Randomized Phase II Study of Duligotuzumab + FOLFIRI vs + FOLFIRI in 2nd-line Patients with KRAS Wild-Type Metastatic Colorectal

A. Hill,1 M. Findlay,2 M. Burge,3 C. Jackson,4 P.G. Alfonso,5 L. Samuel,6 V. Ganju,7 M. Karthaus,8 A. Amatu,9 M. Jeffery,10 M. DiBartolomeo,11 J. Bridgewater,12 A. Coveler,13 M. Hidalgo,14 A.V. Kapp,15 R.I. Sufan,15 B. McCall,15 E. Penuel,15 A. Pirzkall,15 J. Tabernero16 1Tasman Oncology Research, Southport, Australia; 2Faculty of Medical & Health Sciences, Auckland University, Auckland, NZ; 3Royal Brisbane and Women’s Hospital, Herston, Australia; 4Dunedin School of Medicine, University of Otago, NZ; 5Gregorio Marañón Hospital Madrid, Spain; 6Aberdeen Royal Infirmary, Aberdeen, UK; 7Peninsula Oncology Centre, Frankston, Australia; 8Staedtisches Klinikum Muenchen GmbH - Klinikum Neuperlach, Munich, Germany; 9Niguarda Cancer Center, Ospedale Niguarda Ca’ Granda, Milan, Italy; 10Canterbury Regional Cancer and Haematology Service, Christchurch, NZ; 11Fondazione IRCCS Instituto Nazionale dei Tumori, Milan, Italy; 12University College London Cancer Institute, London, UK; 13University of Washington, Seattle, WA; 14Centro Integral Oncologico Clara Campal, Madrid, Spain; 15Genentech, South San Francisco, CA; 16Vall d’Hebron Institute of Oncology, Barcelona, Spain Patient Characteristics Clinical Activity Figure 4. Time on Study Treatment. BACKGROUND Table 1. Patient Demographics and Disease Characteristics. •• PFS in the RAS WT subgroup was median 6.9 vs. 5.7 months for duligotuzumab vs. cetuxumab •• Anti-EGFR mAbs are established in mCRC treatment as single agent or in combination with Duligotuzumab Cetuximab •• In RAS WT patients, ORR was lower in the duligotuzumab arm 17% vs. 33.3% Characteristic All Duligotuzumab (n=53) Cetuximab (n=44) 1 CR chemotherapy + FOLFIRI + FOLFIRI PR PR •• Subgroup analysis by patient and disease characteristics, stratification factors, and biomarker PR PR

PR PR •• Retrospective analyses from clinical studies suggest a potential role for HER3 signaling in context of All RAS WT All RAS WT All RAS WT subsets did not show benefit for duligotuzimab PR PR PR PR PR 2-4 PR initial anti-EGFR activity and/or emergence of resistance after treatment (n=68) (n=53) (n=66) (n=45) (n=134) (n=98) PR • Similar results were observed in all randomized patients PR PR • PR •• Duligotuzumab (MEHD7945A) is a novel dual-action humanized IgG1 antibody that blocks ligand Age (years), median (range) 61 (21–85) 62 (21–85) 62 (26–82) 65 (39–82) 62 (21–85) 63 (21–85) Confirmed CR PR PR Best response of CR Table 3. Summary Outcomes. PR binding to EGFR and HER3, inhibiting all major ligand-dependent HER complex signaling Sex (male) 36 (52.9%) 27 (50.9%) 46 (69.7%) 34 (75.6%) 82 (61.2%) 61 (62.2%) PR Confirmed PR PR Best response of PR 5 • Preclinical activity Race (white) 60 (88.2%) 46 (86.8%) 59 (89.4%) 42 (93.3%) 119 (88.8%) 88 (89.8%) Duligotuzumab Cetuximab a PR Best response of SD • All Randomized RAS WT Hazard Ratio or Odds Best response of PD PR + FOLFIRI + FOLFIRI PR PR Best response of UE -- In vivo, duligotuzumab shows activity in CRC KRAS wildtype xenograft models equal or superior ECOG PS (0) 30 (44.8%) 24 (45.3%) 39 (61.9%) 25 (56.8%) 69 (53.1%) 49 (50.5%) Patients Ratio (90% CI) PR (n=53) (n=45) NA a PR CR PR −− First CR or PR to cetuximab, no effect in KRAS mut models, and is additive in combination with chemotherapy BRAF mutation 8 (13.8%) 8 (15.1%) 8 (16.3%) 8 (17.8%) 16 (15.0%) 16 (16.3%) PR First PD a PFS events 39 (74%) 33 (73%) PIK3CA mutation 9 (15.5%) 8 (15.1%) 4 (8.2%) 2 (4.4%) 13 (12.1%) 10 (10.2%) Adverse event -- In CRC cell lines, standard chemotherapy may modulate the HER3/NRG network Median PFS, mo (90% CI) 6.9 (4.9–7.6) 5.7 (5.5–7.7) Death 1.40 (0.94–2.08) Physician decision Triple WT (RAS, BRAF and PR 36 (64.3%) 36 (67.9%) 33 (64.7%) 33 (73.3%) 69 (64.5%) 69 (70.4%) PR Radiographic progression of disease a OS events 18 (34%) 19 (42%) PR Symptomatic deterioration without PIK3CA) radiographic progression of disease Figure 1. Study Design. Median OS, mo (90% CI) NE (12.0–NE) 12.4 (10.0–NE) Withdrawal by subject Received prior 32 (47.1%) 23 (43.4%) 31 (47.7%) 18 (40.0%) 63 (47.4%) 41 (41.8%) 0.85 (0.49–1.46) Other Time to PD on 1L oxaliplatin- ORR, % (90% CI) 17.0 (9.8–26.6) 33.3 (21.8–46.1) 0.41 (0.18–0.91) Arm A 53 (77.9%) 38 (71.7%) 50 (75.8%) 34 (75.6%) 103 (76.9%) 72 (73.5%) 0 50 100 150 200 250 300 350 400 450 0 50 100 150 200 250 300 350 400 450 500 550 based chemo (≤ 6 months) CI = confidence interval. NE = not estimated. Study day Duligotuzumab: 1100 mg IV q2w Study day 1:1 Randomization WT = wildtype; ECOG = Eastern Cooperative Oncology Group. a Unstratified. FOLFIRI: q2w a The denominator for the all-randomized population is the number of patients with mutation data. Stratified by Figure 2. Progression-Free Survival. • Prior bevacizumab Safety •• Overall time on study treatment less on duligotuzumab vs. cetuximab study arm (yes vs. no) Treat to progression or •• Observation is in line with lower median FOLFIRI dose intensities and the median time to first AEs unacceptable toxicity •• Discontinuation due to AEs were higher in cetuximab arm (13.8%) vs. the duligotuzumab arm (8.8%) • Disease progression within for which irinotecan or 5-FU was modified being earlier (median 22 vs. 35 days in safety-evaluable •• SAEs occurred in 52.2% and 55.6% of patients on duligotuzumab vs cetuximab arm; no noticeable patients) in the duligotuzumab arm 6 months of first-line differences in nature of SAEs were observed. The SAEs regardless of attribution in ≥ 5% of patients oxaliplatin-containing Arm B were pyrexia (6.2%), diarrhea and pulmonary embolism (5.4% each). chemotherapy (yes vs. no) Cetuximab: 400 mg/m2 IV loading dose, 2 • Grade 5 AEs in 5 patients: 2 cetuximab, 3 duligotuzumab; none assessed as treatment-related 250 mg/m IV q1w • CONCLUSIONS FOLFIRI: q2w Table 2. All AEs Regardless of Attribution in ≥20% of Patients and Grade ≥ 3 AEs in ≥ 5% of •• No benefit for duligotuzumab + FOLFIRI relative to cetuximab + FOLFIRI in RAS wildtype patients Patients. (72 PFS events, 37 OS events) FOLFIRI = folinic acid (leucovorin), 5-fluorouracil, and irinotecan, and; IRF = independent review facility; IV = Duligotuzumab Cetuximab •• HER3 protein or gene expression levels did not select for benefit with duligotuzumab, nor did NRG intravenous; PD = progressive disease; q1w = weekly; q2w = every 2 weeks. Characteristic + FOLFIRI + FOLFIRI or AREG expression (N=67) (N=63) •• Based on these results, the role of HER3 appears limited in CRC Any Grade ≥ Grade 3 Any Grade ≥ Grade 3 • Overall, the frequency of AEs was largely similar to cetuximab + FOLFIRI; less rash and IRRs, but OBJECTIVES All 67 (100.0%) 57 (85.1%) 63 (100.0%) 56 (88.9%) • more diarrhea, mucosal inflammation and hypokalemia in duligotuzumab arm Rash and related termsa 53 (79.1%) 5 (7.5%) 56 (88.9%) 14 (22.2%) Primary Objectives • More frequent lower grade GI toxicities may have contributed to the noticeably lower cumulative Diarrhea 61 (91.0%) 12 (17.9%) 42 (66.7%) 9 (14.3%) • dose intensity and shorter duration of chemo on duligotuzumab than on cetuximab arm with a •• Evaluate the efficacy of FOLFIRI + duligotuzumab versus FOLFIRI + cetuximab in patients with Fatigue 44 (65.7%) 6 (9.0%) 36 (57.1%) 10 (15.9%) KRAS wild-type mCRC and in those patients expressing low levels of HER3. Based on emerging possible contribution to the lower ORR on duligotuzumab arm Nausea 33 (49.3%) 3 (4.5%) 32 (50.8%) 2 (3.2%) new data6-11 during the study, the primary analysis was conducted in the RAS WT population instead, as described in Study Design Section. Mucosal inflammation 27 (40.3%) 7 (10.4%) 23 (36.5%) 3 (4.8%) Stomatitis 20 (29.9%) 6 (9.0%) 25 (39.7%) 4 (6.3%) REFERENCES Secondary and Exploratory Objectives Alopecia 18 (26.9%) - 22 (34.9%) 1 (1.6%) Biomarker Analyses 1. Prenen H, et al. Target Oncol 2013; 8:83-96. •• Evaluate the safety and tolerability of FOLFIRI + duligotuzumab versus FOLFIRI + cetuximab Neutropenia 18 (26.9%) 14 (20.9%) 21 (33.3%) 16 (25.4%) •• HER3 measured by qRT-PCR or IHC in RAS WT patients did not predict response to treatment in KRAS wild-type mCRC patients previously treated with first-line oxaliplatin-containing 2. Scartozzi M, et al. Oncologist 2011; 16:53-60. Paronychia 21 (31.3%) 2 (3.0%) 18 (28.6%) 1 (1.6%) •• No correlation between NRG1 expression and response to treatment was observed chemotherapy, focusing on NCI CTCAE Grade ≥3 AEs and SAEs 3. Yonesaka K, et al. Sci Transl Med 2011; 3:99ra86. Hypokalemia 24 (35.8%) 3 (4.5%) 13 (20.6%) 4 (6.3%) •• No correlation between NRG1 and EGFR ligands was observed •• Assess the effect of concomitant FOLFIRI on the pharmacokinetics of duligotuzumab and vice versa 4. Ho-Pun-Cheung A, et al. Int J Cancer 2011; 128:2938-46. Abdominal pain 13 (19.4%) 1 (1.5%) 22 (34.9%) 4 (6.3%) 5. Crocker LM, et al. 2012 AACR Annual Meeting, #1212. •• Evaluate the incidence and impact of anti-duligotuzumab antibodies Decreased appetite 16 (23.9%) 2 (3.0%) 14 (22.2%) 2 (3.2%) Figure 3. Best Response Less in Duligotuzumab vs. Cetuximab Arm; No Relationship with 6. Douillard JY, et al. NEJM 2013; 369(22):2159-60. Dry skin 17 (25.4%) - 13 (20.6%) - HER Expression by qRT-PCR or IHC. 7. Schwartzberg L, et al. JCO 2013; 32(21):2240-7. IRR b 13 (19.4%) 1 (1.5%) 17 (27.0%) 5 (7.9%) STUDY DESIGN 8. Stintzing S, et al. EJC 2013; suppl 3; abstr 445. Vomiting 17 (25.4%) 4 (6.0%) 12 (19.0%) 3 (4.8%) • This open-label, randomized Phase II study enrolled patients with KRAS exon 2 WT mCRC who 9. Bokemeyer C, et al. JCO 2014; 32:5s, suppl; abstr 3505. • Hand-foot syndrome 13 (19.4%) 1 (1.5%) 13 (20.6%) 2 (3.2%) progressed on/after oxaliplatin-containing chemotherapy 10. Ciardiello F, et al. JCO 2014; 32:5s, suppl; abstr 3506. Pulmonary embolism 7 (10.4%) 7 (10.4%) 7 (11.1%) 7 (11.1%) •• Patients received a combination of duligotuzumab (1100 mg IV, q2w) or cetuximab (400 mg/m2 load, 11. Peeters M, et al. JCO 2014; 32:3s, suppl; abstr LBA387. 250 mg/m2 IV, q1w) + FOLFIRI (q2w) until progression or intolerable toxicity Intestinal obstruction 1 (1.5%) 1 (1.5%) 4 (6.3%) 4 (6.3%) a •• An early per-protocol interim safety analysis occurred after an initial 6 and then 20 patients in each Rash and related MedDRA terms = rash, dermatitis acneiform, rash maculo-papular, acne, dermatitis, rash macular, treatment arm received 2 cycles of treatment rash erythematous, rash pruritic, dermatitis atopic, dermatitis bullous, dermatitis exfoliative, rash generalized, rash papular, and rash pustular. ACKNOWLEDGMENTS •• Tumor samples were mandatory and underwent biomarker analysis for ERBB3, NRG1 and EGFR b Any AE occurring during infusion or within 24 hours and suspected to be caused by duligotuzumab or cetuximab. •• Thank you to the many patients for their participation in this trial ligand expression by qRT-PCR, and ERBB3 by IHC •• Genentech provided support for the preparation of this poster •• The primary efficacy analysis was conducted on the RAS wildtype patient subgroup using an allele- Pharmacokinetics specific PCR mutation panel that detects mutations in KRAS and NRAS in exons 2 (G12 and G13) •• Mean observed duligotuzumab exposures from study GO28074 (DARECK) were very similar to and exon 3 (Q61). This was done to account for recent data which identified additional mutations in GO28076 (MEHGAN) and in line with expected exposures from Phase Ia and population PK model KRAS or NRAS exons 2, 3, 4 as negative predictive biomarkers for EGFR inhibition in 1st or 2nd Mobile Friendly e-Prints line mCRC.6-11 •• No evidence of trends in PFS or OS with duligotuzumab exposure, based on exposure-response analyses; indicating that dose was close or at the top of exposure-response curve •• Lower FOLFIRI cumulative dose intensity (%) was observed on duligotuzumab vs. cetuximab arm Scan this QR code on your smart device to download, or RESULTS (Irinotecan: 66.1 [24–100] vs. 75.0 [32–104], 5-FU infusion 66.9 [24–99] vs. 74.0 [32–107]) receive by email, an electronic copy of this poster. •• As of 21 Aug 2014, 11 patients remained active on study •• No ATAs were detected

2015 AACR Annual Meeting, April 18-21, 2015, Philadelphia, PA