February 2004, VOLUME 182
NUMBER 2

Recommend & Share

February 2004, Volume 182, Number 2

Musculoskeletal Imaging

Radiographic Diagnosis of Tarsal Coalition

+ Affiliation:
1Both authors: 1A71 School of Medicine, 30 N 1900 E, Salt Lake City, UT 84132-2140.

Citation: American Journal of Roentgenology. 2004;182: 323-328. 10.2214/ajr.182.2.1820323

ABSTRACT
Next section

OBJECTIVE. The objective of this study was to evaluate the sensitivity of unenhanced radiographic diagnosis of tarsal coalition.

MATERIALS AND METHODS. The study had two phases. The first was a retrospective case and control review. Radiographs of 37 feet (15 talocalcaneal coalitions and 15 calcaneonavicular coalitions) and of 17 patients with foot pain and no coalition used as controls were reviewed independently by three observers who had no prior knowledge of the cases. Each observer reviewed the cases for individual signs of coalition and then decided if coalition was present. The second phase of the study was a prospective evaluation by a single observer of 150 consecutive weightbearing foot radiographs obtained to evaluate nontraumatic foot pain. Patients diagnosed as positive for coalition underwent CT.

RESULTS. On retrospective review of unenhanced radiographs, observers achieved 100% sensitivity and 88% specificity in the diagnosis of talocalcaneal coalitions. Sensitivity and specificity for calcaneonavicular coalitions ranged from 80% to 100% and 97% to 98%, respectively. Several previously unpublished radiographic signs increased sensitivity of diagnosis. For calcaneonavicular coalition, the new signs were altered navicular morphology and visualization of the bar on the anteroposterior radiograph. For talocalcaneal coalition, the new signs were a dysmorphic sustentaculum tali, nonvisualization of the middle subtalar facet, and shortening of the talar neck. In the prospective phase of the study, three talocalcaneal coalitions were detected with no false-positive results.

CONCLUSION. Routine anteroposterior and lateral unenhanced radiographs are a valuable screening tool for tarsal coalition, even when used by inexperienced observers. The newly described signs increase sensitivity of radiographic diagnosis.

Introduction
Previous sectionNext section

Tarsal coalition is an abnormal bony, cartilaginous, or fibrous union between two or more bones of the hind- and midfoot. It is a fairly common condition, generally estimated to affect up to 1% of the population [1, 2]. It can occur as an isolated anomaly or in association with other congenital disorders. The most common coalitions are either between the anterior process of the calcaneus and the navicular bone or between the talus and calcaneus, usually involving the middle facet. Coalitions restrict normal subtalar motion (eversion, inversion, and anterior gliding) and can result in flatfoot deformity, pain, tarsal tunnel syndrome, tenderness, and peroneal tendon spasm [1, 2]. Coalitions are traditionally described as presenting in the second decade of life, but many cases present later in adulthood [2].

The first reports of radiographic diagnosis of tarsal coalition relied on the axial view of the calcaneus to evaluate talocalcaneal coalition and the oblique view of the foot to evaluate calcaneonavicular coalition [1]. Subsequently, a variety of secondary signs of coalition were described on lateral radiographs, including talar beak [3, 4], C sign [5], and “anteater” sign [5, 6]. These secondary signs reflect abnormal bony orientation or abnormal motion or both of the tarsus because of the presence of a coalition.

The recent orthopedic and radiologic literature emphasizes the need to use CT and MRI in diagnosing coalition [2, 711]. One recent review of three cases of talocalcaneal and two cases of calcaneonavicular coalition reported that four of the five cases had normal findings on radiographs [8]. This current trend toward relying on MRI and CT for the diagnosis of tarsal coalition was the impetus for this study reassessing the utility of unenhanced radiographs.

Although CT and MRI are highly accurate in the diagnosis of tarsal coalition, they are expensive and usually are requested only if the diagnosis is suggestive of tarsal coalition clinically. Clinical diagnosis of coalition is not simple and requires skilled examiners. Because anteroposterior and lateral radiographs are obtained in most patients with foot pain, screening for coalition with these routine unenhanced radiographic views is desirable. Oblique radiographs are not routinely obtained at many institutions and therefore were not included in our series.

figure
View larger version (114K)
Fig. 1D. Lateral radiography of foot used to evaluate subtalar coalition. Lateral radiograph of 55-year-old man with tarsal coalition and equivocal talar beak (arrow) shows that projection from head of talus could be interpreted as either beak or osteophyte. Correct diagnosis was made on basis of dysmorphic sustentaculum tali and positive C sign.

Materials and Methods
Previous sectionNext section
Phase 1

We performed a retrospective study of each of the previously described signs as seen on anteroposterior and lateral radiographs, as well as additional signs noted by the senior author. The study population consisted of 28 patients and included 37 feet. There were 15 cases each of middle facet talocalcaneal coalition (11 nonbony and four bony) and 15 cases of calcaneonavicular coalition (13 nonbony and two bony). Seventeen feet without coalitions were included as controls. Three observers reviewed the radiographs without prior knowledge of the patients, history, or advanced imaging findings.

Patients were gathered from a retrospective chart review of an 18-month period at our institution. Radiographs of all patients whose charts had a diagnosis of tarsal coalition were reviewed by the researchers. Only cases of coalition that had CT, MRI, or surgical confirmation were included in the study. Randomly selected radiographs with a diagnosis of flatfoot, cavus foot, or normal findings were included in the study on the basis of negative findings on CT (nine cases), negative findings on MRI (two cases), absence of coalition based on examination of the hindfoot by a pediatric orthopedic surgeon (four cases of cerebral palsy with flexible flatfoot), or lack of symptoms with negative clinical findings on examination by an orthopedic surgeon specializing in the foot and ankle (two cases). Patients were excluded from the study if they had other known congenital anomalies to prevent introducing other confounding factors. Patients were 10–58 years old (average, 29 years).

We evaluated the following established radiographic signs: pes planus, talar beak, C sign, and anteater sign. Only six patients with coalition had anteroposterior radiographs of the ankle, and none of the findings showed the “ball-in-socket talus,” which can be a sign of limited hindfoot motion [2]. Because of the small number of anteroposterior radiographs available for review, this sign was not included as part of the radiographic analysis.

New signs that were evaluated were a short talar neck, visibility of the joint space of the middle subtalar facet, “brick” sign, broad mediolateral dimension of the navicular bone, and a tapered lateral aspect of the navicular bone.

The three observers were given access only to the anteroposterior and lateral foot radiographs because our goal was to ascertain accuracy of diagnosis on routine foot series without the help of the admittedly useful axial and oblique views. All except the radiographic series of five feet were weightbearing. Alignment was not assessed on nonweightbearing radiographs. The senior author did not participate in the review because she had prior knowledge of the patients.

Definition of Terms

Established terms.—“Pes planus” or flatfoot is present when the axis of the talus extends below the axis of the first metatarsal, whereas “pes cavus” is present when the axis of the talus extends above the axis of the first metatarsal [12].

The term “talar beak” refers to a flaring of the superior margin of the talar head, seen on lateral radiographs (Fig. 1A). It is usually easy to differentiate from the hooklike osteophyte of the talar head (Fig. 1B), which can form as a result of osteoarthritis, and from the talar ridge, occurring more proximally at the site of attachment of the ankle capsule (Fig. 1C) to the talar neck [4].

figure
View larger version (163K)
Fig. 1A. Lateral radiography of foot used to evaluate subtalar coalition. Lateral radiograph of 25-year-old man with subtalar coalition shows talar beak (solid straight arrow) as upwards flaring of anterosuperior aspect of talar head. Note C sign (curved arrows), which is continuous cortical contour extending from medial aspect of talus to sustentaculum tali. Middle subtalar facet is not seen. Note that sustentaculum tali (open arrow) has curved undersurface instead of normal flat contour. Talar neck appears shorter than normal.

figure
View larger version (178K)
Fig. 1B. Lateral radiography of foot used to evaluate subtalar coalition. Lateral radiograph of 55-year-old man with talonavicular osteoarthritis shows that talar osteophyte (straight arrow) hooks forward over talonavicular joint. Middle subtalar joint (curved arrow) is clearly seen and open. C sign is absent.

figure
View larger version (164K)
Fig. 1C. Lateral radiography of foot used to evaluate subtalar coalition. Lateral radiograph of 34-year-old woman with normal variant of talar ridge (for comparison with talar beak, A) shows that talar ridge (straight arrow) is located at mid neck of talus, at site of attachment of anterior tibiotalar joint capsule. Note also that middle subtalar facet is open with short gap (curved arrow) between cortical outlines of medial talus and sustentaculum tali (i.e., C sign is absent).

The C sign is present when a continuous arc is seen on lateral radiographs between the medial cortex of the talus and the inferior cortex of the sustentaculum tali [5] (Fig. 1A).

The “anteater” sign is an elongated, broadened anterior process of the calcaneus, visible on lateral radiographs (Fig. 2A). This is distinguished from the normal triangular configuration of the anterior process [7] (Fig. 2B).

figure
View larger version (168K)
Fig. 2A. Lateral radiographs used to evaluate calcaneonavicular coalition. Lateral radiograph of 32-year-old woman with calcaneonavicular coalition shows that “anteater” sign is present. Anterior process of calcaneus (between arrows) is enlarged and elongated and has blunt tip like anteater's snout.

figure
View larger version (164K)
Fig. 2B. Lateral radiographs used to evaluate calcaneonavicular coalition. Lateral radiograph of 18-year-old man with severe flexible flatfoot shows normal anterior process for comparison with A. Because of flatfoot deformity, anterior process of calcaneus overlies neck of talus. Despite alteration of normal anatomic relationships between talus and calcaneus, anterior process of calcaneus forms short triangle with sharp tip (arrow).

New terms.—On a well-positioned lateral radiograph in a healthy individual, the posterior and middle facets of the subtalar joint are visible. “Nonvisualized middle facet,” nonvisualization of the articular cortexes and joint space of the middle subtalar facet, is a sign of subtalar coalition (Figs. 1A and 1E).

figure
View larger version (127K)
Fig. 1E. Lateral radiography of foot used to evaluate subtalar coalition. Lateral radiograph of 35-year-old man with tarsal coalition without talar beak shows that talus appears normal and C sign is absent (arrow). Correct diagnosis was made on basis of dysmorphic sustentaculum tali and nonvisualization of middle subtalar facet.

The normal sustentaculum tali is shaped like a flat brick. In contrast, when talocalcaneal coalition is present, “dysmorphic sustentaculum tali” occurs when the sustentaculum tali is enlarged and has an ovoid configuration on lateral radiographs (Figs. 1A and 1E).

The presence of a “short talar neck” is a subjective assessment based on comparing the length of the talar neck to the anterior portion of the calcaneus (Fig. 1A).

In cases of calcaneonavicular coalition, “calcaneonavicular bar,” which is a visible bony bar or an anomalous articulation, may be seen between the navicular bone and the anterior process of the talus. Visualization of the bar has been previously described on oblique radiographs, which were not included in this study [1, 3]. Our study evaluated the visibility of a bar on anteroposterior radiographs (Figs. 3A,3B,4A,4B,5A,5B).

figure
View larger version (172K)
Fig. 3A. Anteroposterior radiograph used to evaluate calcaneonavicular coalition. Anteroposterior radiograph of 25-year-old man with nonbony calcaneonavicular coalition shows that navicular bone is broader in mediolateral dimension than head of talus and its lateral aspect is tapered relative to medial portion. Abnormal articulation (arrows) between calcaneus and navicular bone is clearly visible.

figure
View larger version (149K)
Fig. 3B. Anteroposterior radiograph used to evaluate calcaneonavicular coalition. Anteroposterior radiograph of 40-year-old woman with flexible flatfoot for comparison with A shows that space is clearly seen between lateral aspect of navicular bone (short arrow) and anterior process of calcaneus (long arrow).

figure
View larger version (157K)
Fig. 4A. Calcaneonavicular coalition without “anteater” sign in 25-year-old man with painful flatfoot. On lateral radiograph, anterior process of calcaneus is obscured by flatfoot deformity.

figure
View larger version (179K)
Fig. 4B. Calcaneonavicular coalition without “anteater” sign in 25-year-old man with painful flatfoot. On anteroposterior radiograph, navicular bone is broad in mediolateral dimension, and articulation with anterior process of calcaneus (arrow) is visible and enables diagnosis of calcaneonavicular coalition.

figure
View larger version (147K)
Fig. 5A. Calcaneonavicular coalition without “anteater” sign in 40-year-old woman. On lateral radiograph, anterior process is obscured by midfoot osteoarthritis.

figure
View larger version (159K)
Fig. 5B. Calcaneonavicular coalition without “anteater” sign in 40-year-old woman. On anteroposterior radiograph, coalition (arrow) is readily apparent.

The “wide navicular” is defined as the proximal articular cortex of the navicular bone, wider than the articular cortex of the talar head (Figs. 3A,3B,4A,4B,5A,5B).

The “laterally tapering navicular was considered to be present when the lateral portion of the navicular bone was appreciably smaller from proximal to distal ends than the medial portion (Figs. 3A,3B,4A,4B,5A,5B).

The three observers were given the radiographs to review independently. They had no prior knowledge of the cases and no history was given. They did not know how many coalitions or normal feet were included in the study. The observers were asked to assess the presence of each of the signs listed previously and to give a diagnosis of no coalition, talocalcaneal coalition, or calcaneonavicular coalition. Degree of confidence was recorded as possible, probable, or definite coalition. Observer 1 was a junior musculoskeletal radiologist, observer 2 was a junior radiology resident, and observer 3 was a junior orthopedic resident. The residents were given a half-hour tutorial in recognizing signs of coalition. The musculoskeletal radiologist was not given a tutorial but was given a brief list and description of the radiographic signs. Results were tabulated according to observer, sign, degree of confidence, and diagnosis.

Phase 2

A second phase of the study was then performed to further test the diagnostic criteria. One hundred fifty consecutive weightbearing radiographs of the foot obtained by a single orthopedic surgeon to evaluate for nontraumatic foot pain were reviewed prospectively by the senior author for signs of coalition. All radiographs were weightbearing, and all included weightbearing anteroposterior radiographs of the ankle as well as the foot.

Results
Previous sectionNext section
Phase 1

In the retrospective phase of the study, radiographic diagnosis was highly concordant between observers. A diagnosis of possible, probable, or definite tarsal coalition was reached by all observers in 15 (100% sensitivity) of 15 cases of talocalcaneal coalition and in 14 (93% sensitivity) of 15 cases of calcaneonavicular coalition. The musculoskeletal radiologist achieved 100% sensitivity for calcaneonavicular coalition.

Specificity for talocalcaneal coalition was 88%, and specificity for calcaneonavicular coalition averaged 97% (observer 1, 97%; observer 2, 100%; and observer 3, 94%). Each individual sign of coalition was assessed separately for its utility. The results are summarized in Table 1. Note that sensitivity and specificity achieved by the use of all signs was much better than for any individual sign.

TABLE 1 Accuracy of Unenhanced Radiographic Signs of Tarsal Coalition Assessed on Routine Anteroposterior and Lateral Radiographs

Pes planus was present in less than half the retrospective cases for which weightbearing radiographs were available. It was diagnosed in 40% of weightbearing radiographs of coalition overall: 30% of talocalcaneal coalitions and 48% of calcaneonavicular coalitions.

Phase 2

In the prospective phase of the study, the senior researcher found three talocalcaneal coalitions in 150 consecutive patients. All were confirmed on CT. There were no false-positive diagnoses. One patient in this group had a ball-in-socket talus, which has been reported as a sign of coalition [2]. Given the absence of other signs of coalition, the radiographs were considered negative for coalition, and the absence of coalition was confirmed on subsequent CT. This patient had a diagnosis of fibular hemimelia.

Discussion
Previous sectionNext section

Early reports of imaging diagnosis of tarsal coalition emphasized the axial (also known as the Harris) projection of the hindfoot to diagnose talocalcaneal coalition and the oblique projection of the foot to diagnose calcaneonavicular coalition. Conway and Cowell [3] reported that “the roentgenographic demonstration of the abnormal articulation between the calcaneus and the navicular is dependent on a proper oblique view of the foot.” However, neither the axial nor the oblique views of the foot are part of the routine foot series at many institutions; therefore, it is desirable to be able to make the diagnosis on the basis of the anteroposterior and lateral views.

Most calcaneonavicular coalitions can be diagnosed on lateral radiographs by using the anteater sign, but the three new signs that we have described increase sensitivity. Calcaneonavicular coalition was identified using the newly identified signs despite the absence of the anteater sign in two patients.

The actual calcaneonavicular bar can be seen on anteroposterior radiographs in a substantial number of coalition cases if careful attention is paid to the lateral margin of the navicular bone. The shape of the navicular bone is altered by the presence of a coalition. Therefore, in addition to showing the calcaneonavicular bar, the anteroposterior radiograph often reveals that the proximal portion of the navicular bone is wider than the talonavicular joint and that the lateral aspect of the navicular bone has a tapered, elongated appearance.

When all radiographic signs were used, radiographic sensitivity for calcaneonavicular coalition was 100% for the musculoskeletal radiologist, with a 97% specificity. The junior residents after only 30 min of training could detect 80% of calcaneonavicular coalitions, with a 98% specificity. This result corresponds favorably with the reported accuracy of CT and MRI for calcaneonavicular coalition [11].

The most commonly used radiographic signs of talocalcaneal coalition are the C sign and the talar beak. We found an 88% sensitivity and 87% specificity for the C sign, lower than the 98% sensitivity and specificity reported recently by Sakellariou et al. [6], perhaps reflecting our use of less experienced observers. Brown et al. [13] recently reviewed the accuracy of the C sign and reported that it had only a 40% sensitivity for talocalcaneal coalition. In addition, they found that it was present in 20% of patients with flatfoot but no tarsal coalition. Although we found a higher sensitivity (88%) for the C sign, we also found a 20% incidence of C sign in patients who had a flatfoot deformity but no tarsal coalition. The talar beak as an isolated sign of coalition has a low sensitivity and is sometimes equivocal.

Talocalcaneal coalitions are usually associated with overgrowth of the medial aspect of the talus and an abnormal morphology of the sustentaculum tali, readily detectable on CT or MRI. The abnormal morphology of the sustentaculum tali is often evident on unenhanced radiographs, with a sensitivity of 82%. Unfortunately, this sign had a specificity of only 70% and varied depending on the angle of the radiographic beam. Visualization of the middle facet joint space is also dependent on the angle of the radiographic beam.

Ball-in-socket configuration of the talus has been reported as a sign of tarsal coalition. We could not adequately evaluate that sign on unenhanced radiographs in our retrospective series because an insufficient number of patients had anteroposterior radiographs of the ankle. However, review of the CT and MRI studies in our series showed no patients with ball-in-socket talus. In the experience of the senior author, the ball-in-socket talus is generally absent in isolated tarsal coalition. It is usually associated with more severe congenital anomalies that alter hindfoot motion to a greater degree. In the second prospective portion of our study, all patients underwent anteroposterior radiography of the ankle, and ball-in-socket talus was absent in the three patients with coalition. One patient presented with ball-and-socket talus, without fifth ray and limb-length discrepancy. Using the criteria in this article, we found radiography to be negative for coalition. On CT, no coalition was seen.

When all radiographic signs were used, the sensitivity in diagnosis of talocalcaneal coalition was 100%. Specificity remains acceptable for a screening study at 88%. False-positive findings were seen primarily in patients who had severe flatfoot deformity and osteopenia related to cerebral palsy and reflect the difficulty in visualization of hindfoot structures in these patients.

The mean age of patients in the study was 29 years (≤ 58 years). Tarsal coalition is traditionally described as presenting in the second decade of life, but many patients become symptomatic later, so the radiologist must be alert to the possibility of this diagnosis in older patients.

The absence of pes planus cannot be used as an indication that coalition is absent because pes planus was present in less than half of our cases. Many patients with coalition have normal alignment, and some even have a cavovarus foot.

We purposely chose to have the radiographs reviewed by junior residents and a musculoskeletal radiologist. We believed that the best test of secondary signs of coalition was to have the signs applied by observers who had no preconceived notions about tarsal coalition and had no other criteria to use other than the ones taught them for the study. Both resident observers had no previous knowledge of radiographic signs of coalition, but after a short training session, they could suggest the diagnosis with a high level of accuracy. This result shows that radiographic diagnosis of coalition does not require advanced training or subspecialty expertise.

The excellent specificity in the blinded, retrospective study may have been at least partially attributable to the high pretest probability of coalition, given the relatively small number of controls. This outcome was the reason for the second phase of the study. In the prospective analysis, there were no false-positive radiographic findings. Although without performing CT on all patients, one cannot say that there were no false-negative diagnoses, the prevalence of coalition in this population was within the expected range.

Routine anteroposterior and lateral unenhanced radiographs are a sensitive screening test for both talocalcaneal and calcaneonavicular coalitions, even with observers who had no prior experience. Detection on unenhanced radiographs is especially valuable in cases in which coalition might not be suspected clinically: older patients, patients being seen by a primary care physician who may be less familiar with diagnosis of coalition, or patients with atypical symptoms.

Address correspondence to J. R. Crim.

References
Previous sectionNext section
1. Harris RL, Beath T. Etiology of peroneal spastic flat foot. J Bone Joint Surg Br 1948; 30:624 –634 [Google Scholar]
2. Varner KE, Michelson JD. Tarsal coalition in adults. Foot Ankle Int 2000; 21:669 –672 [Google Scholar]
3. Conway JJ, Cowell HR. Tarsal coalition: clinical significance and roentgenographic demonstration. Radiology 1969; 92:799 –811 [Google Scholar]
4. Resnick D. Talar ridges, osteophytes and beaks: a radiologic commentary. Radiology 1984; 151:329 –332 [Google Scholar]
5. Lateur LM, Van Hoe LR, Van Ghillewe KV, Gryspeerdt SS, Baert AL, Dereymaeker GE. Subtalar coalition: diagnosis with the C sign on lateral radiographs of the ankle. Radiology 1994; 193:847–851 [Google Scholar]
6. Sakellariou AS, Sallomi D, Janzen DL, Munk PL, Claridge RJ, Kiri VA. Talocalcaneal coalition: diagnosis with the C-sign on lateral radiographs of the ankle. J Bone Joint Surg Br 2000; 82:574 –578 [Google Scholar]
7. Oestrieich AE, Mize WA, Crawford AH, Morgan RC Jr. The “anteater nose”: a direct sign of calcaneonavicular coalition on the lateral radiograph. J Pediatr Orthop 1987; 7:709 –711 [Google Scholar]
8. Stoskopf CA, Hernandez RJ, Kelikian A, Tachdjian MO, Dias LS. Evaluation of tarsal coalition by computed tomography. J Pediatr Orthop 1984; 4:365 –369 [Google Scholar]
9. Wechsler RJ, Karasick D, Schweitzer ME. Computed tomography of talocalcaneal coalition: imaging techniques. Skeletal Radiol 1992; 21:353 –358 [Google Scholar]
10. Wechsler RJ, Schweitzer ME, Deely DM, Horn BD, Pizzutillo PD. Tarsal coalition: depiction and characterization with CT and MR imaging. Radiology 1994; 193:447 –452 [Google Scholar]
11. Emery KH, Bisset GS, Johnson ND, Nunan PJ. Tarsal coalition: a blinded comparison of MRI and CT. Pediatr Radiol 1998; 28:612 –616 [Google Scholar]
12. Mann RA. Flatfoot in children and young adults. In: Mann RA, Coughlin MJ. Surgery of the foot and ankle, 6th ed. St. Louis, MO: Mosby, 1999:723 –767 [Google Scholar]
13. Brown RR, Rosenberg ZS, Thornhill BA. The C sign: more specific for flatfoot deformity than subtalar coalition. Skeletal Radiol 2001; 30:84 –87 [Google Scholar]

Recommended Articles

Radiographic Diagnosis of Tarsal Coalition

Full Access, , , , ,
American Journal of Roentgenology. 2010;195:567-576. 10.2214/AJR.10.4406
Abstract | Full Text | PDF (1538 KB) | PDF Plus (1225 KB) 
Full Access, , , ,
American Journal of Roentgenology. 2013;201:W394-W400. 10.2214/AJR.12.9860
Abstract | Full Text | PDF (1081 KB) | PDF Plus (1159 KB) 
Full Access, ,
American Journal of Roentgenology. 2003;181:1449-1462. 10.2214/ajr.181.6.1811449
Citation | Full Text | PDF (48998 KB) | PDF Plus (2194 KB) 
Full Access, , , ,
American Journal of Roentgenology. 2011;197:W720-W729. 10.2214/AJR.10.5880
Abstract | Full Text | PDF (1120 KB) | PDF Plus (1098 KB) 
Full Access, ,
American Journal of Roentgenology. 2014;203:516-530. 10.2214/AJR.14.12969
Abstract | Full Text | PDF (1240 KB) | PDF Plus (1360 KB) 
Full Access,
American Journal of Roentgenology. 2011;197:1309-1321. 10.2214/AJR.11.7420
Abstract | Full Text | PDF (1195 KB) | PDF Plus (1215 KB)