Musculoskeletal Imaging • Original Research

Finkenstaedt et al.
Ankle MRI

Musculoskeletal Imaging
Original Research
Downloaded from www.ajronline.org by Kaohsiung Medical University on 08/31/18 from IP address 163.15.154.53. Copyright ARRS. For personal use only; all rights reserved

The Calcaneal Crescent in Patients

With and Without Plantar Fasciitis:
An Ankle MRI Study
Tim Finkenstaedt 1,2 OBJECTIVE. The bundled, crescent-shaped trabeculae within the calcaneal tuberosity—
Palanan Siriwanarangsun1,3 which we term and refer to here as the “calcaneal crescent”—may represent a structural adap-
Sheronda Statum1,4 tion to the prevailing forces. Given Wolff law, we hypothesized that the calcaneal crescent
Reni Biswas1,4 would be more robust in patients with plantar fasciitis, a syndrome in part characterized by
Karen E. Anderson5 overload of the Achilles tendon–calcaneal crescent–plantar fascia system, than in patients
without plantar fasciitis.
Won C. Bae1,4
MATERIALS AND METHODS. MR images of 37 patients (27 women and 10 men;
Christine B. Chung1,4 mean age ± SD, 51 ± 13 years; mean body mass index [BMI, weight in kilograms divided by
Finkenstaedt T, Siriwanarangsun P, Statum S, et al. the square of height in meters], 26.8 ± 6.3) referred for workup of foot or ankle pain were retro-
spectively evaluated by two blinded readers in this study. Patients were assigned to two groups:
Keywords: Achilles tendon, bone adaption, calcaneal group A, which was composed of 15 subjects without clinical signs or MRI findings of Achil-
tuberosity, calcaneus, MRI, plantar fascia, plantar les tendon–calcaneal crescent–plantar fascia system abnormalities, or group B, which was com-
fasciitis, tuber calcanei posed of 22 patients with findings of plantar fasciitis. The thickness and cross-sectional area
doi.org/10.2214/AJR.17.19399
(CSA) of the Achilles tendon, calcaneal crescent, and plantar fascia were measured on proton
density (PD)-weighted MR images. The entire crescent volume was manually measured using
Received December 9, 2017; accepted after revision OsiriX software on consecutive sagittal PD-weighted images. Additionally, contrast-to-noise ra-
March 20, 2018.
tio (CNR) as a surrogate marker for trabecular density and the mean thickness of the calcaneal
Based on a presentation at the Society of Skeletal crescent were determined on PD-weighted MR images. The groupwise difference in the mor-
Radiology 2018 annual meeting, Austin, TX. phologic measurements were evaluated using ANOVA with BMI as a covariate. Partial corre-
This study was supported by the Radiological Society of
lation was used to assess the relationships of measurements for the group with plantar fasciitis
North America (Research Fellow Grant number RF1730), (group B). Intraclass correlation coefficient (ICC) statistics were performed.
Swiss National Science Foundation (grant number RESULTS. Patients with plantar fasciitis had a greater CSA and volume of the calcaneal
P2SKP3_168412), and Swiss Society of Radiology. crescent and had lower CNR (i.e., denser trabeculae) than those without Achilles tendon–cal-
The contents of this article are solely the responsibility caneal crescent–plantar fascia system abnormalities (CSA, 100.2 vs 73.7 mm2, p = 0.019; vol-
of the authors and do not necessarily represent the ume, 3.06 vs 1.99 cm3, p = 0.006; CNR, –28.40 vs –38.10, p = 0.009). Interreader agreement
official views of the U.S. Department of Veterans Affairs was excellent (ICC = 0.85–0.99).
or the Swiss National Science Foundation. CONCLUSION. In patients with plantar fasciitis, the calcaneal crescent is enlarged
1
Department of Radiology, University of California,
compared with those without abnormalities of the Achilles tendon–calcaneal crescent–plan-
San Diego, School of Medicine, San Diego, CA. tar fascia system. An enlarged and trabeculae-rich calcaneal crescent may potentially indi-
2
cate that abnormally increased forces are being exerted onto the Achilles tendon–calcaneal
Institute of Diagnostic and Interventional
crescent–plantar fascia system.
­ adiology, University Hospital Zurich, University of
R
Zurich, Switzerland.
he Achilles tendon–calcaneal pearance in the sagittal plane [2, 3]. There-

T
3
Department of Radiology, Siriraj Hospital, ­Mahidol
University, Bangkok, Thailand.
crescent–plantar fascia system fore, we would like to introduce the term
was first described by Arandes “calcaneal crescent” for the first time to de-
and Viladot in 1953 and com- scribe this osseous component of the Achil-
4
Department of Radiology, VA San Diego Healthcare
System, 3350 La Jolla Village Dr, MC-114, San Diego, CA
prises the Achilles tendon, the calcaneus and les tendon–calcaneal crescent–plantar fascia
92161. Address correspondence to C. B. Chung
(cbchung@ucsd.edu). its trabecular system, the plantar fascia, and system that is located within but is not con-
the intrinsic plantar foot muscles [1]. The gruent with the calcaneal tuberosity [2].
5
Private Podiatry Practice, San Diego, CA. bony connection between the insertion to the The biomechanical role of the calcaneal
AJR 2018; 211:1–8 Achilles tendon and the insertion to the plan- crescent is complex, especially if the 3D shape
tar fascia within the calcaneal tuberosity and forces involved in the dynamic gait cycle
0361–803X/18/2115–1
(synonym: tuber calcanei) consists of coarse are considered. Two main forces—that is, com-
© American Roentgen Ray Society trabeculae that have a crescent-shaped ap- pression and tensile forces—are exerted onto

AJR:211, November 2018 1

 Finkenstaedt et al.

the calcaneal crescent (Fig. 1). The compres- velopment, the initially nonossified calcaneal Study Population
sion force is derived from the body weight via apophyseal precursor is rounded, unlike the MR images of 37 patients referred for workup
the talus and from the prevailing ground reac- secondary calcaneal ossification centers that of foot or ankle pain were retrospectively evalu-
tion force (GRF) [4]. The GRF is applied over have a caplike appearance with increasingly ated in this study. Patients were assigned to two
Downloaded from www.ajronline.org by Kaohsiung Medical University on 08/31/18 from IP address 163.15.154.53. Copyright ARRS. For personal use only; all rights reserved

the entire bottom of the foot, and its distribu- linearly orientated crescent-shaped trabeculae groups. Group A was composed of 15 patients
tion during dynamic gait introduces a range [15]. Fusion of the apophysis begins at about without clinical signs and MRI findings of abnor-
of compression loads that are absorbed by the ages 12–15 years and completes at ages 15–20 malities of the Achilles tendon–calcaneal cres-
windlass mechanism of the foot [5, 6]. These years [16]. cent–plantar fascia system (mean age ± SD, 47 ±
loading forces are in opposition to the two Plantar fasciitis is the most common cause 15 years; body mass index [BMI, weight in kilo-
major tensile forces from the Achilles tendon of inferior heel pain [17, 18] and originates grams divided by the square of height in meters],
and the plantar fascia. While the calf muscles most likely from altered biomechanics (e.g., 23.7 ± 3.5). Twelve patients in group A were fe-
linked to the Achilles tendon actively contract, flattened longitudinal arch [19] or increased male (age, 47 ± 17 years; BMI, 23.1 ± 3.6), and
the tensile force on the plantar fascia passive- Achilles tendon tension [20]), leading to three were male (age, 47 ± 15 years; BMI, 26.2 ±
ly increases when loading is applied during the overuse and microtrauma with subsequent 2.0). Group B was composed of 22 patients with
gait cycle [7, 8]. Resulting from the addition of inflammation at its insertion [21, 22]. Al- clinical signs or MRI findings (or both) of plantar
both of these tensile force vectors, an additional tered morphology of the calcaneal crescent fasciitis (age, 54 ± 10 years; BMI, 28.8 ± 7.0). Of
net compression force is exerted onto the calca- may indicate abnormally increased forces the 22 patients in group B, 15 were female (age,
neal crescent. Thus, the tendotuberosity trabec- of the Achilles tendon–calcaneal crescent– 56 ± 11 years; BMI, 26.7 ± 6.0), and seven were
ulae [9] of the calcaneal crescent are exposed to plantar fascia system. Thus, the purpose of male (age, 51 ± 8 years; BMI, 33.3 ± 7.0).
both compression and tensile forces [3]. this study was to investigate morphologic The clinical diagnosis of plantar fasciitis was
According to a study by Kachlik et al. [2] differences of the calcaneal crescent in pa- made by a podiatrist with 26 years of experience.
that was performed on 10 specimens, the cal- tients with plantar fasciitis and in those with- Patients in group B with plantar fasciitis typical-
caneal crescent constitutes a subcortical bone out plantar fasciitis. Given Wolff law, we hy- ly had clinical symptoms such as worsening pain
framework that consists of 8–12 layers of pothesized that the calcaneal crescent would on the undersurface of the heel on palpation, on
cancellous trabeculae [10]. The clinical sig- be more robust in patients with plantar fas- weight-bearing, and when walking after a period
nificance of the calcaneal crescent has been ciitis, a syndrome in part characterized by of rest and characteristic pain that eases with walk-
minimally considered in the literature. Al- overload of the Achilles tendon–calcaneal ing. These symptoms had to be present for at least
though the Achilles tendon is often directly crescent–plantar fascia system, compared 4 weeks for a clinical diagnosis of plantar fasciitis.
continuous with the plantar fascia via a thick- with patients without plantar fasciitis. For patients in group B with plantar fasciitis, the
ened periosteum [11], most mechanical stress symptoms had lasted 14 ± 7 weeks (mean ± SD)
is transmitted through the calcaneal crescent Materials and Methods before the MRI examination was performed. The
[3]. The calcaneal crescent likely represents a This retrospective study was approved by the MRI signs of plantar fasciitis are described later
structural adaption of the cancellous bone to institutional review board of the University of Cal- in this article. Cases with compelling clinical ev-
the prevailing tensile and compression forc- ifornia, San Diego Human Research Protections idence for plantar fasciitis were included and as-
es described, a principle known as Wolff law Program and is compliant with the regulations of signed to group B regardless of the MRI findings.
[12–14]. The entire calcaneal tuberosity rep- the HIPAA. The requirement for informed patient Exclusion criteria were age less than 21 years;
resents an apophyseal equivalent. During de- consent was waived. distinct hind foot deformity; history of calcaneal

Fig. 1—Drawing and inset radiograph of 54-year-

old healthy woman (patient 13 in Table 1) show
calcaneal crescent within Achilles tendon–calcaneal
crescent–plantar fascia system. During standing,
vertical loading force is applied through tibia onto
talus causing, along with ground reaction force (GRF)
applied over entire bottom of foot, compression
force to calcaneal crescent. Furthermore, vertical
loading causes passive stress to longitudinal arch
and, consecutively, tensile forces on plantar fascia
(i.e., windlass mechanism). Contraction of calf
muscles actively generates Achilles tendon tension.
Net compressive force (dashed arrow) results from
addition of both tensile force vectors being exerted
onto calcaneal crescent. In sum, rotational moment
arm is created from application of calf muscle
contraction (Achilles tension) and downward weight
vectors. This torque is being countered mainly by
tension applied onto plantar fascia. During gait cycle,
proportions of respective prevailing forces vary, and
several other complex structures are involved in
active and passive load-sharing system of foot. Solid
arrows show direction of forces.

2 AJR:211, November 2018

 Ankle MRI

fracture, acute calcaneal stress fracture, Haglund MRI signs of plantar fasciitis were a tear or plantar nel ankle coil. The scanning protocol included
deformity, severe Achilles insertion ossifications, fascia thickness more than 5 mm, signal-intensity acquisition of sagittal non–fat-saturated (NFS)
diffuse idiopathic skeletal hyperostosis involving alterations within or adjacent to the plantar fascia, proton density (PD)-weighted spin-echo imag-
the calcaneus, or rheumatoid arthritis; major me- and a bony spur or bone edema of the plantar fas- es (TR/TE, 3250/30; FOV, 130 mm; slice thick-
Downloaded from www.ajronline.org by Kaohsiung Medical University on 08/31/18 from IP address 163.15.154.53. Copyright ARRS. For personal use only; all rights reserved

chanical axis deviation of the lower limbs (e.g., cia insertion [25]. For patients in control group A, ness, 2.5 mm; acquisition matrix, 512 × 256), cor-
coxa vara or valga, genu varum or valgum); neu- the symptoms lasted 19 ± 9 weeks (mean ± SD) be- onal NFS PD-weighted images (TR/TE, 2475/30;
rologic or other pathologic entities hampering the fore the MRI examination was warranted and per- FOV, 130 mm; slice thickness, 2.5 mm; acquisi-
physiologic gait pattern (e.g., neuropathic foot, formed. All MRI examinations were reviewed for tion matrix, 448 × 224), axial NFS PD-weighted
damage to major nerves of the lower extremities, the findings of plantar fasciitis and Achilles tendon images (TR/TE, 2200/35; FOV, 130 mm; thick-
CNS disturbances, long-standing diabetes or neu- abnormalities by a radiologist with 18 years of ex- ness, 2.5 mm; matrix, 384 × 256), and sagittal
romuscular disease); and unusual physical activi- perience in musculoskeletal imaging who did not NFS PD Cube (GE Healthcare) images (TR/TE,
ty level (e.g., highly competitive athletes, long-dis- participate in the later image analysis. 2000/25; FOV, 140 mm; thickness, 0.6 mm; ma-
tance runners, bedridden patients). Additionally, trix, 256 × 256). Furthermore, an axial fat-satu-
patients with significant abnormalities of the tib- Final Diagnoses of Group A Patients Without rated T2-weighted sequence (TR/TE, 3600/68;
ialis posterior tendon as a dynamic stabilizer of Plantar Fasciitis FOV, 130 mm; thickness, 2.5 mm; matrix, 384
the longitudinal foot arch [23, 24] were excluded. The final diagnoses made by the treating podi- × 224), a coronal fat-saturated T2-weighted se-
All these criteria were assessed by the referring atrist, after consideration of the clinical and MRI quence (TR/TE, 3925/70; FOV, 130 mm; thick-
podiatrist who also considered radiographs when findings, for the 15 patients assigned to group A ness, 2.5 mm; matrix, 384 × 224), and a sagit-
applying the exclusion criteria. If a patient with a without abnormalities of the Achilles tendon–cal- tal fat-saturated T2-weighted sequence (TR/TE,
diagnosis of plantar fasciitis had one of the abnor- caneal crescent–plantar fascia system were the 3600/68; FOV, 130 mm; thickness, 2.5 mm; ma-
malities mentioned in the exclusion criteria, then following: four patients (27%) had no significant trix, 448 × 224) were acquired for complete as-
that patient was excluded. Patients were assigned findings; five (33%) had abnormalities of the ex- sessment of the MRI examination.
to control group A if neither clinical nor MRI tensor, flexor, or peroneal tendons (except tibialis
findings were suggestive of an abnormality of the posterior tendon, see exclusion criteria); two (13%) Image Analysis
Achilles tendon–calcaneal crescent–plantar fascia had an abnormality of the deltoid ligament; two All measurements were performed by two
system (e.g., Achilles tendon or plantar fascia ab- (13%) had an abnormality of the anterior talofibu- board-certified radiologists with 7 and 6 years of
normality); for the purpose of this study, we refer lar ligament; one (7%) had mild degenerative os- experience in musculoskeletal imaging. Volumet-
to these patients as “healthy.” teoarthritis of a few forefoot joints; and one (7%) ric measurements of the calcaneal crescent were
The following MRI findings were consistent had a sprain of the os peroneum. performed by reader 1 only. Both readers were
with an Achilles tendon abnormality: tear or fu- blinded to patient identification and clinical data
siform thickening of the tendon, sagittal diameter Imaging as well as to the results of the other reader. The
more than 1 cm with or without pathologic signal- All the ankle and foot MRI examinations were thickness (in millimeters) of the Achilles tendon
intensity alterations of the tendon, and a bony spur performed on a 3-T MRI scanner (MR Discovery (Fig. 2A) was measured on the level of the tibiota-
or bone edema of the Achilles tendon insertion. 750, GE Healthcare) using a dedicated 8-chan- lar joint on sagittal NFS PD-weighted images in

A B C
Fig. 2—Images of 45-year-old woman (patient 35 in Table 1) with left-sided plantar fasciitis are shown to illustrate size measurements.
A, Sagittal proton density (PD)-weighted MR image shows thickness of Achilles tendon (green line, 5.5 mm) measured on level of tibiotalar joint. Blue line indicates
section plane shown in D.
B, Axial PD-weighted MR image shows cross-sectional area (CSA) measurement of Achilles tendon (green outline, 67.6 mm2) measured on level of tibiotalar joint.
C, Sagittal non–fat-saturated (NFS) PD-weighted MR image shows mean thickness measurement of calcaneal crescent (3.6 mm) that is based on initially manually drawn
ROI and is then calculated by custom Matlab software (MathWorks). Software uses myriad data points to measure distance between cranial boundary (color-coded) and
caudal boundary of crescent. Larger distance between both lines is represented by yellow or white color of cranial boundary according to color-coded y-axis. Blue line
indicates section plane shown in E.
(Fig. 2 continues on next page)

AJR:211, November 2018 3

 Finkenstaedt et al.
Downloaded from www.ajronline.org by Kaohsiung Medical University on 08/31/18 from IP address 163.15.154.53. Copyright ARRS. For personal use only; all rights reserved

D E F
Fig. 2 (continued)—Images of 45-year-old woman (patient 35 in Table 1) with left-sided plantar fasciitis are shown to illustrate size measurements.
D, Axial PD-weighted MR image shows CSA measurement of calcaneal crescent (green outline, 76.5 mm2). Blue line in A indicates section plane of this image.
E, Coronal PD-weighted image shows CSA measurement of bigger central cord (arrow) and smaller lateral cord of plantar fascia (green outline, in sum 91.3 mm2). Image
plane was selected at slice with biggest craniocaudal thickness of plantar fascia. Blue line in C indicates section plane of this image.
F, Sagittal NFS PD-weighted Cube (GE Healthcare) image shows entire volume of calcaneal crescent (3.50 cm3) that was calculated by OsiriX software (Pixmeo) after
multiple consecutive ROIs were manually drawn on adjacent sagittal NFS PD-weighted Cube images.

the middle of the Achilles tendon. Care was taken selecting the slice with the greatest craniocaudal Results
to avoid the inclusion of the plantaris tendon or thickness of the plantar fascia and the CSA of the Interreader Agreement
the focal convexity that is occasionally caused at plantar fascia (central and lateral cord) was mea- Measurements of the thickness and CSA
the anterior border of the tendon by insertion of sured on the same slice (Fig. 2E). The volume (in of the Achilles tendon and plantar fas-
soleus fibers [26]. The cross-sectional area (CSA cubic centimeters) of the entire calcaneal crescent cia (central and lateral cords) as well as the
[in square millimeters]) of the Achilles tendon was manually measured using OsiriX volumetric mean thickness, CSA, and CNR of the cal-
(Fig. 2B) was measured on the same level on axi- tool (Pixmeo, version 5.8.2) on consecutive sagit- caneal crescent showed excellent interreader
al NFS PD-weighted images. The mean thickness tal NFS PD Cube images (Fig. 2F). agreement (ICC range, 0.85–0.99; minimum
(in millimeters) and contrast-to-noise ratio (CNR) ICC, 0.85 for CSA measurements of the plan-
of the calcaneal crescent were calculated using Statistical Analysis tar fascia; maximum ICC, 0.99 for CSA mea-
custom Matlab software (version R2015b, Math- Statistical analysis was performed using com- surements of the calcaneal crescent).
Works) (Fig. 2C) on a predefined sagittal NFS mercially available software (Systat 10.2, Systat
PD-weighted image for an indirect evaluation of Software). Categoric variables are expressed as Patients Without Plantar Fasciitis Versus
trabecular density as follows: signal intensity of frequencies and percentages; continuous vari- Patients With Plantar Fasciitis
the calcaneal crescent [defined here as signal1]) ables are expressed as mean ± SD (range). Two The descriptive statistics for the various
subtracted from the signal intensity of the middle readers performed each measurement, and a measurements of patients without abnormal-
of the calcaneus (defined here as signal2) divided mean value was obtained. The t test was used ities of the Achilles tendon–calcaneal cres-
by the SD of the air (noise) [27–29]. We define to evaluate the groupwise difference (i.e., dif- cent–plantar fascia system (group A) and
“trabecular density” as the conglomerate num- ference between group A without abnormalities patients with plantar fasciitis (group B) are
ber and thickness of trabeculae in a given volume. of the Achilles tendon–calcaneal crescent–plan- summarized in Table 1. No significant sex
The signal intensity measurements of the middle tar fascia system and group B with plantar fasci- differences were found between both groups
of the calcaneus and the air were performed us- itis) of BMI and age, and the Pearson chi-square (group A, 12/15 [80%] women; group B, 15/22
ing uniformly square ROIs. The predefined im- test was used to evaluate the groupwise differ- (68%) women; χ2 = 0.43). The mean age of
ages were selected during a consensus reading of ence of sex. ANOVA was performed to evalu- group A was not statistically different from
both readers 4 weeks before the main readout to ate the groupwise difference between the various that of group B (group A vs B, 47 ± 15.2 years
prevent recall bias. The readers were also blinded measurements outlined, with BMI as a covari- vs 54 ± 10.1 years, respectively; p = 0.087),
to the clinical data of the patients during the con- ate. A partial correlation adjusted for BMI was whereas there was a statistical difference be-
sensus reading, too. A lower CNR of the calcaneal used to assess the relationship between the dif- tween the BMI of the groups (group A vs
crescent was regarded as indicating a greater tra- ferent measurements for patients of group B with group B, 23.7 ± 3.5 vs 28.8 ± 7.0; p = 0.014).
becular density. The CSA of the calcaneal cres- plantar fasciitis. Interreader agreement regarding The maximum thickness and the CSA of
cent was measured on axial NFS PD-weighted the various measurements described was evalu- the plantar fascia were greater in group B
images in the superior half of the crescent on the ated with intraclass correlation coefficient (ICC) than group A (maximum thickness, 5.6 vs 3.3
first cranial slice that was not affected by partial statistics. The ICC was defined as follows: poor mm, p < 0.001; CSA, 90.1 vs 47.7 mm2, p <
volume effects (Fig. 2D). The maximal thickness (ICC < 0.7), fair (0.7–0.79), good (0.8–0.89), or 0.001) after adjusting for BMI. Patients with
(in millimeters) of the plantar fascia was mea- high (≥ 0.9) [30]. A p value of < 0.05 was used to plantar fasciitis (group B) had a significantly
sured on coronal NFS PD-weighted images by denote statistical significance. greater CSA, greater volume of the calcaneal

4 AJR:211, November 2018

 TABLE 1: Demographic Characteristics of Patients and Morphologic Measurements of the Three Main Structures
Constituting the Achilles Tendon–CalcanealAnkle MRI
Crescent–Plantar Fascia System, by Patient Group and Patient
Patient Achilles Tendon Calcaneal Crescent Plantar Fascia
Plantar Thickness Mean Volume Maximum
No. Sex Age (y) BMI Fasciitis (mm) CSA (mm2) Thickness (mm) CSA (mm2) (mm3) CNR Thickness (mm) CSA (mm2)
Group Aa
Downloaded from www.ajronline.org by Kaohsiung Medical University on 08/31/18 from IP address 163.15.154.53. Copyright ARRS. For personal use only; all rights reserved

1 M 63 27.8 N 6.8 98.5 3.0 56.0 1.44 −18.69 4.2 45.1

2 F 46 19.2 N 5.5 55.5 3.1 102.5 2.56 −57.30 4.6 41.4
3 F 41 25.8 N 5.6 54.6 3.0 70.8 1.89 −39.10 3.7 54.5
4 M 45 27.1 N 4.4 50.7 3.4 68.4 0.88 −30.00 3.2 41.7
5 F 58 27.0 N 4.6 61.7 3.6 76.4 2.31 −21.11 2.5 38.9
6 F 31 30.1 N 5.8 54.4 2.7 50.1 1.29 −27.63 3.1 89.1
7 F 44 21.6 N 5.4 59.5 3.7 52.7 2.08 −26.37 3.3 52.4
8 F 30 18.8 N 4.1 40.3 3.1 54.5 1.21 −37.70 3.0 40.1
9 F 38 25.1 N 4.8 57.6 3.1 87.0 1.71 −1.44 3.6 65.4
10 F 24 24.0 N 5.8 64.7 3.3 74.0 2.93 −36.44 2.9 46.2
11 F 59 21.0 N 6.1 53.8 3.4 55.2 1.62 −14.81 3.4 46.3
12 F 68 24.9 N 5.8 60.3 3.3 81.8 2.04 −19.80 3.2 43.8
13 F 54 20.2 N 5.3 67.2 2.7 69.4 1.99 −37.15 2.2 24.1
14 F 76 19.7 N 4.8 62.0 3.7 109.0 3.40 −28.96 3.3 36.2
15 M 33 23.8 N 6.4 85.0 3.2 98.6 2.43 −29.49 3.0 50.7
Mean — 47 23.7 — 5.4 61.7 3.2 73.7 1.99 –28.40 3.3 47.7
SD — 15.2 3.5 — 0.8 14.0 0.3 18.9 0.68 13.00 0.6 14.7
Group Bb
16 F 45 30.7 Y 4.9 44.5 2.5 46.4 1.06 −24.86 4.7 66.5
17 F 55 19.8 Y 5.2 56.5 4.0 90.5 2.59 −58.65 8.9 175.2
18 F 54 29.2 Y 7.3 83.1 4.0 88.0 2.72 −52.15 7.9 95.3
19 F 75 21.9 Y 5.5 52.6 3.5 99.1 5.08 −24.83 3.9 56.9
20 F 37 24.3 Y 4.5 66.9 2.4 95.5 2.80 −49.05 5.1 65.4
21 M 54 30.3 Y 6.4 64.2 3.4 87.0 2.30 −43.19 5.5 60.9
22 F 58 18.3 Y 3.7 57.9 2.8 69.9 2.26 −44.49 4.5 71.0
23 M 59 45.9 Y 7.0 82.9 4.1 151.2 4.73 −50.95 6.4 120.1
24 M 61 30.5 Y 6.1 92.5 3.2 140.6 4.77 −44.17 5.9 112.6
25 F 70 29.1 Y 6.1 61.5 3.3 80.0 1.96 −19.67 4.2 59.7
26 F 49 27.4 Y 5.2 66.9 3.4 81.0 2.60 −36.81 6.2 122.2
27 F 61 26.4 Y 5.5 65.7 3.7 112.4 3.25 −22.29 8.6 156.8
28 F 47 33.3 Y 5.3 54.3 3.1 115.3 2.51 −38.49 5.9 95.6
29 F 49 40.7 Y 5.6 58.7 3.8 83.8 2.81 −34.71 6.5 83.7
30 F 56 19.2 Y 5.4 72.4 3.4 107.9 2.84 −41.81 3.1 33.1
31 F 69 30.7 Y 6.2 70.1 3.0 96.7 2.57 −35.92 5.9 64.9
32 M 57 34.0 Y 5.7 82.1 4.2 150.0 4.58 −36.39 6.6 109.3
33 M 44 28.5 Y 6.1 76.0 4.0 125.1 4.48 −30.03 5.9 106.9
34 M 41 39.0 Y 5.4 55.2 3.4 74.7 2.86 −29.87 4.8 79.1
35 F 45 22.2 Y 5.5 67.6 3.6 76.5 3.50 −54.28 4.8 91.3
36 F 66 27.1 Y 5.7 60.2 2.5 76.4 1.72 −30.24 4.5 69.1
37 M 44 25.1 Y 4.5 57.5 3.7 155.6 3.23 −36.32 4.2 86.0
Mean — 54 28.8 — 5.6 65.8 3.4 100.2 3.06 –38.10 5.6 90.1
SD — 10.1 7.0 — 0.8 11.8 0.5 29.2 1.10 11.00 1.5 33.6
Note—Dash (—) indicates not applicable. BMI = body mass index (weight in kilograms divided by the square of height in meters), CSA = cross-sectional area, CNR =
contrast-to-noise ratio, N = no, Y = yes.
aTwelve women and three men who do not have plantar fasciitis and have no abnormalities of the Achilles tendon–calcaneal crescent–plantar fascia system.
bFifteen women and seven men who have clinical or MRI findings (or both) consistent with plantar fasciitis.

AJR:211, November 2018 5

 Finkenstaedt et al.

crescent (Fig. 3), and lower CNR (i.e., denser the mean thickness of the calcaneal crescent Our study showed that patients with plan-
trabeculae) than group A (CSA, 100.2 vs 73.7 (r = 0.49, p = 0.021 and r = 0.52, p = 0.012, tar fasciitis had a significantly greater vol-
mm2, p = 0.019; volume, 3.06 vs 1.99 cm3, respectively) after adjusting for BMI. For pa- ume of the calcaneal crescent than patients
p = 0.006; CNR, –28.40 vs –38.10, p = 0.009) tients without abnormalities of the Achilles without Achilles tendon–calcaneal cres-
Downloaded from www.ajronline.org by Kaohsiung Medical University on 08/31/18 from IP address 163.15.154.53. Copyright ARRS. For personal use only; all rights reserved

(Fig. 4) after adjusting for BMI. tendon–calcaneal crescent–plantar fascia cent–plantar fascia system abnormalities
system (group A), none of these correlations (3.06 vs 1.99 cm3) after adjusting for BMI.
Correlation Between Measurements in were significant. Furthermore, the approximate estimation of
Patients With Plantar Fasciitis the trabecular density in the calcaneal cres-
In patients with plantar fasciitis (group B), Discussion cent using CNR measurements suggested a
the CSA of the Achilles tendon was positively There is no prior work, to our knowl- distinctly greater trabecular density in pa-
correlated with the CSA (r = 0.59, p = 0.004) edge, that has explored the role of the calca- tients with plantar fasciitis. Our latter result
and volume (r = 0.58, p = 0.005) of the calca- neal tuberosity in plantar fasciitis. We have is consistent with a recent study of distance
neal crescent, and the thickness of the Achil- introduced the term “calcaneal crescent” to runners by Best et al. [9] in which high-res-
les tendon was positively correlated with the describe the bundled, crescent-shaped tra- olution extremity CT showed an increase of
mean thickness of the calcaneal crescent (r = beculae within the calcaneal tuberosity. The trabecular thickness in a dose-dependent re-
0.47, p = 0.028) after adjusting for BMI. morphologic differences of the calcaneal lationship. Basically, association does not
Furthermore, in patients with plantar fas- crescent in patients with and in those with- necessarily imply causation. However, tak-
ciitis, the maximum thickness and CSA of out plantar fasciitis and related biomechani- ing the well-known Wolff law into account
plantar fascia were positively correlated with cal considerations are discussed. that trabeculae increase in thickness and

A B C

D E F
Fig. 3—Comparison of calcaneal crescent in healthy patient and calcaneal crescent in patient with plantar fasciitis.
A–F, MR images of 30-year-old healthy woman (patient 8 in Table 1) (A–C) and 59-year-old man with left-sided plantar fasciitis (patient 23 in Table 1) (D–F). Sagittal proton
density (PD)-weighted MR images (A and D) show different morphology of bundled, crescent-shaped trabeculae (arrows, A and D) within calcaneal tuberosity; in D,
prominent calcaneal crescent with coarse, hypointense trabeculae and plantar spur are visible. Sagittal PD-weighted Cube (GE Healthcare) images (B and E) show one
step of volumetric assessment of calcaneal crescent (green outline). After multiple ROI measurements on consecutive slices were performed, OsiriX software (Pixmeo)
calculates entire volume of calcaneal crescent that is shown by wireframe 3D volume-rendered images shown in C and F. Volume of calcaneal crescent in healthy patient
(C) is 1.21 cm3, and volume of calcaneal crescent in patient with plantar fasciitis (F) is 4.73 cm3. Apparently, volume of calcaneal crescent is distinctly greater in patient 23
with plantar fasciitis.

6 AJR:211, November 2018

 Ankle MRI

160 0
150 5
140 –10
130 4

Volume (mm3)
120 –20
Downloaded from www.ajronline.org by Kaohsiung Medical University on 08/31/18 from IP address 163.15.154.53. Copyright ARRS. For personal use only; all rights reserved

CSA (mm2)

110
3

CNR
100 –30
90
80 2 –40
70
60 1 –50
50 p = 0.019 p = 0.006 p = 0.000
40 0 –60
Healthy Plantar Fasciitis Healthy Plantar Fasciitis Healthy Plantar Fasciitis
Patients Patients Patients

A B C
Fig. 4—Box-and-whisker plots show calcaneal crescent morphologic measurements in patients with plantar fasciitis (n = 22) and healthy patients without plantar
fasciitis (n = 15). Morphologic data were adjusted for body mass index as covariate. Whiskers show minimum and maximum values.
A, Plot shows cross-sectional area (CSA) of calcaneal crescent is distinctly greater in patients with plantar fasciitis than in healthy patients (mean, 100.2 vs 73.7 mm2).
B, Plot shows volume of calcaneal crescent is distinctly greater in patients with plantar fasciitis than in healthy patients (mean, 3.06 vs 1.99 cm3).
C, Plot shows contrast-to-noise ratio (CNR) of calcaneal crescent used for approximate trabecular density evaluation was lower in patients with plantar fasciitis than in
healthy patients (mean, –28.40 vs –38.10). Lower CNR of calcaneal crescent (i.e., more hypointense calcaneal crescent) was regarded as sign of greater trabecular density.

number under mechanical stress [13, 14], for chronic biomechanical stress until they foot loading and ultimately the bone geom-
then a conceivable hypothesis is that ab- feature an increase in size [40–43]. etry and trabecular density. For patients in
normally increased tensile and compressive The clinical significance of our basic re- the group A, the symptoms warranting the
forces prevail within the Achilles tendon– search is that our study is an initial morpho- MRI examination lasted a mean of 19 ± 9
calcaneal crescent–plantar fascia system in biomechanical description of the calcane- weeks before the MRI examination was per-
patients with plantar fasciitis. From a physi- al crescent in patients with plantar fasciitis. formed. Within this symptomatic time pe-
cal standpoint, a calcaneal crescent that has Our study may provide the foundation for riod, a less profound bony adaption appears
remodeled to increase in size and posterior further longitudinal studies to elucidate the possible. However, even if disuse osteoporo-
protrusion increases the torque of the ankle bony adaptation process of the calcaneal sis might have occurred to some extent, it is
joint and the lever arm of the Achilles ten- crescent over time. Our study provides evi- unlikely that disuse osteoporosis would af-
don [31, 32] (Fig. 1). This is in line with sev- dence for the conventional explanation of the fect the volume and CSA measurements or
eral studies showing long-standing physical pathogenesis of plantar fasciitis (i.e., chron- the CNR calculations because our CNR cal-
activity leads to an increase of bone geom- ic overuse with microtrauma as a result of culation approach is inherently less prone to
etry [33–35]. Furthermore, long-standing altered biomechanics). Thus, it supports being influenced by the impact of osteoporo-
physical activity leads to changes in bone the current mainstay of conservative treat- sis and image noise. Third, the evaluation of
microarchitecture: Although tensile forc- ment of plantar fasciitis that consists of re- the trabecular density of the calcaneal cres-
es are known to increase the collagen con- duction of tension originating from the calf cent could be estimated only indirectly by the
tent of bone and thus raise the tensile bone muscles and the plantar fasciitis by continu- CNR measurement on NFS PD sequences,
strength, repetitive cyclic deformation caus- ous stretching exercises and weight loss [21]. because no further dedicated sequences for
ing compression forces leads to an increase However, our study design does not allow bone trabecula evaluation were acquired for
of osteoblast activity with a higher calcium distinction whether, if present, the greatest this retrospective study. Fourth, the measure-
apatite concentration of the bone [36–39]. pathologic force originates from the proxi- ments of the calcaneal crescent did not in-
As an ancillary result, our study showed mal side of the Achilles tendon–calcaneal volve the cortical bone. Basically, a great pro-
that in patients with plantar fasciitis the size of crescent–plantar fascia system (i.e., Achil- portion of force is transmitted to the cortical
the calcaneal crescent is positively correlated les tendon) or the distal side (i.e., plantar fas- bone [44, 45]. However, the inclusion of the
with the size of the Achilles tendon and plan- cia) or from the loading force (primary influ- thin cortical bone in the volume or CSA mea-
tar fascia, but no correlation could be found enced by body weight). surements would presumably not change the
in the group A. This first relationship could Our study has limitations. First, there were results greatly and would have required an
be explained by several studies showing that only 15 patients without Achilles tendon– additional measuring technique specifically
in response to mechanical stress (e.g., resis- calcaneal crescent–plantar fascia abnormali- for the cortical bone. Because bone remod-
tance training) the tendon stiffness and the di- ties in our study population, which may ham- eling begins in the trabeculae bone and the
ameter of the tendon collagen fibrils increase per the transferability of our results. Second, cortical bone is the boundary condition of the
[40–42]. The results of group A showed that even if these patients without abnormalities calcaneal crescent, we believe that our mea-
the size of the Achilles tendon and plantar had no clinical or MRI findings related to the surements are representative for the calcaneal
fascia are not linked to the size of the calca- Achilles tendon–calcaneal crescent–plantar crescent. Fifth, the initial consensus reading
neal crescent, which may be explained by the fascia system and we have thoroughly ap- between both readers may have introduced
idea that such a relationship is valid only be- plied the extensive study exclusion criteria, it potential bias, although it took place 4 weeks
yond a certain threshold. Healthy tendons are might be possible that other underlying ab- before the main reading to prevent recall bias.
known to have a high capacity to compensate normalities affect the gait biomechanics or Finally, for size assessment, the exact delin-

AJR:211, November 2018 7

 Finkenstaedt et al.

eation of the anterocranial boundary of the 12. Ruff C, Holt B, Trinkaus E. Who’s afraid of the big signal-to-noise ratio and contrast-to-noise ratio
calcaneal crescent was difficult in some cas- bad Wolff? “Wolff’s law” and bone functional ad- for FMRI data. PLoS One 2013; 8:e77089
es. Nevertheless, the ICC statistics suggested aptation. Am J Phys Anthropol 2006; 129:484–498 29. Kaufman L, Kramer DM, Crooks LE, Ortendahl
reliable measurements between readers. 13. Wolff J. The law of bone remodelling. Berlin, Ger- DA. Measuring signal-to-noise ratios in MR im-
Downloaded from www.ajronline.org by Kaohsiung Medical University on 08/31/18 from IP address 163.15.154.53. Copyright ARRS. For personal use only; all rights reserved

In conclusion, in patients with plantar many: Springer-Verlag, 1986 aging. Radiology 1989; 173:265–267
fasciitis, the calcaneal crescent is enlarged 14. Huiskes R, Ruimerman R, van Lenthe GH, 30. Meyers CR, Blesh TE. Measurement in physical
compared with those without abnormalities ­Janssen JD. Effects of mechanical forces on main- education. New York, NY: Ronald Press, 1962
of the Achilles tendon–calcaneal crescent– tenance and adaptation of form in trabecular bone. 31. Raichlen DA, Armstrong H, Lieberman DE. Cal-
plantar fascia system. An enlarged and tra- Nature 2000; 405:704–706 caneus length determines running economy: im-
beculae-rich calcaneal crescent may poten- 15. Baker BJ. Osteology of infants and children. College plications for endurance running performance in
tially indicate abnormally increased forces Station, TX: Texas A&M University Press, 2005 modern humans and Neanderthals. J Hum Evol
exerted onto the Achilles tendon–calcaneal 16. Rossi I, Rosenberg Z, Zember J. Normal skeletal 2011; 60:299–308
crescent–plantar fascia system. development and imaging pitfalls of the calcaneal 32. Arangio GA, Chopra V, Voloshin A, Salathe EP. A
apophysis: MRI features. Skeletal Radiol 2016; biomechanical analysis of the effect of lateral column
Acknowledgment 45:483–493 lengthening calcaneal osteotomy on the flat foot. Clin
We thank Robert Healey, biomechanical 17. Riddle DL, Schappert SM. Volume of ambulatory Biomech (Bristol, Avon) 2007; 22:472–477
engineer, for the critical review of the biome- care visits and patterns of care for patients diag- 33. Rantalainen T, Duckham RL, Suominen H,
chanical concepts stated in this manuscript. nosed with plantar fasciitis: a national study of ­Heinonen A, Alen M, Korhonen MT. Tibial and
medical doctors. Foot Ankle Int 2004; 25:303–310 fibular mid-shaft bone traits in young and older
References 18. Lawrence DA, Rolen MF, Morshed KA, ­Moukaddam sprinters and non-athletic men. Calcif Tissue Int
1. Arandes RV. Biomecánica del calcaneo. Med Clin H. MRI of heel pain. AJR 2013; 200:845–855 2014; 95:132–140
(Barc) 1953; 21:21–25 19. Van Boerum DH, Sangeorzan BJ. Biomechanics 34. Ducher G, Daly RM, Bass SL. Effects of repetitive
2. Kachlik D, Baca V, Cepelik M, Hajek P, Mandys and pathophysiology of flat foot. Foot Ankle Clin loading on bone mass and geometry in young
V, Musil V. Clinical anatomy of the calcaneal tu- 2003; 8:419–430 male tennis players: a quantitative study using
berosity. Ann Anat 2008; 190:284–291 20. Cheung JT, Zhang M, An KN. Effect of Achilles MRI. J Bone Miner Res 2009; 24:1686–1692
3. Gierse H. The cancellous structure in the calca- tendon loading on plantar fascia tension in the 35. Tan VP, MacDonald HM, Kim S, et al. Influence
neus and its relation to mechanical stressing. Anat standing foot. Clin Biomech (Bristol, Avon) 2006; of physical activity on bone strength in children
Embryol (Berl) 1976; 150:63–83 21:194–203 and adolescents: a systematic review and narrative
4. Kirby KA. Biomechanics of the normal and ab- 21. Thomas JL, Christensen JC, Kravitz SR, et al.; synthesis. J Bone Miner Res 2014; 29:2161–2181
normal foot. J Am Podiatr Med Assoc 2000; American College of Foot and Ankle Surgeons 36. Havaldar R, Pilli SC, Putti BB. Insights into the
90:30–34 Heel Pain Committee. The diagnosis and treatment effects of tensile and compressive loadings on hu-
5. Fuller EA. The windlass mechanism of the foot: a of heel pain: a clinical practice guideline—revision man femur bone. Adv Biomed Res 2014; 3:101
mechanical model to explain pathology. J Am 2010. J Foot Ankle Surg 2010; 49(suppl 3):S1–S19 37. Clarke B. Normal bone anatomy and physiology.
­Podiatr Med Assoc 2000; 90:35–46 22. Grasel RP, Schweitzer ME, Kovalovich AM, et al. Clin J Am Soc Nephrol 2008; 3(suppl 3):S131–S139
6. Bolgla LA, Malone TR. Plantar fasciitis and the MR imaging of plantar fasciitis: edema, tears, and 38. Lisková M, Hert J. Reaction of bone to mechani-
windlass mechanism: a biomechanical link to occult marrow abnormalities correlated with out- cal stimuli. Part 2. Periosteal and endosteal reac-
clinical practice. J Athl Train 2004; 39:77–82 come. AJR 1999; 173:699–701 tion of tibial diaphysis in rabbit to intermittent
7. Keller TS, Weisberger AM, Ray JL, Hasan SS, 23. Kohls-Gatzoulis J, Angel JC, Singh D, Haddad F, loading. Folia Morphol (Praha) 1971; 19:301–317
Shiavi RG, Spengler DM. Relationship between Livingstone J, Berry G. Tibialis posterior dys- 39. Sverdlova N. Tensile trabeculae: myth or reality?
vertical ground reaction force and speed during function: a common and treatable cause of adult J Musculoskelet Neuronal Interact 2011; 11:1–7
walking, slow jogging, and running. Clin Biomech acquired flatfoot. BMJ 2004; 329:1328–1333 40. Michna H, Hartmann G. Adaptation of tendon
(Bristol, Avon) 1996; 11:253–259 24. Balen PF, Helms CA. Association of posterior collagen to exercise. Int Orthop 1989; 13:161–165
8. Kelly LA, Lichtwark G, Cresswell AG. Active tibial tendon injury with spring ligament injury, 41. Brumitt J, Cuddeford T. Current concepts of mus-
regulation of longitudinal arch compression and sinus tarsi abnormality, and plantar fasciitis on cle and tendon adaptation to strength and condi-
recoil during walking and running. J R Soc MR imaging. AJR 2001; 176:1137–1143 tioning. Int J Sports Phys Ther 2015; 10:748–759
I­nterface 2015; 12:20141076 25. Theodorou DJ, Theodorou SJ, Farooki S, 42.
Franceschi F, Papalia R, Paciotti M, et al.
9. Best A, Holt B, Troy L, Hamill J. Trabecular bone ­Kakitsubata Y, Resnick D. Disorders of the plan- Obesity as a risk factor for tendinopathy: a system-
in the calcaneus of runners. PLoS One 2017; tar aponeurosis: a spectrum of MR imaging find- atic review. Int J Endocrinol 2014; 2014:670262
12:e0188200 ings. AJR 2001; 176:97–104 43. Scott A, Backman LJ, Speed C. Tendinopathy: up-
10. Lohrer H, Arentz S, Nauck T, Dorn-Lange NV, 26. Schweitzer ME, Karasick D. MR imaging of disor- date on pathophysiology. J Orthop Sports Phys
Konerding MA. The Achilles tendon insertion is ders of the Achilles tendon. AJR 2000; 175:613–625 Ther 2015; 45:833–841
crescent-shaped: an in vitro anatomic investiga- 27. Fayed N, Morales H, Modrego PJ, Pina MA. Con- 44. Choi K, Goldstein SA. A comparison of the fa-
tion. Clin Orthop Relat Res 2008; 466:2230–2237 trast/noise ratio on conventional MRI and choline/ tigue behavior of human trabecular and cortical
11. Shaw HM, Vazquez OT, McGonagle D, Bydder G, creatine ratio on proton MRI spectroscopy accu- bone tissue. J Biomech 1992; 25:1371–1381
Santer RM, Benjamin M. Development of the hu- rately discriminate low-grade from high-grade 45. Oftadeh R, Entezari V, Sporri G, et al. Hierarchical
man Achilles tendon enthesis organ. J Anat 2008; cerebral gliomas. Acad Radiol 2006; 13:728–737 analysis and multi-scale modelling of rat cortical and
213:718–724 28. Welvaert M, Rosseel Y. On the definition of trabecular bone. J R Soc Interface 2015; 12:20150070

8 AJR:211, November 2018