The aim of this study was to compare accuracy of an image guided percutaneous core needle biopsy (PCNB), using ultrasound or computed tomography, to PCNB without image guidance in the diagnosis of palpable soft tissue tumors.

One hundred forty patients with a suspected soft tissue sarcoma underwent a percutaneous core needle biopsy with or without image guidance. One hundred eleven patients had subsequent surgical excision. The accuracy of guided PCNB and blind PCNB was calculated by comparing the histological results of the needle biopsy to the surgical specimen.

The diagnostic accuracy of blind percutaneous core needle biopsy was 78% (36 of 46 biopsies) and significantly lower (p ≤ 0.025) in comparison to image guided percutaneous core needle biopsy which was 95% (62 of 65 biopsies).

We suggest that image guidance improves the diagnostic accuracy of PCNB especially for small size deep sited suspected soft tissue tumours.

Introduction: Within a study group of 102 consecutive patients diagnosed with chondrosarcoma of the femur, tibia or humerus, an association with previously treated breast cancer was noted. We researched this proposed relationship.

Methods: We retrospectively reviewed the records of all patients diagnosed histologically with chondrosarcoma of the femur, tibia or humerus over a six-year period at a supra-regional bone tumour unit. We identified those patients who had previously been treated for breast cancer.

Results: There were 58 female and 44 male patients. The study group contained six females (10%, mean age 53 years) who had previously been treated for breast cancer, a higher proportion than would be expected. They were referred following identification of a solitary area of increased activity on routine screening with isotope bone scan, presumed to be a solitary bony metastasis. Most (86%) of this breast carcinoma sub-group had developed low-grade bone chondrosarcoma (Trojani grade 0.5-I) and only one case (14%) had developed high-grade chondrosarcoma (Trojani grade II–III).

Discussion: A suspicious long bone lesion on bone scan in a patient with a past medical history of breast cancer must, therefore, not be assumed to be a metastasis without further investigation; the possibility of a chondral lesion should be considered. It is important that patients receive a full multidisciplinary team investigation prior to treatment in order to obtain the correct tissue diagnosis, as the management of these conditions is often different. Our study suggests there may be a relationship between patients previously treated for breast cancer and the development of subsequent chondrosarcoma.

Introduction: Chondrosarcoma is the second most common primary malignant bone tumour. Distinguishing between grades is not necessarily straightforward and may alter the management of the disease. We evaluated the correlation between the pre-operative needle biopsy and excision biopsy histological grading of chondrosarcoma of the femur, tibia and humerus.

Methods: A consecutive retrospective series of 100 patients with a histological diagnosis of chondrosarcoma made at a supra-regional bone tumour unit was reviewed. Twenty-one patients were excluded because 20 had only excision biopsy, due to radiological confidence in the diagnosis, and one had only the pre-operative biopsy on record, thus this series included 79 available cases. The remaining patients underwent a pre-operative needle biopsy.

Results: In 11 instances, there was a discrepancy in histological grade. Therefore, there was an 86% (68 out of 79) accuracy rate for pre-operative histological grading of chondrosarcoma, based on needle biopsy. However, the accuracy of the diagnostic biopsy to distinguish low-grade from high-grade was 90% (71 out of 79).

Discussion: From this series we conclude that accurate image-guided biopsy is a very useful adjunct in determining histological grade of chondrosarcoma and the subsequent treatment plan. At present, a multidisciplinary approach, comprising experienced Orthopaedic Surgeons, Radiologists and Pathologists offers the most reliable means of accurately diagnosing and grading chondrosarcoma of long bones.

Introduction: The Musculoskeletal Tumour Society recommends that patients with musculoskeletal tumours are treated in specialist centres. Core needle biopsy is an effective method of obtaining tissue diagnosis but a dilemma arises when the material is non-diagnostic. Our aim was to evaluate the management of non-diagnostic biopsies.

Method: We retrospectively reviewed all core needle biopsies performed between 2003 and 2009 in our regional centre. Non-diagnostic biopsies were identified and management reviewed.

Results: 4,520 core needle-biopsies were performed of which 120 (2.6%) were non-diagnostic. Of these 85 (70%) were treated definitively on the basis of existing imaging, 8 (7%) required further imaging before treatment and 27 (23%) had a repeat biopsy.

Of the 27 repeat biopsies a positive histological diagnosis was obtained in 22 patients. The remaining 5 were again non-diagnostic giving a total of 98 patients being treated definitively without a tissue diagnosis.

Of these 98 cases, 39 (40%) were treated non-operatively, 37 (38%) had curettage and 22 (22%) underwent wide excision.

In the curettage group 33 out of 37 patients had a benign tumour on final histology. Four patients turned out to have intermediate/high grade tumours and subsequently underwent wide excision.

In the wide excision group, 17 out of 22 patients had an intermediate/high grade tumour on final histology. Five patients underwent an unnecessarily wide excision of a benign lesion.

None of the patients treated non-operatively turned out to have a tumour.

Conclusion: After non-diagnostic core-needle biopsy, the patient can safely be managed without tissue diagnosis, with low error rate, provided they have been subjected to a multidisciplinary discussion.

Objectives:

To determine the diagnostic performance of image-guided percutaneous core needle biopsy (CNB) in patients presenting with pathologic fractures of the appendicular skeleton.

A retrospective audit identified 129 consecutive patients presenting with pathological fractures to a specialist orthopaedic oncology unit over a 9 year period. All patients underwent percutaneous CNB using CT (n=98), fluoroscopy (n=15) or US (n=16) guidance. In all cases MRI or CT was available prior to biopsy to assess the presence and degree of extra-osseous tumour mass. The resulting sample was classified as diagnostic (Group 1) or non-diagnostic (Group 2) on histopathological study. Diagnostic performance was evaluated on the basis of the diagnostic yield and accuracy; these were related to the site of the lesion and presence/absence of extra-osseous mass.

Of 129 biopsies, 99 (77%) were classified as Group 1 and 30 (23%) as Group 2. The commonest sites of pathological fracture without associated soft tissue component and resulting in a non-diagnostic biopsy were the proximal femur and proximal humerus. The average cross-sectional diameter of lesions in Group 1 was 5.7 x 5.9cm. Of the 30 lesions comprising Group 2, no soft tissue component was identified on pre-biopsy cross-sectional imaging in 27 lesions (90%) whereas the remaining 3 (10%) showed a smaller extra-osseous soft tissue component compared to the lesions in Group 1.

Image-guided percutaneous CNB is a reliable method for obtaining a tissue diagnosis in patients presenting with a pathologic fracture of the appendicular skeleton with high accuracy rate. However, those lesions which are purely intra-osseous or have only very small extra-osseous components are likely to be associated with a non-diagnostic biopsy, and should be considered for a primary open procedure.

Within a study group of 102 consecutive patients diagnosed at a supra-regional bone tumour unit with chondrosarcoma of the femur, tibia or humerus, an association with previously treated breast cancer was noted.

There were 58 female patients and 44 male patients. The study group contained six females (10%, mean age 53 years) who had previously been treated for breast cancer, a higher proportion than would be expected. They were referred following identification of a solitary area of increased activity on routine screening with isotope bone scan, presumed to be a solitary bony metastasis.

Most (86%) of this breast carcinoma sub-group had developed low-grade bone chondrosarcoma (Trojani grade 0.5-I) and only one case (14%) had developed high-grade chondrosarcoma (Trojani grade II-III).

A suspicious long bone lesion on bone scan in a patient with a past medical history of breast cancer must, therefore, not be assumed to be a metastasis without further investigation; the possibility of a chondral lesion should be considered. It is important that patients receive a full multidisciplinary team investigation prior to treatment in order to obtain the correct tissue diagnosis, as the management of these conditions is often different.

Our study suggests there may be a relationshipbetween patients previously treated for breast cancer and the development of subsequent chondrosarcoma.

Chondrosarcoma is the second most common primary malignant bone tumour. Distinguishing between grades is not necessarily straightforward and may alter the management of the disease. We evaluated the correlation between the pre-operative needle biopsy and excision biopsy histological grading of chondrosarcoma of the femur, tibia and humerus.

A consecutive retrospective series of 100 patients with a histological diagnosis of chondrosarcoma was reviewed. Twenty-one patients were excluded because 20 had only excision biopsy and one had only the pre-operative biopsy on record, thus this series included 79 available cases. In 11 instances, there was a discrepancy in histological grade.

Therefore, there was an 86% (68 out of 79) accuracy rate for pre-operative histological grading of chondrosarcoma, based on needle biopsy. However, the accuracy of the diagnostic biopsy to distinguish low-grade from high-grade was 90% (71 out of 79).

Introduction: Pathological fractures commonly undergo biopsy to establish tissue diagnosis and plan definitive surgical management.

Methods: 129 patients undergoing image-guided needle biopsy of a pathological fracture between 1998 and 2007 were prospectively identified. Imaging was by CT, ultrasound or fluoroscopy. Biopsy was by Jamshedi, Temno or Trucut needle. The outcome measure was ability to make a tissue diagnosis by this method.

Results: The median age at diagnosis was 43 years. 59% were male, 41% female. The four most common sites of pathological fracture were the femur (35%), humerus (28%), tibia (12%) and pelvis (11%).

The five most common histopathological diagnoses were chondrosarcoma (9%), osteosarcoma (9%), meta-static renal carcinoma (8%), giant-cell tumour (6%), lymphoma (5%).

77% of biopsies yielded a tissue diagnosis. The remaining 23% underwent open biopsy, repeat image-guided needle biopsy or were not further investigated.

In the 30 cases (23%) of non-diagnostic biopsies 80% of these lesions had no extra-osseous component to them and the remaining 20% had a very small extra-osseous component.

Discussion: A tissue diagnosis of a pathological fracture can be obtained by primary image-guided needle biopsy in 77% of cases referred to a specialist bone tumour service. The majority (80%) of unsuccessful biopsies were of lesions with little or no extra-osseous component to the lesion.

The aim of this study was to compare the accuracy of image guided (ultrasound or CT) percutaneous needle biopsy to percutaneous needle biopsy without image guidance in diagnosis of soft tissue tumours.

Eighty-eight consecutive patients with soft tissue lesion who were referred to the soft tissue tumour unit underwent percutaneous needle biopsies of their lesion either with image guidance or without. Sixty-one out of these 88 patients subsequently underwent excision of their lesion and the excised specimen was then subjected to histological examination. The accuracy of image guided percutaneous needle biopsy and percutaneous needle biopsy without image was then calculated by comparing the histological results of the needle biopsy to that of excision biopsy.

The diagnosis accuracy of image guided percutaneous needle biopsy was 92% (34 out 37) compared to 79% (22 out of 28) for percutaneous needle biopsy without image. In 3 out of the 28 patients who had percutaneous needle biopsy without image guidance, there was insufficient material obtained from the needle biopsy to allow a histological diagnosis. This was not the case with any of the patients who had image guided percutaneous needle biopsy.

Conclusion: Using image guidance, either USS or CT scan, improves the diagnostic accuracy of percutaneous needle biopsy and should be the gold standard technique in management of soft tissue tumours. However, if the lesion is palpable and not mobile, the accuracy of percutaneous needle biopsy without image guidance can be up to 79%.

Background: It has been reported that there is poor correlation between neurological injury and degree of bony retropulsion in thoracolumbar burst fractures¹. Wilcox et al² showed biomechanically that there was poor concordance between the extent of post impact spinal canal occlusion and the maximum amount of occlusion that occurred at the moment of impact. In the current study we examined the possibility that variation in the termination of the conus medullaris may offer protection from neurological injury in a proportion of these fractures.

Methods: A retrospective study was made of 39 patients (26M:13M, mean age 35.9 years, range 15 – 75 years) presenting with a single level thoracolumbar burst fracture (T12–L2) between 1998 and 2001. A whole spine MRI scan was performed on all patients and the level of the conus noted. Age, sex, injury severity score (ISS), neurological status (ASIA motor score) and the transverse spinal canal area (TSCA) of the vertebral levels either side of the fractured vertebra was measured. A predicted TSCA for the injured level was then calculated from the mean of the TSCA’s of the adjacent levels. The actual TSCA of the injured level was calculated and this enabled a percentage decrease of the TSCA to be worked out from the predicted value. Analysis was made of the presence or absence of neurological injury in relation to canal compromise and involvement of the conus.

Results: Eighteen patients with neurological compromise and 21 with intact neurology (the age and sex distribution in the two groups were similar). The mean ± SD ASIA motor score of the patients studied was 90.4 ± 23. Mean ISS was 20.2 in the neurologically injured and 10.5 in the intact (p=0.0005). Mean TSCA of the canal was 218mm² in the intact and 150mm² in the injured groups (p=0.006) and mean %TSCA was 70 and 49 respectively (p=0.007). The conus lay between T12 and L2 in all patients. When the conus lay cranial to the fracture (n=13), 38% were neurologically intact. When the conus lay at the level of the fracture (n=26), 62% were intact (NS). Neurological deficit did not occur in the absence of neurological compression on MRI.

Conclusion: Our study showed that the risk of neurological injury from a thoracolumbar burst fracture was not decreased when the conus lay outside the fracture zone. However, there was a statistically significant difference in percentage of canal compromise when the patients with neurological impairment were compared with those that were neurologically intact.

Background: The incidence of neurological symptoms due to spinal stenosis in patients with achondroplasia is reported to be as great as 38%¹. These symptoms most commonly occur in the 4^th decade and myelography and CT myelography are most commonly described in evaluation of the stenosis. Difficulty arises in localisation of stenosis in patients presenting with neurological deficit². The value of MRI of the cervicomedullary junction has been reported in achondroplasia but it has not yet been evaluated in the investigation of spinal stenotic symptoms. The aim of this study was to review our experience of whole spine imaging in patients with achondroplasia that presented with symptoms and signs of neurological deficit.

Methods: We retrospectively reviewed the clinical notes and radiological imaging of 10 consecutive achondroplastic patients (3F:7M, mean age 31.7 years, range 13 to 60yrs) that presented to our unit with neurological compromise between 1998 and 2003. All patients had whole spine MRI at the time of presentation. Recorded from the notes were age and sex, and whether symptom pattern was radiculopathy, claudication or paresis. All radiological levels of stenosis on MRI were documented.

Results: Four patients presented with spinal paresis, four with neurogenic claudication, and two with radiculopathy. MRI confirmed that each patient had at least one region (cervical, thoracic or lumbar) of significant spinal stenosis. In six of the patients an additional region of significant stenosis was identified. All ten patients had lumbar stenosis but this was only the primary site in six of the ten. In the other four patients two had the dominant stenosis in the thoracic spine, one in the cervical spine and one at the foramen magnum – the clinical symptoms correlated with the dominant site in each of these four cases.

Conclusion: MRI was a useful tool for assessment of neurological compromise in the patients with achondroplasia in our study. All ten patients had classical lumbar stenosis on MRI but this was only the dominant site of stenosis in six of the ten cases. The MRI and clinical findings need to be evaluated together to ensure correct surgical treatment.

Background: There is conflicting evidence of correlation between clinical outcome and severity of thoracolumbar spine fractures in neurologically intact patients1. Kalyan et al2 presented the results of their prospective study of thoracolumbar spinal fractures and concluded that the clinical outcome was consistently better predicted by the severity of disc injury than that of the bony fracture. They suggested that if severe disc injury was present, treatment of the disc injury may result in a better short term clinical outcome. The aim of this study was to detail the incidence and type of disc injury in patients with thoracolumbar spinal fractures with intact neurology.

Methods: Retrospective analysis of a prospectively collected spinal injury database at a regional spinal injuries unit. Only patients with a thoracolumbar spinal fracture and intact neurology were included. Retrospective analysis of magnetic resonance imaging (MRI) findings. One hundred and thirty nine neurologically intact patients (89M:50F, mean age 36 years, range 15 – 77yrs) with a thoracolumbar (T11 –L5) spinal fracture were admitted to our spinal unit over the last 11 years (1994 – 2004). Patient data was collected prospectively onto a spinal injuries database. All of these patients had an MRI scan on admission. All types of thoracolumbar fracture were included in this study and the presence or absence of an associated disc injury was recorded retrospectively from the MRI study. The type of disc injury was also recorded.

Results: The incidence of severe disc injury adjacent to a thoracolumbar spinal fracture in our cohort of patients was 43% (60 patients). Disc prolapse or extradural herniation was seen in 16 patients (11%). Intraosseous disc herniation into an adjacent vertebral body was seen in 20 patients (14%). Internal disc disruption was seen in 24 patients (17%). Disc injury was found at more than one level in 21 patients (15%).

Conclusion: The incidence of severe disc injury in our study of neurologically intact patients with a thoracolumbar fracture was considerable (43%). Kalyan et al² suggested that treatment directed at addressing the disc injury in these patients may promote earlier pain relief and also earlier return to pre-morbid activities. If this is the case, then the decision making regarding operative versus non-operative management, in a patient with a thoracolumbar fracture and intact neurology, should be based on the severity of the disc injury as well as the bony injury.

Background: It is very difficult to ascertain how much of the degeneration seen in a post-traumatic spine was due to pre-existing disease and how much was due to the trauma. The aim of the current study was to determine the prevalence of pre-existing disc degeneration on MRI in a region of the spine injured by spinal trauma.

Methods: Prospective whole spine MRI study in 118 patients with spinal trauma. One hundred and eighteen consecutive patients (79M:39F, mean age 36years, range 13 – 90 yrs) admitted to our spinal unit for treatment of their acute spinal trauma were investigated with whole spine magnetic resonance imaging (MRI). Whole spine MRI was already the standard protocol for all patients admitted with spinal trauma to rule out co-existing pathology and multiple spinal fractures¹. Patient data, including age, sex, and mechanism of injury was recorded prospectively onto a spinal injury database. We divided the spine into cervical (C1–C7), thoracic (T1–T12), and lumbosacral (L1–S1) regions. For the purpose of this study we documented the presence or absence of pre-existing degenerative disc disease in all regions of the spine. Of particular interest was any pre-existing degenerative disc disease in the region affected by the spinal trauma. The number of spinal levels affected by pre-existing disc disease within each region was also detailed.

Results: The two most common mechanisms of injury were flexion-compression (68 patients) and flexion-distraction (24 patients). Seventy-one of the patients sustained burst fractures and the vast majority of these were in the thoracolumbar region. Thirty-seven patients (31%) had degenerative disc disease in the same region of the spine that was injured in the spinal trauma. Seven patients had pre-existing cervical degeneration in the presence of cervical spine trauma and thirty had pre-existing lumbosacral degeneration in patients that had lumbosacral trauma. Twenty four patients (20%) had more than one level of degenerative change within the same region as their spinal trauma. Eighteen patients (15%) had degeneration in a different region of the spine to the one that was injured.

Conclusion: Thirty one percent of the patients in our study had pre-existing degenerative disc disease in the same region as their spinal trauma despite the average age of our patients being only 36yrs. This has important medicolegal implications because it means that a large % of patients who sustain spinal trauma have pre-existing degenerative changes which are not the result of their injury.

Introduction: The poor correlation between neurological injury and degree of retropulsion in thoracolumbar burst fractures has been identified, but not adequately explained. We have examined the possibility that variation in the termination of the conus medullaris may offer protection from neurological injury in a proportion of these fractures.

Methods: A retrospective study was made of 39 patients presenting with single level thoracolumbar burst fractures between June 1998 and April 2001. Admission MRI was performed on all patients. Age, sex, ISS, neurological status, mode of treatment and any neurological recovery were recorded. From the MRI scans the levels of the conus and the fracture were noted. Transverse Spine Area(TSA) was measured at the cranial, caudal and injured levels. A predicted TSA and % TSA for the injury level was calculated from the mean of the two other levels. Analysis was of severity of neurological injury in relation to canal compromise and involvement of the conus.

Results: 26 male and 13 female patients of mean age 35.9 (SD 17) years and mean ASIA motor score 90.4 (SD 23) were studied. Neither sex nor age distribution differed between 18 neurologically injured and 21 intact patients. Mean ISS was 20.2 in the neurologically injured and 10.5 in the intact (p=0.0005). Mean TSA of the canal was 218mm2 in the intact and 150mm2 in the injured groups (p=0.006) and mean %TSA was 70 and 49 respectively (p=0.007). The conus lay between T12 and L2 in all. When the conus lay cranial to the fracture (n=13), 38% were neurologically intact. When the conus lay at the level of the fracture (n=26), 62% were intact (NS). Neurological deficit did not occur in the absence of neurological compression on MRI.

Conclusions: Neurological injury is not less likely when the conus lie outside the fracture zone. Canal compromise is a highly significant factor in neurological injury.

We compared the accuracy of image guided (ultrasound or CT) percutaneous core needle biopsy to percutaneous core needle biopsy without image guidance in diagnosis of soft tissue tumours. 140 patients with soft tissue lesion who were referred to a London bone and soft tissue tumour unit underwent percutaneous core needle biopsies of their lesion either with or without image guidance.111 of these 140 patients subsequently had surgical excision. The accuracy of image guided percutaneous biopsy and percutaneous biopsy without image was then calculated by comparing the histological results of the needle biopsy to that of the resection.

The diagnosis accuracy of unguided biopsy was 78% (36 out of 46) compared to 95% (62 out of 65) in image guided. In 6 out of the 46 patients who had unguided biopsy, there was insufficient material obtained from the needle biopsy to allow histological diagnosis. This was not the case with any of the patients who had image guided core needle biopsy.

Using image guidance, either USS or CT scan, improves the diagnostic accuracy of percutaneous core needle biopsy and must be considered in management of patients with soft tissue tumours.

Objective: Axially loaded MRI simulates imaging of the lumbar spine in the standing position and is useful in the assessment of spinal stenosis[1]. This study determines the ability of axially loaded spinal MRI to assess Cobb angle in patients with idiopathic scoliosis.

Design: Prospective study. Newly diagnosed patients with idiopathic scoliosis were referred for MRI of the whole spine. Cobb angle measurements were made from erect AP spinal radiographs prior to MRI. Coronal MR images of the thoracic and/or lumbar spine were obtained prior to and following loading of the spine in an MR compatible compression device (Dynawell). Cobb angle measurements were made on unloaded and loaded MRI studies using the same reference points as on radiographs. Radiographic and MRI Cobb angle measurements were compared. Informed consent was obtained from all patients and the study was approved by the local Ethics Committee.

Subjects: Five patients, all females with mean age 14 years (range 12–16 years) were included in the study. Outcome Measures: Six curves were compared on pre-referral erect radiographs, unloaded and loaded MRI studies, 2 in the thoracic region and 4 in the thoracolumbar region.

Results: Curve characteristics and Cobb angle measurement on radiographs vs. axial unloaded and loaded MRI were as follows: Curve 1; T4-T12, 45°, 36° and 41°. Curve 2; T10-L4, 52°, 22° and 30°. Curve 3; T10-L4, 45°, 36° and 38°. Curve 4; T6-T10, 42°, 22° and 22°. Curve 5; T11-L3, 43°, 32° and 43°. Curve 6; T11-L3, 34°, 11° and 31°

Conclusions: Axial loading increases MRI Cobb angle measurements compared to unloaded studies. Initial results suggest that axial loaded MRI using the Dynawell Compression device may allow comparative measurement of Cobb angle to erect radiographs in the thoracolumbar region, but not in the thoracic region. This is likely related to the loading characteristics of the compression device, which is designed to concentrate loading in the lumbar region. Modification to include loading of the thoracic spine may improve results. The technique has the potential to replace radiography and thus reduce radiation burden to young adolescents with some types of idiopathic scoliosis.

Aim: To test the null hypothesis that plain X-rays can provide the same assessment of sacral screw placement as CT.

Introduction: Engaging the anterior cortex of the sacrum provides additional strength to fixation and is a goal of surgery. The sacrum with its unique anatomy makes it a difficult bone to assess screw placement radiologically. This study examines the positioning of sacral screws as seen on X-rays and compares the result with spiral CT “gold standard”.

Materials and methods: Inclusion criteria: Sacral fixation using Diapason (Stryker) Titanium pedicle screws by one surgeon. Spiral CT, plain AP and lateral X-rays of the sacrum. Exclusion criteria: X-rays with more than three level fixation.

There were 66 patients (132 S1 screws). Surgical technique engaged the anterior cortex to enhance fixation. Two independent observers (a musculoskeletal radiologist and spinal fellow) who were blinded to outcome, reported findings in forms with constrained fields. Assessment of plain X-ray and CT was at separate times not less than three weeks apart. Variables noted: Screw position in pedicle, screw tip position, and angle of screw (sagittal on axial CT scans).

AP X-ray was divided, for each screw, into nine zones based on the first sacral foramina. The position of the screw tip in the zones was noted. The lateral X-ray was divided into three zones to note the tip of the screw in relation to the cortex. The extent of screw protrusion was measured. X-ray technique: Supine AP centred on fusion and lateral X-ray standing, X-ray source 200 cm from the film. CT: Images acquired on Picker PQ 6000 spiral CT with collimated thickness of 3 mm, pitch 1.25 and reconstructive index of 1.Para-sagittal and coronal reconstructions. Spiral CT was used to note the position of the screw within the pedicle and the relation of the screw tip to the anterior cortex. For screws within the pelvis any structure in close proximity was noted.

Results: On CT 10% of the screws had breached the pedicle compared with 2% on the plain X-rays. Anterior cortical perforation had been achieved in 48 out of 132 screws on CT. The sensitivity of the plain X-rays to perforation was 40% with a specificity of 92%. There was an average under estimation of the extent of screw perforation by 4.4 mm (95% confidence ±1 mm). There was a correlation between the position of the screw tip on the AP X-ray and the sensitivity of the lateral X-ray to detect a perforation. The sensitivity ranged from 52% for zone 1 to 15% in zone 8. 15/31 perforations were missed in zone 1, compared with 11/13 in zone 8. For screws penetrating 5 mm or more, in zone 8, 9 out of 10 were missed on lateral X-rays.

Eighty-five screws were placed at an angle of less than or equal to 25° to the sagittal; this included 28 out of 34 screws placed in zone 8. The inter-observer variance of screw angle measurement was 1.1° and intra-observer difference 1.7°. Overall 95% confidence of a single measurement was ±3.3°.

Conclusion: Plain X-rays and CT do not provide the same assessment of sacral screw placement. This is particularly true for sagitally placed screws with screw tips in zones 7–8.

Objective: To document the incidence of adjacent superior segment facet joint violation following transpedicular instrumentation in the lumbar spine as it has been postulated that this can lead to long term deterioration There has been no study so far determining this incidence.

Design: Patients undergoing lumbar fusion were prospectively evaluated with a CT scan and plain radiographs six months following surgery. These were blindly and independantly evaluated by a consultant radiologist and a research fellow.

Subjects: 106 patients (212 top level facet screws) between 1996 and 1999 were evaluated. All patients had their screws and instrumentation inserted through a Wiltse muscle splitting approach and a lateral entry point in the pedicle so as to reduce the risk of facetal impingement. ¹

Outcome Measures: Kappa co-efficient and chi-squared analysis.

Results: The Kappa co-efficient for the CT scan and plain radiographs were 0.88 and 0.39 respectively. On the CT scan both observers noted facet joint impingement in just over 20% of the screws and just over 30% of the patients. The impingement was independent of the level and diagnosis (p> 0.05) and it occurred with uniform incidence in each of the year.

Conclusion: This study raises the theoretical possibility of long term deterioration in the clinical results following the use of transpedicular instrumentaion.

Objectives: To determine if high quality, thin slice (1–3)mm CT scan images would allow proper evaluation of interbody fusion through titanium cages in view of the fact that there are no universally accepted radiological criteria. ¹

Design: Patients undergoing interbody lumbar fusion were prospectively evaluated with a CT scan and plain radiographs six months following surgery. These were blindly and independently evaluated by a consultant radiologist and a research fellow. They were assessed for bridging bony trabeculation both through and surrounding the cages as well as for changes at the cage endplate interface.

Subjects: Fifty-three patients (156 cages) undergoing posterior lumbar interbody fusion using titanium inter-body cages were evaluated. Posterior elements were used to pack the cages and no graft was packed outside the cages.

Outcome Measures: Kappa co-efficient and chi-squared analysis.

Results: On the CT scan both observers noted bridging trabeculation in 95%of the cages-Kappa 0.85, while on radiographs they were present in only 4%-Kappa 0.74. Both observers also identified bridging trabeculation surrounding the cages on the CT scans in 90%-Kappa 0.82, while on the radiographs this was 8%-Kappa 0.86. Radiographs also did not identify all the loose cages.

Conclusions: High quality CT scan images can demonstrate bridging bony trabeculation following the use of titanium interbody cages. It also demonstrated consistent bone growth outside the cages inspite of not using any bone graft.