Endocrine Abstracts

Context: Although Trabecular Bone Score (TBS) has been widely and independently associated with fragility fractures, its diagnostic power in certain populations remains to be established. Women initiating aromatase inhibitors for breast cancer should be screened for bone health to prevent cancer treatment-induced bone loss as much as possible. TBS is regarded as an adjunctive tool to evaluate these patients, but thresholds of pharmacologic interventions are still a gray zone, especially in younger women.

Objective: To test the diagnostic power of TBS to predict prevalent vertebral fractures at the start of endocrine therapy for breast cancer. Low TBS was defined as <1.2.

Design: Cross-sectional, real-life.

Setting: University hospital.

Patients: We retrospectively evaluated 223 consecutive ambulatory women with newly diagnosed early breast cancer from February 2021 to November 2022. They all underwent a prespecified protocol which included mineral biochemistries with bone turnover markers, bone mineral density with TBS, and endocrine consultation. Lateral X-rays images of the spine performed just before breast surgery were reviewed to screen for asymptomatic vertebral fractures.

Results: Mean age of the patients was 61±12 years. Overall, 65 patients (29%) had normal bone density (BMD), 90 patients (40%) had osteopenia, and 68 patients had osteoporosis (31%), with 35/223 (15.7%) women having low TBS. The prevalence of low TBS progressively increased with decreasing BMD (6.2% in normal BMD, 10% with osteopenia, and 32.3% with osteoporosis; P<0.001). Vertebral fracture prevalence was 24/223 (10.8%). BMD category was highly associated with prevalent vertebral fractures (P<0.001), but this was not the case for low TBS alone (P=0.597). A low TBS highly correlated with advanced age (P<0.001), independent of other clinical or biochemical parameters (age, BMD T-score, renal function, 25-OH vitamin D, PTH). ROC analyses showed that TBS Area Under Curve (AUC) was 0.647 (P=0.021), comparable (P=0.561) to that of femur neck T-score (AUC=0.688, P=0.003) for detecting prevalent vertebral fractures. Notably, the diagnostic power of TBS was much higher in women younger than 65 years, with TBS AUC being 0.769 (P=0.032), with femur neck BMD being much less accurate (AUC 0.564, P=0.339). By contrast, TBS AUC dropped to 0.484 (P=NS) in older women, whereas femur neck T-score retained its expected diagnostic accuracy (AUC 0.690, P=0.008).

Conclusions: TBS rather than BMD could better predict prevalent vertebral fractures in women younger than 65 initiating aromatase inhibitors, as compared to older women. In the latter group, femur neck BMD T-score was confirmed to be more accurate.