Bone DEXA interpretation only as great as the sum of its parts

Oct 30, 2003

Radiologists can sometimes overlook the obvious when it comes to reading bone densitometry (DEXA) exams, according to Dr. Leon Lenchik from Wake Forest University School of Medicine in Winston-Salem, NC.

"After six years of lecturing on (DEXA), I’m always surprised to hear that radiologists forget to look at the DEXA image," said Lenchik, who is an associate professor of radiology at the university. "You’d think that radiologists would remember that this is a quantitative assessment that contains an image so (they should) pay attention to it. But for some reason radiologists and non-radiologists too, tend to ignore the image."

This can be problem because a DEXA image offers critical information on patient positioning, scan analysis, and image artifacts, Lenchik explained in a presentation at the 2003 International Skeletal Society meeting in San Francisco. "The image contains information that isn’t going to be revealed if you just look at the numbers," he said.

The key to successful interpretation of a DEXA report is to look at all the elements together, rather than focusing on a single component such as bone mineral density (BMD) or the T-score. Bone densitometry exams are done for multiple reasons:

Diagnosing osteoporosis
Assessing the risk of fractures, bone disorders
Helping to select patients as candidates for drug therapy
Monitoring drug therapy

As a result, reading DEXA reports requires a certain degree of multitasking. "There’s a reason that all of those numbers appear on the DEXA report. They are not there to confuse you," Lenchik said. He outlined a practical approach to DEXA scan interpretation, starting with basic patient information.

Patient history

"We like to begin with a (patient) questionnaire followed by radiographs, if necessary, and, of course, look at the DEXA printout," Lenchik said. The questionnaire should be in a yes-or-no format, with questions about drug history, fracture history, and previous surgical procedures such as vertebroplasty.

"This is probably the easiest way to improve the quality of your DEXA report, to know something about the patient," he said. "One of the things that you want to make sure of is that your technologist enters the correct demographic information into the computer, because the disease course will change if you enter the wrong age or gender."

Protocol

At Lenchik’s institution, the DEXA protocol calls for scanning both hips. Two skeletal sites are measured: the posterior-anterior (PA) lumbar spine and the proximal femur. For patients who cannot undergo the latter (due to a bilateral hip replacement, for example), a forearm scan is obtained.

The patient positioning, scan protocols, and analysis recommendations set out by DEXA unit manufacturers must be followed. Lenchik said there are analytical differences between the two main DEXA scanners on the market: GE Medical System’s Lunar densitometer offers Lateral Vertebral Assessment (LVA), while Hologic’s devices records Instant Vertebral Assessment (IVA).

Positioning

On a properly positioned PA spine scan, the spine appears straight or aligned with the long axis of the scan table. In addition, both iliac crests are visible, with the scan starting in the middle of L5 and terminating in the middle of T12. Be sure to number the vertebral bodies correctly, he said.

For forearm scans, the forearm is straight and the distal end of the radius and ulna are visible. However, the distal region should not include the articular surface of the distal radius.

For correct positioning on a hip scan, the femoral shaft is straight, the hip is internally rotated, and the scan includes the ischium and greater trochanter. Lenchik noted that femoral neck placement is manufacturer specific: The Lunar system measures the midportion of the femoral neck while the Hologic system measures the base of the femoral neck.

These differences make it particularly important to follow manufacturer guidelines, Lenchik said. If positioning errors are made, the scan should be redone so that that the appropriate scan analysis (region-of-interest or ROI; location) is obtained.

Artifacts

The presence of artifacts can have an adverse affect on BMD measurements, Lenchik said. Artifacts that can lead to a false elevation of BMD include degenerative disease, spinal fracture, vertebroplasty, and Paget’s disease. A false decrease in BMD can come up in a report if the patient has undergone laminectomy.

Finally, two things that can have an unpredictable effect on BMD are patient motion and the presence of barium in the soft tissues.

"Depending on how much overlap there is between contrast agent and bone, the effect on BMD may be unpredictable," Lenchik said. "Keep in mind that there are artifacts that can draw BMD in either direction and certainly try not to scan your patient if they’ve had recent gastrointestinal contrast (72 hours or less)."

Numbers and images

DEXA results are expressed in absolute BMD units (grams per square centimeter), T-scores, and Z-scores. The T-score refers to the number of standard deviations that the measured BMD is above or below, compared with the BMD of a young, normal gender-matched reference population.

"A pitfall here is that people focus too much on the T-score. Now the T-score is important, obviously, for diagnosis, except in kids, where we don’t use it. But the Z-score is also important," Lenchik said.

The Z-score is the number of standard deviations that the measured BMD is above or below the BMD age-matched and sex-matched reference population. The Z-score is particularly useful for guiding lab workups to exclude secondary causes of osteoporosis.

As for BMD scores, "basically, when looking at the BMD results, you want to make sure that the bone density increases as you go down the spine, and that the T-scores are all within one standard deviation of each other," he said. "When the T-scores seem to jump dramatically from one vertebra to another, then that can help you identify artifacts."

Of course, looking at the image generated as part of the DEXA scan is crucial. "If you are a clinician or a radiologist who seems to think that the numbers are the means to DEXA interpretation...you can miss a fracture even if you apply the rules (regarding T-scores and BMD)," he said.

ROI

In 2001, the International Society for Clinical Densitometry (ISCD) released a position statement for clinical DEXA scans. Lenchik served on that statement committee, which came up with some general rules.

"The question that frequently arises is ‘Which ROI are you talking about?’ " Lenchik said. "There are a lot of different regions of the spine and hip. We decided to use L1-L4, provided there are no focal artifacts. If there is an artifact, then you have to exclude that vertebra. Try to avoid using a single vertebra or cherry picking for the lowest value."

On PA spine scans, there should be an incremental increase in BMD from L1-L4. Individual vertebral T-scores should be one standard deviation.

For diagnosing osteoporosis, the lower of the T-scores of the PA spine and hip should be used. In the spine, it is preferable to use the weighted average of L1-L4. In the hip, using the lowest T-score of the total femur, femoral neck, and trochanter is suggested.

Lenchik emphasized that discrepancies between measured sites may indicate a true difference in bone density -- or there may be false elevation due to degenerative bone disease, fractures, or postoperative changes. Low BMD does not automatically imply postmenopausal changes.

Monitoring

Whether the patient is being tracked for possible osteoporosis or for therapeutic changes, matching the baseline scan with the follow-up is vital, Lenchik said.

"One helpful bit of advice is to look at the measured area (on the follow-up scan). Make sure that it’s within 2%-3% of the baseline. Because if the area is different, even if the ROI looks the same, there may be a technical issue. You would not trust the follow-up results in the patient," he said.

Also, absolute BMD should be analyzed for monitoring patient response to therapy, not T-scores, as the latter does indicate biologic changes. In addition, the reader should be alert to small changes in BMD versus large changes.

The trick for assessing changes is to calculate the least significant change (LSC). According to the ISCD guidelines:

"Each bone density facility should determine its own precision and LSC in order to determine when changes in bone mineral density is significant at a 95% confidence level (2.77 times the precision error). In patients being monitored for the effect of therapy, only a significant decrease in bone mass on serial testing (greater than the LSC) warrants a change in the therapeutic regimen."

The final report

Unfortunately, there is no consensus among clinicians on how to present a DEXA report. An unpublished survey of 15,000 U.S. radiologists, commissioned by the American Society for Bone and Mineral Research, found that most exclude BMD from their reports, and avoid making any suggestions regarding therapy, Lenchik said.

However, in Lenchik’s estimation, a solid DEXA report should contain the following elements: