Cost-Effectiveness of Immediate MR Imaging Versus Traditional Follow-Up for Revealing Radiographically Occult Scaphoid Fractures

Theodore. Dorsay 1 Nancy M. Major Clyde. Helms Received May 15, 2001; accepted after revision June 25, 2001. Presented at the annual meeting of the merican Roentgen Ray Society, Seattle, pril May 2001. 1 ll authors: Department of Radiology, Duke University Medical Center, Erwin Rd., Rm. 1504, Durham, NC 27710. ddress correspondence to T.. Dorsay. JR 2001;177:1257 1263 0361 803X/01/1776 1257 merican Roentgen Ray Society Cost-Effectiveness of Immediate MR Imaging Versus Traditional Follow-Up for Revealing Radiographically Occult Scaphoid Fractures OJECTIVE. For suspected scaphoid fractures with no radiographic evidence of fracture, treating symptoms with immobilization and radiographic follow-up has long been the standard of care. Modified MR imaging of the wrist is offered at our institution in screening for radiographically occult scaphoid fractures at the time of initial presentation to the emergency department. We show the advantages and comparative costs of this modified protocol versus a traditional protocol. MTERILS ND METHODS. Our modified protocol consists of coronal thin-section T1-weighted and fast spin-echo T2-weighted MR images with fat saturation. review of the literature was performed to assess the accuracy of clinical examination, radiography, and other modalities in the evaluation of scaphoid fractures of the wrist. Charges for this procedure are compared with charges for traditional follow-up. RESULTS. Three of four patients with positive results at clinical examination and negative findings on initial radiographs will be needlessly immobilized and monitored. The charges to the patient at our institution for screening MR imaging of the wrist are $770. The total charges to the patient with the traditional protocol, which would not be necessary with screening MR imaging, are $677 or more if a diagnosis is not made at this time. one scanning or routine MR imaging is often eventually used. CONCLUSION. Cost analysis at our institution suggests the two protocols are nearly equivalent from a financial standpoint. The loss of productivity for patients who are unnecessarily in casts or splints may be substantial. Screening MR imaging of the wrist in this setting is becoming accepted at our institution in a manner similar to screening MR imaging of the hip. M R imaging screening for occult fractures of the hip has been shown to be a cost-effective procedure [1, 2]. Referring clinicians often encounter resistance to a similar screening MR imaging study to rule out occult scaphoid fractures of the wrist (negative radiographic findings with a high level of clinical suspicion). We present data from our institution showing the advantages and comparative costs of this route versus traditional casting and follow-up radiography. Materials and Methods review of the literature of the last 6 years was undertaken to evaluate the positive predictive value of the clinical examination for scaphoid fracture. In addition, we attempted to assess the negative predictive value of initial negative radiographic findings in the setting of strong clinical suspicion. Modified MR imaging of the wrist is offered at our institution at the time of initial presentation to screen for radiographically occult scaphoid fractures. MR imaging is performed on a 1.5-T magnet (General Electric Medical Systems, Milwaukee, WI) with a dedicated wrist coil. This shortened study consists of 3-mm (0.3-mm gap) coronal T1-weighted spinecho (TR range/te range, 500 800/20 30; 2 excitations) and T2-weighted fast spin-echo (TR/TE range, 4000/80 120; 2 excitations) images with fat saturation and a matrix of 256 192. Fracture is identified as a linear area of low signal intensity on T1- and T2- weighted images (Fig. 1). Similar sequences have proven accurate for detecting fracture in both the hip and the wrist [1 4]. We compared charges for the clinical protocol when using screening MR imaging with traditional charges for follow-up, which include charges for the initial orthopedic consultation in the emergency department, casting in the emergency department, a follow-up orthopedic appointment in 7 10 days, repeated wrist radiography, and, often, office fluoroscopy. ctual case scenarios are presented to emphasize particular points. Cases were drawn from our MR imaging database for the 6-year period 1995 2000. During JR:177, December 2001 1257

Dorsay et al. that time, 334 wrist studies were performed; seventeen (5%) were performed to rule out scaphoid fracture. Eight (47%) of these 17 were performed in the most recent year, 2000, which indicates increasing use of MR imaging in this scenario. Results Our review of the literature for the last 6 years showed the positive predictive value of the clinical examination (clinical suspicion that warranted a wrist series or images of the scaphoid bone to rule out scaphoid fracture) varies widely (Table 1). The positive predictive value ranged from 13% to 69%, with a weighted average of 21%. These figures suggest that four of five patients will not have a scaphoid fracture. Our review of the recent literature to evaluate the negative predictive value of negative initial radiographs in the setting of strong clinical suspicion, found a range of 50 87%, with a weighted average of 74% (Table 2). These figures imply that three of every four patients with negative findings on initial radiographs will likely undergo needless immobilization. The charge (professional and technical) to the patient at our institution for screening MR Fig. 1. 20-year-old man with injury sustained while boxing 2 3 months previously who presented with continued generalized wrist pain and stiffness. Scaphoid fracture of wrist was not visible on initial radiograph and remained radiographically occult during entire course of therapy., Coronal T2-weighted fast spin-echo MR image (TR/TE, 4000/72) with fat saturation shows nondisplaced fracture as distinct linear focus of low signal intensity (arrow ) with surrounding bright edema., T1-weighted spin-echo coronal MR image (500/11) obtained at same time as shows that fracture line (arrows) is slightly less evident in surrounding low-signal-intensity edema. TLE 1 Positive Predictive Value (PPV) of Clinical Examination in Recent Studies of Scaphoid Fractures Study Year Positive Clinical Examination a True-Positive b PPV of Clinical Examination (%) Grover [19] 1996 221 29 13 Thorpe et al. [15] 1996 59 8 14 Munk et al. [20] 1995 1052 150 14 Gaebler et al. [4] 1996 32 6 19 Parvizi et al. [21] 1998 251 56 22 Tiel-van uul et al. [13] 1995 160 70 44 Roolker et al. [5] 1996 11 5 45 Roolker et al. [7] 1999 100 49 49 Tiel-van uul et al. [12] 1996 27 15 56 Cook et al. [16] 1997 10 6 60 Lepisto et al. [18] 1995 16 11 69 Total 1939 405 Weighted average PPV c 21 a Number of patients with strong suspicion of scaphoid fracture at clinical examination. b Number of cases with scaphoid fracture confirmed by gold standard (MR imaging, bone scanning, or long-term radiographic follow-up [varied among studies]). c Total positive clinical examinations divided by total true-positive scaphoid fractures. imaging of the wrist is $770. The reduction from a charge of $1256 for standard wrist MR imaging is achieved primarily through reduction of the technical component. The clinical protocol for treatment of suspected scaphoid fracture is summarized in a flowchart (Fig. 2). 1258 JR:177, December 2001

MR Imaging of Radiographically Occult Scaphoid Fractures TLE 2 Study Negative Predictive Value (NPV) of Negative Findings on Wrist Radiographs in Recent Studies of Patients Having Clinical Examination with Strong Clinical Suspicion of Scaphoid Fracture Year a Number of patients with strong suspicion of scaphoid fracture at clinical examination but negative findings on wrist radiographs. b Number of cases with negative radiographic findings and with negative findings for scaphoid fracture confirmed by gold standard (MR imaging, bone scanning, or long-term radiographic follow-up [varied among studies]). c Total positive clinical examinations, negative radiography divided by total true-negative scaphoid fractures. The potential added charges to the patient who follows a traditional protocol that are negated by the application of screening MR imaging of the wrist in the emergency or acute care setting are summarized in Tables 3 and 4. The following actual case scenarios are offered. Charges that may have been avoided by initial screening MR imaging are in parentheses or brackets. Positive Clinical Examination, NPV of Clinical Negative True-Negativeb Examination (%) Radiography a Cook et al. [16] 1997 8 4 50 Tiel-van uul et al. [12] 1996 16 9 56 Lepisto et al. [18] 1995 13 8 62 Roolker et al. [7] 1999 71 51 72 Tiel-van uul et al. [13] 1995 125 90 72 ayer et al. [14] 2000 40 32 80 retlau et al. [17] 1999 47 38 81 Fowler et al. [3] 1998 46 40 87 Total 366 272 Weighted average NPV c 74 Fig. 2. Flowchart shows clinical pathways available to clinician in setting of positive (+) clinical examination (CE) for scaphoid fracture (e.g., fall on outstretched hand and snuffbox tenderness). indicates negative. Case Scenario 1 Day 1. The patient was a 23-year-old woman who flipped and fell backward onto her outstretched arm and presented with elbow and wrist pain (Fig. 3). Radiographs at the urgent care center showed a nondisplaced fracture of the radial neck. No wrist fracture was identified, but snuffbox tenderness was present. The patient s wrist was splinted ($154), and she was referred to an orthopedist. Day 6. t orthopedic follow-up ($135), the patient had swelling and point tenderness in the snuffbox region. The orthopedic report stated: Given the snuff box tenderness, the patient was placed in a short arm thumb spica bandage [$159] for possible scaphoid fracture. I will have the patient return to the clinic in one week. t that point, we will remove the thumb spica and reexamine her wrist. If she s continuing to have scaphoid tenderness, repeat x-ray will be obtained to determine whether or not there is any fracture line evident within the scaphoid. If there s no fracture line evident and she s continuing to have tenderness, then we will obtain either a bone scan or MR imaging to fully evaluate for nondisplaced scaphoid fracture. Day 13. The patient s cast was removed. She continued to have tenderness in the snuffbox region. Radiographs ($130) of the wrist showed a mild degree of irregularity on the radial aspect of the scaphoid bone but no obvious fracture line. The patient s wrist was placed in a thumb spica cast. MR imaging was ordered. Day 24. Complete wrist MR imaging ($1256) showed increased T2 and decreased T1 signals throughout the scaphoid bone, which are compatible with edema. focal area of linear decreased signal intensity on T1- and T2-weighted images extended through the proximal pole of the scaphoid bone, which is compatible with a nondisplaced fracture. Day 27. The patient returned to the clinic ($72) for the MR imaging results, which showed a scaphoid fracture. The cast was not changed because it was in good condition. Day 41. The cast was removed. The patient continued to have some tenderness and swelling around the scaphoid bone on the dorsal aspect of a wrist. Radiographic findings were again negative. The patient s wrist was re-placed in a short arm thumb spica splint. Summary. This scenario shows a situation in which 24 days were required to arrive at the diagnosis of scaphoid fracture. lthough the wrist was treated appropriately in the face of snuffbox tenderness, additional charges of $578 were incurred before definitive diagnosis was made at MR imaging. The additional expense of full MR imaging of the wrist ($486) versus a screening protocol was JR:177, December 2001 1259

Dorsay et al. incurred. s late as day 41, follow-up radiographs showed negative results. Case Scenario 2 Day 1. The patient was a 16-year-old boy who fell onto his outstretched hand, incurring a dorsiflexion injury to the right wrist (Fig. 4). He was seen at a family medical clinic, where radiographs were obtained that raised the question of a possible scaphoid fracture. The patient s hand was splinted ($154), and he was referred to an orthopedist. Day 2. t orthopedic consultation ($135), the patient had mild tenderness over the region of the scaphoid bone, but only with firm pressure. Under office fluoroscopy ($150), a fracture was not seen. The clinical impression questioned a nondisplaced scaphoid fracture versus a sprain of the right wrist. Fig. 3. 23-year-old woman who fell backward onto outstretched arm and presented with wrist pain., Initial radiograph fails to reveal scaphoid fracture., T2-weighted fast spin-echo coronal MR image (TR/TE, 4000/72) with fat saturation obtained 24 days after initial radiograph shows diffuse heterogeneous high signal intensity (arrow ) throughout scaphoid bone. C, T1-weighted coronal MR image (500/14) through scaphoid bone obtained at same time as shows subtle linear focus of low signal intensity (arrow ) across mid waist of scaphoid bone. TLE 3 Possible dditional Charges to Patient with Traditional Protocol a Procedure Cost Emergency department splint $154 Wrist radiographs, 3 views $85 Scaphoid radiographic series $45 Orthopedic services Emergency department $108 consultation Outpatient office visit b $135 Wrist fluoroscopy in office $150 Total $677 a For working up scaphoid fracture with negative initial radiographic findings. b New patient, moderate severity, 30 min. Possible Charges of Traditional Follow-Up Protocol for TLE 4 Radiographically Occult Scaphoid Fracture Versus Modified Screening MR Imaging of the Wrist Procedure Cost Traditional protocol Follow-up, 1st appointment $677 Follow-up, 2nd appointment $547 Nuclear bone scanning $639 Routine MR wrist imaging $1,256 Screening MR wrist imaging $770 Since the wrist was minimally symptomatic, it was suggested to try a little immobilization with a short arm thumb spica cast [$159] and return in ten to 14 days for the re-x-ray. If the radiographs are completely normal and he has no tenderness would then discontinue the cast. If the radiographs are still suspicious but not confirmatory, would recommend CT scan. Day 10. The orthopedic report stated: I was notified of the patient s dissatisfaction with his current splint. He was sent to hand therapy to have a different type of splint made. [H]is MR imaging was scheduled for the 25th and the patient did not want to wait that long because of his inability to write in his current splint. MR imaging is moved up to the 20th. I suggested they put the cast back on as a splint until he can have the study and if the study is negative, then he does not need anything. Day 13. MR imaging of the wrist showed a nondisplaced fracture through the scaphoid waist. Day 16. The orthopedic report stated: In light of the MR images revealing a scaphoid fracture, the patient was brought back to the clinic today for examination, fluoroscopy and casting. On the office fluoroscopy [$150], the scaphoid is definitely fractured. Summary. This scenario shows that the orthopedic surgeon s initial direction was to perform radiographic follow-up and splinting only, then to perform CT if still unsure. The orthopedic surgeon was possibly swayed by initial negative findings on office fluoroscopy. MR imaging was recommended as a more sensitive study, and a full study was scheduled. Possibly because of inexperience with MR imaging, the clinician performed one further office fluoroscopy that confirmed the fractured scaphoid on day 16. Screening MR imaging ($770) would have prevented $589 in office charges plus an additional $486 from the cost of full wrist MR imaging that was performed. Case Scenario 3 Day 1. The patient was a 14-year-old boy who fell on his right face and ulnar wrist (Fig. 5). C 1260 JR:177, December 2001

MR Imaging of Radiographically Occult Scaphoid Fractures On presentation to the emergency department, he was found to have fractures of his right facial bones and the ulnar styloid of his right wrist. His right wrist was then placed in a splint ($154), and an orthopedic consultation was ordered. Day 10. fter orthopedic evaluation, the right wrist splint was removed. The right wrist had tenderness over the ulnar styloid, and the distal radial ulnar joint showed trepidation with stress. Radiographs of the right wrist ($85) again showed a nondisplaced fracture of the ulnar styloid. The clinician s impression was right wrist sprain with associated ulnar styloid fracture, possible disruption of the distal radioulnar joint, and possible disruption of the triangular fibrocartilage complex. The orthopedic report stated: The patient should continue to wear the brace for protection. I have suggested that we obtain MR imaging of the right wrist to further clarify his pathology. He Fig. 4. 16-year-old boy who fell onto outstretched hand and incurred dorsiflexion injury to right wrist. Patient was seen at family medical clinic, where radiographs were obtained that raised possibility of scaphoid fracture., Radiograph obtained at initial visit is inconclusive., T2-weighted fast spin-echo coronal MR image (TR/TE, 4233/72) with fat saturation through scaphoid bone, obtained with flex coil through casting material 13 days after initial radiography, shows obviously poor signal-to-noise ratio. Even with this limitation, linear low-signal-intensity focus (straight arrow ) surrounded by high-signal-intensity edema is seen. Contusion of distal radial metaphysis is incidentally observed (curved arrow ). C, T1-weighted coronal MR image (500/14) through scaphoid bone obtained at same time as shows that low-signal-intensity fracture line is not well visualized in equally low-signal-intensity region of edema (white arrow ). Contusion of distal radial metaphysis (black arrow ) is again noted. C C Fig. 5. 14-year-old boy who fell on wrist., Initial radiograph reveals no evidence of scaphoid fracture. Subtle band of sclerosis (arrows) that most likely represents torus fracture was missed on initial interpretation., T2-weighted fast spin-echo coronal MR image (TR/TE, 4000/74.2) with fat saturation through scaphoid bone, obtained 11 days after initial radiography, reveals linear lowsignal-intensity focus (black arrow ) surrounded by high-signal-intensity edema. Diffuse bright edema of distal radial metaphysis suggests torus fracture (white arrow ). C, T1-weighted coronal MR image (500/14) through scaphoid bone obtained at same time as shows low-signal-intensity fracture line is not well visualized in equally lowsignal-intensity region of edema (white arrow ). Diffuse low-signal-intensity edema of distal radial metaphysis again suggests torus fracture (black arrow ). JR:177, December 2001 1261

Dorsay et al. will return after the study for final diagnosis and disposition. Day 12. The orthopedic report stated: Complete MR imaging ($1298) of the right wrist and hand is reviewed. He is noted to have a nondisplaced scaphoid body fracture. The triangular fibrocartilage has a small nondisplaced tear. His ulnar styloid fracture is visible. Day 17. The orthopedic report stated: The patient s right wrist is somewhat improved in terms of local tenderness. He s placed into a short arm thumb spica cast. He will follow-up in this office in four weeks for removal of his cast and then x-ray of hand and wrist out of cast. Summary. This scenario shows the inability of even the orthopedic surgeon to initially suspect the presence of scaphoid fracture in the setting of a generalized wrist injury (ulnar styloid fracture and possible distal radioulnar disruption). MR imaging clearly showed not only the ulnar styloid fracture but also the radiographically occult scaphoid fracture and triangular fibrocartilage injury. Distal radioulnar joint disruption was discounted on the basis of MR imaging. The patient s wrist was not fully encased in a cast until day 17. Case Scenario 4 The patient was a 20-year-old man who presented with mild snuffbox tenderness after a fall on outstretched hands during basketball practice (Fig. 6). Initial radiographic findings were negative. The patient s wrist was splinted. MR imaging of the wrist was performed later that day, revealing a normal scaphoid. The patient returned to normal activity without negative consequences. Summary. This scenario shows the usefulness of immediate MR imaging in avoiding inconvenience to the patient and decreasing hindrances to productivity. Discussion Isolated nondisplaced fracture of the carpal scaphoid bone is a frequently seen injury in the emergency department and acute care setting. pproximately 7% of scaphoid fractures are not visible on initial radiographs [3, 4]. If the injury is correctly diagnosed within the first week of trauma and the wrist immobilized, the more serious complication of nonunion can be prevented [3 7]. Colles-type short arm (thumb) spica cast is adequate for treating nondisplaced scaphoid wrist fractures. For suspected scaphoid fractures with no radiographic evidence of fracture, treating symptoms with radiographic follow-up has long been the standard of care [8, 9]. s our review of the literature indicated, positive findings at clinical examination and subsequent negative wrist series may lead to unnecessary outpatient follow-up appointments, unnecessary immobilization, and additional exposure to radiation. This additional radiation exposure includes office fluoroscopy, which is often used by the consulted orthopedic surgeon when follow-up radiographs have negative results, adding to patient expense and radiation dosage. To our knowledge, no substantial work in the literature addresses the sensitivity and specificity of fluoroscopy in the setting of negative radiographic findings. Frequently, further negative imaging results occur at the first follow-up appointment in the continued setting of a positive physical examination. Tiel-van uul et al. [10 13] found 25% of scintigraphically confirmed scaphoid fractures followed up for at least 1 year could not be confirmed by radiography. second follow-up appointment may incur charges (up to $547) nearing those of the initial appointment by virtue of similar repeated diagnostic procedures and recasting. one scanning, which is more sensitive and more specific, has been advocated by many as an alternative to immobilization and follow-up [10 14]. The cost at our institution ($639) would approach that of unnecessary clinical imaging and therapeutic follow-up (Tables 3 and 4). In practice, however, this examination is frequently ordered as a more complex study (e.g., with flow studies) with increased charges. dditionally, we can find no case of nuclear bone scanning being used on an emergent basis, thus again resulting in casting of the patient s wrist for at least some time. Many recent studies suggest equal sensitivity and better (sometimes 100%) specificity for MR imaging in accurately detecting radiographically occult scaphoid fractures [3, 4, 15, 16]. On MR imaging, the fracture line will appear as a discrete focus of linear low T1 and T2 signals. dditionally, a diffuse low T1 signal and a high T2 signal are seen in the surrounding scaphoid bone, representing marrow edema. However, bone marrow edema alone does not represent fracture but rather contusion. The T1- and T2-weighted images are often complementary in cases with subtle findings. Diffuse bright edema on T2-weighted MR images (Fig. 3) and low-signal-intensity edema on T1-weighted images (Fig. 5) may obscure a subtle fracture line that is seen on the other sequence. Fowler et al. [3] compared the sensitivity and specificity of MR imaging for detection of radiographically occult scaphoid fractures with that of bone scintigraphy. In 40 patients, bone scanning and MR imaging findings were in agreement. In three patients, a discrepancy existed between the imaging modalities; in all three cases, MR imag- Fig. 6. 20-year-old man with mild snuffbox tenderness after falling on outstretched hands during basketball practice., Initial wrist radiograph shows negative findings., T1-weighted coronal MR image (TR/TE, 500/11) through scaphoid bone, obtained 2 days after initial radiographs, shows no evidence of trauma. T2-weighted images were similarly unremarkable. 1262 JR:177, December 2001

MR Imaging of Radiographically Occult Scaphoid Fractures ing was found to be more sensitive and more specific. Thorpe et al. [15] concluded that MR imaging showed better interobserver agreement for scaphoid injury and fewer false-positive results than bone scanning. Cook et al. [16] studied 18 skeletally immature patients who had presented to the emergency department within 2 days of acute wrist trauma. Those authors reported normal findings on initial MR images to have a negative predictive value of 100%. Of 10 patients with a scaphoid abnormality on MR imaging, six had fractures and four had regional bone marrow edema. Initially, eight of 10 fractures were radiographically occult, although some eventually became evident on later studies. Those lesions with bone marrow edema did not progress to fractures. Clearly, in the setting of continued clinical symptoms and negative radiographic findings, MR imaging not only can exclude scaphoid fractures if not present but also can elucidate other causes for pain. retlau et al. [17] found that MR imaging also revealed a fracture of the capitate bone in one patient, a fracture of the triquetrum in two patients, and a bone bruise of one or more of the carpal bones in eight patients. Lepisto et al. [18] found MR imaging showed a wide spectrum of additional lesions in the wrist that were not detected by routine radiographic analysis. These lesions included seven fragmented triangular fibrocartilages, four torn scapholunate ligaments, and one torn triquetral lunate ligament. one bruises of other carpal bones and seven other carpal fractures were also detected. Cook et al. [16] found seven patients with evidence of extensor tenosynovitis on MR imaging. Thorpe et al. [15] found frequent occurrences of ligamentous injury and carpal instability at MR imaging that were not evident on scintigraphy. further advantage of MR imaging is the ability to image through casting material. This need occurs frequently when imaging takes place after the patient s initial presentation. Imaging through casting material requires a larger coil than the wrist coil (usually a flex or knee coil), resulting in a diminished signal-to-noise ratio. lthough not esthetically as pleasing, the images are diagnostic in our experience (Fig. 3). Fast short tau inversion recovery images may be substituted for fast spin-echo images with fat saturation if necessary. Cost and practicality naturally preclude the use of MR imaging as a first-line diagnostic study; however, the limited wrist protocol we have described that is now being used at our institution brings costs in line with the costs of bone scanning and the traditional first follow-up appointment (Tables 3 and 4). dditionally, this limited MR imaging can be performed in less than 30 min (including setup; 10 min of actual scanning time), which allows flexibility for same-day imaging in our institution. We have found increasing acceptance among the emergency department physicians, which is similar to our experience with MR imaging for occult hip fractures. We conclude that this limited wrist MR imaging protocol (in the setting of suspected scaphoid fracture and negative initial radiographic findings) is nearly as cost-effective as traditional follow-up and immobilization. ssuming the current standard of care (watchful immobilization) is applied, most patients with positive findings at clinical examination and negative radiographic results ( 75% by our review) will undergo needless casting (and possibly recasting), which compromises both lifestyle and productivity. Follow-up appointments additionally result in lost wage-earning hours (often a full day) or increased day-care expenses. When lost productivity and income to the patient are considered, screening MR imaging is favored at initial presentation. This practice has always been the case for the elite athlete, as in case scenario 4. The presence of multiple scanners and technologists make possible the rapid scheduling of this 30-min (or less) procedure. In the outpatient setting, performing MR imaging in the same day may be slightly more difficult. s two of our case scenarios showed, a savings of time and money is possible even if the MR imaging is performed before the first orthopedic appointment. References 1. Deutsch L, Mink JH, Waxman D. Occult fractures of the proximal femur: MR imaging. Radiology 1989;170:113 116 2. Haramati N, Staron R, arax C, Feldman F. Magnetic resonance imaging of occult fractures of the proximal femur. Skeletal Radiol 1994;23:19 22 3. Fowler C, Sullivan, Williams L, McCarthy G, Savage R, Palmer. comparison of bone scintigraphy and MRI in the early diagnosis of the occult scaphoid waist fracture. Skeletal Radiol 1998;27:683 687 4. Gaebler C, Kukla C, reitenseher M, Trattnig S, Mittlboeck M, Vecsei V. Magnetic resonance imaging of occult scaphoid fractures. J Trauma 1996;41:73 76 5. Roolker W, Tiel-van uul MM, ossuyt PM, et al. Carpal box radiography in suspected scaphoid fracture. J one Joint Surg r 1996;78:535 539 6. Roolker L, Tiel-van uul MM, ossuyt PP, Dijkstra PF, van Grieken HJ, roekhuizen TH. The value of additional carpal box radiographs in suspected scaphoid fracture. Invest Radiol 1997; 32:149 153 7. Roolker W, Maas M, roekhuizen H. Diagnosis and treatment of scaphoid fractures: can nonunion be prevented? rch Orthop Trauma Surg 1999;119:428 431 8. Schubert HE. Scaphoid fracture: review of diagnostic tests and treatment. Can Fam Physician 2000;46:1825 1832 9. Ohiorenova D, Whitwell DJ. Occult scaphoid fracture: need to avoid complacency case report and literature review. J Trauma 1996;41: 1054 1055 10. Tiel-van uul MM, van eek EJ, roekhuizen H, akker J, os KE, van Royen E. Radiography and scintigraphy of suspected scaphoid fracture: a long-term study in 160 patients. J one Joint Surg r 1993;75:61 65 11. Tiel-van uul MM, Roolker W, roekhuizen H, Van eek EJ. The diagnostic management of suspected scaphoid fracture. Injury 1997;28:1 8 12. Tiel-van uul MMC, Roolker W, Verbeeten W Jr, roekhuizen H. Magnetic resonance imaging versus bone scintigraphy in suspected scaphoid fracture. Eur J Nucl Med 1996;23:971 975 13. Tiel-van uul MM, roekhuizen TH, van eek EJ, ossuyt PM. Choosing a strategy for the diagnostic management of suspected scaphoid fracture: a cost-effectiveness analysis. J Nucl Med 1995;36:45 48 14. ayer LR, Widding, Diemer H. Fifteen minutes bone scintigraphy in patients with clinically suspected scaphoid fracture and normal x-rays. Injury 2000;31:243 248 15. Thorpe P, Murray D, Smith FW, Ferguson J. Clinically suspected scaphoid fracture: a comparison of magnetic resonance imaging and bone scintigraphy. r J Radiol 1996;69:109 113 16. Cook P, Yu JS, Wiand W, Cook J II, Coleman CR, Cook J. Suspected scaphoid fractures in skeletally immature patients: application of MRI. J Comput ssist Tomogr 1997;21:511 515 17. retlau T, Christensen OM, Edstrom P, Thomsen HS, Lausten GS. Diagnosis of scaphoid fracture and dedicated extremity MRI. cta Orthop Scand 1999;70:504 508 18. Lepisto J, Mattila K, Nieminen S, Sattler, Kormano M. Low field MRI and scaphoid fracture. J Hand Surg r 1995;20:539 542 19. Grover R. Clinical assessment of scaphoid injuries and the detection of fractures. J Hand Surg r 1996;21:341 343 20. Munk, Frokjaer J, Larsen CF, et al. Diagnosis of scaphoid fractures: a prospective multicenter study of 1,052 patients with 160 fractures. cta Orthop Scand 1995;66:359 360 21. Parvizi J, Wayman J, Kelly P, Moran CG. Combining the clinical signs improves diagnosis of scaphoid fractures: a prospective study with follow-up. J Hand Surg r 1998;23:324 327 JR:177, December 2001 1263