Abstract
Background:
Patellofemoral dislocations are often associated with chondral injuries. In adults, diagnostic arthroscopy (DA) has identified chondral injuries in up to one-third of cases where magnetic resonance imaging (MRI) was negative. However, the value of DA in detecting cartilage injuries in pediatric patients undergoing medial patellofemoral ligament reconstruction (MPFLR) has not been well studied.
Purposes:
To evaluate the utility of DA in diagnosing cartilaginous injuries in pediatric patients undergoing MPFLR with negative preoperative MRI findings and to compare postoperative outcomes between patients with and without chondral injury findings on DA.
Study Design:
Cohort study (diagnosis); Level of evidence, 3.
Methods:
A retrospective review was conducted on patients aged 0 to 18 years who underwent combined MPFLR and DA procedures from 2015 to 2023 at a single institution. The exclusion criteria were as follows: revision MPFLR; tibial tubercle osteotomy; multiligament reconstructions; and previous meniscal injury. MRI and DA reports were reviewed for chondral findings. Patients were classified into 3 groups: +MRI/+DA, –MRI/+DA, and –MRI/–DA. Subgroup analysis was performed within the –MRI/+DA group to compare outcomes in the immediate postoperative recovery period between those who underwent a chondroplasty versus those who did not.
Results:
Of the 155 patients included, follow-up averaged 7.8 months (median, 5 months [range, 0-41 months]). Among 89 patients with negative MRI findings, 39 (43.8%) had chondral injuries detected on DA. Targeted treatment was performed in 22 (56.4%) of these patients, a lower rate than in those with positive MRI findings (80.3%; P = .02). No significant differences were observed in knee flexion at 2 weeks postoperatively: +MRI/+DA (77.9°± 23°), –MRI/+DA (66.5°± 30.2°), and –MRI/–DA (73.3°± 30°) (P = .14). Similarly, no significant differences were found at 12 weeks: +MRI/+DA (135.3°± 13°), –MRI/+DA (127.8°± 21°), and –MRI/–DA (134.1°± 13.4°) (P = .10). Range of motion and complication rates were comparable across all groups (P > .05). A subgroup analysis in the –MRI/+DA cohort comparing patients who underwent an arthroscopic procedure (n = 22) with those who did not (n = 17) showed no significant differences in range of motion or complication rates (P > .05).
Conclusion:
Diagnostic arthroscopy detected cartilaginous injuries in 43.8% of pediatric patients with negative preoperative MRI undergoing MPFLR. This highlights the diagnostic value of DA in identifying injuries that would otherwise remain undetected.
Keywords
Patellofemoral instability is common in adolescents and pediatric patients. 13 Epidemiologic studies have found that the annual incidence of first-time patellar dislocation is highest in patients aged 14 to 18 years of 147.7 per 100,000 person-years1,14,20,22 Anatomic and biomechanical parameters affecting patellofemoral instability include ligamentous laxity, trochlear dysplasia, patella alta, patellar tilt, and an increased tibial tuberosity-trochlear groove distance. 5 Nonoperative treatment was historically the mainstay of treatment for first-time patella dislocation; however, recent literature has shown a high risk of recurrent instability. 14 Given this, there has been a trend toward surgical stabilization in patients with repeat dislocations. Specifically, the medial patellofemoral ligament (MPFL) reconstruction (MPFLR) technique has become a popular surgical choice to reduce the risk of recurrent patellar dislocations. 23
There has been an increasing awareness regarding chondral injuries associated with patellofemoral dislocations. Magnetic resonance imaging (MRI) is considered the gold standard for identifying concomitant cartilaginous injury. 29 A recent systematic review demonstrated that 85% of patients with a first-time patellar dislocation had signs of patellar chondral damage on MRI. 18 Shultz et al 26 investigated the benefit of diagnostic arthroscopy (DA) at the time of MPFLR in the adult population, revealing chondral findings on DA in 31.7% of patients with a negative preoperative MRI. However, there exists little data on the utility of DA in the setting of MPFLR in the pediatric population, despite the higher incidence of patellar dislocations and potentially greater risk of associated chondral injuries in this group.9,25
The primary objective of this study was to determine the frequency with which diagnostic arthroscopy identifies patellofemoral chondral lesions not visualized on preoperative MRI in pediatric patients undergoing MPFLR. The secondary objective was to assess whether the identification and subsequent treatment of these lesions influence short-term postoperative outcomes. We hypothesized that diagnostic arthroscopy would detect chondral lesions in a substantial proportion of patients with negative MRI, but that these additional findings would not significantly affect short-term clinical outcomes.
Methods
Patient Selection and Outcome Measures
This was a retrospective cohort study conducted at a single pediatric tertiary referral center, approved by our institutional review board (IRB) (IRB No. IRB-24-107). The study reviewed pediatric patients who underwent MPFLR and DA, performed by 2 fellowship-trained sports medicine orthopaedic surgeons (H.M. and J.V.) between 2015 and 2023. The inclusion criteria were patients <18 years old who had undergone both MPFLR and DA in the same procedure. These were determined using the presence of concomitant procedure codes (CPT-27427). The exclusion criteria were revision MPFLR, concomitant tibial tubercle osteotomy, multiligamentous reconstruction, arthrotomy procedures, <1 month of follow-up, previous meniscal or cartilage injury, or absence of MRI findings. A total of 253 patient records were reviewed, and 155 patients met the inclusion criteria (Figure 1).
Patient selection diagram. DA, diagnostic arthroscopy; MRI, magnetic resonance imaging; MPFLR, medial patellofemoral ligament reconstruction; TTO, tibial tubercle osteotomy.
The demographic variables collected were sex, age, height, weight, laterality, and primary sport. Those who did not specify a sport were categorized as “miscellaneous.” Both surgeons included in this study routinely perform concomitant diagnostic arthroscopies. MRI and operative reports were analyzed to identify chondral injuries preoperatively and at the time of surgery. MRI image analysis was not part of the study methodology, as chondral findings were extracted from the official radiologist reports. However, reviewing the MRI images is part of standard preoperative practice, and the operating surgeons routinely evaluate the images themselves before surgery. Arthroscopic findings were categorized as chondromalacia, osteochondral lesions, or general chondral defects. The grade of chondromalacia, when available, was recorded; otherwise, injuries were labeled as “general chondral defects.” Cartilaginous procedures were also recorded from the operative report. These were grouped into chondroplasty or microfracture. Knee range of motion and complications were collected from postoperative clinical notes. Complications were categorized as persistent pain, subsequent dislocation, revision or hardware failure, and manipulation under anesthesia.
Patients were divided into groups based on the presence of chondral findings on the MRI and the presence of chondral findings on the DA. These groups were as follows: (1) positive MRI findings of chondral injury, positive DA findings of chondral injury (+MRI/+DA); (2) negative MRI findings of chondral injury, positive DA findings of chondral injury (–MRI/+DA); (3) negative MRI findings of chondral injury, negative DA findings of chondral injury (–MRI/–DA). For the –MRI/+DA group, a subgroup analysis was performed to compare outcomes between those who underwent a targeted arthroscopic procedure and those who did not, including range of motion and postoperative complications (eg, persistent pain, instability, and hardware failure).
Surgical Procedure
All procedures were conducted with the patient in the supine position under general anesthesia, with a nonsterile tourniquet placed around the proximal thigh and a femoral nerve block for pain management. Each surgery began with a diagnostic arthroscopy through the standard inferomedial and inferolateral parapatellar portals. These were treated based on the prerogative of the surgeon. Subsequently, the MPFLR was performed.
Both surgeons performed the procedure in a similar fashion, although one used a single medial incision centered between the patella and the medial femoral condyle, and the other used a 2-incision technique. After dissection, the medial aspect of the patella was exposed. In all cases, a gracilis or semitendinosus allograft was used. Two suture anchors were used to secure the center of the graft to the patella. Intraoperative fluoroscopy was utilized to identify the Schottle point, and a femoral tunnel was drilled distal to the physis. The 2 ends of the allograft were sutured using a whip-stitch technique. A soft tissue tunnel was created, and the 2 graft ends were passed between the medial retinaculum and the medial capsule of the knee, exiting the retinaculum at the femoral tunnel entry point. The graft was then passed into the femoral tunnel and secured under correct tension with an interference screw, with the knee at 30°. The knee was then cycled through its full range of motion, confirming the patella was tracking centrally within the trochlear groove.
Postoperative Protocol
A soft dressing and a hinged knee brace were used. Early nonweightbearing range of motion from 0° to 90° was encouraged immediately. Weightbearing with the brace locked in extension was allowed either immediately or at 2 weeks postoperatively, at the surgeon’s preference. At 6 weeks, patients were allowed to return to normal ambulation. Postoperative rehabilitation protocol did not differ if chondral work was performed.
Statistical Analysis
A 3-group chi-square test was used to analyze categorical variables (sex, laterality, and sport). For categorical variables that did not meet the assumptions of the chi-square test, Fisher exact tests were used for secondary analyses. For continuous variables (age, body mass index [BMI], or number of patellar dislocations), the Shapiro-Wilk test was first used to assess normality. If all groups demonstrated normal distributions, a 1-way analysis of variance was performed; otherwise, a Kruskal-Wallis test was used. If necessary, a post-hoc comparison was done using the Dunn test with Bonferroni correction. For comparisons between the 2 groups, an unpaired Student t test was used for normally distributed data, and a Mann-Whitney U test was used when normality assumptions were not met. For non-normally distributed variables, results are presented as medians with ranges; for normally distributed variables, means with standard deviations are reported. Statistical significance was set at α = .05. All data were analyzed using Microsoft Excel Version 16.78 (Microsoft Corp) and R Version 4.3.1 (R Foundation for Statistical Computing).
Results
Patient Characteristics
A total of 155 pediatric patients were included in this study, with a mean follow-up time of 7.8 months (median, 5 months [range, 0-41 months]). 66 patients were in the +MRI/+DA group, 39 patients were in the –MRI/+DA group, and 50 patients were in the –MRI/-DA group. All patient demographic data are presented in Table 1. There were no statistically significant differences between groups for the following variables: age, BMI, laterality, and the number of sports played (P > .05). Women were majorly represented in the –MRI/+DA group (71.2%) and the –MRI/-DA group (76%), as compared with the +MRI/+DA group (51.5%). A statistically significant difference in the time interval between MRI and DA was noted among the 3 groups (median: +MRI/+DA 36.5 months, –MRI/+DA 80 months, and –MRI/-–DA 61.0 months; P = .01). Post-hoc analysis showed that the +MRI/+DA group differed significantly from both the –MRI/+DA group (P = .01) and the –MRI/-DA group (P = .048), whereas there was no significant difference between the –MRI/+DA and –MRI/–DA groups (P≥ .999). There was a difference in the number of dislocations across the –MRI/–DA, –MRI/+DA, and +MRI/+DA groups (medians, 3, 2.5, 3.5; H = 9.19; P = .0101), and post-hoc testing found the +MRI/+DA group had more dislocations than the –MRI/+DA group (P = .011).
Comparison of Patient Characteristics a
Data are presented as n (%), median (range), or n. Bold P values indicate statistical significance. DA, diagnostic arthroscopy; MRI, magnetic resonance imaging.
Refers to presence (positive) or absence (negative) of chondral injury in the respective MRI or DA.
Findings of Chondral Injury
Of the 89 total patients with negative MRI reports, 39 (43.8%) had positive chondral findings during DA, representing the –MRI/+DA cohort (Table 2). In the same group (–MRI/+DA), the rate of undergoing arthroscopic treatment for chondral lesions was 56.4%, compared with 80.3% in the +MRI/+DA group (P = .04) (Table 3).
Incidence of Chondral Injury Findings a
DA, diagnostic arthroscopy; MRI, magnetic resonance imaging.
Arthroscopic Findings and Arthroscopic Procedure Completed to Treat Chondral Injury a
Data are presented as n (%). DA, diagnostic arthroscopy; MRI, magnetic resonance imaging.
Percentage of the total number of patients.
Bolded values indicate statistical significance with P < .05
The most common procedure performed for the treatment of these chondral lesions in both +MRI/+DA and –MRI/+DA groups was a shaving chondroplasty, and the most common arthroscopic finding was chondromalacia (Table 3). There were no statistically significant differences between the types or locations of procedures performed in these 2 groups (P > .05). Additionally, there were no statistically significant differences in the grade of chondromalacia. However, there was a 22.7% rate of osteochondral lesions in the +MRI/+DA group, while the –MRI/+DA group had none (P < .01). Similarly, there was a 13.6% rate of a general chondral defect in the +MRI/+DA group, while the –MRI/+DA group had none (P = .02).
Surgical Outcomes and Complications
Across all 3 groups, there were no significant differences in postoperative knee range of motion (flexion and extension) (P > .05) (Table 4). All groups demonstrated improvements from a 2-week range of motion to a 12-week range of motion. At 2 weeks postoperatively, there was no statistically significant difference in knee flexion among the groups: +MRI/+DA (median, 85° [range, 30°-120°]), –MRI/+DA (median, 70° [range, 10°-100°]), and –MRI/–DA (median, 80° [range, 20°-140°]) (P = .60). Similarly, by 12 weeks, knee flexion remained comparable across groups: +MRI/+DA (median, 140° [range, 90°-155°]), –MRI/+DA (median, 135° [40°-145°]), and –MRI/–DA (median, 140° [100°-160°]) (P = .16). The rate of complications between all groups was similar (P = .67). Further, there were no statistical differences in the rate of types of complications between all 3 groups (P > .05).
Comparison of Patient Outcomes a
Data are presented as median (range), n (%), or n. DA, diagnostic arthroscopy; MRI, magnetic resonance imaging; ROM, range of motion.
A separate subgroup analysis was performed to compare surgical outcomes between those who had an arthroscopic procedure done (n = 22) and those who did not (n = 17) (Table 5). Similarly, no significant differences were observed in knee range of motion (flexion and extension), rate of complications, or type of complications (P > .05).
Subgroup Analysis Comparing Patient Outcomes Within Negative MRI, Positive DA Group a
Data are presented as median (range) or n. DA, diagnostic arthroscopy; MRI, magnetic resonance imaging; ROM, range of motion.
Discussion
The results of this study demonstrate that in a cohort of pediatric patients undergoing MPFLR with negative MRI findings for chondral lesions, 43.8% had chondral lesions discovered during DA. Of these patients, 56.4% subsequently received arthroscopic treatment for the identified lesion. In comparison, 80.3% of patients who had chondral lesions noted on MRI before surgery received arthroscopic treatment. Furthermore, postoperative knee range of motion and complication rates were similar across all patient groups, regardless of whether a chondral injury was present or the type of intervention.
Our study results parallel those seen in Shultz et al, 26 where the authors found that DA revealed chondral findings in 31.7% of cases with a negative preoperative MRI in the adult population. Concomitant chondral injury in the pediatric population has a wide range, from 23% to 85%.6,7,15,16,18,24,25,30 Our findings of an overall rate of chondral injury of 67.7% are consistent with the pediatric literature.
There is a wide range of sensitivities (6%-99%) and specificities (50%-99%) regarding the utility of MRI in diagnosing cartilage injuries.8,10,11,17,28,31 MRI may fail to detect low-grade cartilaginous injuries. A specific study by Pihlajamäki et al 21 found that the sensitivity of 1.0-T MRI was low for grade 1 lesions (13%) but higher for grade 2, 3, or 4 lesions (83%) compared with DA. 21 However, a systematic review by Chen et al 3 demonstrated that the pooled sensitivity of quantitative, 3T, 1.5T, and <1.5T MRIs was 0.82, 0.79, 0.67, and 0.55, respectively, indicating that higher field strength MRIs have greater diagnostic potential for chondral injuries. In the present study, there were no significant differences in the grade of chondromalacia between the +MRI/+DA group and the _MRI/+DA group. The –MRI/+DA group underwent intervention for their chondral findings at a lower rate than that of the +MRI/+DA group, with rates of 56.4% and 80.3%, respectively. This points toward the diagnostic potential of DA in pediatric patients undergoing MPFLR. However, while diagnostic arthroscopy may identify chondral injuries not visible on MRI, our study does not provide sufficient long-term follow-up to guide treatment recommendations. As such, decisions to intervene should be based on intraoperative assessment of lesion characteristics, with further research needed to determine the clinical significance of treating these findings.
Interestingly, the interval between MRI and diagnostic arthroscopy was longest in patients with false-negative MRI findings (–MRI/+DA group). One possible explanation is that MRI sensitivity for detecting knee chondral injuries may change over time due to tissue alterations.12,28 However, this interpretation is limited by the fact that most patients in our cohort experienced multiple dislocations, making it difficult to accurately define the interval between injury and MRI. An alternative explanation is that additional dislocation events occurred between the MRI and DA, resulting in progressive chondral damage that was not present or severe enough at the time of the MRI to be detected. 2 This difference in time intervals was statistically significant between the +MRI/+DA and –MRI+DA cohorts, but not between the –MRI/+DA and –MRI/–DA cohorts. The high variability within each group, as reflected by the large range of values, may have limited our ability to detect a more definitive relationship. Nonetheless, this finding highlights the importance of considering repeat MRI if a substantial delay exists between initial imaging and planned surgical intervention, which is common in pediatric orthopaedic surgery practices.
Low-grade concomitant cartilage injuries in children can often be treated nonoperatively. Surgical stabilization is reserved for patients with unstable cartilage flaps, detached osteochondral fragments, mechanical blocks to motion, and/or failed nonoperative management. 4 Importantly, 17 of the 39 patients in the –MRI/+DA group did not undergo arthroscopic intervention, which may suggest that the chondral findings observed during DA were minor and did not warrant surgical treatment. In our study, both surgeons used DA before their MPFLR, which allowed not only to evaluate the utility of DA in diagnosing cartilage but also to determine whether DA affected treatment. In our series, this resulted in the diagnosis of cartilage injuries in 39 (43.8%) patients and treatment in 22 (14.2%) patients who would have otherwise gone undetected. In our study, we found no significant differences in range of motion or short-term complications across any of the groups at the 2-week postoperative or 12-week postoperative time points, implying minimal morbidity associated with the addition of the DA. This raises the question of whether routine DA at the time of MPFLR can be recommended as an intervention that improves diagnosis and alters treatment with minimal risk.
Drawbacks of adding DA to MPFLR include financial implications that may outweigh any benefits, particularly in cases where the chondral lesion is minor or absent. 27 In this study, there were 50 patients (32.3%) who had absent chondral injury on MRI and subsequent DA. The additive cost of DA can be a financial and time burden to both patients and health care facilities and institutions. 19 However, identifying chondral lesions during DA may offer the surgeon valuable real-time information that can guide intraoperative decision-making and patient counseling, particularly when preoperative MRI is inconclusive. While this study found no differences in short-term clinical outcomes across the groups, there is a need for further evaluation of mid-term and long-term outcomes for the pediatric cohort undergoing MPFLR with and without DA.
Limitations
This study has several limitations. First, the retrospective nature of the cohort introduces the potential for selection and information bias. However, this was addressed by using consistent inclusion criteria and detailed patient records. Second, this study was conducted at a single institution with only 2 surgeons, which may limit the generalizability of the findings to other centers with varying surgical techniques and patient populations. To mitigate this, surgical protocols were standardized, thereby enhancing internal validity and reducing variability within the procedures. The relatively small sample size of 155 patients further limits the study’s power to detect subtle differences in outcome measures, particularly in subgroup analyses. Additionally, the reliance on MRI reports and operative notes may introduce variability in the assessment of chondral injuries, as imaging interpretations and arthroscopic findings can differ based on the physician’s experience and subjective grading of chondral defects. This also resulted in a lack of standardized reporting and details on the chondral defects, including size. This was partially mitigated by ensuring that both surgeons were fellowship-trained in sports medicine. In addition, detailed MRI sequencing information, including the use of T2-STAR or slice thickness, was not consistently available, limiting our ability to evaluate how MRI quality may have affected chondral lesion detection. Furthermore, other biases may have been introduced, as all patients underwent DA before MPFLR, thus eliminating the ability to have a control group of patients who did not undergo DA before MPFLR. Despite this, the sequential approach reflects real-world clinical practice, adding relevance to the findings. Another potential limitation is the lack of pathoanatomic characterization within the study cohorts, as these data were not available. Additionally, the follow-up period was relatively short, which limits the ability to assess longer-term functional outcomes and the durability of surgical results. Finally, this study lacked patient-reported outcome measures, which limited the ability to evaluate patient satisfaction and function, as well as any data on return to sport.
Conclusion
Diagnostic arthroscopy detected cartilaginous injuries in 43.8% of pediatric patients with negative preoperative MRI undergoing MPFLR. This highlights the diagnostic value of DA in identifying injuries that would otherwise remain undetected.
Footnotes
Final revision submitted July 23, 2025; accepted August 19, 2025.
One or more of the authors has declared the following potential conflict of interest or source of funding: S.V.T. has received hospitality payments from Stryker Corp and ImpactOrtho. E.L. has received support for education from Medical Device Business Services, Inc, and hospitality payments from Stryker Corp and ImpactOrtho. N.V. has received hospitality payments from Depuy Synthes Sales Inc. H.M. has received consulting fees from Medical Device Business Services, Inc, and hospitality payments from ImpactOrtho, Arthrex, DePuy Synthes Sales Inc, and Stryker. J.V. has received hospitality payments from OrthoPediatrics Corp, Smith & Nephew Inc, Stryker Corp, and ImpactOrtho. AOSSM checks author disclosures against the Open Payments Database (OPD). AOSSM has not conducted an independent investigation on the OPD and disclaims any liability or responsibility relating thereto.
Ethical approval for this study was obtained from Phoenix Children’s Hospital (# IRB-24-107).
