1. Correlation of Hand-Foot Skin Reaction (HFS) with Treatment Efficacy in Pancreatic Cancer (PC) Patients (pts) Treated with Gemcitabine/Capecitabine plus Erlotinib: A Subgroup Analysis from the AIO-PK0104 Randomized, Cross-Over Phase III Trial in Advanced PC M. Haas 1, S. Boeck 1, R. P. Laubender 2, D. P. Modest 1, U. Vehling-Kaiser 3, D. Waldschmidt 4, E. Kettner 5, A. Märten 6, C. Winkelmann 7, S. Klein 8, G. Kojouharoff 9, T. C. Gauler 10, L. Fischer von Weikersthal 11, M. Clemens 12, M. Geissler 13, T. F. Greten 14, S. Hegewisch-Becker 15, O. Rubanov 16, U. Mansmann 2 and V. Heinemann 1 1 Department of Hematology and Oncology, Klinikum Grosshadern and Comprehensive Cancer Center, LMU Munich; 2 Institute for Medical Informatics, Biometry and Epidemiology, LMU Munich; 3 Practice for Medical Oncology, Landshut; 4 Department of Hematology and Oncology, University of Cologne; 5 Department of Hematology and Oncology, Klinikum Magdeburg; 6 University of Heidelberg; 7 Department of Internal Medicine, Krankenhaus Lutherstadt-Wittenberg; 8 Department of Internal Medicine IV, Klinikum Bayreuth; 9 Practice for Medical Oncology, Darmstadt; 10 West German Tumor Center, University Hospital Essen; 11 Gesundheitszentrum St. Marien, Amberg; 12 Mutterhaus der Boromaeerinnen, Trier; 13 Department of Gastroenterology and Oncology, Klinikum Esslingen; 14 Medical School Hannover; 15 Private Practice for Oncology, Hamburg; 16 Practice for Medical Oncology, Hameln; all in Germany Background: AIO-PK 0104 investigated the efficacy and safety of gemcitabine/erlotinib (G/E) followed by capecitabine (C) vs. C/E followed by G. The present subgroup analysis evaluated the correlation between C-associated skin toxicity and outcome parameters in PC. Methods: Within this multicenter phase III trial, pts with confirmed advanced PC were randomly assigned to 1 st - line treatment with either C (2,000 mg/m 2 /d, d1-14 q d21) plus E (150 mg/d, arm A) or G (1,000 mg/m 2 over 30 min weekly x 7, then d1, 8, 15 q d28) plus E (150 mg/d, arm B). A cross-over to either G (arm A) or C (arm B) was performed after treatment failure (e. g. disease progression or unacceptable toxicity). Time to treatment failure after 1 st - and 2 nd -line therapy (TTF2) was the primary study endpoint. Treatment-related skin toxicity was evaluated separately for each treatment arm/each regimen based on NCI-CTCv2. Results: Of 279 eligible pts, 47 had locally advanced, 232 had metastatic disease and 141 pts received second- line chemotherapy. For the present subgroup analysis data on skin toxicity were available from 255 pts. For the 73 pts (29%) with a HFS (any grade documented at any time during the treatment strategy), TTF2 and OS both were significantly prolonged compared to pts without HFS (7.4 vs 4.0 months, p<0.001 and 9.7 vs 5.5 months, p=0.002, respectively). Considering HFS during 1 st -line treatment in 123 pts within the CE arm, these results could be confirmed for the 47 pts (38%) with a documented HFS of any grade (TTF2: 7.6 vs. 3.2 months, p<0.001; OS: 10.2 vs. 4.4 months, p=0.001). In pts receiving 1 st -line treatment with G/E (n=132) no difference in outcome was observed for pts with (n=13) or without (n=119) HFS of any grade (TTF2: 5.7 vs. 4.2 months, p=0.375; OS: 8.4 vs. 6.6 months, p=0.505). Conclusions: The current subgroup analysis of AIO-PK0104 supports the assumption of a correlation between HFS in PC pts treated with capecitabine or capecitabine/erlotinib and efficacy endpoints like TTF2 and OS. A capecitabine-associated HFS thus might be predictive for efficacy in patients with advanced PC. Trial Design and Treatment Patient CharacteristicsCorrelation of HFS with Efficacy – 1 st Line Therapy Cumulative Number of Patients with HFS During Treatment with Capecitabine – 1 st Line Therapy Treatment cycle 123456712 New pts. with HFS 920732411 1st line: capecitabine + erlotinib 1st line: gemcitabine + erlotinib 2nd line: gemcitabine 2nd line: capecitabine TTF 1 TTF 2 R Patients with data on HFS (n=255, safety population) Correlation of HFS During 1 st Line Therapy with Efficacy - TTF2 Conclusions Frequency of HFS - 1 st Line Therapy Delay of Chemotherapy and Dose Reductions Due to HFS – 1 st Line Therapy Correlation of HFS with Efficacy – 1 st and 2 nd Line Therapy Parameter Arm A Cap + E  Gem (n=123) Arm B Gem + E  Cap (n=132) Overall (n=255) No% % % Age (years) Median 636563 Range 38-7532-78 Gender Male7863745615260 Female4537584410340 Stage of disease Locally advanced221821164317 Metastatic101821118421283 Performance status KPS 60-80%453747369237 KPS 90-100%7763826415963 Missing134 Treatment with 2 nd line chemotherapy 6250765813854 N=281 1:1 Arm A Arm B Arm A: capecitabine 2.000 mg/m 2 /d p.o. d1-14, q d21 plus erlotinib, 150 mg/d p.o.; followed by gemcitabine Arm B: gemcitabine 1.000 mg/m 2 over 30 min i.v. weekly x 7, then d1, 8, 15 q d28 plus erlotinib, 150 mg/d p.o.; followed by capecitabine TTF 1: time-to-treatment failure after 1 st line therapy TTF 2: time-to-treatment failure after 1 st and 2 nd line therapy Parameter Arm A Cap + E (n=123) Arm B Gem + E (n=132) Overall (n=255) No% % % HFS-Grade 076611199019577 123191083313 2161332197 386--83 Any grade (1-3)473813105019 Parameter Arm A Cap + E Arm B Gem + E Overall No% % % No (pts)123132255 Delay of treatment - overall4738665011344 - due to HFS97--94 No (pts)122130252 Dose reductions - overall292469539839 - due to HFS108-- 4  Treatment delays and dose reductions overall were more common during treatment with gemcitabine and erlotinib (arm B)  HFS was responsible for delayed application or dose reductions of capecitabine in < 10% HFS- grades 01-30 EndpointSubgroupNoMonthsP-valueHR (95% CI) TTF1 Arm A: Cap+E76472.04.0<0.001 0.42 (0.29-0.62) Arm B: Gem+E119133.55.70.322 0.75 (0.42-1.33) TTF2 Arm A: Cap+E76473.27.6<0.001 0.46 (0.32-0.67) Arm B: Gem+E119134.25.70.375 0.77 (0.43-1.37) OS Arm A: Cap+E76474.410.20.001 0.52 (0.35-0.76) Arm B: Gem+E119136.68.40.505 0.81 (0.44-1.51)  HFS during 1 st line treatment with capecitabine and erlotinib was significantly correlated with efficacy endpoints HFS-grades01-30 EndpointSubgroupNoMonthsP-valueHR (95% CI) TTF2All patients182734.07.4<0.001 0.61 (0.47-0.81) OSAll patients182735.59.70.002 0.64 (0.48-0.86) Methods  Definition of hand-foot-skin reaction (HFS): NCI-CTC version 2  Retrospective, post-hoc subgroup analysis based on toxicity data from the prospective, multicenter randomized AIO-PK0104 study (Boeck S et al, ASCO 2010: LBA#4011)  Safety population including 255 of the 274 eligible study patients  Separate subgroups for HFS during 1 st and 2 nd line therapy  Correlation of HFS with the primary (TTF2) and secondary study endpoints (TTF1, OS) for treatment efficacy  Analyses were repeated for patients who were on study for at least 3 cycles and therefore had at least one tumour assessment (i.e. patients who were defined as assessable for treatment efficacy)  Median time to capecitabine-attributed HFS was two cycles  In 36 out of 47 patients (77%) HFS was documented for the first time within the first three treatment cycles  The occurrence of HFS during treatment with capecitabine plus erlotinib was correlated with efficacy endpoints (e. g. TTF1, TTF2 and OS)  Capecitabine-attributed HFS occurred early within the treatment course (after a median of two cycles)  Capecitabine-associated HFS might be predictive for treatment efficacy in patients with advanced pancreatic cancer  The occurrence of HFS documented at any time during 1 st or 2 nd line therapy (in both arms) also was significantly correlated with treatment efficacy #4023 Hand-Foot Skin Reaction Arm A: capecitabine + erlotinib → gemcitabine No HFS: 3.2 months HFS grades 1-3: 7.6 months HR 0.46 (95%CI 0.32-0.67), p<0.001 Overall: 47 pts. with HFS Correlation of HFS During 1 st Line Therapy with Efficacy – Patients on Study for ≥ 3 Cycles Only HFS-grades01-30 EndpointSubgroupNoMonthsHR (95% CI) TTF1Arm A: Cap+E36343.04.5 0.49 (0.30-0.80) TTF2Arm A: Cap+E36344.48.3 0.53 (0.33-0.87) OSArm A: Cap+E36346.710.4 0.62 (0.37-1.01)  In this subgroup analysis, only patients still on protocol after completion of 3 cycles (time of first tumour assessment) were included (i.e. patients who were defined as assessable for treatment efficacy)  Differences in outcome between HFS + and HFS – patients remained apparent