Hypothesis: In patients receiving osteoporosis-dose bisphosphonates or denosumab, a model that updates periodontal inflammation status and extraction timing over follow-up will discriminate medication-related osteonecrosis of the jaw (MRONJ) better than models based mainly on antiresorptive duration.

Why this is plausible:

• MRONJ usually emerges at the intersection of drug exposure and local wound-healing stress rather than from duration alone.
• AAOMS 2022 and the Khan international consensus both emphasize dental extraction, active oral disease, and nonhealing bone as central clinical triggers.
• Observational osteoporosis cohorts suggest denosumab and other potent antiresorptives do not create uniform risk; dental context appears to concentrate events.

Testable design:

• Prospective multicenter cohort of adults on osteoporosis-dose antiresorptives, enriched for glucocorticoid-treated autoimmune disease.
• Time-updated covariates: extraction/implant timing, periodontal status, denture trauma, smoking, diabetes control, steroid exposure, antiresorptive class and duration.
• Primary endpoint: adjudicated MRONJ at 12-24 months.
• Compare dynamic model vs duration-only baseline model using AUROC, calibration slope, and decision-curve analysis.

Falsifiable prediction:

• The dynamic oral-context model will improve AUROC by at least 0.10 and show better calibration in the upper-risk decile than duration-only triage.

Key references:

1. Ruggiero SL, et al. J Oral Maxillofac Surg. 2022;80:920-943. doi:10.1016/j.joms.2022.02.008
2. Khan AA, et al. J Bone Miner Res. 2015;30:3-23. doi:10.1002/jbmr.2405
3. Yamada S, et al. Sci Rep. 2022;12:7991. doi:10.1038/s41598-022-11615-9