Agreement and Disagreement
To the Editor:
We would like to respond to the letter written by Dr. Jean Weigert in reference to our manuscript published in the May 1997 edition of the Journal of Bone and Mineral Research. Overall, we believe that Dr. Weigert has made some valuable comments. We are in strong agreement that quantitative computer tomography (QCT) is a powerful, but seriously underutilized, tool in diagnosis of osteoporosis and follow-up of therapy. However, we feel that certain points need clarification.
First, it is not necessarily true that QCT is the most accurate method of measuring bone mineral density (BMD). In fact, the absolute physical accuracy errors for QCT, primarily related to variable marrow-fat content and to beam hardening,1, 2 are comparable if not slightly higher than those of dual X-ray absorptiometry (DXA). However, because it provides a measurement of purely trabecular BMD, it certainly is the most sensitive technique for assessment of vertebral fracture risk,3, 4 as Dr. Weigert indicates. Vertebral fracture discrimination and physical accuracy are, however, separate issues.
Second, new three-dimensional image analysis algorithms used in conjunction with enhanced quality control measures ensure that “when Mrs. Jones has her QCT in San Francisco, California, it can be accurately compared to one in West Hartford, Connecticut.” This may be an overly strong statement, given differences between calibration standards, CT systems, and image analysis approaches. Even when one is comparing spinal BMD measurements between two scanners of the same make and model, using the same calibration phantom, there are likely to be slight differences. This is simply due to small differences in scanner calibration, as well as interoperator variation in the use of the analysis software. This is the case with DXA as well. The key point is that the comparison between Mrs. Jones's measurement in two different clinical sites may be sufficiently accurate for diagnosis, but an error of a few percent may confound a longitudinal measurement.
Finally, we agree with Dr. Weigert about the importance of educating the referring physicians about the applicability of QCT to their patient populations. This education process needs to take into account the relative advantages and sources of error in QCT and DXA measurements. This will enable the physician to make the appropriate decision as to which diagnostic procedure is most suitable for a given patient.
