analyses.⁵⁷ All analyses were conducted using SAS Statistical Software, Version 9.4 (SAS Institute, Cary, NC, USA) with statistical significance based on p<.05.

RESULTS

Baseline Characteristics, qMRI, and Non-Imaging Outcomes

The mean ± standard deviation (SD) age of the patients in the three studies was 17.6 ± 7.1 years. Sixty-four (54%) of the patients were female. Ninety-eight (82%) of the patients were Caucasian (non-Hispanic). The mean ± SD of the body-mass index (BMI) was 23.9 ± 3.4 (kg/m²). The mean ± SD of the MTS was 5.2 ± 2.4 degrees. Table 2 presents the qMRI, patient-reported, clinical, and functional outcomes recorded at 6 to 9 months postoperatively, considered potential predictors of revision surgery.

Table 2.

qMRI and non-imaging outcome measures (n=119).

Outcomes at 6 to 9 months after surgery
Postoperative qMRI Outcomes
Average Cross-Sectional Area (mm2), mean ± SD	60.3 ± 14.6
Normalized Average Signal Intensity, mean ± SD	1.3 ± 0.2
Predicted Failure Load (N), mean ± SD	705 ± 207
Postoperative Patient-reported, Clinical & Functional Outcomes
IKDC subjective score, mean ± SD	78.7 ± 13.4
SSDKT at 6-month (mm), mean ± SD a	2.5 ± 2.8
Quadriceps Strength (% contralateral), mean ± SD b	94.4 ± 17.1
Hamstrings Strength (% contralateral), mean ± SD b	91.6 ± 20.2
Hamstrings-Quadriceps Ratio (%), mean ± SD	47.3 ± 15.2
Single-Leg Hop Ratio (% contralateral), mean ± SD c	84.9 ± 17.6

^an=115

^bn=118

^cn=101

SSDKT=Side-to-side difference in KT (surgical - contralateral)

Univariate Predictors of 2-Year Revision

Sixteen (13%) of the 119 analyzed patients required revision surgery prior to the 2-year follow-up visit. Outcomes significantly associated with increased risk of revision at 2 years were higher IKDC Subjective Score at 6 to 9 months postoperatively (p=.035) and lower predicted failure load on the 6-to-9-month postoperative qMRI (p=.014; Table 3). Each 10-unit increase in IKDC subjective score (total range 100 points) corresponded to a 65% increase in the odds of revision ACL surgery. Additionally, a 100 Newton (N) increase in qMRI predicted failure load corresponded to a 34% decrease in the odds of revision. The average CSA and normalized SI on the 6-to-9-month qMRI were marginally associated with increased odds of revision (p<.06), with lower average CSA and higher normalized SI associated with increased risk of 2-year revision (Table 3).

Table 3.

Unadjusted odds ratios associated with revision by 2-year follow-up. P-values in bold type are significant.

Predictor	Odds Ratio	95% CI	P-Value
Postoperative Patient-Reported, Clinical & Functional
Outcomes at 6 to 9 months after surgery
IKDC Subjective Score (per 10 unit increase)	1.66	1.04, 2.65	.035
SSDKT (per 1 mm increase)	1.09	0.91, 1.32	.35
Quadriceps Strength (per 10% increase)	1.06	0.77, 1.46	.71
Hamstring Strength (per 10% increase)	1.17	0.92, 1.49	.19
Hamstring-Quadricep Ratio (per 10% increase)	1.06	0.77, 1.47	.71
Single-Leg Hop Ratio (per 10% increase) c	1.27	0.90, 1.80	.18
Postoperative MRI Outcomes
Average Cross-Sectional Area (mm2) (per 10 mm2 increase)	0.64	0.41, 1.01	.054
Signal Intensity	1.63	0.99, 2.69	.057
Predicted Failure Load (N) (per 100 N increase)	0.66	0.48, 0.92	.014

SSDKT=Side to side difference in KT (surgical – contralateral)

Multivariable Predictors of 2-Year Revision

After adjusting for patient age and MTS in the multivariable analysis, the 6 to 9 months postoperative qMRI predicted failure load was the only significant predictor of 2-year revision risk after BEAR (Table 4). Each 100N increase in predicted failure load at 6 to 9 months corresponded to a 29% decrease in the odds of revision ACL surgery by 2 years postoperatively (Table 4).

Table 4.

Adjusted odds ratios associated with revision surgery on multivariable stepwise logistic regression. P-values in bold type are significant.

Predictor	Odds Ratio	95% CI	p-value
Age (years)	0.65	0.47, 0.90	.009
MTS (degrees)	1.25	0.97, 1.61	.079
Predicted Failure Load (N) per 100N increase	0.71	0.51, 0.99	.044

MTS=medial tibial slope

DISCUSSION

The significant finding of the study was that an increase in predicted failure load of the healing ACL, as measured via qMRI at 6 to 9 months post-surgery, was associated with decreased odds of revision surgery by 2 years. Predicted failure load remained a significant predictor after adjusting for baseline variables (i.e., age, posterior tibial slope). Patient-reported, clinical, and functional outcomes at 6 to 9 months did not contribute to the predictive ability of revision in the multivariable model.

Traditionally, return to sport decisions are based on time, with athletes typically being returned to sport at 6 to 9 months after surgery.⁵⁸ More recently, some surgeons have also included a battery of patient-reported, clinical, and functional outcomes to guide the postoperative care plan; in particular the time for safe return to sports. ^7,8,53 Acceptable scores on assessments such as subjective rating scales, knee laxity tests, and functional hop testing have been used to determine a patient’s readiness to return to play.⁸ These outcomes remain controversial as studies have shown reinjuries occurring shortly after using these assessments to decide if a patient is ready to return to sport.^11,38,42,55 Moreover, these assessments lack sufficient resolution to directly assess the healing ligament and are often influenced by factors unrelated to the ACL structure. For example, physical examinations of the knee (e.g., the Lachman and pivot shift tests) can be influenced by the injury or hypertrophy of secondary stabilizers of the knee,^19,47 as well as by age,³⁰ sex,³⁰ and bony anatomy,³⁵ and are also prone to observer bias. Functional testing (e.g., hop testing and balance testing) can be influenced by the quality of the rehabilitation program, patient compliance, and/or fear of reinjury.^10,37 Likewise, patient-reported outcomes after ACL surgery have been shown to be influenced by self-esteem levels,¹⁰ body mass index,²⁰ and smoking.²⁰ All of these support the need to improve current clearance protocols, including the addition of non-invasive qMRI approaches to directly assess the healing ligament. These techniques have been shown to identify detailed structural changes in healing ligament and provide an objective evaluation of ligament maturation and remodeling after surgery.^26,27,49 Our current observations are in agreement with a recent study that focused on signal intensity, indicating an increased risk of hamstrings autograft failure in patients with a higher signal intensity ratio at 1 year.⁴¹ The current study used a combination of SI and CSA to measure predicted failure load to predict the risk of revision surgery, which was evaluated in both univariate and multivariable settings.

In the current study, among all the patient-reported, clinical, and functional outcomes, only the IKDC Subjective Outcome Score at 6 to 9 months was associated with subsequent failure in a univariate analysis. Interestingly, the results of this study demonstrated that a higher IKDC Subjective Score was predictive of a higher risk of revision after BEAR, while prior work has demonstrated the opposite for ACL reconstruction patients, where a higher IKDC has been associated with a lower risk of revision surgery. ^10,31,56 One possible explanation for this finding is that patients with a higher IKDC score felt better sooner and may have returned to activities earlier as a result.¹⁰ Further work to better quantify postoperative activity levels in BEAR patients is planned and should verify or invalidate this hypothesis.

A previous porcine study, which utilized a similar sequence to that used in the current study, demonstrated that size (volume) and normalized SI were independent predictor variables of the maximum failure load,⁵ and that an increase in tissue volume and a decrease in normalized SI were related to a higher maximum failure load. Furthermore, it was determined in the porcine study that the combination of both variables in the maximum failure load prediction model provided a significant improvement in the prediction of the maximum failure load (R²=.73), compared to either variable by itself (R²≤.56).⁵ These prior findings align with those of the current study as it would be expected that an increase in the healing ACL failure load would lessen the risk of revision surgery. These findings complement prior reports of temporal changes in ACL or graft normalized SI within 2 years after surgery, corresponding to tissue healing,^26,27 and those associated with high graft SI to incomplete healing and integration,^18,28,41 as well as graft rupture within 1 year after ACLR.⁴¹

To our knowledge, postoperative qMRI parameters have not been used to measure the risk of ACL revision surgery in BEAR patients. This study, therefore, provides new insight by suggesting that the predicted failure load, as calculated using qMRI techniques at 6 to 9 months, is associated with the risk of revision by 2 years post-surgery. This result indicates that for every 100N increase in predicted failure load measured at 6 to 9 months, the odds of revision by 2 years significantly decrease by 29%. This result provides evidence that a direct assessment of ACL integrity provides valuable data to predict the risk of ACL restoration failure. Future prospective studies are needed to assess whether the use of qMRI assessment can reduce ACL revision surgery when making return-to-sport decisions. A strength of this study was the low number of patients lost to follow-up, with 120 of the original 124 patients receiving the 6-to-9-month MRI and 119 patients providing information on revision surgery at the 2-year follow-up. An additional strength of this study was that the qMRI assessor was blinded to postoperative outcomes and did not know which patients had reinjured their ACL at the time of the analysis. The qMRI data were collected at the 6-to-9 month time point and analyzed prior to obtaining the revision data for all patients.

This study also has limitations. The total sample size (n=119) of this cohort and the number of revisions (n=16) was relatively small. Accordingly, derived confidence intervals for the estimated odds-ratios were quite wide. A larger study would lead to more precise point estimates and may potentially identify additional factors with smaller effect sizes that may also be associated with failure. Nonetheless, both the univariate and multivariable analyses showed that the qMRI predicted failure load was associated with subsequent revision surgery. In addition, this study only evaluated the outcomes of patients undergoing the BEAR procedure. However, the qMRI methods used in this study could potentially be applied to ACL reconstruction or primary repair patients as well. Therefore, it would be interesting to determine the relationships between the qMRI parameters and graft failure after ACL reconstruction or primary repair in future studies. While ACL reconstruction patients were included in the BEAR I and II trials, the sample size (n=10 and n=35) was too small to be considered for this analysis.^33,34 Another limitation was that the MR images were only acquired on scanners by one manufacturer. Future studies should explore if these same equations and qMRI methods would be applicable when different manufacturer magnets are utilized. Lastly, the imaging was performed using a Tim Trio scanner in the single site BEAR I and II trials while both the Tim Trio and Prisma scanners were used in the BEAR III trial. However, a recently validated harmonization procedure was used to standardize the qMRI images between scanners.¹⁶ Future large-scale prospective studies are needed to confirm the current findings and assess the utility of qMRI in revision prediction after ACL reconstruction.

The current study provides an example of how qMRI parameters can be used to assess the integrity of a healing ACL after the BEAR procedure. Several hurdles, including cost, must be overcome to translate qMRI in routine clinical practice to predict the integrity of the healing ligament or graft. To aid in translation, we have developed harmonization procedures to standardize qMRI results between scanners.¹⁶ We have also engineered automatic ACL segmentation techniques to reduce the segmentation time from hours to seconds to obtain the qMRI parameters required for the prediction.^14,15 An automatic pipeline that includes image harmonization, ACL segmentation, and the predicted equations to determine the failure properties of the healing ligament is under development. We intend to extend the post-processing pipeline to include ACL grafts to increase generalizability, and to help facilitate the clinical translation of qMRI. Quantitative MRI provides a tool to evaluate rehabilitation progress and assist with return to sport decisions when needed in light of the additional costs of an MRI.

In conclusion, quantitative magnetic resonance imaging obtained at 6 to 9 months after bridge-enhanced ACL restoration surgery determined that the predicted failure load at that time point was a significant predictor of revision ACL surgery within 2 years of the BEAR procedure. Taken together, these data suggest that a direct measure of the structural properties of the healing ligament using qMRI techniques may be beneficial to assess rehabilitation progress and inform return-to-sport decisions.