Assessing the effects of immune checkpoint inhibitors on bone utilizing machine learning-assisted opportunistic quantitative computed tomography.

Abstract

Immune checkpoint inhibitors (ICIs) are widely used in cancer treatment, yet their impact on bone health remains unclear. This study aimed to perform a retrospective cohort study utilizing routine CT scans from patients with melanoma to perform opportunistic QCT analysis to investigate the effects of ICI treatment on skeletal health, including volumetric BMD (vBMD) measurements and osteoarthritis (OA) parameters. A previously established machine learning-assisted opportunistic QCT pipeline was used to estimate lumbar spine vBMD from baseline and 12-mo follow-up CT scans in patients with melanoma treated with ICI therapy and those not treated with ICI therapy. Facet joint OA, osteophyte formation, and endplate sclerosis were also graded. Independent and paired t tests were used to determine any differences in vBMD and OA parameters between ICI users and non-ICI users. Multivariable linear regression models were used to control for confounding variables. Non-ICI users had a significant decrease in vBMD of -6.96 mg/cm3 from baseline to follow-up, whereas the ICI users had no significant change. There was a significant difference in change in vBMD from baseline to follow-up between the 2 groups, with the non-ICI users experiencing a 11.22 mg/cm3 larger decrease in vBMD. After adjusting for baseline age, sex, baseline vBMD, and change in OA parameters, this difference remained significant at -13.04 mg/cm3. Among the ICI users, those who had a decline in vBMD had a lower baseline vBMD compared with those who had increased vBMD. Neither group showed a significant change in OA parameters over the follow-up period, nor a difference in change between ICI and non-ICI users, even after adjusting for sex, age, and baseline OA parameters. While the effects of ICI treatment on vBMD may vary based on baseline bone health, ICIs do not significantly impact OA parameters in the short term.