3D QCT: A Useful Tool in Following Therapy
Jean M. Weigert, M.D., Imaging Center of West Hartford, CT
Christopher E. Cann, Ph.D., University of California, San Francisco
Jean M. Weigert, M.D., Imaging Center of West Hartford, CT
Christopher E. Cann, Ph.D., University of California, San Francisco
There are two basic clinical indications for performing bone densitometry: accurate measurement of bone mineral density in order to identify patients with low bone mass and increased risk of fracture, and precise monitoring of patients placed on therapy to determine its effect (1–3).
The goal of therapy is to increase bone mass or at least stop its loss, by inhibiting resorption and/or stimulating bone formation. Reduction in the incidence of fractures is the most important clinical outcome (1,4), and while it may not be the direct consequence of increased bone mass in individual patients, studies have shown this relationship in populations treated with various agents. Therapies with documented results include hormonal replacement therapy, alendronate, calcitonin, and more recently raloxifene. Many other therapies are in the experimental stages including newer bisphosphonates, other selective estrogen receptor modulators (SERMs), parathyroid hormone and its analogs, and fluoride compounds.
The ability to monitor subtle changes in a precise and accurate manner is very important both clinically and in research. Quantitative computed tomography (QCT), and in particular 3D QCT may be a very good modality to do just that. Conventional, or 2-dimensional QCT has been available and widely used since the early 1980s. It utilizes traditional computed tomography technology to acquire CT images for bone density measurements. This is done using a lateral scout image to localize specific slices (10 mm thick each) through four vertebral bodies. A region of interest (ROI) in the vertebra is compared to the density of ROIs in a standard and a bone density measurement calculated in mg/cc. The radiation dose is approximately 50 µSv (5). In vivo precision is about 3 mg/cc, or 1.9% in normals and 3% in osteoporotics (6). This technology, when performed using strict quality control, can be very useful in diagnosing osteoporosis and monitoring patients on therapy.
Recent advances in QCT technology have improved our ability to measure BMD precisely and accurately. 3D QCT takes advantage of the new helical CT scanners to obtain a data set in 30–40 seconds (traditional CT scanners can also be used although they take 2–3 minutes to acquire the data). After a lateral scout is obtained, 3 mm contiguous slices are acquired creating a set of about 25 images. The 3D volumetric data are displayed in axial, coronal and sagittal planes, a precise region of interest defined, and bone density calculated accurately normalized to published data (7). The radiation dose is comparable to 2D QCT, about 60 µSv. The precision in clinical practice is improved a factor of 3 to about 1 mg/cc (0.7% in normals and 1.1% in osteoporotics), based on short term in vivo precision at our site of 0.8 mg/cc and long term in vitro precision from our QA phantom of 0.5 mg/cc. These improvements have made 3D QCT an ideal modality not only to diagnose patients accurately, but to monitor those patients on therapy over periods as short as 6–12 months.
In our clinical practice we have defined two groups of patients, those who have a significant change in QCT BMD between two measurements (“Responders”) and those whose change is non-significant (“Non-Responders”). Responders are those who show a change greater than 3 mg/cc, with those showing less than 3 mg/cc change classified as Non-Responders. This definition is based on the statistical definition that a “real” difference must be at least 2.8 times the precision of the measurement. Note that the definition does not identify those patients who might have a positive clinical response to a therapy by simply maintaining bone instead of losing or gaining it. We have followed 64 postmenopausal women, 50–82 years old, who were given therapy for an 8–18 month period after an initial QCT BMD measurement. Their histories varied and choice of therapy was made by themselves and their clinicians. 49 were placed on alendronate only, 5 on HRT, 7 already on HRT were also given alendronate, 3 were placed on calcitonin, and 1 already on HRT was also placed on calcitonin. All had supplemental calcium and vitamin D. The mean BMD on entry was 76 mg/cc with a mean T-score of -3.5.
There was significant individual variation in response to therapy with all the treatment groups, indicating that no one treatment worked equally well for all women taking it (Figure 1). It also appeared that there was no added effect of alendronate for patients already on HRT therapy in this small group. The mean increase on HRT was 40% higher than in the alendronate group, but it was not a statistically significant difference. There was a decrease in all patients on calcitonin alone. Because this was a retrospective study of patient- and physician-chosen therapies, there was no “control” group, so even in the “non-responders” with less than 3 mg/cc increase in BMD the therapy may have prevented bone loss.
Mean annualized BMD changes measured by QCT between 2 exams 8–18 months apart. The standard deviations are comparable for the different therapies, showing that the individual variation in response to each therapy is similar. The effect of alendronate alone appears to be less than that for HRT, but the difference is not significant in these small groups. There is also no measurable added effect of alendronate to the effect of HRT.
Figure 1
In review of the patients taking alendronate alone, 75% were classified as responders and 25% non-responders in the first 8–18 months of therapy. The average increase was 11% per year, which represented an average increase in T-score of 0.4 units. Excluded from analysis were those patients who discontinued therapy. Non-response was not evaluated in terms of those patients who were not compliant or not taking adequate calcium and vitamin D. Even among those patients classified as non-responders, no patient showed a measurable loss of BMD (Figure 2).
Individual changes in trabecular density by QCT from first to second exam. Bars represent 2.8 times the precision, or the region of no statistically significant change. Changes are given in mg/cc—baseline BMD for the group as a whole averaged 76 mg/cc.
Figure 2
11/12 women (>90%) using HRT with or without alendronate increased BMD, on average 15% per year, while the remaining patient did not lose bone. The success of this therapy may minimize the need for early followup BMD measurements in these patients. If patients enter therapy with very low bone mass, the need for additional therapy can be determined by QCT at the end of one year, while it may take 2–3 years to make this determination with DXA.
No patients on calcitonin showed improvement. All lost significant BMD (except one patient on concurrent HRT). These findings require further investigation, given that there was no control group, and the fact that these patients were the ones who could not tolerate HRT or alendronate so were given calcitonin as an alternative therapy.
Monitoring the effects of therapy using QCT instead of DXA allows individual treatment decisions to be made earlier. We saw that 75% of patients placed on alendronate had a measurable increase in BMD on an average 1 year followup, whereas it takes 3 years to show a comparable improvement using DXA (8). The average improvement with QCT was 11% in that first year, compared to an average with DXA of 3–4%. Other studies predict that the 3–4% improvement with DXA should lead to a 10% reduction in fractures, not the 50% seen (9). The average increase by QCT, 0.4 T-score units, predicts a 40% decrease in fractures based on odds ratios (10). Therefore, BMD changes measured by QCT may be a better indicator of reduced fracture risk than changes measured by DXA.
3D QCT may also be useful as a tool for research on emerging therapies that may have different effects on different bone regions. For example, PTH has been shown to increase trabecular bone while it may increase cortical porosity, reducing the density of cortical bone (11). Using DXA, increased trabecular density with increased cortical porosity in the proximal femur results in little change in total BMD, while 3D QCT can isolate the trabecular and cortical bone components and determine their independent responses to these therapies.
We conclude that 3D QCT is a precise and accurate method for following patients treated for osteoporosis using a variety of therapies. Responders versus non-responders can be detected early thereby maximizing the effectiveness of therapy. This may also be a more cost effective way to monitor those patients utilizing fewer tests over a shorter time period to determine who should or should not remain on these drugs. Newer 3D techniques to measure isolated cortical and trabecular bone in the hip are a useful modality in research settings to evaluate the mechanisms of drug therapy.
References
Presented at the 4th Annual Scientific Meeting, International Society of Clinical Densitometry, Orlando, FL, January 15-18, 1998.