Patellar Tendon–Lateral Trochlear

Ridge Distance
A Novel Measurement of Patellofemoral Instability
R. Justin Mistovich,* MD, John W. Urwin,y BS, Peter D. Fabricant,z MD, MPH,
and J. Todd R. Lawrence,y§ MD, PhD
Investigation performed at Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA

Background: Abnormalities in the trochlea-patella-tibia relationship have been shown to be risk factors for recurrent patellofe-
moral instability, although no current measurements quantify patellar containment in the trochlea. Standard measurements,
such as tibial tubercle–trochlear groove (TT-TG) distance, do not account for the containment of the patella by the trochlea.
Our goal was to develop a measurement to assess how well the trochlea contained the extensor mechanism.
Hypothesis: A novel measurement describing the amount of the patellar tendon lateral to the lateral trochlear ridge (PT-LTR)
would be a reliable measurement and significantly greater among patients with patellofemoral instability.
Study Design: Cohort study (diagnosis); Level of evidence, 2.
Methods: The authors analyzed radiology records from 2005 to 2014 for patients aged 5 to 18 years with and without patellofe-
moral dislocations who had knee magnetic resonance imaging (MRI). Two blinded reviewers evaluated 215 MRI studies. Standard
and novel morphology measurements were calculated for each knee and compared in a case-control design. Interobserver reli-
ability of each measure was assessed by the intraclass correlation coefficient. Predictability for patellofemoral dislocation was
calculated with 2-tailed independent-samples Student t tests. Discriminative capacity was calculated with receiver operating
characteristic analyses and area under the curve (AUC). An optimal measurement cutoff with resultant sensitivity and specificity
was calculated.
Results: Standard measurements of TT-TG distance, tangential axial width of the patella (TAWP), and tangential axial trochlear
width (TATW) had excellent agreement between raters; lateral femoral condyle length had good agreement; and the novel
measurement—width of the tendon beyond the lateral femoral condyle (PT-LTR)—also had excellent agreement. These under-
went predictability and discriminative capacity analyses. TT-TG, TAWP, TATW, and PT-LTR were significant predictors of patel-
lofemoral instability. In receiver operating characteristic analysis, TAWP had an AUC of 0.65, below the 0.8 threshold. TATW had
an AUC of 0.814 and, when \32.5 mm, was 76% sensitive and 77% specific for dislocations. TT-TG demonstrated an AUC of
0.806. TT-TG 13.5 mm was 76% sensitive and 76% specific for dislocations. PT-LTR demonstrated an AUC of 0.876 and, when
5.55 mm, was 73% sensitive and 89% specific for patellofemoral dislocation.
Conclusion: PT-LTR is reliable, predictable, and discriminative for patellofemoral dislocations. This measurement had sensitivity
similar to that of TT-TG but with higher specificity.
Keywords: patellofemoral dislocation; MRI measurements; pediatric knee dislocation; TT-TG

Patellofemoral dislocation is one of the most common trau-
matic knee injuries among pediatric and adolescent athletes,
carrying an annual incidence of 43 per 100,000—more than
double the rate for adults.1,13 There is a high rate of associ-
ated injuries and long-term complications, with up to 25%
of patients having an associated osteochondral fracture.3
Affected individuals are at risk of continued impairment,
with long-term follow-up demonstrating that as many as
The American Journal of Sports Medicine
1–7
DOI: 10.1177/0363546518809982
Ó 2018 The Author(s)
71% of patients suffer a recurrent dislocation while facing
a higher incidence of patellofemoral arthritis.5,7,14,16,17
The stability of the patellofemoral articulation is a com-
plex interplay between static and dynamic forces generated
by the surrounding osseous and soft tissue structures.
Abnormalities in just 1 or several factors can predispose
patients to patellofemoral instability. The overall lateral
force vector on the patella, established mainly by lateraliza-
tion of the tibial tubercle and other rotational deformities
about the knee, must be balanced by soft tissue and bony
restraints: primarily the medial patellofemoral ligament
and the patella-trochlear architecture, respectively. Ana-
tomic factors noted to be associated with patellar instability

1
2 Mistovich et al The American Journal of Sports Medicine

thus include (1) those that increase the degree of laterally
directed force and, therefore, the tibial tubercle–trochlear
groove (TT-TG) distance and (2) those that indicate less
soft tissue or bony restraints, such as trochlear dysplasia
and patella alta.10
Surgical decision making is often based on simply iden-
tifying factors out of the range of normal, rather than
understanding how those factors in combination contribute
to the overall instability picture. For instance, a TT-TG dis-
tance out of the normal range is often used to indicate a tib-
ial tubercle osteotomy. However, it was recently noted that
having a very lateralized tibial tubercle is not pathologic if
there is a large and deep trochlear groove to balance this
force.8 What therefore seems to be important is not specif-
ically how far the distal tibial tubercle attachment is rela-
tive to the center of the trochlear grove (TT-TG distance)
but whether this laterally directed force can be dynami-
cally balanced during knee range of motion by the bony
and soft tissue restraints. Thus, when patients with patel-
lar instability are assessed, it would seem that the real
question should be, can the bony and soft tissue con-
straints balance the net lateral vector force generated by
the extensor mechanism? Asked another way, can the
trochlea and the medial and lateral soft tissue restraints
contain the extensor mechanism?
Given the dynamic interplay of biomechanical factors,
no single measurement has been identified that describes
every contributing component of patellar instability.
Thus, we sought to (1) identify and assess novel magnetic
resonance imaging (MRI)–based alignment measures that
may better evaluate the containment of the extensor mech-
anism within the trochlear architecture and (2) compare
those measures with existing ones. We hypothesized that
our novel measurements—specifically, the width of patel-
lar tendon beyond the lateral trochlear ridge—would
have an improved ability to predict patellofemoral disloca-
tions among children and adolescents when compared with
standard measures of lateralization.
METHODS
After obtaining approval from the institutional review
board at our tertiary care children’s hospital, we queried
radiology records for patients over a 10-year period from
2005 to 2014 who carried a diagnosis of a patellofemoral
dislocation and had knee MRI. Over the same period, we
also identified a control group of patients who had a knee
MRI but did not have a documented history of a patellofe-
moral dislocation. All MRIs was obtained via an identical
protocol with the knee in extension and immobilized in
a coil on a variety of 1.5- to 3-T scanners. The charts for
these patients were then reviewed to confirm the presence
or absence of a patellofemoral dislocation and ensure that
patients were assigned to the appropriate groups.
Patients were also excluded per the following findings:
poor-quality MRI imaging that was not interpretable, age
\5 years or .18 years, and MRI not accessible or measur-
able in the picture archiving and communication system.
While some patients ultimately developed recurrent insta-
bility and some eventually had multiple MRI scans, we
measured only those that were obtained after a first-time
patellar dislocation. A de-identified randomized patient
list was then compiled, blinding reviewers to the diagnosis.
Standard measurements recorded included the follow-
ing (Figure 1):
 TT-TG distance
 Tibial tuberosity–posterior cruciate ligament distance
 Tangential axial width of the patella (TAWP)
 Tangential axial trochlear width (TATW)
 Lateral trochlear inclination
 Length of medial trochlear facet
 Length of lateral trochlear facet
 Trochlear facet asymmetry (calculated by dividing the
length of the medial femoral condyle by the length of
the lateral femoral condyle [LFC])
Novel measurements were as follows (Figure 2):
 Axial width of patellar tendon
 Axial width of patellar tendon beyond lateral trochlear
ridge (PT-LTR)
 Lateral trochlear ridge–trochlear groove distance
Novel Measures
Axial Width of the Patellar Tendon. This measurement
is calculated on axial MRI cuts (Figure 2A). By scrolling
proximally to distally, the first cut demonstrating the ten-
don free from bone is identified. A straight line measuring
the distance from the medial edge of the tendon to its lat-
eral edge is drawn and measured.
Axial Width of the Patellar Tendon Beyond the Lateral
Trochlear Ridge. This measurement is also calculated on
axial MRI cuts (Figure 2B). The first distal cut demonstrat-
ing the patellar tendon without bone is again identified,
and the most lateral point of the patella is marked. The
apex of the lateral trochlear ridge is identified and marked
with a line perpendicular to the posterior condylar axis. If
this cannot be identified on the same cut, this may be
marked on the most visible cut, then translated to the

§
Address correspondence to J. Todd R. Lawrence, MD, PhD, School of Medicine, University of Pennsylvania, Wood Building, Second Floor, 34th Street
and Civic Center Boulevard, Philadelphia, PA 19104-4399, USA (email: lawrencej@email.chop.edu).
*School of Medicine, Case Western Reserve University, Cleveland, Ohio, USA.
y
Division of Orthopaedics, Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA.
z
Division of Pediatric Orthopaedic Surgery, Hospital for Special Surgery, New York, New York, USA.
One or more of the authors has declared the following potential conflict of interest or source of funding: J.T.R.L. receives royalties from Sawbones Inc
and has received a grant from the OMeGA Medical Grants Association. R.J.M. is a paid consultant for Orthopediatrics; receives educational support from
Arthrex Inc; and holds stock or stock options in Right Mechanics. 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.
AJSM Vol. XX, No. X, XXXX Pediatric Patellofemoral Instability 3

Figure 1. (A) Tibial tubercle–trochlear groove distance (TT-TG). (B) Tibial tuberosity–posterior cruciate ligament distance
(TT-PCL). (C) Tangential axial width of the patella (TAWP). (D) Tangential axial trochlear width (TATW). (E) Lateral trochlear incli-
nation. (F) Length of medial trochlear facet. (G) Length of lateral trochlear facet.
4 Mistovich et al
Figure 2. (A) Axial width of patellar tendon (WPT). (B) Axial width of patellar tendon beyond lateral trochlear ridge (PT-LTR). (C)
Lateral trochlear ridge–trochlear groove distance.
tendon cut, similar to calculating TT-TG. The distance
between the points is then measured in line with the ten-
don itself. Distances lateral to the trochlear ridge are
assigned a negative value.
Lateral Trochlear Ridge–Trochlear Groove Distance.
This measurement is also calculated on the axial MRI
cuts (Figure 2C). The posterior condylar axis is drawn.
Two lines perpendicular are also drawn: 1 intersecting
the trochlear groove and 1 intersecting the apex of the lat-
eral trochlear ridge. The distance between these, parallel
to the posterior condylar axis, is measured.
Two raters evaluated the studies. A total of 215 ratings
were evaluated for reliability by both raters. Statistical
analysis is described in detail in the corresponding section.
In summary, further analysis was performed only on meas-
ures that passed the initial reliability screen. A predictabil-
ity screen was then performed to determine the ability of
measurements to identify differences between the experi-
mental and control groups. Measurements demonstrating
statistically significant differences between groups then
underwent an analysis of discriminative capacity with
the receiver operating characteristic (ROC).
Statistical Analysis
Statistical analyses (SPSS v 22; IBM) were performed by
members of the research team with advanced training
in biostatistics (P.D.F.). Interobserver reliability for each
measurement was assessed by calculating the intraclass
The American Journal of Sports Medicine
correlation coefficient via a 2-way random effects model
with absolute agreement between raters 1 and 2 (Table 1).
Intraclass correlation coefficients were interpreted with the
criteria of Landis and Koch.2,9 Differences between disloca-
tors and controls were assessed with 2-tailed independent-
samples Student t tests, and assessment of discriminatory
capacity was performed with ROC analyses. For variables
with an area under the curve (AUC) .0.80, a measurement
cutoff with resultant sensitivity and specificity was recorded.
The AUC for each ROC curve was calculated, and the predic-
tor variable discrimination adequacy was determined by
comparing those values with currently accepted standards
of discriminative capability (AUC  0.80).2
RESULTS
There were 178 patients in the experimental group and 37
in the control group. For patients in the dislocation group,
MRI was performed at a mean 6 SD 16.8 6 23.8 days after
injury: 38.4% were performed within 1 week, 28.9% in the
second week, and the remainder after 2 weeks from injury.
As judged by a trained pediatric radiologist, 57 (32%) of the
178 dislocators had no effusion; 45 (25%), small effusion; 28
(16%), moderate effusion; and 48 (27%), large effusion.
For nondislocators, 9 patients had MRI performed
39.9 6 51.1 days after a traumatic injury. The major indi-
cations for control studies receiving MRI included knee
pain, history of trauma, and knee mass or swelling.
AJSM Vol. XX, No. X, XXXX Pediatric Patellofemoral Instability 5

TABLE 1
Reliability, Predictability, and Discriminative Capacity Resultsa

Mean 6 SD, mm

Measurement ICCb Nondislocators Dislocators P Value AUCc

Tibial tuberosity–trochlear groove distance 0.822 11.46 6 3.92 16.77 6 5.02 \.001 0.806
Tibial tuberosity–posterior cruciate ligament distance 0.553 — — — —
Tangential axial width of the patella 0.861 40.67 6 5.58 37.91 6 4.97 \.001 0.650
Tangential axial trochlear width 0.856 35.88 6 5.83 26.51 6 9.08 \.001 0.814
Lateral trochlear inclination 0.389 — — — —
Length of medial trochlear facet 0.583 — — — —
Length of lateral trochlear facet 0.654 — — — —
Trochlear facet asymmetry 0.250 — — — —
Axial width of patellar tendon 0.103 — — — —
Axial width of patellar tendon beyond lateral trochlear ridge 0.856 1.54 6 3.47 9.43 6 6.32 \.001 0.876
Lateral trochlear ridge to trochlear groove distance 0.594 — — — —

a
AUC, area under the curve; ICC, intraclass correlation coefficient.
b
ICCs .0.7 are considered reliable (in bold).9
c
AUCs .0.8 are considered discriminative (in bold).2

Of the standard measurements, TT-TG, TAWP, and
TATW had excellent correlation between raters. LFC
length had good correlation, but the remainder of the
established measurements had only fair to poor correlation
between raters. Of the novel measurements, the width of
the tendon beyond the LFC (PT-LTR) had excellent corre-
lation, whereas the lateral trochlear ridge–trochlear
groove) distance and the axial width of the patellar tendon
had only fair and poor correlation, respectively. Thus, as
TT-TG, TAWP, TATW, and PT-LTR had sufficient reliabil-
ity when measured by different raters, further analysis
was limited to these measurements to assess their ability
to predict patients with patellar instability.
A predictability screen was performed to identify statisti-
cally significant differences in measurements between the
dislocator group and the control group with independent-
samples Student t tests (Table 1). TT-TG, TAWP, and
TATW, as well as the new measurement, PT-LTR, each
demonstrated statistically significant differences for
patients in the control group compared with those who
had sustained a dislocation. As noted previously, the tibial
tubercle was significantly lateralized relative to the troch-
lear groove for instability patients versus control patients,
with the mean TT-TG distance being 17 mm in the disloca-
tors and 11 mm in controls (P \ .001). The width of the
trochlea was significantly narrower in the dislocators versus
the controls (27 mm vs 36 mm, P \ .001), and the patellar
width was significantly smaller in the dislocators than the
controls (38 mm vs 41 mm, P \ .001), although the clinical
significance of this is unclear. Patients with instability were
further noted to have significantly more patellar tendon
width outside the lateral trochlear ridge (PT-LTR), indicat-
ing less containment of the extensor mechanism, with dislo-
cators having 9 mm outside the groove versus controls
having only 2 mm (P \ .001).
Finally, the discriminative capability of those measure-
ments that demonstrated satisfactory reliability and predict-
ability was further assessed with ROC analyses (Figure 3).
Tangential axial width of the patella had an AUC of 0.65.
As this value was below the threshold (\0.8), no further anal-
ysis was performed. TATW had an AUC of 0.814. At values
\32.5 mm, it was 76% sensitive and 77% specific for
a dislocation.
TT-TG demonstrated an AUC of 0.806, and a TT-TG dis-
tance 13.5 mm was 76% sensitive and 76% specific for
a patellar dislocation. The width of the patellar tendon
beyond the lateral trochlear ridge (PT-LTR) demonstrated
an AUC of 0.876. At values 5.55 mm, this measurement
was 73% sensitive and 89% specific for a patellofemoral
dislocation.
DISCUSSION
Patellofemoral instability remains a common yet inher-
ently complex problem. Greater than 100 surgical proce-
dures have been described to treat this condition.6,15
Similarly, many measurements have been proposed to
quantify patients at risk of instability and assist surgeons
with operative planning. However, each measurement
does not fully account for the multiple potential anomalies
within the patellofemoral joint that can result in instabil-
ity or how these anomalies in concert result in the final
containment of the patella in the trochlea.
Recent studies identified the importance of trochlear dys-
plasia in contributing to patellofemoral instability. In a study
of pediatric and adolescent patients after initial traumatic
patellofemoral dislocations treated nonoperatively, skeletally
immature patients with trochlear dysplasia had only a 31%
chance of success, as opposed to a 62% overall rate of success
with nonoperative treatment in the entire cohort.11 Addition-
ally, in examining risk factors for failure of isolated MPFL
reconstruction, a recent study identified trochlear dysplasia
as the most important predictive factor. Interestingly, in
this study, TT-TG distance in isolation did not affect out-
comes of isolated medial patellofemoral ligament reconstruc-
tion; only for those patients with accompanying trochlear
dysplasia did TT-TG become important.8
6 Mistovich et al
Figure 3. Discriminative capacity: axial width of patellar tendon beyond lateral trochlear ridge and tibial tubercle–trochlear groove
distance.
Further studies recently identified additional limita-
tions of TT-TG distance in predicting the dynamic relation-
ship between the patella and the trochlea. In a study
comparing static TT-TG measurements of the knee in
extension with tracking measured on dynamic MRI, the
authors found that .60% of patients with TT-TG distances
15 mm actually tracked neutrally or medially when the
measurement was performed dynamically.3
In addition to existing limitations with TT-TG to quan-
tify patellar containment, current classifications and meas-
urements of trochlear dysplasia have limitations. Trochlear
dysplasia is classically described by a sulcus angle .145°.12
Additional radiographic findings of trochlear dysplasia
noted on lateral radiographs include the crossing sign,
a supratrochlear spur, and a hypoplastic medial facet.4
While the Dejour classification descriptively classifies troch-
lear dysplasia, it does not provide a quantitative numerical
value or, perhaps most important, relate it to the ability of
the trochlea to contain the extensor mechanism.
In an effort to solve some of the limitations with preexist-
ing measurements, our study attempted to better quantita-
tively describe axial pathology predisposing pediatric and
adolescent patient populations to patellofemoral instability.
We found that our novel measurement of the PT-LTR was
reliable, predictable, and discriminative for patients with
patellofemoral dislocations. This measurement had sensitiv-
ity similar to that of TT-TG and was more specific. We spec-
ulate that PT-LTR may provide advantages over algorithms
that utilize TT-TG distance alone or ones that pair TT-TG
with a categorical description of trochlear dysplasia, as it
provides a static quantification of the relationship between
The American Journal of Sports Medicine
the extensor mechanism and the region of the trochlea ulti-
mately responsible for its containment. Further studies are
needed to better understand the direct effect of variables
such as patella alta and trochlear dysplasia on this mea-
surement. We suspect that patients with such conditions
may be more susceptible to loss of patellar containment,
which could be noted with PT-LTR but may result in normal
or near-normal TT-TG measurements (Figure 4). However,
further studies are needed to investigate this hypothesis. If
dynamic MRI assessments ever become widely available,
utilizing PT-LTR to assess containment at different degrees
of knee flexion may prove to be an even more useful tool.
Our study does have some weaknesses. While our catch-
ment area is quite large, patients in 1 region may have
unique variants that are not universally found across
more broad populations. Additionally, a broader multicenter
study with a larger patient population may have identified
discernable differences in some measurements that our
sample size was unable to identify. Furthermore, as we
focused on axial measurements only, studies of sagittal
and coronal plane measurements may find further value
in assessment of these patients. As with other measure-
ments (eg, TT-TG), it is possible that knee flexion could
influence the measurement of PT-LTR. While this is a possi-
ble source of variation, all MRI in this study was performed
with standard protocols. Additionally, even if there were
slight variations in the knee flexion angle, this would be
expected only to increase the variance of the measurements
and bias our results toward the null hypothesis, thus mak-
ing our observed association an underestimate of the true
association. Another potential limitation is the matching
AJSM Vol. XX, No. X, XXXX
Figure 4. Case examples for measurement of PT-LTR vs
TT-TG. Axial magnetic resonance images from (A) a
16-year-old female control patient with knee pain but no
patellar instability and (B) a 16-year-old female patient with
patellar instability. (A) The TT-TG distance for the control
patient was 14 mm with a PT-LTR of 2 mm. (B) The TT-TG
distance for the patient with patellar instability was similar
at 15 mm, but the PT-LTR was markedly different at
20 mm. The patient with patellar instability went on to
become a recurrent dislocator and required surgical interven-
tion to address her instability. PT-LTR, axial width of patellar
tendon beyond lateral trochlear ridge (arrows on the lower
portion of panels on the tibial cuts); TT-TG, tibial tubercle–
trochlear groove (arrows on the upper portion of the panels
on the femoral cuts).
of the control comparison group. All differences between the
groups are potentially subject to bias. However, given that
the described anatomic parameters mirror what other stud-
ies reported and that the anatomic differences described by
the novel measurements should contribute to instability, it
is more likely that any differences in the groups represent
the anatomic differences between dislocators and normal
knees and not a selection bias.
In conclusion, no single measurement can be universally
applied to all patients with patellofemoral instability. None-
theless, the length of the patellar tendon lateral to the
trochlear ridge (PT-LTR) considers trochlear morphology
For reprints and permission queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav.
Pediatric Patellofemoral Instability 7
and extensor mechanism lateralization in a single measure-
ment near the level of the patellar-trochlear interface, giv-
ing a quantitative description of patellar containment that
is reproducible, predictive, and highly discriminative for
patellofemoral instability.
REFERENCES
1. Beasley LS, Vidal AF. Traumatic patellar dislocation in children and
adolescents: treatment update and literature review. Curr Opin
Pediatr. 2004;16(1):29-36.
2. Bradley A. The use of the area under the ROC curve in the evaluation of
machine learning algorithms. Pattern Recognition. 1997;30(7):1145-1159.
3. Carlson VR, Sheehan FT, Shen A, Yao L, Jackson JN, Boden BP. The
relationship of static tibial tubercle-trochlear groove measurement and
dynamic patellar tracking. Am J Sports Med. 2017;45(8):1856-1863.
4. Dejour D, Le Coultre B. Osteotomies in patello-femoral instabilities.
Sports Med Arthrosc. 2007;15(1):39-46.
5. Fithian DC, Paxton EW, Stone ML, et al. Epidemiology and natural history
of acute patellar dislocation. Am J Sports Med. 2004;32(5):1114-1121.
6. Gausden EB, Fabricant PD, Taylor SA, et al. Medial patellofemoral
reconstruction in children and adolescents. JBJS Rev. 2015;3(10):
01874474-201510000-00001.
7. Hinton RY, Sharma KM. Acute and recurrent patellar instability in the
young athlete. Orthop Clin North Am. 2003;34(3):385-396.
8. Kita K, Tanaka Y, Toritsuka Y, et al. Factors affecting the outcomes of
double-bundle medial patellofemoral ligament reconstruction for
recurrent patellar dislocations evaluated by multivariate analysis.
Am J Sports Med. 2015;43(12):2988-2996.
9. Landis JR, Koch GG. The measurement of observer agreement for
categorical data. Biometrics. 1977;33(1):159-174.
10. Lewallen L, McIntosh A, Dahm D. First-time patellofemoral disloca-
tion: risk factors for recurrent instability. J Knee Surg. 2015;28(4):
303-309.
11. Lewallen LW, McIntosh AL, Dahm DL. Predictors of recurrent insta-
bility after acute patellofemoral dislocation in pediatric and adoles-
cent patients. Am J Sports Med. 2013;41(3):575-581.
12. Malghem J, Maldague B. Depth insufficiency of the proximal troch-
lear groove on lateral radiographs of the knee: relation to patellar dis-
location. Radiology. 1989;170(2):507-510.
13. Nietosvaara Y, Aalto K, Kallio PE. Acute patellar dislocation in chil-
dren: incidence and associated osteochondral fractures. J Pediatr
Orthop. 1994;14(4):513-515.
14. Palmu S, Kallio PE, Donell ST, Helenius I, Nietosvaara Y. Acute patel-
lar dislocation in children and adolescents: a randomized clinical trial.
J Bone Joint Surg Am. 2008;90(3):463-470.
15. Parikh SN, Lykissas MG. Classification of lateral patellar instability in
children and adolescents. Orthop Clin North Am. 2016;47(1):
145-152.
16. Sanders TL, Pareek A, Johnson NR, Stuart MJ, Dahm DL, Krych AJ.
Arthritis after lateral patellar dislocation. Am J Sports Med. 2017;
45(5):1012-1017.
17. Stefancin JJ, Parker RD. First-time traumatic patellar dislocation:
a systematic review. Clin Orthop Relat Res. 2007;455:93-101.