Trabecular Bone Score vs. DEXA: What’s the Difference?

DEXA (dual-energy X-ray absorptiometry) measures how much mineral is packed into your bones, while the trabecular bone score (TBS) evaluates the internal structural quality of bone using the same scan image. Think of it this way: DEXA tells you how dense the wood in a beam is, while TBS tells you whether that wood has a tight, resilient grain or is riddled with gaps. Both pieces of information matter for predicting whether a bone will break, and they capture different aspects of skeletal health that don’t always move in the same direction.

What DEXA Measures and What It Misses

DEXA works by passing two X-ray beams of different energy levels through your body and measuring how much each beam is absorbed by bone versus soft tissue. The result is a number called bone mineral density, or BMD, expressed as grams of mineral per square centimeter. That number is then compared to reference populations to produce a T-score, which is the basis for diagnosing osteoporosis (a T-score of −2.5 or lower) or osteopenia (between −1.0 and −2.5).

The problem is that BMD is a two-dimensional projection of a three-dimensional structure. The scan flattens everything in its path into a single plane, so it cannot distinguish between a bone that is genuinely dense throughout and one that simply has more material stacked in the beam’s line of sight. Inaccuracies in DEXA arise partly from differences between the soft tissue inside and outside the bone within the scan area, which can skew the density reading.¹ More practically, BMD captures mineral quantity but says nothing about how that mineral is arranged. Two people can have identical BMD values yet very different fracture risks, because the internal scaffolding of their bones differs.

How TBS Reads Bone Texture from the Same Scan

TBS is not a separate scan. It is a software analysis applied to the lumbar spine image your DEXA machine already produces. The algorithm evaluates pixel-by-pixel gray-level variations across that image, essentially reading the “texture” of the projected bone.² A healthy trabecular network, with its dense mesh of tiny bone struts, produces a more varied, complex texture pattern. Degraded bone with wider gaps between thinner struts produces a more homogeneous, smoother-looking image.

The key insight is that TBS can distinguish between DEXA scans that show similar BMD measurements.³ Two spine scans might both read a T-score of −1.5, but TBS could flag one as having degraded microarchitecture and the other as normal. This textural approach has been validated against actual bone structure: measures of trabecular thickness, number, spacing, and connectivity all correlate with what TBS captures from the flat DXA image.⁴

TBS values are reported on a continuous scale. Generally, a score above 1.35 suggests normal microarchitecture, between 1.20 and 1.35 indicates partial degradation, and below 1.20 signals degraded structure. These thresholds give clinicians a quick way to stratify patients beyond what BMD alone reveals.

Diabetes and the Paradox of “Strong-Looking” Bones

Type 2 diabetes offers the clearest illustration of why BMD alone can be misleading. People with type 2 diabetes fracture more often than their DEXA scans would predict, because their BMD is frequently normal or even higher than average. The mineral is there, but the internal architecture is compromised. A large study of Manitoba women found that diabetes was associated with higher BMD at all measured sites yet lower TBS at the lumbar spine.⁵ Women with diabetes in the lowest third of TBS values had roughly 2.6 times the odds of being in that low group compared to those without diabetes, even after adjustments. TBS independently predicted fractures in the diabetic group, with an adjusted hazard ratio of about 1.27 per standard deviation decrease. Adding BMD to the prediction model actually made the apparent effect of diabetes on fracture worse, while adding TBS reduced it, suggesting TBS was capturing the bone quality deterioration that BMD was masking.

Reviews of the broader literature confirm that TBS appears more accurate than BMD for capturing fracture risk in diabetic bone disease, particularly in postmenopausal women with type 2 diabetes.⁶ This is one of the strongest clinical arguments for looking beyond density alone.

Glucocorticoid Use and Hidden Bone Damage

Long-term steroid medications (prednisone, prednisolone, and similar drugs) are another situation where DEXA can give false reassurance. Glucocorticoids damage bone quality rapidly, often before measurable mineral loss shows up on a standard scan. In a study of patients on glucocorticoid therapy for autoimmune diseases, degraded microarchitecture as measured by TBS was far more common than osteoporosis as defined by BMD. Among those with vertebral fractures, about 76% had degraded TBS but only 38% met the BMD threshold for osteoporosis.⁷

A separate study comparing older women on glucocorticoids to those who had never taken them found that TBS Z-scores were substantially lower in the treated group (−0.80 versus 0), while their lumbar spine BMD Z-scores were essentially identical. BMD alone could not discriminate between the two groups, but TBS could, with reasonable accuracy.⁸ A controlled study of women receiving glucocorticoid therapy likewise found that TBS was significantly lower while BMD showed no significant difference, reinforcing that TBS and adjusted fracture risk tools may serve as better assessment adjuncts in this population.⁹

Fracture Risk Prediction and the FRAX Connection

FRAX is the most widely used tool for estimating your 10-year probability of a major osteoporotic fracture. It combines clinical risk factors like age, weight, smoking history, and family history, optionally with BMD. TBS can be fed into the FRAX calculation as an adjustment factor, and doing so generally improves the accuracy of the prediction.¹⁰

That said, the improvement is not dramatic in every population. A community-based cohort study of Korean women found that adding TBS to FRAX did not predict major osteoporotic fractures better than FRAX without BMD alone.¹¹ This doesn’t mean TBS is useless in those populations, but it does suggest that the benefit of adding TBS to fracture prediction varies by ethnicity, age group, and the underlying cause of bone fragility. The biggest gains tend to appear in people whose bone quality is compromised out of proportion to their density, such as those with diabetes or on glucocorticoids.

The Spinal Degeneration Problem

Both DEXA and TBS are measured at the lumbar spine, and the lumbar spine of anyone over about 55 is rarely a pristine structure. Bone spurs (osteophytes), disc degeneration, and vascular calcifications are extremely common in older adults and can artificially inflate BMD readings. One study found that osteophytes and endplate changes were associated with up to a 14% increase in measured BMD values.¹² In women with both osteophytes and endplate sclerosis, spinal BMD was about 19% higher than in unaffected women, and the ability of the spine scan to distinguish fracture patients from non-fracture patients was significantly reduced when degenerative changes were present.¹³

TBS is not immune to this problem. Since it works from the same DEXA image, anything that alters the pixel pattern of the scan can affect TBS as well. However, because TBS reads texture rather than total mineral signal, it is less susceptible to the additive effect of calcified deposits that simply pile more mineral into the beam path. Clinicians still need to be aware that severe degeneration can muddy both measurements, and in heavily affected vertebrae, excluding individual levels from the analysis is standard practice.

Soft Tissue Thickness and the BMI Challenge

A well-documented limitation of TBS is its sensitivity to the amount of soft tissue overlying the spine. Phantom studies have shown that increasing soft tissue thickness artificially lowers TBS values in a statistically significant, dose-dependent way across multiple scanner types.¹⁴ In practical terms, this means that people with higher body mass index may receive artificially low TBS readings that overstate their fracture risk, while very thin patients might get inflated scores.

The latest version of the TBS software (version 4.0) directly corrects for soft tissue thickness, and early results suggest it substantially weakens the correlation between TBS and BMI or abdominal fat. In one study, the correlation between TBS and BMI dropped from statistically significant with the older software version to non-significant with version 4.0.¹⁵ If your TBS was measured before this update, the result may be less reliable if you carry significant abdominal weight. It’s worth asking which software version your facility uses.

How TBS Compares to High-Resolution Bone Imaging

The gold standard for viewing trabecular bone structure in living people is high-resolution peripheral quantitative CT (HR-pQCT), a specialized scanner that can image individual bone struts at the wrist and shin. TBS at the lumbar spine correlates significantly with HR-pQCT measures of trabecular bone density, trabecular number, connectivity, and spacing at peripheral sites.¹⁶ In Chinese American and white women, TBS was associated with all HR-pQCT indices except radial cortical thickness and tibial trabecular thickness, with correlation coefficients ranging from about 0.20 to 0.52.¹⁷

More recent work has confirmed that TBS correlates with both trabecular and cortical bone parameters measured by HR-pQCT, which may explain why TBS contributes to hip fracture risk prediction even though the hip is predominantly cortical bone.¹⁸ These correlations are moderate, not perfect. TBS is a two-dimensional proxy for three-dimensional structure, and it will never replace direct imaging. But HR-pQCT machines cost hundreds of thousands of dollars, measure only the extremities, and exist in a handful of research centers worldwide. TBS extracts useful structural information from a scan that millions of people already get.

TBS in Men

Most osteoporosis research has focused on postmenopausal women, so the evidence base for TBS in men is thinner. What exists is encouraging. In a cohort of elderly Japanese men, those who experienced major osteoporotic fractures had significantly lower TBS. Adding TBS to FRAX improved the reclassification of who would and wouldn’t fracture, even though the overall area under the curve for prediction didn’t change much.¹⁹ In the Manitoba Bone Density Program, spine TBS predicted major osteoporotic fractures and hip fractures independently of clinical FRAX scores in men, and hip fractures independently of both FRAX and BMD.²⁰ The researchers themselves noted that larger studies are needed to confirm these findings, but the direction is consistent: TBS adds information beyond what density provides in men as well as women.

Primary Hyperparathyroidism and Other Endocrine Conditions

Overactive parathyroid glands (primary hyperparathyroidism) cause excess calcium to be pulled from bone, and TBS tends to be reduced in affected patients. In postmenopausal women with primary hyperparathyroidism, both BMD and TBS were significantly lower than in matched controls.²¹ However, TBS doesn’t always perform equally well in this condition. In a cohort of 153 Romanian patients with symptomatic primary hyperparathyroidism, mean TBS fell in the partially degraded range, and patients with vertebral fractures had significantly lower TBS than those without. But TBS was not independently associated with fractures after adjusting for years since menopause.²² The lesson here is that TBS is not a universal fracture predictor; its added value depends on the clinical context. In conditions where the dominant mechanism of bone damage is microarchitectural (as in diabetes or steroid use), TBS shines. In conditions where other factors overwhelm the structural signal, it may not add much.

Monitoring Treatment Response

One practical question patients ask is whether TBS can track how well their osteoporosis treatment is working. Expert consensus from a meeting organized by major European and international osteoporosis organizations concluded that TBS provides useful adjunctive information when monitoring treatment with long-term denosumab and bone-building (anabolic) agents.²³ The same expert group noted that TBS, when combined with BMD and clinical risk factors, can inform both treatment initiation and the choice of therapy. This does not mean TBS replaces BMD for monitoring; rather, it adds a second dimension. If your BMD is rising on treatment but your TBS isn’t improving, that could signal that bone quality isn’t keeping pace with mineral gains.

Cross-Platform Calibration and Practical Access

TBS is a proprietary software analysis, and not every DEXA facility has it installed. Even among those that do, results can vary between scanner brands. Studies comparing TBS across different Hologic and GE-Lunar machines found reasonable agreement when a dedicated TBS calibration phantom was used, but notably lower agreement without it. In one comparison of Hologic and GE-Lunar devices without phantom calibration, the correlation between TBS readings dropped to around 0.7.²⁴ That means if you switch DEXA facilities between scans, your TBS values may not be directly comparable unless both machines have been cross-calibrated. This is similar to the existing challenge with BMD, where switching machines can produce apparent changes that are really just measurement differences, but the issue is more pronounced for TBS.

For routine clinical use, the practical advice is straightforward: get your DEXA scans at the same facility on the same machine whenever possible, and ask whether TBS is available if you have risk factors (diabetes, steroid use, family history of fracture) that might make bone quality a concern beyond what density alone reveals.

TBS in Children and Adolescents

Bone assessment in younger people is an area where TBS research is just getting started. Reference curves for TBS in children and adolescents have recently been developed, showing distinct age- and sex-related patterns tied to puberty. In Mexican children and adolescents, TBS rose steeply during puberty: girls showed a sharper increase beginning around age 9 and peaking by 16, while boys had a more gradual rise starting around 10 to 11 and peaking by 18.²⁵ Similar work in Brazilian children confirmed that TBS varies by age, sex, and pubertal stage, and established normative values for that population.²⁶ These are early steps. Without well-established pediatric reference ranges across diverse populations, TBS in children remains a research tool rather than a clinical one. But the groundwork for eventual clinical use is being laid, particularly for conditions like childhood cancer treatment or chronic inflammatory disease where bone quality may suffer before density drops.