Opportunistic Diagnosis of Osteoporotic Vertebral Fractures on Imaging Studies performed for Alternative Clinical Indications

Introduction

In an era of increasing life expectancy, osteoporosis has become a major global health concern [1,2]. Osteoporosis is a skeletal disorder characterised by compromised bone strength which predisposes to increased fracture risk [2]. At least one third of all post-menopausal women, and one fifth of men older than 50 will suffer an osteoporotic fracture in his/her lifetime [3-5]. The National Osteoporosis Foundation (NOF) estimates that approximately 54 million Americans suffer from osteoporosis resulting in 2 million fractures annually [6]. Population-based studies have demonstrated an increasing prevalence of osteoporotic fractures resulting in hospitalisation, increased morbidity and mortality and placing increasing burden on healthcare systems [7-9]. Vertebral fractures (VF) account for up to 50% of osteoporotic fractures making them the most common fracture subtype [10]. The incidence of vertebral fractures increases with age [10,11]. Up to 26% of Scandinavian women are diagnosed with at least one VF in their lifetime [11]. VFs are a major cause of pain and reduced mobility. Many patients who have sustained a VF suffer with the psychological fear of isolation and loss of independence [12,13]. Additionally, sustaining a VF is an independent risk factor for mortality [14]. Studies show that patients with previous VFs are five times more likely to obtain an additional VF and are twice as likely to suffer a hip fracture with resulting morbidity and mortality [15,16]. Encouragingly, evidence has shown that early intervention with pharmacological agents such as bisphosphonates result in a relative risk reduction of up to 0.6 for vertebral fractures and up to 0.8 for non-vertebral fractures [17]. Therefore, it is vital that VFs are correctly diagnosed so that patients are investigated and treated appropriately. However, there is a discrepancy between best recommended management and real-life clinical practice studies concluding that many patients diagnosed with an osteoporotic fracture are never appropriately investigated or treated for osteoporosis [18-20].

Many imaging studies performed for alternative clinical indications fortuitously include the spine. Radiologists do not always systematically review the spinal vertebra when they are not the specific clinical area of concern [21,22]. This can lead to a missed opportunity to detect vertebral fractures and diagnose osteoporosis [21,22]. VFs are evident on various imaging modalities that are performed for alternative clinical indications but are frequently not reported by radiologists [23,24]. Use of terminology such as ‘wedging’, ‘endplate compression’ and ‘endplate concavity’ in radiology reports can be confusing and may not be clearly understood as vertebral fracture or implication of underlying osteoporosis by the ordering physician. Non-diagnosis or inappropriate reporting of VFs in this way is a missed opportunity to diagnose osteoporosis, to provide appropriate treatment and to reduce patients risk of further osteoporotic fractures [18]. In this paper, we discuss the radiological assessment of VFs and describe how fractures can be diagnosed on the most used imaging modalities including plain film, MRI, CT and bone scans (Figure 1A- 1B).

Figure 1A: Lateral lumbar spine radiograph of an 80-yearold female patient. The radiograph demonstrates several insufficiency compression fractures; severe anterior wedge fracture at T12, mild compression fracture of the L1 and L4 superior endplates and moderate compression fracture at L2.

Figure 1B: Lateral thoracic spine radiograph demonstrates a moderate compression fracture at T7 with secondary kyphosis.

Assessment of Fractures

Genant et al. devised the Semi-Quantitative (SQ) method for describing vertebral fractures [25]. This method has high inter- and intra-observer agreement, even amongst inexperienced reviewers [25]. The method is widely reproducible and is often used in research settings and clinical trials. The SQ method is a relatively straight-forward method to grade fractures and avoids otherwise confusing language which may be misinterpreted. First described on lateral radiographs, the SQ method employs visual inspection to grade vertebral fractures. Grade 0 is normal without loss of vertebral body height. Grade 0.5 are borderline vertebral fractures. Grade 1 fractures show mild deformity with approximately 20 % to 25 % loss of height and 10 % to 20 % reduction in area. Grade 2 fractures are moderately deformed with 25 % to 40 % loss of height and 20 % to 40 % loss of area. Grade 3 vertebral fractures have lost 40 % or more of their height and area. The SQ method is not without its limitations. Employing this method may inadvertently overdiagnoses VFs in patients with congenital or acquired vertebral anomalies [26]. Additionally, employing the SQ method alone would fail to diagnose minor endplate fractures which do not result in loss of vertebral body height. In response, Jiang et al devised the algorithm-based qualitative (ABQ) approach which focuses on vertebral endplate deformities [27]. Using this method, an experienced radiologist needs to assess various aspects of endplate abnormality before diagnosing it as a fracture. Jiang et al showed that using a stringent criteria-based algorithm in this way, the ABQ method is likely to diagnose only one third of fractures that would be diagnosed by the SQ method alone. Similarly, Black et al showed that the SQ method diagnosed three times the number of mild vertebral fractures compared to other quantitative methods [28].

Recognition of Fractures

Imaging Modalities:

A. Plain Films: For clinically suspected VFs, plain films including antero-posterior (AP) and lateral projections are usually the first line of investigation. The lateral film is particularly useful (Fig. 1A and 1B). The radiologist should carefully examine the vertebral body outline, especially the superior and inferior endplates to ensure VFs are not missed. The pedicles are examined for symmetry on the AP film. Subjectively identifying reduced bone density heightens the index of suspicion for VFs as these patients are at much greater risk. Dynamic radiographs of the vertebrae can increase the likelihood of correct diagnosis on plain radiography. This method allows the radiologist to compare supine images with lateral sitting radiographs to evaluate for changes in vertebral body height. The sensitivity and specificity of dynamic radiographs for diagnosing acute VFs is 66% and 96% respectively [29]. While moderate and severe VFs are rarely misdiagnosed, there are several conditions which can be mistaken for mild VFs leading to overdiagnosis. These include developmental short vertebral height, physiological wedging, Scheuermann’s disease, degenerative scoliosis, Schmorl’s nodes and Cupid’s bow deformity (smooth developmental curvature of the inferior endplate of lumbar vertebra) [30]. Possible reasons for underdiagnosis of VFs by non-musculoskeletal radiologists include focusing on other acute imaging findings, lack of specialist knowledge about osteoporosis/ osteoporotic VFs or simply ignoring osteoporotic VFs completely [31].
Vertebrae are included on many plain films when there is no clinical suspicion of VF. Examples include abdominal radiographs for patients with abdominal pain or chest radiographs in patients with cardio-respiratory symptoms. Less commonly, the vertebrae are incidentally imaged during barium investigations, interventional, cardiac, and fluoroscopic procedures. Even if not performed to out rule a VF, each imaged vertebra should be carefully evaluated to ensure no underlying occult VF. Despite the obvious opportunity to diagnose VFs in this way, there is a paucity of published literature in the area. The most studied radiographic technique to incidentally diagnose VFs is the chest radiograph. In a large study of over 10,000 post-menopausal women who underwent a lateral chest x-ray, 41% of radiologists who identified a VF failed to document it in the report summary, and only 36% were put on treatment for osteoporosis on discharge [32]. In a smaller retrospective review of chest x-rays of post-menopausal women, Gerlach showed that 14.1 % had a moderate or severe VF visible on chest radiograph [21]. Unfortunately, less than one quarter of visible VFs were referenced in the radiologists’ summary and only one seventh of these patients received a discharge diagnosis of VF. As a result, only 18% of patients were discharged with appropriate medical therapy for underlying osteoporosis. The lateral chest radiograph on elderly patients is an opportunity to incidentally diagnose VFs by assessing vertebral bodies and clearly reporting them in the final summary [33]. Despite their importance in the initial investigation for suspected VF, many patients with VFs will have no morphological change on plain films. It is important not to dismiss patient symptoms based on normal radiographs since many patients with normal plain films may only have acute changes detectable on MRI [34]. Loss of vertebral height may not be evident at time of acute symptoms but can be evident on subsequent follow-up radiograph.
MRI: MRI is a time-intensive imaging modality with relative contraindications such as claustrophobia, presence of a nonconditional pacemaker and first trimester of pregnancy. MRI has a sensitivity of 100% in detecting spinal trauma and is an excellent method to diagnose and assess VFs [34]. MRI has a sensitivity and specificity of up to 82% and 98% respectively for distinguishing osteoporotic VFs from other types of fracture [35], (Figures 2A-2C). In addition to identifying a VF, MRI may also diagnose other uncommon causes for back pain such as infection or malignancy, and allows assessment of spinal ligaments, spinal cord, surrounding CSF and meninges. The Short Tau Inversion Recovery (STIR) sequence is particularly sensitive to acute fractures as it nullifies marrow fat signal over a large body area such as the entire vertebral column allowing increased visibility of acute pathology such as fracture. STIR sequences in combination with T1 weighted sequences are helpful to differentiate benign osteoporotic VFs from those caused by malignancy [36]. The presence of marrow oedema recognised as high signal on fluid sensitive STIR or T2-weighted fat saturated sequences indicates recent fracture. Marrow oedema is absent in a chronic vertebral fracture. Benign vertebral fractures typically are seen as linear low T1 signal. Malignancy or infection in contrast cause diffuse nonlinear replacement of the normal marrow of the vertebra. For every MRI study performed, initial localizer sequences are utilised by radiographers to plan image acquisition. These localizers are obtained from thick slices and are not suitable for diagnostic detail but do represent an opportunity to diagnose a VF when not suspected. Strong inter-observer agreement has been reported in detecting VFs in the thoracic and lumbar spine on localizer images [37]. In another study, musculoskeletal radiologists examined 856 localizers of patients undergoing breast MRI. The authors concluded that 8.9 % of patients had a VF visible on the MRI localizer, but none were documented in the final report [38]. MRI localizers are a quick and reliable method of diagnosing vertebral fractures when not suspected and may negate the necessity for further imaging or using ionising radiation.

Figure 2: MRI Lumbar Spine with T1, T2 & STIR sequences of an acute mild compression fracture at T10 in a 67-year-old female patient.

Computed Tomography (CT): CT uses high doses of ionising radiation to acquire images. CT imaging is available 24/7 in most tertiary hospitals and offers almost instant acquisition of images. CT has excellent sensitivity and specificity for identifying VFs;100% and 97% respectively [39]. CT of the spine may be requested when a VF is clinically suspected and when the radiograph is normal. Of note, a non-displaced vertebral fracture on a background of osteopenia, may not be evident on CT [39]. In patients with known VF, CT can help to provide additional information such as stability of the fracture and protrusion of bone fragments into the spinal canal. CT can also aid with clinical decisions such as patient suitability for surgical intervention or vertebroplasty. The majority of CTs are performed for clinical indications not specifically related to identification of VFs including Cardiac CT, CTPA and CT thorax to evaluate for thoracic pathology and CT KUB, CT abdomen/ pelvis, CT colonography and CT peripheral angiograms/venograms performed to identify intra-abdominal pathology. Vertebral morphology, particularly on sagittal reformats is well visualised on these CT studies. Modern CT scanners can display vertebrae in the region imaged in excellent bony detail in coronal, sagittal and axial reformats without the requirement for further imaging or radiation exposure to the patient. Of these, the sagittal reconstructions are particularly important to diagnose VFs (Figure 3) [40]. Despite the ability to utilize CT to diagnose occult VFs, CT is often not effectively exploited in this way. A New Zealand study retrospectively reviewed sagittal reconstructions of CT abdomen or thorax in patients over 65 years. 22 of 175 patients had a VF visible on sagittal reconstruction, and 77% of these had previously undiagnosed VF. The authors concluded that reviewing reformatted CT of the abdomen and pelvis improved diagnosis of VFs but are frequently not reported - thereby missing an opportunity to diagnose osteoporosis, treat with appropriate medical therapy and to reduce risk of future osteoporotic fractures and associated morbidity and mortality [41]. Similar to localizers in MRI, CT scout views are obtained prior to final image acquisition. These use low levels of radiation to acquire 2-Dimensional images which are used to plan the final CT image. Lateral CT scout views may show fractures not visible on axial CT images. One study of 300 CT scans involving the thoracic and lumbar spines demonstrated the sensitivity and specificity of diagnosing VFs on scout views to be 98.7% and 99.7% respectively. The authors concluded that scout views should be used to evaluate for VFs on CTs performed for other clinical indications [42].

Figure 3: Sagittal reformatted CT of the Lumbar Spine in an 83-year-old female demonstrating severe compression fracture at L1, moderate compression fracture of T11 and mild compression fracture of L2.

Skeletal Scintigraphy (Bone Scans)

Tc 99m is a radioisotope which can be bound to MDP and injected intravenously. The radioisotope travels through the patient’s bloodstream and binds to remodelling bone. Three hours after injection, patients are placed on a gamma camera which identifies bony hotspots where Tc99m has accumulated. 80% of VFs are visible as hotspots, usually linear in morphology, at 24 hours following injury and almost all return to normal within two years [43]. The major limitation of bone scans is their poor specificity. The most common indication for performing bone scans is to identify osseous metastatic disease in patients with known primary malignancy. However, bone scans are also utilized to identify occult fractures or osteomyelitis. Due to their non-specific nature, hotspots can also be caused by degenerative changes. For this reason, bone scans are often reported in conjunction with other available imaging such as MRI, CT, or plain films (Figures 4 & 5).

Figure 4: Bone scan for completion of staging in a 67-year-old female with non-small cell lung cancer. There are non-specific foci of increased radioisotope uptake in the mid thoracic spine. Comparison was then made to previous staging CT thorax (Fig. 5)

Figure 5: Review of the staging CT thorax confirmed the areas of uptake on bone scan in Fig 4 correlating to previous moderate wedge compression sclerotic vertebral fractures at T6 and T7 secondary to metastatic disease.

Discussion

Osteoporosis is an increasing public health concern and predisposes patients to VFs. Prompt diagnosis and early intervention with appropriate medical treatment is imperative. The literature shows that incidental VFs, on imaging studies performed for alternative clinical indications are underdiagnosed thereby missing an opportunity to diagnose vertebral fracture, diagnose osteoporosis if not previously diagnosed and treat patient appropriately. Untreated and undiagnosed VFs can significantly impact on a patient’s quality of life and life expectancy. Patients with osteoporotic fractures can endure intolerable pain, loss of independence and suffer psychologically due to fear of isolation. Many patients require polypharmacy for pain control, and all are at high risk of future osteoporotic fractures. The mid-thoracic region and thoraco-lumbar junction are the most frequently affected areas and may result in spinal kyphotic deformities. Kyphosis predisposes to loss of balance, muscle wasting, further degenerative changes at adjacent intervertebral joints, restrictive lung disease, inability to work and loss of earnings [44]. Fortuitously many imaging studies including plain radiography, CT, MRI and Bon scans include the thoracic and lumbar spine in the area of imaging. This provides an opportunity to diagnose unsuspected abnormalities of the spine when these imaging studies are performed for alternate clinical indications. Many radiologists however do not systematically review the vertebra in these studies and miss the opportunity to identify abnormalities such as vertebral fractures and osteoporosis. When vertebral morphological abnormalities are identified equivocal language such as ‘loss of height’ or ‘wedging’ to describe VFs can be misleading. This terminology is ambiguous for referring physicians who may not appreciate that these are vertebral fractures and imply underlying osteoporosis. There is no agreed gold standard for diagnosing VFs on imaging. As a result, many VFs are both under and over-diagnosed. One strategy is the semi-quantitative method for grading fractures. Even amongst inexperienced observers, the SQ method demonstrates high levels of agreement [21]. Alternatively, the ABQ method forces the radiologist to answer a number of questions before diagnosing a VF and is arguably more accurate [27]. Whichever method is employed, it remains imperative the reading radiologist clearly states the existence of a VF in the report summary to improve the proportion of patients discharged on appropriate medical therapy. A number of imaging techniques performed for various clinical indications may show VFs in the area imaged. There is under reporting of VFs which are clearly visible on lateral chest radiographs, MRI localizers and CT scout views. Unless sagittal reformats of CT studies are routinely performed often VFs are not visible on standard axial images even to experienced musculoskeletal radiologists. The term ‘inattentional blindness’ refers to an inability to notice unexpected events when immersed in an alternative task. In one experiment, 83% of expert radiologists failed to recognise a gorilla drawn onto a stack of CT images when they were focusing on finding pulmonary nodules [45]. Another phenomenon coined “satisfaction of search” refers to a relative difficulty in identifying further pathological findings following identification of another significant abnormality [46]. These factors are relevant to radiologists when searching for clinically significant pathology not related to the spine on x-ray, MRI, or CT and thus VFs can easily be overlooked. Dedicated education programmes delivered to radiologists and internal medical physicians may help to improve the diagnosis and management of VFs. In one study, recognition of VFs amongst internists almost doubled from 22% to 43% following provision of basic lectures, posters and flyers. The same study demonstrated a significant increase in patients discharged on osteoporosis treatment from 11% to 40% [47]. In another study, there was a marked improvement in the ability of a radiology resident to correctly identify VFs after undergoing specific teaching [48].

Conclusion

In conclusion, VFs are a major health concern in an era of aging population. Many factors have contributed to underdiagnosis and treatment of VFs. When identified by a radiologist ambiguous terminology should be avoided and the SQ method employed. The spine is included in many imaging studies performed for alternative clinical indications. This is a fortuitous opportunity to assess the spinal vertebrae and diagnose fractures when present. Irrespective of the clinical indication or imaging modality, a high index of suspicion for VFs should be always employed. Basic education programmes for radiologists and internists would serve to improve the diagnosis of VFs and treatment of osteoporosis.

For more Articles on : https://biomedres01.blogspot.com/