MRONJ risk associated with combined antiresorptive and vascular endothelial growth factor receptor tyrosine kinase inhibitors Authors: G. Mammone, R. Coropciuc, T. Van den Wyngaert, R. Jacobs, P. Bossier, C. Kerkhofs, L. Wijnen, S. Meerschaert, E. Vissenaekens, P.M. Clement, O. Bechter, P.R. Debruyne, A. Laenen, B. Beuselinck Published: 2026 ABSTRACT BackgroundConcomitant use of vascular endothelial growth factor receptor tyrosine kinase inhibitors (VEGFR-TKIs) and bone resorption inhibitors (BRIs) may increase the risk of medication-related osteonecrosis of the jaw (MRONJ). The relative contribution of VEGFR-TKI, BRI type, and patient-related risk factors remains unclear.MethodsWe retrospectively reviewed patients treated with BRIs, with or without concomitant VEGFR-TKIs. The primary endpoint was MRONJ-free survival, defined as time from BRI initiation to MRONJ diagnosis. Secondary endpoints included MRONJ incidence and skeletal-related events (SREs).ResultsOverall, 233 patients received BRI/VEGFR-TKI combination (study group) and 986 received BRI alone (control group). Median MRONJ-free survival was shorter in the study group than in controls (79 vs 202 months). However, after adjustment for smoking status, age, sex, and BRI type, concomitant VEGFR-TKI use was not independently associated with MRONJ-free survival (HR 1.3, 95% CI 0.8–2.1; p = 0.2), whereas denosumab use and active smoking remained independent predictors. VEGFR-TKI therapy was associated with earlier MRONJ occurrence among BP-treated patients (adjusted HR 3.22, 95% CI 1.08–9.56; p = 0.03), but not among denosumab-treated patients. MRONJ-incidence did not differ between BRI/VEGFR-TKI and BRI groups (14.6% vs 12.0%; p = 0.3), but BRI-exposure was shorter in the combination group. SRE occurrence was 69.5%, similar between BP- and denosumab-treated patients; however, denosumab-treated patients had longer exposure, resulting in a lower SRE rate (0.13 vs 0.39 events/month).ConclusionMRONJ risk depends on BRI type and individual patient profile. VEGFR-TKIs may accelerate MRONJ onset in BP-treated patients, whereas denosumab showed higher MRONJ risk independently of VEGFR-TKI use. Treatment choices should balance MRONJ risk against skeletal disease burden. Introduction Bone resorption inhibitors (BRIs), such as denosumab, a monoclonal antibody targeting the receptor activator of nuclear factor kappa-B ligand, and bisphosphonates (BPs), which suppress osteoclast-mediated bone resorption, have been shown to reduce the risk of skeletal-related events (SREs) across different malignancies (Henry et al., 2011). Medication-related osteonecrosis of the jaw (MRONJ) is defined as exposed bone, or bone that can be probed through an intraoral or extraoral fistula in the maxillofacial region, persisting for at least 8 weeks, in patients with current or prior exposure to BRIs and no history of radiation therapy or metastatic disease involving the jaws (Ruggiero et al., 2022). This clinically relevant adverse event is rare in patients treated for osteoporosis, with an incidence of approximately 0.01%–0.1% per year. In contrast, the incidence increases to 1%–2% among cancer patients after 1–2 years of treatment and may rise further to 1–15% in those with additional risk factors, such as prolonged therapy (Khan et al., 2015), or concomitant administration of vascular endothelial growth factor receptor tyrosine kinase inhibitors (VEGFR-TKIs). Additionally, patients with advanced malignancies treated with BRI and antiangiogenetic therapy who undergo surgery for MRONJ frequently report worse surgical prognosis compared with BRI alone (Zou et al., 2026), likely due to an impaired microcirculation affecting more the patients treated with combinations than patients receiving BRIs alone or healthy volunteers (Huang et al., 2025). The combination of BRIs and VEGFR-TKI is primarily employed in patients with metastatic renal cell carcinoma (mRCC) and thyroid carcinoma with bone metastases (Fusco et al., 2015; Christodoulou et al., 2009; Mauceri et al., 2019). In 2018, we published a series of 90 cancer patients treated concomitantly with VEGFR-TKIs and BRIs, mainly BPs, alongside a control series of 533 patients receiving BRIs without VEGFR-TKIs. Global MRONJ-incidence was about 11% in both groups, despite a threefold shorter median BRI-exposure in the combination group. Concomitant treatment was associated with a shorter MRONJ-free survival even after correction for covariates including smoking status, infectious dental setting and type of medication (Van Cann et al., 2018). Differently from thyroid cancer management, clinical practice treating mRCC patients has changed over the last decade, with the introduction of nivolumab immunotherapy in second-line (Motzer et al., 2015), which potentially reduces the duration of the concomitant BRI/VEGFR-TKI treatment. Furthermore, first-line combinations of immune checkpoint inhibitors and VEGFR-TKIs have become standard of care in mRCC (Motzer et al., 2018, Motzer et al., 2024; Choueiri et al., 2021; Rini et al., 2019), resulting in improved disease control and potentially prolonged concurrent exposure to VEGFR-TKIs and BRIs. In parallel, denosumab has largely replaced BPs and is associated with a higher incidence of MRONJ, even in patients not receiving VEGFR-TKIs (Limones et al., 2020; Stopeck et al., 2016; Higuchi et al., 2018). Taken together, these changes in clinical practice warrant an updated evaluation of the risk of MRONJ in this evolving therapeutic context. Patients and methods Patients with bone metastases from solid tumors treated with BRIs, either BPs or denosumab, between March 2000 and July 2025 at the University Hospitals Leuven (Leuven, Belgium), were identified through the pharmacy records and included as control group. For the study group, we retrospectively reviewed the medical records of patients treated with BPs or denosumab in combination with VEGFR-TKIs between January 2005 and July 2025, following the introduction of VEGFR-TKIs. These patients were treated at the Department of General Medical Oncology of the University Hospitals Leuven (Leuven, Belgium) and at the Department of Medical Oncology of AZ Groeninge (Kortrijk, Belgium). The study protocol was approved by the Ethics Committees of both institutions (UZ Leuven S63833; AZ Groeninge AZGS2014077). Among patients treated with both BRIs and VEGFR-TKIs, treatment initiation was not always synchronous, which may limit the precise assessment of the incremental impact of VEGFR-TKIs on MRONJ-incidence. Therefore, patients receiving concomitant VEGFR-TKIs were subdivided into a “together” group, in which VEGFR-TKIs were initiated within three months of BRI initiation, and an “on-top” group, in which VEGFR-TKIs were initiated more than three months after the first BRI administration. The “together” group allows a more accurate estimation of the MRONJ risk with concomitant VEGFR-TKIs and BRIs. The primary endpoint was MRONJ-free survival, defined as the time from initiation of BRI therapy to the diagnosis of MRONJ. At the onset of MRONJ, patients stopped BRIs. Patients who did not develop MRONJ were censored at the date of death or last follow-up, whichever occurred first. In some patients, BRIs were stopped without occurrence of MRONJ, following the discretion of the treating physician. Secondary endpoints included the overall MRONJ-incidence, the MRONJ-incidence according to year of BRI-exposure, and the occurrence of SREs (see Supplemental patients and methods for further details). MRONJ-free survival was selected as the primary endpoint because MRONJ-incidence is influenced by BRI-exposure, defined as the interval in months between initiation and discontinuation of BRI therapy. A time-to-event analysis appropriately accounts for differences in treatment exposure over time. With the introduction of nivolumab, the duration of concomitant VEGFR-TKI and BRI therapy, defined as the period of simultaneous administration of both agents, became shorter than the total duration of BRI therapy. In the present analysis, we primarily evaluated the impact of the total BRI-exposure. When feasible, we additionally explored the effect of concomitant exposure in the subgroup of patients treated with denosumab, restricting the analysis to events occurring during the period of combined BRI/VEGFR-TKI therapy. Patients were then censored for the outcome of MRONJ-development at the end of the concomitant exposure period. From 2005 onwards, dental check-up before the start of BRIs was implemented as standard of care in the participating centers. The dental check-up before the start of BRIs included checking for caries, dental degeneration, underlying abscesses and gingival pathology. The oral examination was supplemented by a panoramic X-ray, if indicated. No routine X-rays were performed in the control group. If the dental assessment revealed contraindications, BRIs were not started. If dental extractions were necessary, BRI-treatment was started after complete soft-tissue healing. During BRI-treatment, there was no systematic screening for MRONJ, but MRONJ diagnostic was initiated once patients reported symptoms. Diagnosis of MRONJ was confirmed in most cases by an experienced maxillofacial surgeon of the University Hospitals Leuven. In few other cases, the diagnosis was made by the patient's peripheral dentist (see Supplemental patients and methods for more information). MRONJ-free survival from the start of BRI-exposure was estimated using the Kaplan-Meier method and estimates were compared with the Mantel-Cox log-rank test. Fisher exact test was used to compare proportions. A p-value <0.05 was considered statistically significant. Multivariable Cox proportional hazards models were used to analyze the independent prognostic effect of given predictors for the risk of MRONJ, with results presented as hazard ratios (HR) with 95% confidence intervals (CI). Missing values were not imputed. All statistical analyses were based on available data, and cases with missing information for a given variable were excluded from the corresponding analysis. The statistical analyses were conducted using GraphPad Prism version 10.5 (GraphPad Software, La Jolla, USA), SAS software (version 9.4 of the SAS System for Windows) and R software (version 4.5.0). Results Included patients We identified 233 patients treated concomitantly with BRIs and VEGFR-TKIs, including 170 in the “together” group and 63 in the “on top” group. Median follow-up was 21 months. mRCC represented 85% of the underlying malignancies. The most used VEGFR-TKI was sunitinib, followed by pazopanib, axitinib, cabozantinib and sorafenib. Median BRI-exposure was 13 months. In the “together” group 105 patients received denosumab, with a median total and concomitant exposure of 12 and 7 months, respectively (58% of total exposure). 65 patients received BPs; of these 5 switched to denosumab after a median BP-exposure of 21 months, followed by an additional median exposure of 6 months. Median total and concomitant BRI-exposure were 4 and 3 months (75% of total exposure). Within the “on top” subgroup, 39 patients started VEGFR-TKIs, after a median denosumab-exposure of 15 months. Median total and concomitant denosumab-exposure were 31 and 10 months respectively (32% of total exposure). 24 patients started VEGFR-TKIs after a median BP-exposure of 12 months. Median total and concomitant BRI-exposure were 25.5 and 7.5 months, respectively (29% of total exposure). One patient after a BP-exposure of 24 months switched to denosumab with a further 14 months of treatment. The control cohort comprises 986 patients, mainly with breast and prostate cancer, Median follow-up was 29 months and the median BRI-exposure 22 months. Within this cohort, patients were treated with BPs (n = 457), denosumab (n = 415) or BPs followed by denosumab (n = 114). Median exposures were 18, 12 and 39 months, respectively. Fig. 1 provides an overview of the study and control groups, together with the main comparisons performed in the analysis. Patient characteristics are reported in Table 1.Fig. 1Study design and summary of subgroup comparisons. Panel A: Composition of the study groups, including patients treated with BRIs and concomitant VEGFR-TKIs. Panel B: Composition of the control groups, including patients treated with BRIs without VEGFR-TKIs. Panel C: Summary of the main comparisons between study and control groups. The color code indicates whether concomitant VEGFR-TKI exposure was independently associated with MRONJ risk in multivariable analysis: blue indicates no independent association, whereas red indicates an independent association. Hazard ratios and 95% confidence intervals were estimated using Cox proportional hazards models.°Together: VEGFR-TKIs were initiated within three months of BRI initiation.*On top: VEGFR-TKIs were initiated more than three months after the first BRI administration.Abbreviations: BP = bisphosphonate; BRI = bone resorption inhibitor; Den = denosumab; MRONJ = medication-related osteonecrosis of the jaw; TKI = VEGFR-TKI; VEGFR-TKI = vascular endothelial growth factor receptor tyrosine kinase inhibitor.Fig. 1Table 1Patient characteristics.Control groupsStudy groupsPooled control groupDENBPsBPs > DENPooled study groupStudy group togetherDEN + TKIs togetherTKIs on top of DENBPs + TKIs togetherTKIs on top of BPsNumber of included patients (n)986415457114233170105396524Gender: male (%)334/986(33.9%)259/415(62.4%)162/457(35.4%)16/114(14%)160/233(68.6%)116/170(68.2%)74/105(70.4%)27/39(69.2%)42/65(64.6%)17/24(70.8%)Median age at start BRI (range)64(24–91)65(24–91)63(24–89)64.5(29–84)63(21–88)63(21–88)64(33–88)66(34–80)62(21–83)62.5(31–85)Median months of BRI (range)22(1–251)23(1–151)18(1–251)39(3–150)13(1–121)8(1–113)12(1–113)31(9–121)4(1–38)25.5(7–86)Median months of concomitant BRI + VEGFR-TKI (range)––––7(1–113)6(1–113)7(1–113)10(1–57)3(1–37)7.5(2–45)Total duration of BRI-exposure (months)30,49312,63312,71851424875256920471589522717Total duration of concomitant BRI and VEGFR-TKI exposure (months)––––266117861376562410313Survival status: alive (%) at study closure150/986(15.2%)90/415(21.7%)55/457(12.0%)5/114(4.4%)32/233(13.7%)22/170(12.9%)20/105(19%)9/39(23%)2/65(3.1%)1/24(4.2%)Median follow-up (months) since start of BRI (range)29(1–247)25(1–151)23(1–247)51(2–180)21(0–235)13(1–235)17(1–118)46(10–150)8(1–235)47(8–170)Smoking status (n. %)Never468/848(55.2%)205/379(54%)201/370(54.3%)67/99(67.7%)72/185(38.9%)52/144(36.1%)33/92(35.8%)13/29(44.8%)19/52(36.5%)7/12(58.3%)Former236/848(27.8%)111/379(29.2%)105/370(28.3%)17/99(17.2%)81/185(43.8%)65/144(45.1%)41/92(44.5%)12/29(41.4%)24/52(46.1%)4/12(33.4%)Active144/848(17%)63/379(13.9%)64/370(17.3%)15/99(15.2%)32/185(17.3%)27/144(18.8%)18/92(19.5%)4/29(13.8%)9/52(17.3%)1/12(8.3%)Data on MRONJ Number of MRONJ (n. %)119/986(12%)66/415(15.9%)32/457(7.0%)21/114(18.4%)34/233(14.6%)20/170(11.7%)15/105(14.3%)7/39(17.9%)5/65(7.7%)7/24(29.2%) Smoking status (n. %)Never58/111(52.2%)31/64(48.4%)17/28(60.7%)9/19(47.4%)13/30(43.4%)10/20(50%)8/15(53.3%)1/5(20.0%)0/5(0%)2/5(40.0%)Former27/111(24.3%)17/64(26.5%)5/28(17.8%)6/19(31.6%)14/30(46.6%)9/20(45%)6/15(40%)3/5(60.0%)3/5(60.0%)2/5(40.0%)Active26/111(23.4%)16/64(36.3%)6/28(21.4%)4/19(21.1%)3/30(10%)1/20(5%)1/15(6.7%)1/5(20.0%)2/5(40.0%)1/5(20%) Dental extraction prior to MRONJ (n. %)55/105(52.4%)31/58(53.4%)16/27(59.2%)7/18(38.9%)13/28(46.4%)11/19(57.8%)10/14(71.4%)1/5(20.0%)1/5(20%)1/4(25%) Median age at diagnosis MRONJ (range)67(46–89)67(48–89)60.5(48–69)73(46–83)64(31–85)64(31–85)64(58–85)69(53–84)54(31–78)54(53–75) Gender: male (n. %)37/119(31.1%)20/66(30.3%)4/32(12.5%)7/21(33.3%)24/34(70.5%)15/20(75%)13/15(86.6%)4/7(57.1%)2/5(40%)5/7(71.4%) Stage (n. %)138/76(50%)22/43(51.1%)11/21(52.4%)5/11(45.4%)13/25(52%)8/16(50%)8/12(66.6%)3/6(50%)0/4(0.0%)2/3(66.7%)223/76(30.3%)14/43(32.5%)4/21(19.0%)4/11(36.3%)11/25(44%)7/16(43.7%)3/12(25%)3/6(50%)4/4(100%)1/3(33.3%)311/76(14.5%)5/43(11.6%)5/21(23.8%)1/11(9%)0/25(0.0%)0/16(0.0%)0/12(0.0%)0/6(0.0%)0/4(0.0%)0/3(0.0%) Grade (n. %)131/81(38.3%)22/51(43.1%)8/19(42.1%)1/10(10%)8/24(33.3%)6/16(37.5%)6/12(50%)2/6(33.4%)0/4(0.0%)0/2(0.0%)222/81(27.1%)16/51(31.3%)3/19(15.8%)3/10(30%)9/24(37.5%)6/16(37.5%)4/12(33.3%)1/6(16.6%)2/4(50%)2/2(100%)310/81(12.3%)8/51(15.6%)0/19(0.0%)1/10(10%)6/24(25%)3/16(18.7%)1/12(8.3%)3/6(50.0%)2/4(50%)0/2(0.0%)414/81(17.3%)2/51(3.9%)5/19(26.3%)5/10(50%)0/24(0%)0/16(0%)0/12(0.0%)0/6(0.0%)0/4(0.0%)2/2(100%)Underlying malignancy (n; %) Renal cell carcinoma36/986(3.4%)23/415(5.5%)10/457(2.2%)3/114(2.6%)199/233(85.4%)150/170(88.2%)92/105(87.6%)28/39(71.7%)58/65(89.2%)21/24(87.5%) Sarcoma (including gastro-intestinal stromal tumors)20/986(2.3%)6/415(1.4%)13/457(2.8%)1/114(0.9%)11/233(4.7%)5/170(2.9%)1/105(0.95%)4/39(10.3%)4/65(6.2%)2/24(8.4%) Thyroid carcinoma36/986(3.6%)17/415(4%)16/457(3.5%)3/114(2.6%)16/233(6.8%)10/170(5.8%)8/105(7.6%)6/39(15.4%)2/65(3.1%)– Neuro-endocrine carcinoma3/986(0.2%)1/415(0.2%)2/457(0.4%)–2/233(0.9%)2/170(1.2%)2/105(0.02%)––– Paraganglioma1/986(0.1%)1/415(0.2%)––2/233(0.9%)2/170(1.2%)1/105(0.95%)–1/65(1.5%)– Breast carcinoma566/986(57.7%)225/451(54.2%)249/457(54.4%)92/114(80%)1/233(0.4%)––––1/24(4.1%) Prostate carcinoma189/986(18.6%)74/415(17.8%)104/457(22.9%)11/114(9.6%)–––––– Lung carcinoma (small cell and non-small cell)44/986(4.7%)34/415(8.2%)6/457(1.3%)4/114(3.5%)–––––– Malignant melanoma25/986(2.7%)12/415(2.8%)13/457(2.8%)––––––– Bladder carcinoma12/986(1.2%)3/415(0.7%)9/457(1.9%)––––––– Head and neck squamous cell carcinoma16/986(1.7%)6/415(1.4%)10/457(2.2%)–––––––Other tumor types38/986(3.8%)13/415(3%)25/457(5.5%)–2/233(0.9%)1/170(0.6%)1/105(0.9%)1/39(2.6%)––BRI (n; %) Bisphosphonates457/986(46.3%)–457/986(100%)83/233(35.6%)60/170(35.3%)––60(92.3%)23(96%) Denosumab415/986(42.1%)415/986(100%)–144/233(61.8%)105/170(61.8%)105(100%)39(100%)–– Bisphosphonates followed by denosumab114/986(11.6%)–114/986(100%)6/233(2.6%)5/170(2.9%)––5(7.7%)1(4%)VEGFR-TKI (n; %) Sunitinib––––102/233(43.7%)79/170(46.5%)30/105(28.5%)6/39(15.3%)49/65(75.3%)17/24(70.8%) Pazopanib––––62/233(27%)43/170(25.3%)36/105(34.2%)16/39(41%)7/65(10.7%)3/24(12.5%) Axitinib––––63/233(27%)51/170(30%)45/105(42.8%)11/39(28.2%)6/65(9.2%)1/24(4%) Sorafenib––––28/233(12%)18/170(10.6%)10/105(9.5%)3/39(7.6%)8/65(12%)7/24(29%) Vandetanib––––4/233(1.7%)3/170(1.7%)3/105(2.8%)1/39(4.2%)–– Regorafenib––––1/233(0.42%)1/170(0.6%)––1/65(1.5%)– Tivozanib––––1/233(0.42%)1/170(0.6%)1/105(0.9%)–– Cabozantinib––––53/233(22.7%)40/170(23.5%)39/105(37%)13/39(33.3%)1/65(1.5%)0/24(0%) AEE––––1/233(0.4%)––1/39(4.2%)–– Lenvatinib––––10/233(4.3%)4/170(2.3%)4/105(3.8%)6/39(15.3%)–– Nintedanib––––2/233(0.85%)1/170(0.6%)1/105(0.9%)1/39(4.2%)––Abbreviations: BP: bisphosphonates; BRIs: bone resorption inhibitors. DEN: denosumab; MRONJ: Medication Related Osteonecrosis Of The Jaw. SG: Study Group: TKIs: tyrosine kinase inhibitor; VEGFR: vascular endothelial growth factor receptor. Study design and summary of subgroup comparisons. Panel A: Composition of the study groups, including patients treated with BRIs and concomitant VEGFR-TKIs. Panel B: Composition of the control groups, including patients treated with BRIs without VEGFR-TKIs. Panel C: Summary of the main comparisons between study and control groups. The color code indicates whether concomitant VEGFR-TKI exposure was independently associated with MRONJ risk in multivariable analysis: blue indicates no independent association, whereas red indicates an independent association. Hazard ratios and 95% confidence intervals were estimated using Cox proportional hazards models. °Together: VEGFR-TKIs were initiated within three months of BRI initiation. *On top: VEGFR-TKIs were initiated more than three months after the first BRI administration. Abbreviations: BP = bisphosphonate; BRI = bone resorption inhibitor; Den = denosumab; MRONJ = medication-related osteonecrosis of the jaw; TKI = VEGFR-TKI; VEGFR-TKI = vascular endothelial growth factor receptor tyrosine kinase inhibitor. Table 1Patient characteristics.Control groupsStudy groupsPooled control groupDENBPsBPs > DENPooled study groupStudy group togetherDEN + TKIs togetherTKIs on top of DENBPs + TKIs togetherTKIs on top of BPsNumber of included patients (n)986415457114233170105396524Gender: male (%)334/986(33.9%)259/415(62.4%)162/457(35.4%)16/114(14%)160/233(68.6%)116/170(68.2%)74/105(70.4%)27/39(69.2%)42/65(64.6%)17/24(70.8%)Median age at start BRI (range)64(24–91)65(24–91)63(24–89)64.5(29–84)63(21–88)63(21–88)64(33–88)66(34–80)62(21–83)62.5(31–85)Median months of BRI (range)22(1–251)23(1–151)18(1–251)39(3–150)13(1–121)8(1–113)12(1–113)31(9–121)4(1–38)25.5(7–86)Median months of concomitant BRI + VEGFR-TKI (range)––––7(1–113)6(1–113)7(1–113)10(1–57)3(1–37)7.5(2–45)Total duration of BRI-exposure (months)30,49312,63312,71851424875256920471589522717Total duration of concomitant BRI and VEGFR-TKI exposure (months)––––266117861376562410313Survival status: alive (%) at study closure150/986(15.2%)90/415(21.7%)55/457(12.0%)5/114(4.4%)32/233(13.7%)22/170(12.9%)20/105(19%)9/39(23%)2/65(3.1%)1/24(4.2%)Median follow-up (months) since start of BRI (range)29(1–247)25(1–151)23(1–247)51(2–180)21(0–235)13(1–235)17(1–118)46(10–150)8(1–235)47(8–170)Smoking status (n. %)Never468/848(55.2%)205/379(54%)201/370(54.3%)67/99(67.7%)72/185(38.9%)52/144(36.1%)33/92(35.8%)13/29(44.8%)19/52(36.5%)7/12(58.3%)Former236/848(27.8%)111/379(29.2%)105/370(28.3%)17/99(17.2%)81/185(43.8%)65/144(45.1%)41/92(44.5%)12/29(41.4%)24/52(46.1%)4/12(33.4%)Active144/848(17%)63/379(13.9%)64/370(17.3%)15/99(15.2%)32/185(17.3%)27/144(18.8%)18/92(19.5%)4/29(13.8%)9/52(17.3%)1/12(8.3%)Data on MRONJ Number of MRONJ (n. %)119/986(12%)66/415(15.9%)32/457(7.0%)21/114(18.4%)34/233(14.6%)20/170(11.7%)15/105(14.3%)7/39(17.9%)5/65(7.7%)7/24(29.2%) Smoking status (n. %)Never58/111(52.2%)31/64(48.4%)17/28(60.7%)9/19(47.4%)13/30(43.4%)10/20(50%)8/15(53.3%)1/5(20.0%)0/5(0%)2/5(40.0%)Former27/111(24.3%)17/64(26.5%)5/28(17.8%)6/19(31.6%)14/30(46.6%)9/20(45%)6/15(40%)3/5(60.0%)3/5(60.0%)2/5(40.0%)Active26/111(23.4%)16/64(36.3%)6/28(21.4%)4/19(21.1%)3/30(10%)1/20(5%)1/15(6.7%)1/5(20.0%)2/5(40.0%)1/5(20%) Dental extraction prior to MRONJ (n. %)55/105(52.4%)31/58(53.4%)16/27(59.2%)7/18(38.9%)13/28(46.4%)11/19(57.8%)10/14(71.4%)1/5(20.0%)1/5(20%)1/4(25%) Median age at diagnosis MRONJ (range)67(46–89)67(48–89)60.5(48–69)73(46–83)64(31–85)64(31–85)64(58–85)69(53–84)54(31–78)54(53–75) Gender: male (n. %)37/119(31.1%)20/66(30.3%)4/32(12.5%)7/21(33.3%)24/34(70.5%)15/20(75%)13/15(86.6%)4/7(57.1%)2/5(40%)5/7(71.4%) Stage (n. %)138/76(50%)22/43(51.1%)11/21(52.4%)5/11(45.4%)13/25(52%)8/16(50%)8/12(66.6%)3/6(50%)0/4(0.0%)2/3(66.7%)223/76(30.3%)14/43(32.5%)4/21(19.0%)4/11(36.3%)11/25(44%)7/16(43.7%)3/12(25%)3/6(50%)4/4(100%)1/3(33.3%)311/76(14.5%)5/43(11.6%)5/21(23.8%)1/11(9%)0/25(0.0%)0/16(0.0%)0/12(0.0%)0/6(0.0%)0/4(0.0%)0/3(0.0%) Grade (n. %)131/81(38.3%)22/51(43.1%)8/19(42.1%)1/10(10%)8/24(33.3%)6/16(37.5%)6/12(50%)2/6(33.4%)0/4(0.0%)0/2(0.0%)222/81(27.1%)16/51(31.3%)3/19(15.8%)3/10(30%)9/24(37.5%)6/16(37.5%)4/12(33.3%)1/6(16.6%)2/4(50%)2/2(100%)310/81(12.3%)8/51(15.6%)0/19(0.0%)1/10(10%)6/24(25%)3/16(18.7%)1/12(8.3%)3/6(50.0%)2/4(50%)0/2(0.0%)414/81(17.3%)2/51(3.9%)5/19(26.3%)5/10(50%)0/24(0%)0/16(0%)0/12(0.0%)0/6(0.0%)0/4(0.0%)2/2(100%)Underlying malignancy (n; %) Renal cell carcinoma36/986(3.4%)23/415(5.5%)10/457(2.2%)3/114(2.6%)199/233(85.4%)150/170(88.2%)92/105(87.6%)28/39(71.7%)58/65(89.2%)21/24(87.5%) Sarcoma (including gastro-intestinal stromal tumors)20/986(2.3%)6/415(1.4%)13/457(2.8%)1/114(0.9%)11/233(4.7%)5/170(2.9%)1/105(0.95%)4/39(10.3%)4/65(6.2%)2/24(8.4%) Thyroid carcinoma36/986(3.6%)17/415(4%)16/457(3.5%)3/114(2.6%)16/233(6.8%)10/170(5.8%)8/105(7.6%)6/39(15.4%)2/65(3.1%)– Neuro-endocrine carcinoma3/986(0.2%)1/415(0.2%)2/457(0.4%)–2/233(0.9%)2/170(1.2%)2/105(0.02%)––– Paraganglioma1/986(0.1%)1/415(0.2%)––2/233(0.9%)2/170(1.2%)1/105(0.95%)–1/65(1.5%)– Breast carcinoma566/986(57.7%)225/451(54.2%)249/457(54.4%)92/114(80%)1/233(0.4%)––––1/24(4.1%) Prostate carcinoma189/986(18.6%)74/415(17.8%)104/457(22.9%)11/114(9.6%)–––––– Lung carcinoma (small cell and non-small cell)44/986(4.7%)34/415(8.2%)6/457(1.3%)4/114(3.5%)–––––– Malignant melanoma25/986(2.7%)12/415(2.8%)13/457(2.8%)––––––– Bladder carcinoma12/986(1.2%)3/415(0.7%)9/457(1.9%)––––––– Head and neck squamous cell carcinoma16/986(1.7%)6/415(1.4%)10/457(2.2%)–––––––Other tumor types38/986(3.8%)13/415(3%)25/457(5.5%)–2/233(0.9%)1/170(0.6%)1/105(0.9%)1/39(2.6%)––BRI (n; %) Bisphosphonates457/986(46.3%)–457/986(100%)83/233(35.6%)60/170(35.3%)––60(92.3%)23(96%) Denosumab415/986(42.1%)415/986(100%)–144/233(61.8%)105/170(61.8%)105(100%)39(100%)–– Bisphosphonates followed by denosumab114/986(11.6%)–114/986(100%)6/233(2.6%)5/170(2.9%)––5(7.7%)1(4%)VEGFR-TKI (n; %) Sunitinib––––102/233(43.7%)79/170(46.5%)30/105(28.5%)6/39(15.3%)49/65(75.3%)17/24(70.8%) Pazopanib––––62/233(27%)43/170(25.3%)36/105(34.2%)16/39(41%)7/65(10.7%)3/24(12.5%) Axitinib––––63/233(27%)51/170(30%)45/105(42.8%)11/39(28.2%)6/65(9.2%)1/24(4%) Sorafenib––––28/233(12%)18/170(10.6%)10/105(9.5%)3/39(7.6%)8/65(12%)7/24(29%) Vandetanib––––4/233(1.7%)3/170(1.7%)3/105(2.8%)1/39(4.2%)–– Regorafenib––––1/233(0.42%)1/170(0.6%)––1/65(1.5%)– Tivozanib––––1/233(0.42%)1/170(0.6%)1/105(0.9%)–– Cabozantinib––––53/233(22.7%)40/170(23.5%)39/105(37%)13/39(33.3%)1/65(1.5%)0/24(0%) AEE––––1/233(0.4%)––1/39(4.2%)–– Lenvatinib––––10/233(4.3%)4/170(2.3%)4/105(3.8%)6/39(15.3%)–– Nintedanib––––2/233(0.85%)1/170(0.6%)1/105(0.9%)1/39(4.2%)––Abbreviations: BP: bisphosphonates; BRIs: bone resorption inhibitors. DEN: denosumab; MRONJ: Medication Related Osteonecrosis Of The Jaw. SG: Study Group: TKIs: tyrosine kinase inhibitor; VEGFR: vascular endothelial growth factor receptor. Patient characteristics. Abbreviations: BP: bisphosphonates; BRIs: bone resorption inhibitors. DEN: denosumab; MRONJ: Medication Related Osteonecrosis Of The Jaw. SG: Study Group: TKIs: tyrosine kinase inhibitor; VEGFR: vascular endothelial growth factor receptor. Active smoking, older age, and male sex are associated with higher MRONJ risk As reported in Table 1, the prevalence of active smokers (range 8.3–19.2%) and the proportion of patients with a dental extraction before the development of MRONJ (range 20–71.4%) were very variable between the control and the study groups. Smoking-status was known in 1033 patients. MRONJ-free survival was significantly shorter in active smokers compared to former and never smokers (global log-rank p = 0.007). MRONJ-incidence was 16.5% (95%CI: 10.99–21.97%), 12.9% (95%CI: 9.29–16.57%) and 13.1% (95%CI: 10.26–16.04%), respectively (Suppl. Fig. 1A). MRONJ-free survival was significantly shorter in older patients (global log-rank p = 0.04). When comparing patients over 40 years of age to those aged 20–39 years old, the HR for MRONJ was 2.4 (95% CI: 1.4–3.9; p = 0.001) (Suppl. Fig. 1B). MRONJ-free survival was significantly shorter for men than for women (HR 1.8; 95%CI: 1.2–2.6; p = 0.0005). Global MRONJ-incidence was the same: 13.6% in both groups, but median BRI-exposure was 14 months in men and 25 months in women (Suppl. Fig. 1C). The type of antiresorptive therapy influences MRONJ risk and timing In the denosumab control group (n = 415), after a median BRI-exposure of 23 months, 66 MRONJ events occurred (15.9%; 95% CI 12.6–19.9), with a median MRONJ-free survival of 124 months. In the BP control group (n = 457), after a median BRI-exposure of 18 months, 32 MRONJ events occurred (7.0%; 95% CI 4.9–9.8), with a median MRONJ-free survival of 202 months. Compared with BPs alone, denosumab was associated with significantly shorter MRONJ-free survival (HR 2.0; 95% CI 1.3–3.0; p = 0.0008) and higher MRONJ-incidence (OR 2.5; 95% CI: 1.6–3.9; p < 0.001) (Suppl. Fig. 2A–C; Suppl. Table 1). We then assessed whether concomitant VEGFR-TKI therapy further increased MRONJ risk. In the overall concomitant-cohort, patients receiving BRIs and VEGFR-TKIs (n = 233) had shorter MRONJ-free survival than controls receiving BRIs alone (n = 986; median 79 vs 202 months; HR 1.9; 95% CI 1.2–3.0; p = 0.0007). However, this association lost statistical significance after adjustment for covariates, including smoking status, type of BRI, age at start BRI and sex (HR 1.3; 95%CI 0.8–2.10; p = 0.2). In contrast, denosumab remained independently associated with increased MRONJ risk (HR 1.5; 95% CI 1.0–2.2; p = 0.04), together with smoking status and age at BRI initiation (Fig. 2A, B: Suppl. Table 2). Similar findings were observed in the subgroup of patients who initiated BRIs and VEGFR-TKIs “together”. No significant differences in MRONJ incidence were observed between the overall concomitant-treatment cohort and controls, nor between the “together” cohort and controls, despite shorter BRI-exposure in the study group in both comparisons. (Suppl. Fig. 3A–D, Suppl. Table 2).Fig. 2Panel A. MRONJ-free survival analysis in patients treated concomitantly with BRIs and VEGFR-TKI compared to those treated with BRIs alone. Panel B: Forest plot for multivariable analysis: all pooled study groups vs all pooled control groups.Abbreviations: BP = bisphosphonate; BRI = bone resorption inhibitor; Den = denosumab; MRONJ = medication-related osteonecrosis of the jaw; TKI = VEGFR-TKI; VEGFR-TKI = vascular endothelial growth factor receptor tyrosine kinase inhibitor.Fig. 2 Panel A. MRONJ-free survival analysis in patients treated concomitantly with BRIs and VEGFR-TKI compared to those treated with BRIs alone. Panel B: Forest plot for multivariable analysis: all pooled study groups vs all pooled control groups. Abbreviations: BP = bisphosphonate; BRI = bone resorption inhibitor; Den = denosumab; MRONJ = medication-related osteonecrosis of the jaw; TKI = VEGFR-TKI; VEGFR-TKI = vascular endothelial growth factor receptor tyrosine kinase inhibitor. Overall, these findings suggested that denosumab is associated with higher MRONJ-incidence and earlier MRONJ occurrence compared with BPs alone, and that MRONJ risk in the overall cohort appears to be driven more by BRI type and clinical risk factors than by concomitant VEGFR-TKI use. We therefore performed subgroup analyses according to BRI type. Bisphosphonate and VEGFR-TKI combination increases MRONJ risk In the BP control group MRONJ incidence was 7.0% (95%CI: 4.9–9.8) and median BP-exposure was 18 months (Suppl. Tables 3, 4). Excluding the patients who switched to denosumab, 83 patients were treated with BPs and VEGFR-TKI, including 60 patients in the “together” group and 23 in the “on top” group. In this study group MRONJ-incidence was 9.6% (95%CI: 2.84–15.1), without significant difference with the control group (p = 0.36), although a similar incidence was achieved with a markedly shorter median BP-exposure in the combination group (7 months; p < 0.0001). Consistently, in a time-to event analysis, median MRONJ-free survival was 58 months in the study group and 202 months in the BP-control group (HR 3.6; 95%CI: 1.02–12.5; p < 0.0001). In multivariate analysis (MVA), concomitant VEGFR-TKIs was the only variable that remained independently associated with MRONJ-free survival (HR 3.22; 95%CI: 1.08–9.56; p = 0.03) (Fig. 3A–C).Fig. 3Panel A: MRONJ-incidence by duration of exposure in patients treated with VEGFR-TKIs together and on top of BPs (excluding patients who switched to denosumab) compared to patients treated with BPs during the first, second year of exposure and beyond. Panel B: MRONJ-free survival in patients treated with VEGFR-TKIs together and on top of BPs (excluding patients who switched to denosumab) compared to patients treated with BPs. Panel C: Forest plot for multivariable analysis: patients treated with VEGFR-TKI together and on top of BPs (excluding patients who switched to denosumab) vs patients treated with BPs.Abbreviations: BP = bisphosphonate; MRONJ = medication-related osteonecrosis of the jaw; TKI = VEGFR-TKI; VEGFR-TKI = vascular endothelial growth factor receptor tyrosine kinase inhibitor.Fig. 3 Panel A: MRONJ-incidence by duration of exposure in patients treated with VEGFR-TKIs together and on top of BPs (excluding patients who switched to denosumab) compared to patients treated with BPs during the first, second year of exposure and beyond. Panel B: MRONJ-free survival in patients treated with VEGFR-TKIs together and on top of BPs (excluding patients who switched to denosumab) compared to patients treated with BPs. Panel C: Forest plot for multivariable analysis: patients treated with VEGFR-TKI together and on top of BPs (excluding patients who switched to denosumab) vs patients treated with BPs. Abbreviations: BP = bisphosphonate; MRONJ = medication-related osteonecrosis of the jaw; TKI = VEGFR-TKI; VEGFR-TKI = vascular endothelial growth factor receptor tyrosine kinase inhibitor. These findings were explored in the 65 patients who initiated VEGFR-TKIs and BPs “together”, where 5 cases of MRONJ occurred (7.7%; 95%CI: 2.9–17.8%; p = 0.8) with a median duration of BRI-exposure of 4.0. Median MRONJ-free survival was shorter in this subgroup than in the BP control group: 38 versus 202 months, respectively (log-rank p < 0.0001). Of note, 3 of the 5 MRONJ cases occurred in patients who switched from BPs to denosumab. These events developed shortly after the switch, after a median of 6 months, and followed a longer cumulative BP-exposure, median 15 months. Although the late switch to denosumab may have introduced potential confounding, these patients were retained in the study group to avoid selection bias related to excluding patients with sequential BRI-exposure. However, given the limited number of MRONJ events in this subgroup, these findings should be interpreted with caution. In contrast with the other subgroups, a higher MRONJ-incidence was observed among the 24 patients treated with BP plus VEGFR-TKI “on-top” compared with the control group (29.2%; 95% CI 13.4–51.2%) versus 7.0% (95% CI 4.9–9.8%; p < 0.0001). This finding may be partly explained by longer cumulative BRI-exposure in this subgroup; however, the small sample size limits its interpretability. Denosumab use independently increases MRONJ risk, regardless of VEGFR-TKI use In the denosumab control group, MRONJ-incidence was 15.9% (95% CI: 12.6–19.9). With similar median denosumab-exposure in the study and control groups (18 vs 23 months; p = 0.07), overall MRONJ-incidence among the 144 patients treated with denosumab and VEGFR-TKIs was 15.2% (95% CI: 9.3–21.0), not different from the denosumab control group (p = 0.8) (Suppl. Tables 5, 6). HR for MRONJ-free survival was 1.2 (95%CI: 0.7–1.9) and this difference was not significant. In MVA, only smoking status and older age at start of BRIs were independently correlated with shorter MRONJ-free survival (Fig. 4A–C). Among patients treated with denosumab and VEGFR-TKIs, 13% (n = 14) also received concomitant ICIs. The incidence of MRONJ in this subgroup was 14.3% (n = 2/14).Fig. 4Panel A: MRONJ-incidence by duration of exposure in patients treated with VEGFR-TKI together and on top of denosumab compared to patients treated with denosumab during the first, second year of exposure and beyond. Panel B: MRONJ-free survival in patients treated with VEGFR-TKI together and on top of denosumab compared to patients treated with denosumab. Panel C: Forest plot for multivariable analysis: patients treated with VEGFR-TKI together and on top of denosumab vs patients treated with denosumab.Abbreviations: Den = denosumab; MRONJ = medication-related osteonecrosis of the jaw; TKI = VEGFR-TKI; VEGFR-TKI = vascular endothelial growth factor receptor tyrosine kinase inhibitor.Fig. 4 Panel A: MRONJ-incidence by duration of exposure in patients treated with VEGFR-TKI together and on top of denosumab compared to patients treated with denosumab during the first, second year of exposure and beyond. Panel B: MRONJ-free survival in patients treated with VEGFR-TKI together and on top of denosumab compared to patients treated with denosumab. Panel C: Forest plot for multivariable analysis: patients treated with VEGFR-TKI together and on top of denosumab vs patients treated with denosumab. Abbreviations: Den = denosumab; MRONJ = medication-related osteonecrosis of the jaw; TKI = VEGFR-TKI; VEGFR-TKI = vascular endothelial growth factor receptor tyrosine kinase inhibitor. Similar results were observed in the subgroup of 105 patients who started VEGFR-TKI and denosumab “together”, in which global MRONJ-incidence did not differ between the study group (14.3%; 95%CI: 8.5–22.8%) and the denosumab control group (p = 0.7), with significantly shorter median BRI-exposure in the study group (12 vs 23 months: p < 0.0001). Nevertheless, MRONJ-free survival was comparable and in MVA, only smoking status remained independently associated with MRONJ-free survival (Suppl. Fig. 4A–D). In the “on top” subgroup (n = 39), even if the median BRI-exposure was significantly longer in the study group than in the control group (31 vs 23 months, p = 0.002), the overall MRONJ-incidence in the study subgroup was 17.9% (95%CI: 8.1–34.1) not different from the control group (p = 0.90), as well as MRONJ-free survival (HR 0.8; 95%CI: 0.4–1.7; p = 0.6) (Suppl. Fig. 5A, B). A further confirmation of the enhanced MRONJ risk related to denosumab itself came from the analysis on concomitant exposure. In the denosumab “together” group, concomitant exposure was only 58% of the total exposure. In 30 out of 105 patients, the BRI was continued after the stop of the VEGFR-TKI for a median of 11 months (range 4–86). In the analysis adjusted for concomitant exposure, HR for MRONJ increased to 2.0 (95%CI: 0.9–4.5; p = 0.001), losing significance after correction for covariates, suggesting that denosumab itself may play a more important role than the combination (Suppl. Fig. 6A–C). Denosumab shows a lower frequency of SREs despite longer treatment duration Data on SREs during concomitant VEGFR-TKI administration was available in 128 out of the 170 patients treated “together” with BRIs and VEGFR-TKIs. The observation period covered a median of 6 months (range 1–69) and a total of 1389 months of concomitant therapy. Treatment groups comprised 53 patients receiving BPs, 70 receiving denosumab, and 5 who switched from BPs to denosumab during VEGFR-TKI therapy. Eighty-nine patients (69.5%) experienced at least one SRE during concomitant therapy, with a total of 248 SREs recorded (mean: 2.76 SREs per patient; SRE rate: 0.18 events per month). The distribution was: 41.6% had one SRE, 36.0% had two to four SREs, and 22.5% had five or more SREs (Supplemental Table 7). The most common SRE was the need for radiation therapy (n = 58; 45.3%), followed by pathological fracture (n = 36; 28.1%), hypercalcemia (n = 27; 21.1%), spinal cord compression (n = 11; 8.6%), and orthopedic surgery (n = 10; 7.8%) (Fig. 5A).Fig. 5Panel A: SRE incidence of each subtype in mRCC patients treated concomitantly with BRIs and VEGFR-TKIs. Panel B: SRE-incidence of each subtype in mRCC patients treated concomitantly with BRIs and VEGFR-TKIs comparing the bisphosphonates group and the denosumab group (5 patients were treated sequentially with bisphosphonates followed by denosumab).Abbreviations: BP = bisphosphonate; Den = denosumab; RCC = renal cell carcinoma; SRE = Skeletal related events.Fig. 5 Panel A: SRE incidence of each subtype in mRCC patients treated concomitantly with BRIs and VEGFR-TKIs. Panel B: SRE-incidence of each subtype in mRCC patients treated concomitantly with BRIs and VEGFR-TKIs comparing the bisphosphonates group and the denosumab group (5 patients were treated sequentially with bisphosphonates followed by denosumab). Abbreviations: BP = bisphosphonate; Den = denosumab; RCC = renal cell carcinoma; SRE = Skeletal related events. Among the 5 patients who switched from BPs to denosumab during VEGFR-TKI therapy (median duration: 21 months; total: 91 months), 3 patients (60%) experienced one SRE each: one required palliative radiotherapy, one had a pathological fracture, and one developed hypercalcemia. In the BPs group (n = 53), 39 patients (73.6%) developed at least one SRE, totaling 117 events (mean: 2.93 per patient). With a median treatment duration of 3 months (total: 301 months), the SRE rate was 0.39 events per month. In the denosumab group (n = 70), 47 patients (67.1%) experienced at least one SRE, totaling 128 events (mean: 2.72 per patient). The median treatment duration was 9.5 months (total: 994 months), yielding an SRE rate of 0.13 events per month, 67% lower than with BPs despite three times longer exposure (Fig. 5B). Discussion In this large real-world cohort, we confirm that MRONJ remains a clinically relevant complication in patients receiving BRIs, with risk influenced by treatment duration, BRI type, clinical factors and concomitant VEGFR-TKI therapy in selected subgroups. Previous retrospective studies suggested an increased MRONJ risk in cancer patients receiving BRIs combined with VEGFR-targeted therapies, with reported incidences ranging from 6% to 17% (Higuchi et al., 2018; Beuselinck et al., 2012; Smidt-Hansen et al., 2013; McKay et al., 2014; Guillot et al., 2019; Stellato et al., 2025; Srivastava et al., 2021). Mechanistically, VEGFR inhibition may impair local angiogenesis and mucosal repair, thereby facilitating MRONJ development in BRI-exposed patients. In a mouse model, Zhao et al. linked this effect to reduced gingival fibroblast proliferation and migration (Zhao et al., 2023). The available evidence on this topic remains limited by small sample sizes, heterogeneous antiangiogenic treatments, scarce data on denosumab, and, importantly, the lack of control groups in most studies. The MRONJ-incidence observed in our combination cohort was 11.7%, within the range described in previous studies. However, when compared with a large control cohort of cancer patients receiving BRIs without VEGFR-TKIs, overall MRONJ-incidence did not significantly differ between the study and control group. In this control cohort, denosumab was associated with a higher MRONJ risk than BPs, a finding that differs from pivotal randomized trials comparing denosumab and zoledronic acid, in which MRONJ-incidence was low, approximately 1–2%, and similar between treatment arms (Lipton et al., 2004). This discrepancy may be explained by shorter follow-up, selected trial populations, and lower cumulative exposure in randomized trials, whereas real-world cohorts with longer exposure may better capture the cumulative MRONJ risk associated with denosumab. The large control cohort and time-to-event approach allowed us to show that earlier MRONJ occurrence with concomitant BRIs and VEGFR-TKIs was evident in unadjusted analyses, but remained significant only among BP-treated patients after adjustment and subgroup analyses. Consistent with our findings, a recent mRCC study reported an overall MRONJ-incidence of 11.6%, with most cases occurring in denosumab-treated patients. In that study, no MRONJ was observed in the small subgroup of 10 patients receiving ICI plus VEGFR-TKI combinations. In our cohort, among 14 patients treated with denosumab, VEGFR-TKIs, and ICIs, MRONJ-incidence was identical to that observed in patients treated with denosumab plus VEGFR-TKIs without ICIs (Stellato et al., 2025). Although based on a very small subgroup, the high proportion of MRONJ cases after switching from BPs to denosumab is noteworthy. Specifically, 3 out of the 5 patients who switched developed MRONJ, suggesting that denosumab may trigger or accelerate MRONJ after prior BP-exposure. A similar observation was reported in the study of Higuchi et al., in which nearly half of MRONJ cases occurred after switching from zoledronic acid to denosumab (Higuchi et al., 2018). On the other hand, SREs were frequent in our population, supporting the clinical need for antiresorptive treatment in patients with bone-metastatic mRCC. In the zoledronic acid trial, SREs occurred in 37% of zoledronic acid-treated patients compared with 74% of placebo-treated patients (Lipton et al., 2004). Previous retrospective studies align with this data (Saad and Lipton, 2005; Velev et al., 2023), while more recent data in the immunotherapy era reported SREs in approximately one-third of mRCC patients with bone metastases (Harada et al., 2021). In our cohort, the observed frequency of SREs over time was lower with denosumab despite longer treatment duration. This finding should be interpreted descriptively, but may reflect the stronger antiresorptive activity of denosumab compared with BPs reported in pivotal trial (Lipton et al., 2012) and/or differences in observation time and systemic treatment efficacy across treatment eras. The higher MRONJ risk associated with denosumab, regardless of VEGFR-TKI use, and with BP-VEGFR-TKI combinations should be balanced against the high burden of SREs, which may substantially impair quality of life. In patients with limited bone disease, without hypercalcemia or high risk of fracture or spinal cord compression, local treatment such as radiotherapy may be considered and BRI initiation delayed or withheld. Conversely, patients with extensive bone disease or high-risk skeletal lesions may derive greater clinical benefit from BRIs despite the potential MRONJ risk. Moreover, treatment choice should also consider the potential complexity of denosumab discontinuation if MRONJ develops. Since denosumab is not incorporated into the bone matrix, its antiresorptive effect may dissipate relatively quickly after discontinuation, leading to a “rebound phenomenon” that can clinically manifest with multiple vertebral fractures (Cheung et al., 2022). Additionally, several MRONJ risk factors are potentially modifiable, including active smoking and invasive dental procedures. In our series, MRONJ occurred after dental procedures in 57.8% of cases in the “together” subgroup and 22.2% in the “on-top” subgroup. Similarly, Guillot et al. reported dental extraction in 43% of patients developing MRONJ (Guillot et al., 2019), supporting the importance of preventive dental strategies. Perioperative antibiotics and the intraoperative use of autologous platelet concentrates have shown promising results in reducing MRONJ occurrence after dental extractions or dentoalveolar surgery (Coropciuc et al., 2023), although further evidence is needed (Fortunato et al., 2020; Govaerts et al., 2020). This study has limitations inherent to its retrospective design, including potential selection bias, incomplete information on dental risk factors, and heterogeneity in treatment exposure and follow-up. In addition, some subgroup analyses included a limited number of MRONJ events and should therefore be interpreted cautiously. Conclusion Our findings suggest that MRONJ risk depends on the type of BRI and the individual patient profile, and that concomitant VEGFR-TKI therapy may further accelerate MRONJ onset in patients receiving BPs. Denosumab was associated with a higher MRONJ risk than BPs, independently of VEGFR-TKI use. Treatment decisions should therefore be individualized, balancing MRONJ risk against skeletal disease burden and considering the expected benefit from antiresorptive therapy and the feasibility of preventive dental strategies. Moreover, our findings support smoking cessation counselling and close multidisciplinary monitoring in patients with MRONJ risk factors. CRediT authorship contribution statement G. Mammone: Writing – review & editing, Writing – original draft, Formal analysis. R. Coropciuc: Data curation. T. Van den Wyngaert: Writing – review & editing, Supervision. R. Jacobs: Data curation. P. Bossier: Data curation. C. Kerkhofs: Data curation. L. Wijnen: Data curation. S. Meerschaert: Data curation. E. Vissenaekens: Data curation. P.M. Clement: Writing – review & editing, Data curation. O. Bechter: Writing – review & editing, Data curation. P.R. Debruyne: Writing – review & editing, Data curation. A. Laenen: Methodology. B. Beuselinck: Writing – review & editing, Supervision, Conceptualization. Consent for publication All authors have read and approved the manuscript for publication. Patients' consent for publication of their data was obtained where applicable. Ethics approval and consent to participate All patients provided their written informed consent to participate. The study was approved by the Ethics Board of both institutions (UZL-S63833 and AZGS2014077). The study was performed in accordance with the Declaration of Helsinki. Funding There was no specific funding for this project. BB holds an FWO Vlaanderen senior clinical research mandate (1801525N). Declaration of competing interest The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: PRD reports a relationship with Astellas Pharma that includes: consulting or advisory. PRD reports a relationship with MSD that includes: consulting or advisory and travel reimbursement. PRD reports a relationship with Biocartis NV that includes: equity or stocks. PRD reports a relationship with Mural Oncology Inc. that includes: equity or stocks. BB reports a relationship with MSD that includes: consulting or advisory, paid expert testimony, and speaking and lecture fees. BB reports a relationship with Bristol-Myers Squibb Company that includes: consulting or advisory, funding grants, and speaking and lecture fees. BB reports a relationship with Ipsen that includes: consulting or advisory and speaking and lecture fees. PRD reports a relationship with Ipsen that includes: consulting or advisory and travel reimbursement. There was no specific funding for this project. BB holds an FWO Vlaanderen senior clinical research mandate (1,801,525 N). If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.