Cancer survivors miss early bone-density testing in Israeli real-world study

A large retrospective study of 21,112 adults with newly diagnosed cancer found that bone-density testing after diagnosis was both uncommon and often delayed. Researchers at Assuta Medical Centers in Israel analyzed dual-energy X-ray absorptiometry, or DXA, use among patients diagnosed between January 2015 and December 2024. Only 35.5% of the cohort underwent DXA at any time after diagnosis, and just 7.3% received an assessment within six months. Among patients who were scanned, the median interval from cancer diagnosis to DXA was 14.8 months. The results identify a substantial gap between recommendations to evaluate skeletal risk in patients at increased risk and the routine delivery of bone-density assessment in a broad oncology population.

Skeletal complications can arise from both malignant disease and its treatment. The study describes cancer treatment–induced bone loss as a multifactorial process involving systemic inflammation, reduced physical activity, nutritional deficiencies and exposure to treatments such as chemotherapy, radiotherapy, endocrine therapy, glucocorticoids and hematopoietic stem-cell transplantation. These factors can act together to increase bone turnover, reduce bone mineral density and elevate fracture risk. Bone loss associated with cancer therapies is particularly well established in hormone-sensitive malignancies, including breast and prostate cancer, where aromatase inhibitors and androgen-deprivation therapy are linked to accelerated bone loss. However, population-based evidence described by the researchers indicates that fracture risk is elevated across a broader range of cancers and can persist for years after diagnosis.

Despite that wider risk, guidelines from major oncology and endocrine organizations recommend DXA for people at increased skeletal risk, especially those exposed to bone-toxic therapies. The researchers also note that osteoporosis risk-assessment tools can underestimate fracture risk in cancer survivors and that malignancy is not consistently incorporated into fracture-prediction models. Their retrospective cohort used the MDClone platform to extract fully anonymized, statistically accurate data from electronic medical records at Assuta, a private healthcare network operated by Maccabi Healthcare Services. Eligible participants were adults with a first documented cancer diagnosis and complete demographic information. Mean age was 60.2 years, 77.0% were women, and mean body-mass index was 26.7 kg/m². Chemotherapy exposure was recorded in 21.4% of patients and radiotherapy in 7.0%.

Assessment patterns differed sharply according to sex and cancer type. Women were more likely than men to undergo DXA, with utilization rates of 41.7% and 15.0%, respectively. Breast cancer had the highest rate of testing at 42.2%, followed by prostate cancer at 30.0%, while rates were generally below 15% for lung, gastrointestinal, ovarian and pancreatic cancers. In a multivariable logistic regression, female sex was associated with greater odds of DXA assessment than male sex, with an adjusted odds ratio of 2.13 and a 95% confidence interval of 1.79–2.50. Each additional year of age was associated with slightly higher odds, while breast and prostate cancer were associated with adjusted odds ratios of 5.82 and 5.29, respectively, compared with other cancer types. These associations describe testing patterns rather than proving why individual scans were ordered or omitted.

Timing was similarly uneven. Only 14.1% of the full cohort underwent DXA during the first 12 months after diagnosis. Among all patients who eventually received a scan, the median time to assessment was 14.8 months. Median timing was 12.9 months for breast cancer, with an interquartile range of 5.8–28.0 months, and 12.7 months for prostate cancer, with an interquartile range of 4.4–28.8 months. Patients with other malignancies waited longer: their median time to DXA was 17.9 months, with an interquartile range of 8.3–35.1 months. Men and women who were assessed had broadly comparable timing, indicating that the largest sex difference concerned whether a scan occurred at all rather than the interval among those who received one.

Bone density was frequently abnormal when the scans finally occurred. The analysis of bone-density outcomes included 5,570 tested patients with complete, evaluable measurements. Using the lowest T-score from the lumbar spine, femoral neck or total hip, 40.2% had normal bone mineral density, 45.2% met the criteria for osteopenia and 14.5% met the criteria for osteoporosis. The mean lowest T-score was −1.72 among patients classified as osteopenic and −2.95 among those classified as osteoporotic. Site-specific results showed the least favorable average T-score at the femoral neck, at −1.14, compared with −0.83 at the total hip and −0.54 at the lumbar spine. Osteopenia was most frequent at the femoral neck, affecting 50.3% of patients with available site-level measurements, while osteoporosis was most common at the lumbar spine at 10.2%.

These measurements show a large burden of low bone mass among people who reached DXA, but they do not establish when the bone loss began or what caused it. Abnormal density may have predated the cancer diagnosis, and the retrospective data could not uniformly capture histology, disease stage, treatment dose or duration, lifestyle factors and other influences on skeletal health. The investigators could not systematically identify hematopoietic stem-cell transplantation, and DXA examinations performed outside Assuta may not have been captured. Fracture outcomes were not analyzed, so the study could not determine whether delayed testing was associated with more fractures. DXA completion was also used as a surrogate for bone-health evaluation, even though some patients without a recorded scan may have met osteoporosis criteria through other clinical information. These restrictions prevent the findings from establishing a causal effect of the testing gap on skeletal outcomes.

The analysis nevertheless points to missed opportunities for earlier skeletal risk identification. The researchers proposed several potential ways to make assessment more systematic, but did not test the effectiveness of these approaches. Electronic oncology records could issue automated DXA-referral prompts when a cancer is diagnosed or when a patient starts treatment likely to affect bone. Bone-health evaluation could be integrated into standardized survivorship-care plans and multidisciplinary tumor reviews, including for cancers less consistently covered by endocrine-therapy guidance. Dedicated referral pathways were also proposed for patients receiving chemotherapy, radiotherapy or glucocorticoids. Because the study documented substantially lower use among men, the authors suggested targeted education and closer collaboration among oncology, endocrinology and specialist bone-health services, while acknowledging the need to determine whether such measures improve assessment and outcomes.

Uncertainty remains particularly relevant for newer systemic therapies, including immune checkpoint inhibitors and targeted agents. The study notes that their long-term effects on bone mineral density and fracture outcomes remain incompletely characterized. A cited meta-analysis of randomized trials involving more than 19,000 patients associated immune checkpoint inhibitor use with a nonsignificant 18% increase in clinical fracture incidence relative to controls, with an odds ratio of 1.18 and a 95% confidence interval of 0.82–1.70. Because fractures were infrequently reported and the confidence interval included no effect, the estimate was imprecise and does not establish a fracture-causing effect. The present study also had broader generalizability limitations: it included one healthcare network in a specific national insurance system, did not consistently capture ethnicity, and may have enrolled a less complex case mix than the wider Israeli cancer population. Within those boundaries, the results show that DXA after cancer diagnosis was selectively delivered and frequently late, with abnormal bone density already common when assessment occurred.

Subject of Research: Real-world DXA utilization, timing and bone-mineral-density outcomes after cancer diagnosis

Article Title: Delayed and selective bone health assessment following cancer diagnosis: a real-world gap in survivorship care

Article References: Rouach, V., Versano, Z., Sasson, M., Greenman, Y., & Makori, A. (2026). Delayed and selective bone health assessment following cancer diagnosis: a real-world gap in survivorship care. Supportive Care in Cancer, 34(11), Article 1124. https://doi.org/10.1007/s00520-026-11340-3

Image Credits: AI Generated

DOI: 10.1007/s00520-026-11340-3

Keywords: cancer survivorship, bone density, DXA, osteoporosis, osteopenia, fracture risk, screening disparities, real-world cohort, cancer treatment, Delayed, selective, bone