|
|
||||||||
Perspective |
1 Department of Radiology, Leiden University Medical Center, Albinusdreef 2, P.
O. Box 9600, NL-2300 RC, Leiden, The Netherlands.
2 Department of Orthopaedic Surgery, Leiden University Medical Center, NL-2300
RC, Leiden, The Netherlands.
3 Present address: Department of Diagnostic Radiology, University Hospital
Ghent, De Pintelaan 185, 9000 Ghent, Belgium.
4 Department of Pathology, Leiden University Medical Center, NL-2300 RC, Leiden,
The Netherlands.
Received March 28, 2002;
accepted after revision May 30, 2002.
Address correspondence to W. R. Obermann.
Introduction
|
|
|---|
|
|
The first report in the literature of technical and clinical success with radiofrequency thermal ablation in the treatment of osteoid osteoma by Rosenthal et al. [5] appeared in 1992. This treatment has been performed in our hospital since 1994. Now, a decade later, CT-guided radiofrequency thermal ablation has been proven to be an accepted, safe, minimally invasive, and cost-effective treatment for osteoid osteoma [5,6,7,8,9,10,11]. The radiologist's role in the management of this condition has evolved from simply confirming the diagnosis of osteoid osteoma (along with his or her orthopedic colleagues) to curing the abnormality.
Although this technique is now routinely used in some tertiary referral centers, it could be offered more widely, because in many centers surgery is still routinely performed. This technique could be performed in centers of reasonable size, preferably in those offering a dedicated musculoskeletal imaging service. Radiologists with expertise in bone biopsy or some experience in interventional radiology are ideally suited to learn the procedure. After performing the treatment in approximately 20 patients, the radiologist will have experience in most variations in location and technique. This initial training could be acquired during a formal musculoskeletal or interventional fellowship at a major center offering this treatment, or, for a more experienced radiologist, during a minifellowship. Thereafter, a steady number of patients per year based on local referral patterns and cooperation with the center's orthopedic surgeons are an important consideration in maintaining skills and justifying the cost of equipment. However, to our knowledge, a detailed description of how to perform the procedure, along with technical tips for optimizing outcome and avoiding pitfalls, is still not available in the literature. We recall the steep and long part of the learning curve required to achieve excellent results with this method. For example, some osteoid osteomas by nature of size or location are more difficult to treat and warrant special consideration. Spinal lesions are an area of special interest yet are ideally suited to this technique, sparing the patient and surgeon a difficult and potentially hazardous operation. Furthermore, the diagnosis and management of residual and recurrent lesions can prove problematic.
This perspective will outline the technique of radiofrequency thermal ablation in the treatment of primary and recurrent osteoid osteoma based on our experience in treating nearly 130 patients since 1994, with a success rate of 92%, defined by the relief of pain. Our purpose was to educate radiologists considering introducing this form of treatment to their center and to offer suggestions for refinements in technique to those already performing it.
The Principle of Thermocoagulation
|
|
|---|
|
|
To perform radiofrequency thermal ablation successfully, one must
understand the concept of a "treatment zone," which may be defined
as the amount of tissue ablated. The maximal size of the treatment zone may be
predicted by the following equations:
![]() | (1) |
![]() | (2) |
For a noninsulated electrode tip length of 5 mm, there is an approximately 1-cm spherical treatment zone of focal osteonecrosis [5, 14, 15].
|
|
|---|
Equipment
CT guidance affords the best available visualization of needle and probe
placement in the lesion nidus. Helical CT with low-dose,
"quick-check" CT fluoroscopy results in savings of time and dose
for the patient [16,
17]. A general anesthetic
allows a pain-free procedure and absolutely stable patient position, although
a spinal anesthetic is an option for patients with lower limb lesions. Early
in our experience, local anesthetic proved unsuccessful because of inadequate
pain control, in spite of adequate anesthetic infiltration in soft tissue and
overlying periosteum. Entering the nidus itself elicits extreme pain, in most
cases resulting in patient movement and loss of position. The time required
for the procedure is typically 90 min, including the time until the patient is
stable under anesthesia.
For most patients with limb lesions, supine positioning is ideal because it affords good access and is the best position for administration of a general anesthetic. The limb may be internally or externally rotated and secured with tape or straps to allow good skin access, easier needle placement, and avoidance of neurovascular structures. Spinal lesions are treated with the patient lying prone with a padded ring beneath the chest for easier ventilation. A laser goniometer (Targo-Beam; Vasculab Medizintechnik, Wismar, Germany) can help guide accurate, first-time needle placement (Fig. 3A,3B).
|
|
We use the Bonopty coaxial bone biopsy system (Radi Medical Systems, Uppsala, Sweden) for lesion access. The system's small-caliber needles and multicapability components such as drill and biopsy cannula make it ideally suited for radiofrequency thermal ablation. The system comprises a 95-mm-long 14-gauge (2.1-mm) Bonopty Penetration cannula, a 100-mm-long 15-gauge (1.7-mm) Bonopty Drill, a 16-mm-long 15-gauge (1.7-mm) Bonopty Extended Drill, and the 160-mm-long 15-gauge (1.7-mm) Bonopty Biopsy cannula. The radiofrequency thermal ablation probe (SMK-TC15; Radionics, Burlington, MA) is a 15-cm-long, straight, rigid electrode with a diameter of 1 mm. An incorporated temperature-measuring device (thermistor) allows precise monitoring of the probe-tip temperature. The probe is introduced using a 145-mm-long dedicated Sluyter-Mehta 20-gauge thermal ablation cannula with a 5-mm-long noninsulated tip (Radionics). The 20-gauge thermal ablation cannula is placed through the bone-penetration cannula at the time of treatment.
The radiofrequency thermal ablation probe is connected to a radiofrequency generator (Radionics-RFG 3C RF-Lesion Generator System), which supplies the monopolar radiofrequency current. The device delivers an AC of approximately 500 kHz in a continuous unmodulated sinusoidal waveform when in "lesion" mode.
Grounding
Grounding consists of a dispersive electrode placed close to the lesion
site to draw current back to the radiofrequency unit. A large-area adhesivegel
grounding pad like that used in the operating room has the advantages of no
skin penetration and reduced current density, resulting in less heating at the
dispersive electrode to avoid tissue burns at this site
(Fig. 4A).
|
Contraindications
The radiofrequency generator can cause unwanted physiologic effects;
therefore, this technique is contraindicated in patients with cardiac
pacemakers. Anesthetic-monitoring equipment has not caused a problem in our
group of patients to date, although experience is limited at this stage.
Potential Complications
Although this technique is minimally invasive, potential complications that
may occur during needle passage include bleeding and nerve injury. These can
be avoided by knowledge of anatomic structures in the region of needle passage
and an alteration in the approach to avoid neurovascular bundles. Soft-tissue
burns, especially skin burns, are a further possible complication. There is a
higher risk of skin necrosis in osteoid osteomas in superficially located
bone, in which extra care is required. This complication is avoided by
withdrawing the outer cannula to approximately 1 cm above the noninsulated tip
of the coagulation cannula. In spite of this precaution, we had one skin burn
complication in a patient with an anterior tibial osteoid osteoma, with
subsequent development of a subcutaneous fistula to the skin, requiring
surgical débridement. The most likely explanation is a defect in the
insulation material covering the thermal ablation cannula that was not
confirmed because the cannula had been discarded immediately after the
procedure. A quick visual check of the insulation material to avoid this rare
occurrence is advised.
Informed Consent
Formal informed consent is obtained before the procedure and should include
an outline of the procedure with specific mention of the low risk of potential
complications as described previously. We specifically state that a second
treatment may be required in the unlikely event that there are residual or
recurrent symptoms. For osteoid osteomas close to joints, especially in the
hand or foot, the patient should be warned that damage to articular cartilage
is a possibility, which may predispose to early degenerative arthritis.
However, this should be balanced by the fact that surgery is potentially even
more damaging. The patient should be advised that after treatment, pain may be
transiently increased for 24-48 hr, for which a moderate-strength oral
analgesic may be required. Informed consent for a general anesthetic is
obtained from the patient by the anesthesiologist on the day of treatment.
|
|
|---|
Eight Steps
Step 1, localization and planning.From a 3- to 4-cm block
of 1-mm-thick slices, the precise lesion size is determined. For osteoid
osteoma of less than 1 cm, a single, central skin entry point is planned, and
the table position noted. For a lesion greater than 1 cm, more than one probe
position must be planned because of the limited size of the treatment zone.
The best approach is then chosen, and the angle of inclination is measured (a
90° true vertical approach is often easiest but not always possible near
vital structures). The aim is to puncture in the scan plane; planning an entry
point perpendicular to the bone surface will help to avoid needle slippage
(Fig. 4B). Safe anatomic
entrance is sometimes through the opposite, normal cortex. This allows
avoidance of neurovascular structures and joints without drilling through
added amounts of hard, reactive bone.
|
Step 2, grounding pad.The grounding pad should be placed close to the planned skin entry point to allow the shortest current path through the patient (Fig. 4A). The area is then sterilized and draped.
Step 3, superficial bone entry with use of tenting.A 2-mm skin incision is made, and the penetration cannula with stylet is then inserted through the soft tissues and into the bone surface (Fig. 4C). If the needle tip slips at this stage, the skin and underlying tissues around the cannula can be gathered up with a pinchlike maneuver to form a "tent." While the operator exerts forward force on the cannula against the bone, the soft tissues are lifted partway over the shaft to enable better purchase for bone entry. This technique of tenting minimizes the tendency of gravity, overlying muscles, and other elastic tissues to displace the needle tip (Figs. 4D and 4E).
|
|
|
Step 4, drilling and milling.The inner stylet is exchanged for the drill (when needed for deeper lesions), and drilling to the edge of the nidus is performed (Fig. 4F). During drilling, position and direction are verified with further scans. The drill has an eccentric mechanism, which creates a larger hole than the actual drill diameter; however when no progress is made in dense bone, the drill can be carefully removed and cleaned. If a slightly wayward direction occurs, the edge of the channel can be "milled" by angulating the drill with forward pressure toward the lesion. The penetration cannula is carefully advanced over the drill so that it sits at least in bone cortex, and the drill is then removed. The penetration cannula is held firmly against the bone during exchange to maintain a stable position. The anchored penetration cannula now serves as a fixed pathway for biopsy and radiofrequency thermal ablation.
|
Step 5, biopsy.Although biopsy is optional because the decision to treat the patient has typically been made before the procedure, we perform it routinely for histologic confirmation of the diagnosis that may prove helpful in cases of residual or recurrent symptoms.
The biopsy cannula is inserted through the penetration cannula, and the biopsy specimen is removed by using gentle suction (Fig. 4G). Tissue material, fixed in formalin, is sent for histology. Appropriate decalcification is mandatory before embedding and cutting slides. In confusing clinical presentations, material is also occasionally sent for microbiologic examination.
|
Step 6, cannula and probe placement.The thermal ablation cannula with a stylet is next inserted through the bone-penetration cannula, and a check scan is made. Generally, the thermal ablation cannula tip should lie in the center of the lesion. Control scans of 2 mm in thickness are obtained 5 mm cranial and 5 mm caudal to the thermal ablation cannula tip position. If the osteoid osteoma nidus is contained in these limits, then the position is satisfactory. Thicker control scans make it easier to visualize the lesion but are less accurate for positioning (Rosenthal DI, personal communication). The radiofrequency thermal ablation probe (the tip protrudes slightly) replaces the stylet, which is removed (Fig. 4H). The penetration cannula is then withdrawn slightly so that its tip is at least 1 cm above the bare tip of the thermal ablation cannula. This position will prevent the current contacting the penetration cannula and resulting in undesired tissue burns or loss of current. A final check scan is obtained to ensure that there is no loss of position during the exchange of the thermal ablation probe for the stylet or on withdrawal of the bone-penetration cannula tip.
|
If the lesion extends beyond the outer confines stated previously, there is no guarantee that the entire lesion will be ablated. A second ablation must then be planned in the same session to ensure that the thermal ablation cannula tip is within 5 mm of the inner edge of the nidus to ensure incorporation in the treatment zone and adequate ablation. A second ablation position can often be achieved through the first access hole by angulating the penetration cannula and milling the edge of the hole with the drill to the desired position. This procedure will require check scans outside the previous scan plane.
Step 7, electrode connection.To avoid any effects of
galvanic potentials or static electricity, the radiologist must first attach
the dispersive electrode cord to the grounding pad by an alligator clip and
plug it into the reference jack on the radiofrequency generator. The
radiofrequency thermal ablation probe may then be connected (this can be done
by the assistant to enable the primary operator to remain sterile). A further
check scan at this point is not usually required unless the patient moves
excessively during the connection maneuver. The radiofrequency generator is
switched on, and electrical impedance is measured. Expected tissue resistance
is 200-600
, which confirms an adequate circuit. When a higher value is
recorded, the thermal ablation probe and cannula should be checked for
external damage. If tissue resistance exceeds 1000
and is associated
with increased current requirement, there is an inadequate circuit. The
temperature at the tip of the thermal ablation probe is the guide to assessing
the current requirement; an increased current requirement is defined by the
need to excessively increase the current output to achieve the desired
temperature of 90°C. In this situation, the possible causes of an
increased requirement include an equipment fault or more commonly a "dry
tip" with poor electrical coupling. Removing the thermal ablation probe
and flushing the thermal ablation cannula with 1 mL of 0.9% saline should
improve electrical conductivity and current flow and correct resistance
values.
Step 8, radiofrequency thermal ablation: 90°C for 4 min.The automatic temperature override control is set to 93°C (the maximal desired temperature). Built-in circuitry will prevent the lesion temperature from exceeding the set value. Radiofrequency thermal ablation is performed by smoothly turning the output control knob for 30-60 sec until the desired temperature of 90°C is displayed. The lesion time control is set to its maximum of 2 min, then repeated (total time, 4 min). Rosenthal et al. [9], who began using an ablation time of 4 min, now prefer 6 min because patients were experiencing a number of recurrences [5, 9]. However, Rosenthal concedes that there is probably little scientific rationale for this change because his group's original experiments had shown that a steady state is reached with little change in treatment-zone size after approximately 3 min (Rosenthal DI, personal communication). We are satisfied with our present protocol of 4 min ablation time. During ablation, the output control button should be adjusted up or down to ensure a stable lesion temperature of 90°C (down-regulation of current load is most often necessary) [5]. If a second ablation is required, the bone-penetration cannula can be repositioned through a separate skin incision if necessary. The previous steps are repeated to ensure ablation of the entire lesion.
Current Intensity
Current intensity is one of the most important variables influencing the
size of the treatment zone. A lack of appreciation of the size of the
treatment zone relative to the size of the osteoid osteoma to be ablated may
result in failure to treat the entire lesion and inevitable residual or
recurrent symptoms. Appropriately applied current will result in a treatment
zone of predicted and desired size. The current is too low when the
temperature at the tip fails to reach 90°C for the 4-min duration of
treatment, producing an undersize zone. If the current is too high or is
applied too rapidly, heating may be so intense that solidification and
charring limit further current flow. These effects result in a suboptimally
small treatment zone. Charring manifests as an abrupt fall in current with a
voltage rise due to increased tissue resistance. Alternatively, areas of
vaporization can result in an irregular-shaped zone that may be larger in
parts, but with other areas that are not ablated
[13]
(Fig. 5). The technique
advocated in this article uses a noncooled electrode tip that restricts the
size of the treatment zone but has the advantage of precise tissue destruction
encompassing the nidus itself, with minimal damage to normal adjacent bone.
The noncooled tip is ideal for most osteoid osteomas. Cooled tips such as
those used in radiofrequency thermal ablation of various liver lesions involve
the use of an infusion of cool saline through the electrode, which has the
advantage of allowing greater heat transmission with higher currents to create
larger treatment zones. However, the size of the treatment zone is also not
completely predictable with a cooled tip (Rosenthal DI, personal
communication). Although larger treatment zones would occasionally be
desirable in treating some osteoid osteomas to avoid multiple probe
placements, we prefer the noncooled tip to have an entirely predictable
treatment-zone size and thus to minimize damage to adjacent nonlesion
tissue.
|
Physiologic Reaction to Radiofrequency Thermal Ablation
Despite the use of general anesthetic, in about 50% of patients, we have
observed a physiologic reaction to entering the nidus or during ablation. This
reaction consists of variable increases in blood pressure, heart rate, and
respiratory rate. The patient may even move and cause a loss of position. The
reaction subsequently normalizes when the lesion is completely destroyed and
seems compatible with theories suggesting a neurogenic origin for the pain
associated with an osteoid osteoma
[18,19,20].
We routinely ask our anesthetist to look for this reaction because it is
useful confirmation that we have entered a nidus.
Postprocedure and Bone Healing
Pain is variable after radiofrequency thermal ablation. Some patients
report pain for up to 1 or 2 days after the procedure. This typically settles
quickly, and analgesia is rarely required. A clinical check is made before
discharge (usually the same day as the procedure). Patients may bear weight
immediately and return to normal activities (including sports). Follow-up
clinical assessment is made at 2 weeks. At this time, patients with persistent
pain requiring a second thermal ablation can be identified. Patients who
experience recurrent pain after a period of immediate pain relief can be
identified either by instructing the patient to come back or by using a
standard follow-up scheme with an increasing interval. Ultimately, resolution
of pain is the primary parameter used to define a successful treatment.
Radiographically, partial or complete infilling of the nidus with sclerotic bone (Fig. 6A,6B,6C) is expected over 2-27 months, although little or no change in lesion appearance is also possible. Eventually the nidus site can become indistinguishable from surrounding bone, and reactive changes in adjacent bone and periosteum also tend to diminish.
|
|
|
|
|
|---|
Spinal Lesions
Spinal lesions compose 10% of cases of osteoid osteoma. Most lesions occur
in the lumbar spine (59%), followed by the cervical (27%), thoracic (12%), and
sacral regions (2%) [1,
3,
4]. Lesions are almost always
in the posterior elements and are located in the pedicles (75%), laminae,
articular processes, and only uncommonly in the vertebral body. More
important, for this form of treatment, lesions do not usually involve the
spinal canal or paraspinal tissues.
The technique used for spinal lesions is essentially the same as that for any other site. The patient is placed prone, and a posterior approach is used (Fig. 6A,6B,6C). For most lesions in the posterior elements, ablation should be straightforward. Pedicle or posterior vertebral body lesions will require considerably more drilling, analogous to routine bone biopsy at these sites. Care should be taken to avoid entering a facet joint or neural foramen. When no complete shelf of protective bone surrounds a lesion, thermal injury to neural or vascular structures may occur. We have not had to treat such a lesion to date; however, Dupuy et al. [21] successfully treated a 1-cm intraspinal osteoid osteoma sited at the junction of the T11 right pedicle and lamina, which abutted the thecal sac with no ill effects. These same researchers measured temperature changes in the adjacent spinal canal while applying radiofrequency to vertebral bodies in pigs. No cytotoxic temperature elevations were recorded in the spinal canal [21].
In an ex-vivo study reported in the same article [21], the authors confirmed that there is decreased heat transmission in cancellous and cortical bone. They raised the possibility that local heat sinks due to the presence of the rich epidural venous plexus and cerebrospinal fluid pulsation may also be factors.
For most spinal osteoid osteomas, radiofrequency thermal ablation is a safe and effective alternative to medical management and preferable to surgery. As mentioned by Dupuy et al. [21] and in spite of their success in one case, this technique may be contraindicated when no intact cortex is evident between the nidus and the spinal cord or nerve root.
Large Lesions
For osteoid osteomas larger than 1 cm in any dimension, radiofrequency
thermal ablation must be performed at more than one probe position, preferably
with overlap of the treatment zones. Large cancellous bone osteoid osteomas
are more likely to be round; therefore, we place the probe in two access
tracts in a craniocaudal direction. In each tract, we perform the ablation in
two separate positions, superficial and deep (four positions and a 16-min
ablation) (Fig.
7A,7B,7C).
Large cortical osteoid osteomas in the long bones tend to be elongated and may
exceed 1 cm in length in a craniocaudal direction. New multidetector CT
scanners with fast multiplanar reconstructions may aid planning by allowing
assessment of the precise lesion configuration at the time of treatment (Fig.
8A,8B).
|
|
|
|
|
Lesions Close to Joints
The hip joint is involved much more commonly than the elbow, wrist, knee,
and foot. Avoiding a transarticular approach will minimize the small risk of
introducing infection, negate cooling effects from any associated joint
effusion, and also reduce the chance of inadvertent ablation of articular
cartilage. For acetabular lesions, needle passage through the hip joint is
usually not required. For lesions of the anterior proximal femoral neck, joint
passage is occasionally required; however, we have encountered no complication
in the few cases in which it was necessary. Internally rotating the lower limb
provides easier access (perpendicular to the cortex) to lesions of the
anterior proximal femoral neck by an anterior approach. The femoral
neurovascular bundle is usually easy to avoid. For posterior proximal femoral
neck lesions, we generally also use an anterior approach. For lesions close to
small joints, particularly in the carpus or tarsus, some degree of articular
cartilage damage is inevitable (as would occur in surgery), and this should be
discussed with the patient at the time of obtaining consent.
Lesions Adjacent to Open Growth Plates
A cranially or caudally angulated approach in these rare cases reduces the
risk of damaging the undulating growth plate
(Fig. 9).
|
Superficial Lesions
A dorsal approach for most carpalmetacarpal and
tarsalmetatarsal lesions protects the important volar structures
(Fig. 10A). After placement of
the thermal ablation probe, we withdraw the bone-penetration cannula to above
the level of the skin to negate the risk of contact with the active tip and
thus avoid skin burns (Fig.
10B). The minimal distance for safety from the tip of the
bone-penetration cannula to the skin is 1 cm. The outer cannula can be taped
to the thermal ablation cannula with sterile tape, or, if taping is not
possible, removed entirely. The thermal ablation cannula is then reinserted
through the skin and drill hole, followed by the thermal ablation probe.
|
|
Diagnosis and Management of Residual and Recurrent Symptoms
Residual symptoms may be defined as pain or impaired function or both,
identical to or resembling the presenting complaints, that persist for more
than 2 weeks after radiofrequency thermal ablation. Recurrent symptoms are
defined as the reappearance of symptoms that follow a symptom-free period
after radiofrequency thermal ablation.
Patients with residual and recurrent symptoms will usually benefit from a second ablation of the apparently incompletely treated nidus. Remnant areas of viable lesion persist peripherally in round lesions or at one or both ends of elongated lesions. We repeat the initial imaging protocol when symptoms are present, and we have found that radiographs and limited-section CT are often sufficient. Radiologic confirmation depends on the timing of recurrent symptomsthe longer the interval, the more the sclerosis is seen at the ablation site on CT. Lesion radiolucency stands out, giving a good target for a repeated procedure. Dynamic contrast-enhanced CT or MR imaging can help when mixed areas of radio-lucency and sclerosis on CT make it difficult to determine the precise location of the recurrence. Rapid enhancement within 3-6 sec after arterial enhancement, followed by rapid washout, in our experience and in that of Von Kalle et al. (Von Kalle T et al., presented at the International Pediatric Radiology meeting, May 2001), is indicative of the presence of viable lesion tissue (Fig. 11A,11B,11C).
|
|
|
Two or more probe placements may be required for the follow-up radiofrequency thermal ablation, which can be expected to result in a cure. If symptoms persist but the lesion continues to look like an osteoid osteoma, a third ablation can be attempted. However, if the lesion now appears atypical, surgical excision is probably warranted when feasible. The initial treatment using this technique is based on clinical and radiologic proof. Although a definitive histologic diagnosis is not required, confirmatory information can be helpful in guiding further management in case of residual or recurrent symptoms.
|
|
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
D. Motamedi, T. J. Learch, D. N. Ishimitsu, K. Motamedi, M. D. Katz, E. W. Brien, and L. Menendez Thermal Ablation of Osteoid Osteoma: Overview and Step-by-Step Guide RadioGraphics, November 1, 2009; 29(7): 2127 - 2141. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. D. Brown and E. vanSonnenberg Issues in Imaging-Guided Tumor Ablation in Children Versus Adults Am. J. Roentgenol., September 1, 2007; 189(3): 626 - 632. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. M. Vanderschueren, A. H. M. Taminiau, W. R. Obermann, A. A. van den Berg-Huysmans, and J. L. Bloem Osteoid Osteoma: Factors for Increased Risk of Unsuccessful Thermal Coagulation Radiology, December 1, 2004; 233(3): 757 - 762. [Abstract] [Full Text] [PDF] |
||||
![]() |
U. Albisinni, E. Rimondi, M. C. Malaguti, R. Ciminari, C. H. Pinto, and W. R. Obermann A Comment on CT-Guided Radiofrequency Thermal Ablation of Osteoid Osteoma Am. J. Roentgenol., August 1, 2004; 183(2): 538 - 538. [Full Text] [PDF] |
||||
![]() |
P O'Donnell and P O'Donnell Evaluation of focal bone lesions: basic principles and clinical scenarios Imaging, December 1, 2003; 15(4): 298 - 323. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |