1. Zoledronate does not reduce the risk of treatment failure in osteosarcoma: results of the French multicentre OS2006 randomised trial L Brugières, MC Le Deley, F Rédini, P Marec-Bérard, H Pacquement, C Lervat, JC Gentet, N Entz-Werlé, B Bui, N Corradini, G de Pinieux, P Petit, K Buffard, JY Blay, S Piperno-Neumann 15-18 October 2014, Berlin

2. Disclosure s  Novartis  Chugaï 15-18 October 2014, Berlin

3. Zoledronate in osteosarcoma Preclinical models Rat -transplantable model of osteosarcoma - Z prevents the formation of bone osteolytic lesions and reduces local tumor growth - IFO+Z enhances tumor regression and tissue repair Lung metastases model in mice (IV injection of POS-1 murine osteosarcoma cells) - Z suppresses lung mets in vivo and prolongs overall survival of osteosarcoma-bearing mice - Z has a direct antitumoral effect on POS-1 cells in vitro Ory et al, Cancer (2005) implantation Z OL (100 µg/kg) sacrifice J0 J7 J14 J21 J28 J35 CZ I I+Z 15-18 October 2014, Berlin

4. OS2006 trial  Randomised phase III trial involving all French paediatric oncology and most adult sarcoma centres (overall 48 centres)  Objective: To evaluate the impact of the addition of a 10-month Zoledronate treatment to chemotherapy and surgery on the event-free survival of osteosarcoma patients  Primary endpoint : EFS 15-18 October 2014, Berlin

5. OS2006 - Inclusion criteria  All newly diagnosed high grade osteosarcoma except : Small cell osteosarcoma Maxillary osteosarcoma Extra-osseous osteosarcoma Patients with primary resection multiple metastases for whom complete removal was not expected to be feasible even after shrinkage with chemotherapy  Age > 5 years and < 50 years 15-18 October 2014, Berlin

6. OS2006 - Chemotherapy MTX-ETO-IFO according to OS94 protocol API-AI according to FSG protocol <18 years18-25 years > 25 years left to the choice of each center (Assi et al Curr Oncol 2010, Piperno-Neumann et al ASCO 2006) (Le Deley et al Eur J Cancer 2007) 15-18 October 2014, Berlin

7. OS2006 - Treatment plan 15-18 October 2014, Berlin MTX-ETO-IFO API-AI Lenograstim 263 µg D6-D12

8. Zoledronate  Randomisation at diagnosis between 2 arms with or without zoledronate  Dose of Zoledronate 10 monthly IV infusions: 4 before/ 6 after surgery Dose >25 years: 4 mg <18 years : 0.05 mg/kg (max 4 mg/ dose) 18-25 y: 0.05 mg/kg for the first 2 courses, then 4 mg Dose reduction if gr 3-4 hypocalcemia  Vitamin D3 and calcium supplementation in both arms 15-18 October 2014, Berlin

9. Statistical considerations  Randomisation stratified on:  age and type of chemotherapy => 4 strata  risk group : non metastatic and resectable versus metastatic or not resectable  centre  Sample size : 470 patients required to achieve  a 80%-power to detect a 13% improvement of 3 y-EFS (from 55% to 68%) in the Zoledronate arm (HR=0.65),  with a two-sided log-rank test (alpha=5%)  and 3 interim analyses  This is the results of the second interim analysis performed after the inclusion of 318 pts 15-18 October 2014, Berlin

10. Participant flow (Apr 2007-Feb 2014) Did not meet eligibility criteria N=70 No zoledronate (Z-) N=158 zoledronate (Z+) N=160 Suspension of randomisation N=17 Not included in the randomised trial N = 116 including 69 refusal Assessed for eligibility N=521 Included in the randomised trial N=318 (73%) Potentially eligible N=434 15-18 October 2014, Berlin

11. Z-Z+Total N=158N=160N=318% Age and planned chemotherapy 10911021969 Less than 18y – MTX-Eto-Ifo 18-25 years – MTX-Eto-Ifo 18193712 18-25 years – API-AI 1211237 > 25 years – API-AI 19203912 Risk group 13212926182 Non metastatic and primary resectable Non metastatic and primary non resectable 1341 Metastatic disease 25285317 Site of the primary (MD=33) Limb 13013226292 Axial 1310238 Size of the primary (MD=44) <10 cm 765813449 >10 cm 627814051 Alkaline phosphatases (MD=41) Normal level 807315355 Above the upper limit (> 1xULN) 596512445 LDH (MD=58) Normal level 787915760 Above the upper limit (> 1xULN) 544910340 Baseline characteristics

12. Zoledronate - Total administration  In Zoledronate arm:  55 patients had an omission of >1 injection, including 4 patients with no Zoledronate injection 4 patients who stopped Zoledronate after 1 st injection  Zoledronate dose = protocol dose (+/-10%) for 851/1013 injections (84%)  No patient allocated to Z- arm received Zoledronate 15-18 October 2014, Berlin

13. Toxicity during treatment 15-18 October 2014, Berlin  No significant increase of toxicity regarding the most frequent expected toxicities (i.e. hematotoxicity, infection, transfusion, ASAT/ALAT elevation, mucositis…)  Hypocalcemia grade 2-4 significantly more frequent in Z+ arm per course: OR = 7.2, 95%CI, 4.9 – 10.6, p<0.001 decreasing risk over time

14. Surgery and histological analysis of the primary tumour 15-18 October 2014, Berlin

15. OS2006 Event-Free (EFS) and Overall Survival (OS) 3-y OS= 78.3% 3-y EFS= 60.3% 15-18 October 2014, Berlin Median follow-up 3.1 years

16. Z+, 3-y EFS=58.3% (49-67) Z-, 3-y EFS= 62.3% (53-71) Z-, 3-y OS= 83.3% (75-89) Z+, 3-y OS=73.2% (64-81) Impact of Zoledronate on outcome Z-, N=158 Z+, N=160Total, N=318 Number of events4957106 Local progression/relapse4610 Metastatic progression/relapse283462 Combined161531 Progression/relapse NOS112 Death as 1st event011 HR=1.31 (CI: 0.79-2.18) p=0.17 15-18 October 2014, Berlin HR=1.42 (CI: 0.70-2.88), p=0.21

17. 15-18 October 2014, Berlin Impact of Zoledronate on EFS Stability of the results Age and chemotherapy group GroupNo. Events / No. Entered Z+Z- Hazard RatioHR [95% CI] Z+ better | Z- better <18y – MTX-Eto-Ifo38/11131/1091.29 [0.70;2.38] 18-25y – MTX-Eto-Ifo6/196/180.99 [0.19;5.08] 18-25y – API-AI6/116/122.35 [0.44;12.6] >25y – API-AI7/206/191.17 [0.28;4.92] Risk group Non metastatic41/12933/1321.50 [0.82;2.73] Metastatic or non resectable16/3216/260.94 [0.36;2.45] Histologic response Good responder26/8421/881.62 [0.73;3.58] Poor responder23/4422/451.20 [0.55;2.61] Overall57/16149/1581.31 [0.79;2.18] 0.11.014.0 Interaction P=0.80 P=0.30 P=0.49

18. OS 2006 Futility analysis  Probability of demonstrating a benefit with zoledronate after inclusion of the 470 pts initially planned <0.0001  Decision to close the trial and to release the results  Final analysis in a few months after updating follow-up of all patients 15-18 October 2014, Berlin

19. OS 2006 - Conclusion  Zoledronate in association with chemotherapy did not reduce the risk of treatment failure  There was no safety concern apart from the expected higher incidence of hypocalcemia  Overall results (EFS, OS) are consistent with previous studies both in children and adult patients  Translational studies on going to try to understand these unexpected clinical results 15-18 October 2014, Berlin

20. Thanks  To patients and families  To Unicancer, SFCE and FSG  To French National Cancer Institute (INca) and La Ligue contre le Cancer  To Novartis, Chugaï  To investigators, and data managers 15-18 October 2014, Berlin

21. Backslides 15-18 October 2014, Berlin

22. Outcome of patients with a localised disease, by chemotherapy group and histological response EFS MTX GR versus PR OS MTX GR versus PR EFS API-AI GR versus PR OS API-AI GR versus PR 15-18 October 2014, Berlin

23. Second interim analysis (26 MARCH 2014)  Trend for a harmful effect of Zoledronate  On EFS and OS  Stable results after exclusion of the 8 patients allocated to Z+ who received 0 to 1 Zoledronate injection  No heterogeneity across the different strata at the time of randomisation  Boundaries defining a significant harm not crossed  Recommendation of early stopping for futility Probability of demonstrating a benefit <0.0001 (even under H1) current statistic test 15-18 October 2014, Berlin Stop for efficacy Stop for harm Stop for futility