Evaluation of osteoporosis is critical in the management of spinal disorders. It not only enables the spine surgeons to distinguish between normal and low BMD and treat osteoporosis effectively, but has a significant role in optimizing perioperative planning. The use of modified pedicle screws, multiple anchor-points, augmentation techniques to enhance the stability of screw-rod constructs and thus achieve higher fusion rates goes hand-in-hand with effective anti-osteoporotic treatment to achieve better surgical outcomes. Precise BMD estimation and increased detection rates help in optimizing care across the population as a whole. DXA is generally considered the gold standard for BMD evaluation as was established by the WHO in 1994. However, in a position paper issued in the year 2000, the WHO committee suggested that the T-scores of hip/femoral neck DXA BMD may be used as gold standard to make a diagnosis of osteoporosis. The committee also stated that irrespective of the femoral neck T-score, an individual with a low BMD at spine or a fragility fracture may be considered osteoporotic. This is due to the discordance in bone density values across multiple sites found in the multisite DXA of an individual which is also evident in this study as depicted in Figure 5A, B, body fat influencing the BMD values of DXA, degenerative changes causing overestimation of spine BMD leading to a notion that spine DXA BMD can be ignored if low BMD is found at other sites.[10–15] This can cause confusion and misinterpretation of DXA T-scores by both the clinicians as well as the patients. Thus, there is a need for development of a protocol to delineate spinal osteoporosis from multisite bone density evaluation by appropriate use of the available densitometry technology on the patient specific basis,[13,14] rather than generalization into routine screening for osteoporosis. Post-menopausal BMD evaluation for better fracture predictability for patients with fractures, monitoring disease progress, or response to therapy should be differentiated from the routine screening for osteoporosis.
Discordance of DXA bone mineral density values in multiple sites across the body. A, DXA report showing osteopenia at spine, osteoporosis at bilateral neck of femur. B, DXA report showing a normal bone density at spine but low BMD at other sites. BMD indicates bone mineral density; Q-CT, quantitative computed tomography.
DXA reflects lesser fracture predictability since it measures the integral BMD unlike Q-CT which calculates the trabecular BMD.[14,16–18] Again, DXA provides an area measure of bone density which is generally less than the volumetric measure given by Q-CT. Adding to this, there were reports suggesting DXA can miss 26% to 60% of osteoporotics and 11% to 18% of the clinically osteoporotic individuals were shown to be normal. In the current study, Q-CT revealed higher detection rates than DXA in a matching study sample. Q-CT engages low dose CT protocols and can offer quick and safe bone density estimation with avoidance of confounds from osteophytes and other degenerative changes of spine, aortic calcification etc. Although there can be a speculation about radiation exposure from Q-CT which is relatively higher than DXA, Q-CT has an effective dose of 25 to 360 μSv which is comparable to routine radiological examinations such as chest radiographs or a single lumbar spine radiograph, but much lower than CT scan. Present day fourth generation DXA scanners with fan beams have an effective dose of 6.7 to 31 μSv.[21,22] However, the patients are exposed to a total effective dose which will be a summation of the regions scanned during their single visit for whole body DXA. Though this exposure is relatively lower, there are concerns on the fan beams of having increased the risk of occupational radiation hazard to operating technicians who may be exposed to scatter from multiple scans in a day. Q-CT has added advantages of a very narrow band (mostly three slices) of exposure to radiation beam than whole body and a very low risk of radiation to the staff due to adequate shielding from radiation.
There is a high prevalence of untreated spinal osteoporosis in the general population and in spite of technological advances, many perioperative spine patients are suspected to be osteoporotic clinically or on cursory assessment of radiographs. Due to the lack of consensus on spinal osteoporosis evaluation, hip or peripheral DXA is being utilized to suspect spinal osteoporosis in view of degenerative changes in spine. Besides, there is evidence that BMD cannot be generalized even for the spine as a whole. Even with regards to a particular region of spine, BMD is reported to vary between individual vertebrae. Thus, it is prudent to measure spine BMD to detect spinal osteoporosis, rather relying on hip or other peripheral site DXA T-scores. Spinal osteoporosis impacts the practice of spine surgeons on many facets like fragility fractures, bone fragility impacting spine instrumentation resulting in pedicle screw loosening, cage subsidence, fusion failures, need for recurrent surgeries, secondary fractures, with associated morbidity and mortality.
There are a few studies showing good correlation between CT Hounsfield units (HU) and BMD measurements as well as higher detection rates of osteoporosis by using CT based modalities than DXA.[15,17,24] However, the routine use of normal CT scans for detecting spinal osteoporosis is associated with many limitations like higher radiation exposure, increased costs and non-availability of experienced radiologists/technicians, and phantom calibrations necessary for BMD estimation in the existing routine CT imaging protocols at the diagnostic centers. These render normal CT scans unsuitable for this purpose. Opportunistic use of already available CT scans may be beneficial if they covered the desired region of spine and are recently done, provided the patient didn't start any anti-osteoporotic medication. But scans done without a reference phantom cannot measure precise BMD and measurements obtained lack comparability. In recent times, HRCT chest is being done as a routine for COVID-19 screening. Although they may be used for opportunistic BMD estimation in the near future to avoid additional BMD scans for various reasons, they may still not give reliable BMD of the cervical and lumbar regions due to the obvious reasons and limitations mentioned above.
The scan costs of both the modalities are comparable at the imaging centers where the study cohorts were evaluated. The costs of equipment of Q-CT and DXA vary with the technical specifications available across various generations of the scanners. Comparison of state-of-art equipment available for both the modalities reveals Q-CT to be more expensive than DXA. However, while DXA machine needs a separate investment, Q-CT can also be extracted from regular CT imaging. Again, the cost differences may further narrow down with more frequent use of Q-CT.
Literature reveals technical superiority of Q-CT over DXA for generalized consideration of osteoporosis across multiple skeletal sites while studying various modalities available.[24,25] However, there is paucity of data based on comparative studies consisting of a significantly large sample size with age-matched, sex-matched study cohorts with specific focus on spinal osteoporosis. Again, random allotment of study individuals to the modalities, comparing the T-scores: their mean, mode as well as distribution in the study cohorts add to strengths of this study. It highlights the statistically significant difference between mean as well as the mode of the T-scores of each study cohort. The bell curves plotted to illustrate the distribution of T-scores in each study group suggest a positively skewed Q-CT curve indicating that the mode was more negative than mean of T-scores. This assumes importance meaning the most commonly occurring value (mode) was more negative than the average (mean), indicating that a large percentage of the Q-CT screened ladies belonged to more extreme (negative) side of the spectrum. From a comparative perspective too, the majority of the observations in the bell curve of Q-CT were towards the left of the curve (more negative values) than relatively symmetrical bell shaped DXA curve where the difference between mean and mode of T-scores was lesser. The curves were divided into three colored zones for better visualization—red zone for osteoporosis, yellow for osteopenia, and green for normal bone density. Q-CT group revealed a higher number of T-scores in red zone than DXA, thus showing a higher sensitivity in detecting severe osteoporosis. DXA curve revealed higher number of observations falling in yellow zone (osteopenia) than red and green. The peak of the curve in Q-CT lies in red zone while that of DXA appears in the yellow zone, meaning the former diagnosed higher prevalence of osteoporosis than the latter. Both the groups demonstrated lesser prevalence of normal bone density (green) than those with low bone density (red and yellow) (Figure 2).
Although the study has a statistically significant sample size of 718, and is age and sex matched with a common skeletal site of interest for BMD estimation, it has a drawback. The compared Q-CT and DXA BMD T-scores belonged to different sets of individuals. However, evaluating the same subject using two scans with the same purpose has practical, ethical, and financial considerations apart from the availability issues of both Q-CT and DXA. There have been studies with both the scans done on the same individuals which proved disagreement between spine T-scores of DXA and Q-CT.[25,26] However, these studies have their own drawbacks; the results being statistically insignificant due to low sample sizes, variation in time elapsed between the scans obtained, possibility of administration of anti-osteoporotic medication between the scans and speculation on radiation exposure.
Spine. 2022;47(6):E258-E264. © 2022 Lippincott Williams & Wilkins