Ultrasonographic-Surgical Correlation of Ankle Tendon Tears

Ultrasonographic-Surgical Correlation of Ankle Tendon Tears Gayle M. Waitches, DO, Matthew Rockett, DPM, Michael Brage, MD, Gary Sudakoff, MD This study evaluates the accuracy of ultrasonography in detecting ankle tendon tears of the peroneal, posterior tibial, and flexor digitorum longus tendons based on operative findings and clinical follow-up. A prospective study was performed in 33 patients with clinically suspected tendon injury. Sixty-eight tendons were evaluated sonographically. The diagnosis of an intrasubstance tear was made when disruption of uniform tendon architecture by hypoechoic linear or globular clefts was observed. Criteria used to diagnose complete tendon rupture included discontinuity or gap within the tendon or complete nonvisualization of the tendon. Treatment decisions were based on a combination of clinical parameters and imaging studies. Twenty-six patients had the presence or absence of Ankle tendon tears can be both debilitating and difficult to diagnose. The most common pathologic tendon condition on the medial side of the ankle is rupture or dysfunction of ABBREVIATIONS PT, Posterior tibial; FDL, Flexor digitorum longus; FHL, Flexor hallucis longus; PB, Peroneus brevis; PL, Peroneus longus; MR, Magnetic resonance Received October 24, 1997, from the Department of Radiology, The University of Washington School of Medicine, Harborview Medical Center, Seattle, Washington (G.M.W.); Harris County Podiatric Surgical Residency, Houston, Texas (M.R.); Department of Orthopedics, The University of Chicago, Chicago, Illinois (M.B.); and Department of Radiology, The Hammond Clinic, Hammond, Indiana (G.S.). Revised manuscript accepted for publication January 11, 1998. Address correspondence and reprint requests to Gayle M. Waitches, DO, University of Washington School of Medicine, Department of Radiology, Harborview Hospital, 325 Ninth Ave., Seattle, WA 98104-2499. tear confirmed at surgery. Five patients had a final diagnosis based on clinical findings, and two were lost to follow-up. Of the 68 tendons evaluated sonographically, 54 were directly inspected at surgery; 20 were found to be torn and 34 were intact. Ultrasonography was able to identify all tears correctly with an accuracy of 93%, a sensitivity of 100%, and a specificity of 88%. The positive and negative predictive values were 83% and 100%, respectively. The combined accuracy, sensitivity, and specificity of ultrasonography in detecting tendon tears in all patients evaluated both surgically and by clinical follow-up were 94%, 100%, and 90%, respectively. KEY WORDS: Ankle, tendons; Musculoskeletal system, ultrasonography; Posterior tibial tendon; Peroneal tendons. the PT tendon. Although medial ankle pain, unilateral flatfoot deformity, hindfoot valgus, and forefoot abduction are the typical clinical findings of PT tendon tear, 1,2 other, less frequent problems can cause similar symptoms. These include rupture of the flexor retinaculum, tarsal tunnel syndrome, and tears or tenosynovitis of the FDL and FHL tendons. 1 Lack of recognition of PT tendon tear can lead to progressive collapse of the medial longitudinal arch of the foot. 1 Peroneal tendon injuries often are more difficult to diagnose. 3,4 Patients typically have pain and weakness of these tendons on focused examination, together with ambulatory ankle instability. However, sinus tarsi syndrome, injury to the calcaneofibular ligament, and subtalar joint instability also can masquerade as peroneal tendon tear. 1 4 Accurate detection and confirmation of tendon tear is necessary to institute appropriate therapy. The imaging modality of choice in detecting ankle tendon tears has primarily been MR imaging. 5 10 1998 by the American Institute of Ultrasound in Medicine J Ultrasound Med 17:249 256, 1998 0278-4297/98/$3.50

250 ANKLE TENDON TEARS J Ultrasound Med 17:249 256, 1998 Although ultrasonography has been applied to the diagnosis of musculoskeletal abnormalities involving the shoulder, knee, and hand, aside from the Achilles tendon, ultrasonography has been a relatively underutilized modality for imaging ankle tendons. 11 27 The purpose of this prospective study was to determine if sonography is accurate in diagnosing ankle tendon tears, using operative findings and clinical follow-up as the gold standard. MATERIALS AND METHODS A prospective study of 33 patients was conducted between October 1995 and May 1997. Patients were referred for sonographic evaluation after examination by experienced orthopedic or podiatric physicians. The sonographic studies were performed prospectively and independently by two radiologists experienced in musculoskeletal imaging. The orthopedic and podiatric surgeons were not blinded to the results of the examinations, and each operative decision was based on a combination of clinical parameters and imaging results. Twenty-six patients underwent surgical tendon exploration, five patients were followed clinically, and two were lost to follow-up. The clinical, surgical, and sonographic findings were correlated in all cases. The duration of clinical follow-up ranged between 15 to 17 months (mean, 16.5 months). The range of time between the sonographic study and subsequent surgery was 1 day to 3 months (mean, 1.5 weeks). Two tendons were evaluated in all patients; the PT and FDL tendons were scanned in those with medial ankle pain, and both peroneal tendons (PB, PL) were scanned in patients with lateral ankle symptoms. One patient had both medial and lateral tendon pairs evaluated. The PT and FDL tendons were inspected directly at the time of medial ankle surgery, and both peroneal tendons were evaluated at the time of lateral ankle surgery. The FHL tendon was not visualized at surgery owing to its containment in a separate sheath, and was therefore not included in the sonographic data. Although the contralateral asymptomatic medial or lateral tendons also were scanned in all patients to provide a comparison of normal, these tendons were not confirmed to be normal at surgery, and therefore these tendons were not included in the data. Ultrasonographic examinations of the ankle were performed by a commercially available Siemens Elegra unit (Siemens, Issaquah, WA) or an ATL Ultra- Mark 9 HDI unit (Advanced Technology Laboratories, Bothell, WA) using either a 7.5 or a 10 MHz linear transducer. In most cases, imaging of the ankle tendons was done without the use of a commercially available acoustic stand-off pad. The medial aspect of the ankle was scanned after placing the patient in either the right or left lateral oblique position, with a small towel securing the ankle. Longitudinal and transverse images were obtained from approximately 8 cm superior to the tip of the malleolus to the navicular tuberosity for the PT tendon or deep to the navicular for the FDL tendon. Refractive artifact was avoided by placing the transducer perpendicular to the tendon throughout its course. To ensure that the correct tendon was evaluated, the foot was inverted and plantar flexed to visualize isolated PT tendon motion. Similar provocative kinetic maneuvers of the FDL tendon included flexion and extension of the lesser digits. Focused attention was always directed to the site of maximum patient complaint. The contralateral asymptomatic tendon also was evaluated in all cases, and hard copy images were obtained. Scanning of the lateral aspect of the ankle was accomplished in similar fashion; transverse and longitudinal images were obtained from approximately 8 cm superior to the tip of the lateral malleolus to either the point of insertion of the PB on the fifth metatarsal base or until the PL tendon coursed under the plantar tunnel. The foot and ankle were placed through active motion during scanning to ensure that the tendons were correctly identified. Eversion and inversion of the foot causes the PB to slide within its sheath while the PL tendon remains relatively motionless. Plantar flexion and dorsiflexion of the first metatarsal bone at the first metatarsal cuneiform joint prompts motion of the PL tendon while the PB remains static. Hard copy images were obtained. The sonographic criterion used to identify a tendon tear is disruption of the uniformly parallel echogenic fibrils by one or more hypoechoic clefts or gaps within a tendon (Figs. 1, 2). 23,26 28 These hypoechoic areas may or may not extend to the tendon s peripheral surface. The sonographic criteria compatible with a complete tendon rupture include nonvisualization of the tendon in the absence of surgery or tendon discontinuity, or both. 23,28 RESULTS Of the 68 tendons that were evaluated sonographically, 54 were inspected directly at the time of surgery (14 PT, 14 FDL, 13 PB, and 13 PL). Twenty of these tendons were found to be torn at surgery (13 PT, four PB, and three PL). Eighteen tendons were partially torn, and two were completely ruptured. Ultrasonography interpreted all surgically proven tendon tears correctly for a sensitivity of 100%, a specificity of 88%, and an accuracy of 93%. The positive and negative predictive values were 83% and 100%, respectively.

J Ultrasound Med 17:249 256, 1998 WAITCHES ET AL 251 A Four tendons that were thought to be torn sonographically were found to be normal at surgery. Two cases occurred in a patient with suspected tears of both PL and PB tendons. Although tendon tears were not detected surgically, the abnormal sonographic findings in this patient were consistent with the final surgical diagnosis of stenosing tenosynovitis. Two other false positive examinations occurred early in the course of our study, when we were less experienced. Both errors were due in part to a failure to corroborate the abnormality in orthogonal planes. Of the 54 tendons inspected surgically, the 30 remaining tendons were considered normal sonographically. Four of these patients were taken to surgery on the basis of abnormal clinical parameters despite normal appearing tendons sonographically; all eight tendons were surgically intact. Two of these patients with intact tendons showed other sonographic peritendinous abnormalities, which were confirmed at surgery. One patient with a hypoechoic fluid collection inferior to the PB muscle at the level of the malleolus had a subtalar joint capsule rupture with a cyst that was closely associated with the peroneal tendon sheath. The second patient had a poorly defined hypoechoic area deep to the inframalleolar portion of the PT tendon, which was interpreted as a probable hematoma at the time of ultrasonography. This patient was found to have a ruptured retinaculum. Of the remaining two of the four patients with normal tendons who were taken to surgery, no detectable abnormalities were B C Figure 1 44 year old woman with history of left medial ankle pain, exacerbated with ambulation. A, Transverse image of the PT tendon just below the malleolus shows an enlarged and hypoechoic tendon, with a central focal area of hypoechogenicity within the substance of the tendon (straight arrow). The curved arrows point to the tibia just below the medial malleolus. B, Longitudinal imaging confirms the abnormality seen on transverse scanning. A linear hypoechoic area is observed coursing through the tendon in longitudinal fashion, consistent with an intrasubstance tear (straight arrow). C, Intraoperative photograph taken at the time of surgical repair demonstrates the partial longitudinal tear within the edematous tendon (small straight arrow).

252 ANKLE TENDON TEARS J Ultrasound Med 17:249 256, 1998 Figure 2 Transverse images of the PB (short open black arrow) and PL (curved black arrow) tendons just below the level of the malleolus. The white arrow points to a longitudinal tear of the PL tendon, which was confirmed at the time of surgical repair. The solid straight black arrow points to the lateral malleolus. noted in one, and the other suffered from sinus tarsi syndrome. The 22 remaining tendons that were considered sonographically normal and were visualized surgically were all found to be normal (Fig. 3). Fourteen tendons in seven patients were examined sonographically without surgical confirmation. Two patients were lost to follow-up. Ten tendons were therefore followed in five patients (two PT, two FDL, three PB, three PL). The final diagnosis in these five patients was based on clinical parameters and longterm follow-up. Six tendons in three patients (three PB, three PL) had no evidence of a tear sonographically. Of these three patients, one had chronic symptoms compatible with diabetic neuropathy, another had pain relief after fusion of the calcaneocuboid joint, and the last had complete cessation of symptoms after a short-term period of conservative management. These three patients are considered true negative cases. Of the remaining two of the five patients followed clinically (four tendons; two PT, two FDL), both were diagnosed as having PT tendon tears sonographically. These patients refused surgery, and both continue to have clinical findings consistent with tear. These cases are considered true positives. In summary, 68 tendons were evaluated sonographically; 54 were inspected at surgery, and 14 were followed clinically. Of the 14 tendons followed clinically, four were not included because two patients were lost to follow-up. Of the remaining 10 tendons evaluated clinically, six were considered true negative and four were considered true positive. The combined accuracy, sensitivity, and specificity of ultrasonography in detecting tendons tears for all patients examined both surgically and by clinical follow-up are 94%, 100%, and 90%, respectively. The negative and positive predictive values are 100% and 86%, respectively (Tables 1, 2). DISCUSSION MR imaging has been the standard imaging modality in the evaluation of tendon abnormalities around the ankle. 5 10 Several authors have reported high accuracy in detecting ankle tendon tears involving the PT and peroneal tendons. 5 8 In 1988, Rosenberg and coworkers reported a sensitivity of 95% and specificity of 100% in the detection of PT tears using surgery as the gold standard. 6 More recently, Khoury and associates reported an MR imaging study of 12 patients with peroneal tendon tears in which MR imaging was Figure 3 A, Longitudinal image of a normal posterior tibial tendon (short arrows). Note the uniformity of the echogenic tendon fibril architecture. B, Transverse image of the normal PT and FDL tendons in the same asymptomatic patient. A B

J Ultrasound Med 17:249 256, 1998 WAITCHES ET AL 253 Table 1: Surgical Sonographic Correlation of Tendon Tears Number Torn Tendons Torn Tendons Tendons (N) (Surgery) (Sonography) PT 14 13 13 FDL 14 0 0 PB 13 4 7 PL 13 3 4 Total 54 20 24 Table 2: Combined Surgical, Clinical, and Sonographic Correlation of Tendon Tears Ultrasonography Combined Surgical Ultrasonography Correlation Clinical Correlation (N = 54) (N = 66) True positive tendon tears 20 24 True negative tendon tears 30 36 False positive tendon tears 4 4 False negative tendon tears 0 0 Sensitivity 100 100 Specificity 88 90 Accuracy 93 94 Positive predictive value 83 86 Negative predictive value 100 100 equally accurate. 7 Ultrasonography, on the other hand, has been a relatively underutilized modality for imaging ankle tendons. 10 27 Aside from Achilles tendon imaging, little information on sonographicsurgical correlation has been published concerning additional ankle tendon tears. 6,26 A review of the literature shows a recent article by Miller and colleagues, who describe the use of ultrasonography in the evaluation of PT tendon pathologic conditions. 26 Coakley and coauthors reported on the use of ultrasonography in evaluating the PT tendon in patients with rheumatoid arthritis and found that tendon thinning and hindfoot deformity were common. 15 A case report by Stephenson and coworkers describes the detection of tenosynovitis of the PT tendon sonographically, concluding that ultrasonography may obviate additional diagnostic tests. 12 There have also been few articles comparing MR imaging with ultrasonography in evaluating ankle tendon injury. 10,26 Davies and coauthors in 1991 compared sonography, CT, and MR imaging in imaging patellar tendinitis and concluded that ultrasonography should be the initial imaging modality in these patients. 10 Our personal experience has been that MR imaging is less sensitive than ultrasonography in excluding ankle tendon tears. Ultrasonography has greater tendinous contrast resolution and can detect subtle tendon texture abnormalities that are not appreciable with MR imaging (Fig. 4). We therefore set out to investigate the overall accuracy of ultrasonography in detecting PT, FDL, and peroneal tendon tears using surgical and clinical correlation. The clinical utility of ultrasonography depends on how accurately it can distinguish tendon tear from other pathologic conditions such as tendinitis and tenosynovitis. Patients with tendon tears usually are treated surgically, whereas those with tendinitis or tenosynovitis are managed conservatively. 6 A complete tendon tear rarely presents a clinical or imaging dilemma. 23,27,28 Sonographic confirmation of an acute rupture often is not necessary, as this entity usually is appreciable clinically. Partial intrasubstance tears, on the other hand, are more common and more challenging clinically to diagnose. Ultrasonography can be most helpful to the clinician in this clinical setting to allow for distinction between tear and other disease processes. Our reported high accuracy and sensitivity in detecting tendon tears may be due in part to the specific imaging criteria used to diagnose tear. We defined partial tears as either globular or linear hypoechoic gaps within the substance of the tendon, which often extend to the tendon periphery. 23,26 28 While applying these specific imaging criteria results in a high sonographic accuracy, we did, however, visualize other sonographic findings in patients with tendon tears (Fig. 5). These include an enlarged and diffusely hypoechoic tendon, a large amount of peritendinous fluid, thinning of tendon contour, decreased mobility, and peritendinous material of mixed echogenicity (usually representing blood, fibrosis, synovial hypertrophy, or tendon debris). These additional findings are probably less specific for tendon tear, as they can also be found in tendinitis and tenosynovitis. We are pursuing further prospective trials to evaluate the accuracy of these findings in diagnosing tears. Tendon ultrasonography is able to assess the severity and extent of tendon injury. Often, the operative approach and type of surgical repair vary with the location and extent of the injury. Information regarding the size of the gap between the tendon ends or the condition of the torn tendon fibers in complete rupture can aid the surgeon in determining surgical approach. Persistent diastasis of tendon ends in plantar flexion often is treated operatively, whereas complete coaptation may be treated conservatively. 27 Ultrasonography also may be helpful in postoperative

254 ANKLE TENDON TEARS J Ultrasound Med 17:249 256, 1998 patients with persistent pain, as approximately 2% of tendons treated surgically, and 10 to 35% of those treated medically, will rerupture. 17 All abnormal intratendinous areas must be corroborated in orthogonal planes to confirm the existence of a tear and avoid false positive diagnoses. We obtain the longitudinal view first because it provides a survey of overall tendon architecture. Any deviation or disruption of longitudinal tendon fibril uniformity should be reevaluated transversely. Two of our four false positive examinations occurred when we failed to confirm a suspected abnormality with transverse scanning. We now realize that there may be occasional slight variability in the normal longitudinal appearance of ankle tendons, which is not necessarily pathologic unless it can be confirmed in an orthogonal plane. This study is limited in that it was neither random nor blinded. The results of the sonographic examinations were made available to the surgeon preoperatively. This invariably produces a selection bias in that patients with abnormal ultrasonograms are more likely to have an operation, leading to a greater than normal prevalence of disease. Despite the fact that more patients with abnormal tendons were taken to surgery, every patient had at least two tendons scanned preoperatively: the questionable abnormal tendon and the adjacent tendon. The bias toward a higher number of positive tendon tears is offset by the simultaneous evaluation of true and false negative results in adjacent tendons. For example, the PB tendons are imaged simultaneously in a patient with suspected PB tear. Both tendons are inspected at surgery, confirming the normal appearance of the unaffected tendon. Moreover, four patients were taken to surgery despite a negative sonogram, and all eight tendons were confirmed to be normal at surgery. Finally, tendon ultrasonography is not meant to be a screening modality in which every patient with nonspecific ankle pain is referred for this examination. The inherent limitations of ultrasonography in evaluating osseous structures dictate that those referred for this modality should have a high pretest probability of tendon disease. We therefore believe that this study provides information regarding not only the positive but also the negative predictive value of ultrasonography in detecting tendon tears. Our study also was limited in that the final diagnosis in five patients was based only on clinical parameters. Some patients may have had tendon tears that which were not accurately detected by physical examination. The literature suggests, however, that the diagnosis of PT tears by an experienced clinician can be extremely accurate and may even be utilized as a standard. 30 Two of five patients examined clinically had medial ankle symptoms highly compatible with PT tear. Both refused surgery because of personal circumstances that do not allow them the anticipated postoperative convalescent period. Both patients continue to have clinical features compatible with tear. The three remaining patients followed clinically had lateral ankle symptoms with negative ultrasonograms and could conceivably have had tendon tears. Although the authors assume that patients with a negative examination who improve clinically do not have a torn tendon, it is possible that some of these patients may have had sonographically occult subtle tears that may have healed with conservative management. Another unavoidable limitation of this study is the possibility that subtle central intrasubstance tears that do not extend to the tendon periphery may not be visible during gross surgical inspection. These tears could go undetected at the time of surgical exploration, leading to a higher sonographic false positive rate. The clinical significance of missing subtle central Figure 4 Ultrasonogram and MR image of a partial longitudinal tear of PT tendon. A, The longitudinal ultrasonogram shows a linear area of hypoechogenicity within the central substance, suggestive of partial tear (white arrows). B, The T2-weighted sagittal MR image shows a normal appearing tendon (black arrows). A partial longitudinal tear was identified at the time of surgery. A B

J Ultrasound Med 17:249 256, 1998 WAITCHES ET AL 255 intrasubstance tears is uncertain when surgical findings are used as the gold standard, as these potential tears may or may not progress to more significant abnormalities. Of our four false positive examinations, two occurred in a patient with suspected tears of both PL and PB tendons. And although tendon tears were not detected surgically, the abnormal sonographic findings in this patient were consistent with the final diagnosis of stenosing tenosynovitis (Fig. 6). It is possible that in the remaining two false positive cases, an occult intrasubstance tear may have been present that was not seen at surgery. Further sonographic and clinical follow-up may be necessary in patients with inconsistent surgical-sonographic correlation and persistent symptoms. Ultrasonography as an imaging modality also has a few limitations. The examination is operator-dependent, with a variable learning curve. We found that the sonographic tendon anatomy and appearances became understandable after approximately 2 to 3 weeks. Another minor limitation is the amount of time required for the study. Including the contralateral ankle, the study can take from 30 to 45 min. However, the standard competitor for this modality is MR imaging, and a 45 min examination is not long in comparison to that. Finally, the major limitation of ultrasonography is its inability to image deeper tissues and bones around the ankle. If pathologic conditions other than ankle tendons are in question, MR imaging may be the more appropriate modality. A Figure 5 43 year old woman with history of progressively worsening right medial ankle pain over the last 2.5 years. The pain was recently exacerbated after tripping and twisting her ankle 4 months prior to presentation. A, Longitudinal scanning of the PT tendon shows mixed echogenicity material (short arrows) surrounding the tendon (long arrow), compatible with blood and fibrotic debris. The distended synovium is marked with calipers. B, Transverse image shows a vague hypoechoic mass-like structure adjacent to the insertion of the PT tendon (long arrows) on the navicular bone (NAV). This was found to represent torn and retracted tendon fibers and accompanying fibrosis. C, T1-weighted longitudinal MR image confirms the existence of multiple severe tendon tears (solid arrows) with surrounding intermediate signal intensity material, which was operatively shown to represent blood and fibrotic tissue. Note the corresponding MR imaging appearance of the retracted tendon fibers and accompanying fibrosis (open arrow) at the navicular tuberosity (arrowhead). C B

256 ANKLE TENDON TEARS J Ultrasound Med 17:249 256, 1998 Figure 6 Example of a false positive case; this is a longitudinal image of the PL tendon, which was interpreted as a tear. The tendon appears thinned and irregular in contour and is surrounded by a large amount of fluid. The suggestion of a thin hypoechoic linear area within the substance of the tendon also is present (arrow). This patient was taken to surgery, and was found to have stenosing tenosynovitis without tear. CONCLUSION In conclusion, we have found that ultrasonography is an accurate and sensitive modality in detecting tears of the peroneal, PT, and FDL tendons. The ability to distinguish tendon tear from other pathologic conditions allows for more appropriate patient management and, it is hoped, more rapid patient recovery. REFERENCES l. Turek SL: Orthopaedics: Principles and Their Application. 4th Ed. Philadelphia, JB Lippincott, 1984, p 1407 2. Sarrafian SK: Anatomy of the Foot and Ankle. Philadelphia, JB Lippincott, 1983, pp 251, 207 3. Sammarco GJ: Peroneus longus tendon tears: Acute and chronic. Foot Ankle Int 16:245, 1995 4. Bassett FH, Speer KP: Longitudinal rupture of the peroneal tendons. Am J Sports Med 21:354, 1993 5. Khoury NJ, El-Khoury GY, Saltzman CL, et al: MR imaging of posterior tibial tendon dysfunction. AJR 167:675, 1996 6. Rosenberg ZS, Cheung Y, Jahss MH, et al: Rupture of posterior tibial tendon: CT and MR imaging with surgical correlation. Radiology 169:229, 1988 7. Khoury NJ, El-Khoury GY, Saltzman CL, et al: Peroneus longus and brevis tendon tears: MR imaging evaluation. Radiology 200:833, 1996 8. Schweitzer ME, Caccese R, Karasick D, et al: Posterior tibial tendon tears: Utility of secondary signs for MR imaging diagnosis. Radiology 88:655, 1993 9. Schweitzer ME, van Leersum M, Ehrlich SS, et al: Fluid in normal and abnormal joints: Amount and distribution as seen on MR images. AJR 162:111, 1994 10. Davies SG, Baudouin CJ, King JB, et al: Ultrasound, computed tomography and magnetic resonance imaging in patellar tendinitis. Clin Radiol 43:52, 1991 11. Fornage BD: Achilles tendon: US examination. Radiology 159:759, 1986 12. Stephenson CA: Sonographic diagnosis of tenosynovitis of the posterior tibial tendon. J Clin Ultrasound 18: 114, 1990 13. Kaplan PA, Anderson JC, Norris MA, et al: Ultrasonography of post-traumatic soft-tissue lesions. Radiol Clin North Am 27:973, 1989 14. Bianchi S, Zwass A, Abdelwahab I, et al: Diagnosis of tears of the quadriceps tendon of the knee: Value of sonography. AJR 162:1137, 1994 15. Coakley FV, Samanta AK, Finlay DB: Ultrasonography of the tibialis posterior tendon in rheumatoid arthritis. Br J Rheumatol 33:273, 1994 16. Kainberger FM, Engel A, Barton P, et al: Injury of the Achilles tendon: Diagnosis with sonography. AJR 155:1031, 1990 17. Blei CL, Nirschl RP, Grant EG: Achilles tendon: US diagnosis of pathologic conditions. Radiology 159:765, 1986 18. Harcke HT, Grissom LE, Finkelstein MS: Evaluation of the musculoskeletal system with sonography. AJR 150:1253, 1988 19. Fornage BD, Rifkin MD: Ultrasound examination of tendons. Radiol Clin North Am 26: 87, 1988 20. Dillehay GL, Deschler T, Rogers LF, et al: The ultrasonographic characterization of tendons. Invest Radiol 19:338, 1984 21. Harcke HT, Grissom LE, Finklestein MS: Evaluation of the musculoskeletal system with sonography. AJR 150:1253, 1988 22. Crass JR, van de Vagte GL, Harkavy LA: Tendon echogenicity: Ex vivo study. Radiology 167:499, 1988 23. van Holsbeeck M, Powell A: Ankle and foot. In van Holsbeeck M, Introcaso JH: Musculoskeletal Ultrasound. St. Louis, Mosby Year Book, 1991, p 221 24. Chhem RK, Beauregard G: Synovial diseases. In van Holsbeeck M, Introcaso JH: Musculoskeletal Ultrasound. St. Louis, Mosby Year Book, 1991, p 43 25. Nazarian LN, Rawool NM, Martin CE, et al: Synovial fluid in the hindfoot and ankle: Detection of amount and distribution with US. Radiology 197:275, 1995 26. Miller SD, Van Holsbeeck M. Boruta PM, et al: Ultrasound in the diagnosis of posterior tibial pathology. Foot Ankle Int 17:555, 1996 27. Thermann H, Hoffman R, Zwipp H, et al: The use of ultrasound in the foot and ankle. Foot Ankle 13:386, 1992 28. Marcelis S, Daenen B, Ferrara MA: Peripheral Musculoskeletal Ultrasound Atlas. New York, Thieme Medical Publishers, 1996, pp 12, 30, 167 29. Saltzman CL, Bonar SK: Tendon problems of the foot and ankle. In Lutter LD, Mizel MS, Pfeffer GB (Eds): Orthopedic Knowledge of Update of the Foot and Ankle. Chicago, American Academy of Orthopedic Surgeons, 1994, p 269 30. Mann RA, Thompson FM: Rupture of the posterior tibial tendon causing flat foot. J Bone Joint Surg [Am] 67:556, 1985