Search

Background ROS1 fusions are found in 1% of lung cancer patients. EUCROSS is the first prospective European trial to evaluate crizotinib in this patient subset. Here we present an updated analysis of the progression-free survival (PFS), overall survival (OS) and molecular characteristics of progression. Methods Multi-centre, single arm phase II trial. Key eligibility criteria: ≥18 years of age, advanced/metastatic lung cancer, centrally confirmed ROS1-rearrangement. Treatment: 250 mg crizotinib BID. Key endpoints of this report: updated PFS (according to RECIST 1.1) and OS and molecular tumour characterization. Results Thirty patients were eligible for response evaluation (N = 30; intention-to-treat population, N = 34). Median follow-up was 44.9 months (95% CI, 39.6-47.4). At data-cut off 19 patients (63.3%) discontinued treatment due to progression or death. Investigator-assessed efficacy: ORR, 70% (95% CI, 51–85; 21/30); disease control rate, 90.0% (95% CI, 73.5-97.9; 27/30); median PFS, 19.4 months (95% CI, 10.1-31.2); 24-months-OS probability, 65.5% (95% CI, 48.3-82.9). Prevalence of co-occurring genetic aberrations by hybrid-capture sequencing was 61%. TP53 mutations were most frequent (28%; 5/18). PFS (p = 0.0219) and OS at 24 months (p = 0.015) were significantly shorter in TP53-mutant patients. Tissue or blood samples were collected in six patients at progression. In three (50%) samples, secondary mutations in ROS1 (i.e. p.G2032R (N = 2), p.L2026M (N = 1)) were detected. In one (17%), a PIK3CA mutation (p.E545K) was identified. In the other two patients (33%) aberrations of unknown significance were detected. Conclusions Updated PFS and OS results confirm the efficacy of crizotinib in ROS1-rearranged lung cancer patients. Patients with co-mutations in TP53 had significantly worse outcomes compared to TP53 wild-type patients. The most common mechanisms of resistance were mutations in ROS1. Next-generation ROS1 inhibitors or the multi-kinase inhibitor cabozantinib may be promising treatment options for these patients. Clinical trial identification NCT02183870. Legal entity responsible for the study University Hospital of Cologne. Funding Pfizer. Disclosure S. Michels: Honoraria (self), Advisory / Consultancy, Research grant / Funding (institution): Pfizer; Honoraria (self), Advisory / Consultancy, Research grant / Funding (institution): Novartis; Research grant / Funding (institution): Bristol-Myers Squibb; Honoraria (self): Boehringer Ingelheim; Research grant / Funding (institution): Janssen; Honoraria (self): Roche. B. Massuti Sureda: Honoraria (self): Merck Sharp Dome; Honoraria (self): Boehringer Ingelheim; Honoraria (self): Roche; Honoraria (self): Pfizer. H. Schildhaus: Honoraria (self): ZytoVision; Honoraria (self): Pfizer; Honoraria (self), Research grant / Funding (institution): Roche; Honoraria (self), Research grant / Funding (institution): Novartis. M. Sebastian: Honoraria (self): Boehringer Ingelheim; Honoraria (self): Eli Lilly; Honoraria (self): Pfizer; Honoraria (self): Roche; Honoraria (self): Novartis; Honoraria (self): AstraZeneca; Honoraria (self): Bristol-Myers Squibb; Honoraria (self): Merck Sharp Dome; Honoraria (self): Takeda; Honoraria (self): Mediolanum; Honoraria (self): AbbVie; Honoraria (self): Celgene. E. Felip: Honoraria (self): AstraZeneca; Honoraria (self): Boehringer Ingelheim; Honoraria (self): Bristol-Myers Squibb; Honoraria (self): Celgene; Honoraria (self): Eli Lilly; Honoraria (self): Guardiant; Honoraria (self): Merck Sharp Dome; Honoraria (self): Merck KGa; Honoraria (self): Novartis. M. Reck: Honoraria (self): Boehringer Ingelheim; Honoraria (self): Bristol-Myers Squibb; Honoraria (self): AstraZeneca; Honoraria (self): Abbot; Honoraria (self): Celgene; Honoraria (self): Merck Sharp Dome; Honoraria (self): Merck KGa; Honoraria (self): Novartis; Honoraria (self): Pfizer; Honoraria (self): Roche. S. Merkelbach-Bruse: Honoraria (self): Pfizer; Honoraria (self): AstraZeneca; Honoraria (self): Bristol-Myers Squibb; Honoraria (self): Roche; Honoraria (self), Research grant / Funding (institution): Novartis. L. Nogova: Honoraria (self), Research grant / Funding (institution): Pfizer; Honoraria (self), Research grant / Funding (institution): Novartis; Honoraria (self), Research grant / Funding (institution): Janssen; Honoraria (self): Boehringer Ingelheim; Honoraria (self): Roche; Honoraria (self): Celgene. J. Wolf: Honoraria (self), Research grant / Funding (institution): Novartis; Honoraria (self), Research grant / Funding (institution): Pfizer; Honoraria (self), Research grant / Funding (institution): Janssen; Honoraria (self): Boehringer Ingelheim; Honoraria (self): AstraZeneca; Honoraria (self), Research grant / Funding (institution): Bristol-Myers Squibb; Honoraria (self): Chugay; Honoraria (self): Ignyta; Honoraria (self): Eli Lilly; Honoraria (self): Merck Sharp Dome; Honoraria (self): Roche. All other authors have declared no conflicts of interest.

Background Post-operative platinum-based chemotherapy is the standard of care for resected NSCLC with nodal involvement. Expression of genes involved in DNA repair may have a prognostic role for the outcome. BRCA1 plays an important role in DNA repair pathways and could have a prognostic impact in this setting. The SCAT trial results found that for low BRCA1 levels subgroup Cis-Gem was superior to Cis-Doc and in high BRCA1 subgroup docetaxel single agent without platinum achieved similar survival to Cis-Doc. Risk of recurrence analysis according BRCA-1 levels has been performed. Methods From Jun/2007 to May/2013 591 patients were screened and 500 were included, 108 in the control arm treated with Cis-Docetaxel and 392 in the experimental arms treated with Cis-Gem, Cis-Doc or docetaxel alone according to terciles of BRCA1 expression level. With a cut-off September 30th 2018 and a median follow-up of 60 months, recurrence patterns were analysed for the whole group according BRCA1 levels in tumor tissue and comparison were made for risk of recurrence, single/multiple recurrence, thoracic/extrathoracic and metastatic sites (liver, bone, brain). Results Cumulative recurrence was evaluable in 232/456 patients (50.8%), 182/354 patients treated in the experimental arm and in 50/102 patients treated in the control arm (RR 1.04; 0.83-1.30) (p = 0.672). The majority of recurrences 159/232 (68.5%) were single site intrathoracic recurrences in 121/232 (52%) while 111/232 were extrathoracic (47.8%). Overall recurrence was 56.5% (113/200 p) for low tercile BRCA1 vs 48.8% (63/129) for intermediate tercile vs 44% (56/127) for high tercile (p = 0.025). No differences were seen between tercile groups for single site (p = 0.35), multiple site (p = 0.26), intrathoracic (p = 0.36), or extrathoracic (p = 0.38). More frequent distant metastatic sites were: bone (42 patients), brain (38 patients) and liver (11 patients). RIsk reduction was seen for brain metastases in patients with higher tercile BRCA1 (p = 0.003). Conclusions For NSCLC resected patients with lymph node involvement risk of recurrence remains high with a cumulative rate > 50%. Relative risk of recurrence was lower for tumors with higher BRCA1 levels. Distant metastases were seen in 47.8% of patients. Brain metastases risk was significant lower for patients with low BRCA1 expression. BRCA1 levels acts as a prognostic factor in early stages NSCLC. Clinical trial identification EudraCT: 2007-000067-15; NCT 00478699. Legal entity responsible for the study Spanish Lung Cancer Group - Grupo Espanol Cancer de Pulmon. Funding Sanofi Aventis. Disclosure All authors have declared no conflicts of interest.

Background Immunotherapy has become the main therapy in advanced NSCLC patients, but recently, combination with chemotherapy in locally advanced stages is showing promising results. Many studies have described peripheral blood immune cells parameters as response biomarkers. In our study, we described the effect of neo-adjuvant chemo-immunotherapy treatment in Complete Blood Count (CBC) and Peripheral blood mononuclear cells phenotype, as well as, the association of these parameters with pathological response. Methods 46 resectable stage IIIA NSCLC patients treated with neo-adjuvant chemo-immunotherapy from NADIM clinical trial were analysed. Samples were extracted before initiating the neo-adjuvant treatment with nivolumab plus carboplatin and at the third cycle before patients underwent surgery. We classified patients in 3 subgroups of pathological response: complete (pCR), mayor ( 10% viable tumour, pIR). Results Absolute numbers of Leucocytes, Eosinophil, Monocytes, Neutrophils, Haemoglobin and Platelets from hemograms were significantly reduced after treatment. However, no changes were observed for Lymphocytes, Basophils, LDH levels or the lung immune prognostic index (LIPI). Additionally, post-treatment neutrophil-to-lymphocyte (NLR), myeloid-to-lymphoid lineage (M:L) and platelets-to-lymphocytes (PLR) ratios were decreased. Remarkably, from all the CBC absolute numbers and ratios, only PLR variation showed differences between pCR and pIR. On the other hand, percentages of T cells, B cells, NK cells and macrophages did not vary after treatment, however activation of CD4 T cells and NK cells were significantly downregulated after treatment, associated to pCR. Also, PD-1 expression on T and NK cells pre-treatment was higher on pCR compared to pIR, reaching statistical significance only on CD4 T cells. Conclusions In our study, we described predictive biomarkers of response to treatment. A higher decrease on PLR post-treatment is associated to pIR. Moreover, higher expression of PD-1 pre-treatment in CD4 T cells, as well, as a reduced activation on CD4 T cells and NK cells, after treatment are associated to pCR. Clinical trial identification NCT 03081689; EudraCT 2016-003732-20. Legal entity responsible for the study Spanish Lung Cancer Group. Funding BMS. Disclosure M. Provencio: Advisory / Consultancy: BMS; Advisory / Consultancy, Travel / Accommodation / Expenses: MSD; Advisory / Consultancy, Travel / Accommodation / Expenses: AstraZeneca; Advisory / Consultancy: Boehringer. All other authors have declared no conflicts of interest.

Background Target therapy guide by biomarkers have become the standard of care for patients with lung cancer (LC). So, identify those molecular alterations is one of the most important care needs nowdays. Our objective was to know the implementation degree of these tests in a large cohort of patients in Spain using the Thoracic Tumor Registry (TTR) of the Grupo Espanol de Cancer de Pulmon (Spanish Lung Cancer Group). Methods The TTR is an observational cohort multicenter study of LC in Spain. The study is conducted according to the Declaration of Helsinki and approved by the institutional review board of each participating institute. The registry was approved by the Spanish Drug Agency as a non-post-authorization, non-interventional study. We analyzed the molecular biomarkers considering all stages of LC. Results A total of 7,872 patients from 58 Spanish centers were enrolled between August 2016 to December 2018. The most frequent screened molecular test was the EGFR mutations, it was performed in 4,456 patients (67.5%). The proportion of biomarker evaluation has varied over time, ranging from 57.9% prior to 2012 up to 73.7% in 2017. The molecular assessment of some biomarkers reached 81.4% of these patients, with some differences between Regional Communities, regarding the molecular tests performed. Three thousand four hundred forty-six (3,446) patients (52.2%) had a stage IV at diagnosis. There was performed some biomarker test in 92% of the 2570 patients with stage IV and non-squamous histology. In those stage IV non-squamous patients, the EGFR and ALK test were performed in 92% and 79% respectively but 2 years ago ALK test was done only in 40% of the patients. ROS1 was studied in 20% of the cases. Conclusions Although no national plan exists for molecular biomarker analysis in Spain, the implementation of biomarkers tests in all the hospitals that contribute to the TTR is high. The increase in the ALK analysis in the last period is relevant. As we have some regional differences, it is important to understand the causes to improve them. Clinical trial identification NCT02941458. Legal entity responsible for the study Spanish Lung Cancer Group. Funding Novartis, Merck Sharp and Dohme, Lilly. Disclosure All authors have declared no conflicts of interest.

Background Tobacco use, mainly as cigarette smoking, is the leading cause of lung cancer. Eighty-five percent of LC occur in smokers. Understanding the Spanish smoking habits allows the government to design health care policies against this consumption. To do so, the Grupo Espanol de Cancer de Pulmon (Spanish Lung Cancer Group) made this analysis within the context of the Thoracic Tumor Registry (TTR). Methods The TTR is an observational cohort multicenter study in Spain. The study is conducted according to the Declaration of Helsinki and approved by the institutional review board of each participating site. The registry was approved by the Spanish Drug Agency, as a non-post-authorization, non-interventional study. Results Data have been collected from 6,600 LC from 58 different hospital sites across Spain. 12% (866 p) were non-smokers, 46% (3,039 p) were former smokers and 39% (2,611 p) were smokers. There were significant differences by gender, more women were non-smokers (37 % vs. 45% in males), meanwhile more former smokers were male (53.4% vs. 27.9% in women) (p-valor Conclusions Tobacco use is the leading cause of LC in Spain accounting 85% of the cases. Consumption has increased in both genders, but specially in women, in our country among lung cancer patients. Tobacco cessation campaigns, especially in women, should be a priority in western countries, like Spain, and it has to be adapted to regional differences in tobacco use. Clinical trial identification NCT02941458. Legal entity responsible for the study Spanish Lung Cancer Group. Funding Novartis, MSD, Lilly. Disclosure M. Provencio: Advisory / Consultancy: BMS; Advisory / Consultancy, Travel / Accommodation / Expenses: MSD; Advisory / Consultancy, Travel / Accommodation / Expenses: AstraZeneca; Advisory / Consultancy: Boehringer Ingelheim. All other authors have declared no conflicts of interest.

Background CT-RT is the standard treatment for unresectable LA-NSCLC. P plus vinorelbine is widely used. Metronomic CT is a frequent administration of low doses of CT. mOV has shown good efficacy and improved safety, and could improve the RT effect. Our goal is to evaluate the efficacy and safety of P-mOV with radical RT in patients (pts) with LA-NSCLC. Methods Pts aged 18-75 years with histologically proven untreated and unresectable LA-NSCLC, adequate bone marrow, hepatic & renal function, ECOG PS0-1, received P 80mg/m2 D1 every 3 weeks combined with mOV 50mg/day on days D1, 3 & 5/weekly, 2 cycles (cy) as induction; patients without progression received 2 more cy of P at the same dose with mOV 30mg/day on D1, 3 & 5/weekly, concurrently with RT (66Gy in 6.5weeks). Primary endpoint was progression-free survival (PFS) by RECIST v1.1; secondary endpoints were: overall response rate (ORR), disease control rate (DCR), overall survival and safety profile. To guarantee an overall type-1 α error no greater than 0.05 and a type II (β) error 0.1 for PFS, a sample size of 67 pts was planned. Results Sixty-seven pts were recruited in 17 Spanish sites from 04/2016 to 06/2017. One of them didn’t meet all the inclusion criteria. We analyzed 66 pts included. Pt characteristics: Male 77.3%; median age 62 (range 33-75); PS 0/1 52/49%; smokers 53%; adenocarcinoma/squamous 43.9/42.4%; stage IIIA/B 42.4/57.6%. Only 32.3% of pts presented any grade 3-4 adverse event, including: neutropenia 20.0%; anemia 4.6%; febrile neutropenia 6.2%; esophagitis 3.1%; pneumonitis 1.5%. There were two deaths non-related to the treatment, during this period. Fifty-one pts have completed the treatment. ORR: 67.7%. DCR: 84.6%. With a median follow-up of 22.3 months (range 1.1-34.9), the median PFS is 11.5 months (CI95%; 9.9-15,4). Conclusions mOV-P with RT is as effective as the standard administration of vinorelbine, improving its safety profile. Clinical trial identification EudraCT 2015-003312-21. Legal entity responsible for the study Spanish Lung Cancer Group. Funding Pierre Fabre. Disclosure All authors have declared no conflicts of interest.

ML18147 evaluated continued bevacizumab with second-line chemotherapy for patients with metastatic colorectal cancer (mCRC) progressing after the standard first-line bevacizumab-containing therapy.Evaluating outcomes according to tumor Kirsten rat sarcoma virus oncogene (KRAS) status was an exploratory analysis. KRAS data were collected from local laboratories (using their established methods) and/or from a central laboratory (mutation-specific Scorpion amplification-refractory mutation system). No adjustment was made for multiplicity; analyses were not powered to detect statistically significant differences.Of 820 patients, 616 (75%) had unambiguous KRAS data; 316 (51%) had KRAS wild-type tumors and 300 (49%) had mutant KRAS tumors. The median progression-free survival (PFS) was 6.4 months for bevacizumab plus chemotherapy and 4.5 months for chemotherapy [P0.0001; HR = 0.61; 95% confidence interval (CI): 0.49-0.77] for wild-type KRAS and 5.5 and 4.1 months, respectively (P = 0.0027; HR = 0.70; 95% CI: 0.56-0.89) for mutant KRAS. The median overall survival (OS) was 15.4 and 11.1 months, respectively (P = 0.0052; HR = 0.69; 95% CI: 0.53-0.90) for wild-type KRAS and 10.4 versus 10.0 months, respectively (P = 0.4969; HR = 0.92; 95% CI: 0.71-1.18) for mutant KRAS. In both analyses, no treatment interaction by KRAS status was observed (PFS, P = 0.4436; OS, P = 0.1266).Bevacizumab beyond first progression represents an option for patients with mCRC treated with bevacizumab plus standard first-line chemotherapy, independent of KRAS status.

Aim: In the SLCG large-scale screening, 16.6% of EGFR mutations were found, median PFS and OS for pts receiving erlotinib was 14 mos and 27 mos, respectively. The HR for the duration of PFS was 1.68 for the presence of L858R mutation in paired serum DNA (P = 0.02) (NEJM 2009; 361:958-67). Mutant EGFR NSCLC selectively activates STAT3 signaling which promotes cell survival and single EGFR inhibition can even enhance STAT3 activation. Growing evidence shows that targeting STAT3 or upstream or downstream components could improve the outcome. Hence, the resistance mechanisms should be revised, T790M is detected in 60% before therapy (reported by SLCG) and MET signaling can cause inter-receptor crosstalk with AXL, EPHA2, JAK1 and CDCP1. BIM mRNA expression is an independent predictive marker of response to erlotinib (reported by SLCG). High BIM can predict response and early adaptive resistance via IL6-STAT3-Bcl2. In those patients with low BIM, STAT3 can be activated through over-expression of AXL and EPHA2. Crosstalk between TGFb and NOTCH signaling can also be involved. Methods: Transcripts involving STAT3 signaling were examined from 80 EGFR mutant NSCLC pts, including BIM, STAT3 and SHP2. Also MET, AXL, Mer, EphA2 and Gas6 were analyzed. Together with TGFb and NOTCH constituents: ADAM17, NUMB, HES1, DUSP1, RBJ, LncRNA-ATB, SLUG, GLI and PKCi. Repurposing drugs, in combination with gefitinib, erlotinib and afatinib, were examined in cell cultures and in subcutaneous and orthotopic xenograft mice models. Results: Data will be presented on the correlation of BIM and SHP2 with regard to the mechanism of activation of STAT3 and the influence in response, PFS and OS in the 80 erlotinib-treated EGFR mutant NSCLC pts, and the influence of the above mentioned signaling constituents. Early data shows synergism of gefitinib with AXL inhibitor in the PC9-R cells. Also, activity of niclosamide and paclitaxel. Conclusions: Based on the results, the SLCG-GFCP will carry out a trial with erlotinib in combination with the most active repurposing drugs found, customized by BIM expression. Disclosure: All authors have declared no conflicts of interest.

Aim: The use of a CGA is recommended to detect a patient's vulnerability but its integration in treatment decision making has never been prospectively evaluated. Our main objective was to test if the use of CGA in the allocation of treatment could improve the management of advanced NSCLC in first line. Methods: Randomized, multicentric, prospective phase III study in patients ≥70 y, PS 0-2 with stage IV NSCLC. Arm A standard algorithm of chemotherapy allocation: carboplatin based doublet in PS ≤ 1 and age ≤75y, docetaxel in PS =2 or age >75y. Arm B treatment allocation based on CGA: carboplatin based doublet for fit patients, mono-therapy for vulnerable patients and BSC for frail patients. Four cycles were given every three weeks. Main endpoint: time to failure free survival (TFFS) Secondary endpoints: overall Response Rate (ORR), overall survival (OS), toxicity, QoL and life expectancy adjusted on QoL. QoL was assessed during treatment by EQ-5D health questionnaire at baseline, weeks 6, 12, 20, 28, and 36. A mixed-effects model was used, to compare the utility score and therefore the QOL between arms A and B. Results: 493 patients were randomized from 01/2010 to 01/2013 by 45 centers. Respectively in arms A and B, 34.4% and 47% of patients received a carboplatin-based doublet, 65.6% and 31.4% received docetaxel and in arm B 21.5% received BSC. Median TFFS was 3.2 m, 95%CI:[2.9; 4.1] for standard arm and 3.1 m, 95%CI:[2.7; 4.4] for experimental arm, p = 0.71. Compliance with QoL was 87% at baseline and respectively 66%, 65%, 54%, 60% and 57% at week 6, 12, 20, 28 and 36. The utility score at each evaluation is superior in the experimental arm than in the standard arm, but this difference is significant only at week 36 (p = 0.02). Using a linear mixed generalized model, the utility score tends to decrease over time and is not significantly different between the two arms (p = 0.85). Life expectancy adjusted on QoL was 130.1 days in standard arm and 133.3 days in experimental arm, p = 0.51. Conclusions: ESOGIA did not show a superiority of a CGA-based treatment allocation. In the experimental arm, more patients received a carboplatin-based doublet and 21% of frail patients received an exclusive BSC management. Despite a trend in favor of CGA arm, there is no significant difference in terms of QoL score and life expectancy adjusted on QoL. Further sudies are needed for GA in NSCLC elderly patients. Disclosure: C. Chouaid: advisory board LILLY; H. Lena: advisory board Lilly. All other authors have declared no conflicts of interest.

Aim: The Spanish Lung Cancer Group (SLCG) undertook an industry-independent, two biomarker-directed, randomized trial in advanced NSCLC p. The study (NCT00617656/GECP-BREC) compared non-selected cisplatin-based chemotherapy with therapy customized according to BRCA1/RAP80 mRNA levels. The trial was closed prematurely due to a detrimental effect in the biomarker-directed arm. Further genetic analysis defined PALB2, the bridging molecule that connects BRCA1 and BRCA2, as a promising biomarker to elucidate DNA repair mechanisms. Methods: We assessed mRNA levels of PALB2, RIF1, 53BP1, RNF8 and ATM (as biomarkers with critical roles in homologous recombination [HR]) in tissue from 177 cisplatin plus docetaxel-treated NSCLC p in the BREC trial. We correlated our findings with PFS, OS and response. Results: Median age 62; 79.1% male; 51.4% adenocarcinoma. Results of PALB2 mRNA expression were intriguing. PFS was 5.6 months (m) for p with high/intermediate (H-I) PALB2 and 4.1 m for p with low (L) PALB2 (p = 0.0018). OS was 13.2 m for p with H-I PALB2 compared to 9.9 for p with L PALB2 (p = 0.0377). In the univariate analysis, H-I PALB2 was a marker of longer PFS (HR = 0.56, 95% CI, 0.38, 0.80; p = 0.002) and OS (HR = 0.64, 95% CI, 0.41, 0.98; p = 0.0394). In the multivariate analysis, only H-I PALB2 was associated with longer PFS (here HR = 0.56 and p = 0.0022) and OS (HR = 0.61 and p = 0.0343). Among 143 p with data for response and PALB2 expression, 49.5% of p with H-I PALB2 were responders, compared to only 28% with L PALB2 (p = 0.0131). No significant differences were found for PFS, OS or response according to expression status of the other 4 biomarkers. Conclusions: Our findings reveal PALB2 to be a predictive and prognostic biomarker in advanced NSCLC p treated with docetaxel plus cisplatin. The BRCA1-PALB2-BRCA2 network is a critical determinant of responsiveness to DNA interstrand cross-linking agents and antimicrotubule agents. In our study, BRCA1 had no effect on treatment outcome of this population. PALB2 was found to be a strong marker to predict sensitivity to antimicrotubule agents, without affecting sensitivity to DNA damage-based chemotherapy. Disclosure: All authors have declared no conflicts of interest.

Aim: Patients (pts) with RAS mutations other than KRAS exon 2 do not benefit from P. We explored the effect of these mutations in the efficacy of P-FOLFOX4 or P-FOLFIRI, as 1st-line treatment in WT KRAS CRC pts with LLD. Methods: Phase II, open-label, multicentre study in which WT KRAS CRC pts ≥ 18 years (y) with ≥4 liver metastases; at least 1 >10 cm; or not resectable were randomized to P-FOLFOX4 or P-FOLFIRI. Objective response rate (ORR), progression-free survival (PFS) and overall survival (OS) were analysed according to RAS status (exons 2, 3, 4 of KRAS/NRAS). Results: 77 pts were analysed (38 P-FOLFOX4 [82% male, median age: 65 y]/ 39 P-FOLFIRI [72% male, 63 y]). It was possible to determine RAS status in 83.1% of pts (82.8% WT, 17.2% mutant). Higher ORR and longer PFS and OS were observed in WT-RAS versus mutant RAS subgroups in both P-FOLFOX4 and P-FOLFIRI groups, although differences were not significant, probably due to small sample sizes (Table). There were no significant differences between P-FOLFOX4 and P-FOLFIRI within each of the RAS strata, either. Peri-operative and overall safety were similar, except for higher grade 3/4 neutropenia (39.5% vs 10.3%; p = 0.003) and neuropathy (13.2% vs 0%;p = 0.025) in the P-FOLFOX4 arm. P-FOLFOX4 % or median (95% CI) P-FOLFIRI % or median (95% CI) P value* P-FOLFOX4 vs P-FOLFIRI Total WT RAS, n 27 26 53 ORR (not confirmed) 77.8 (62.1–93.5) 73.1 (56.0–90.1) 75.5 (63.9–87.1) PFS, mo 12.8 (6.2–22.0) 14.8 (7.1–18.7) 0.675 13.4 (9.9–18.6) OS, mo 39.0 (26.4–NA) 45.8 (32.8–51.5) 0.634 44.7 (32.8–51.5) Mutant RAS, n 4 7 11 ORR (not confirmed) 50 (1.0–99.0) 57.1 (20.5–93.8) 54.6 (25.1–84.0) PFS, mo 15.4 (3.7–24.9) 12.6 (1.9–16.2) 0.699 12.6 (3.7–24.9) OS, mo 31.4 (20.6–NA) 42.4 (13.7–42.4) 0.391 31.4 (13.7–NA) P value * WT vs Mutant RAS PFS 1.000 0.238 0.403 OS 0.843 0.616 0.724 * Wilcoxon test; NA = not achieved; CI = confidence interval; mo = months. Conclusions: In pts with WT KRAS CRC and LLD, P plus chemotherapy provide a high response rate, allowing potentially curative resection. Better outcomes are observed in the WT RAS subgroup, without differences between the two regimens. Supported by Amgen S.A. Disclosure: B. Massuti Sureda: Consultant or advisory role: Amgen Honoraria. Amgen; M. Valladares-Ayerbes: Consultant or advisory role: Merck, Amgen Honoraria: Merck, Amgen Research funding: Merck, Amgen E. Aranda: Advisory Role: Roche, Merck. All other authors have declared no conflicts of interest.

Background Tumor EGFR and KRas mutations are both predictive and prognostic biomarkers in patients with advanced NSCLC. We analyzed the correlation between these biomarkers and treatment outcomes in a phase III trial of 3rd/4th line sorafenib in patients with NSCLC. Methods The global, randomized, placebo-controlled MISSION trial enrolled 703 patients with advanced relapsed/refractory NSCLC of predominantly non-squamous histology. The primary study endpoint was overall survival (OS). EGFR and KRas mutations were analyzed in archival tumor samples and in circulating tumor DNA isolated from plasma. Results Tumor and/or plasma mutation data were available from 347 patients (49%). EGFR and KRas mutations were detected in 89 (26%) and 68 (20%) patients, respectively, and were well balanced between treatment arms. Analysis of the interaction between EGFR mutation status and treatment effect on survival suggested that patients with EGFR mutations (mEGFR) benefitted from sorafenib, while those with wild-type EGFR (wtEGFR) did not (p = 0.023). Median OS was two-fold longer in mEGFR patients receiving sorafenib versus placebo (423 vs 197 days, HR 0.48, p = 0.002). There was no significant difference in OS between patients with wtEGFR receiving sorafenib or placebo (253 vs 256 days, HR 0.92, p = 0.559). An interaction was also seen between EGFR mutation status and the sorafenib effect on PFS (p = 0.015). Patients with mEGFR treated with sorafenib had better outcomes compared to placebo based on Cox regression analysis (HR 0.27, p Conclusion Post-hoc analyses of efficacy outcomes in MISSION suggest that advanced NSCLC patients with EGFR mutations may derive a survival benefit from receiving 3rd/4th line sorafenib. These results must be interpreted with caution due to the small, non-representative nature of the genetic biomarker subpopulation analyzed in this trial. Further prospective investigation may be warranted. Disclosure T.S.K. Mok: Honoraria: AstraZeneca, Roche, Eli Lilly, Merck Serono, Eisai, BMS, BeiGene, AVEO, Pfizer, Taiho, Boehringer Ingelheim, and GSK Biologicals Speaker: Astrazeneca, Roche, Eli Lilly, Boehringer Ingelheim, and Merck Serono Research funding: Astrazeneca. L. Paz-Ares: Dr. Paz-Ares has received honoraria from Bayer HealthCare Pharmaceuticals, Lilly, Roche and Pfizer. Y. Wu: Dr. Wu has received lecture fees from Roche, AstraZeneca, Eli Lilly, and Pfizer. V. Hirsh: Member of the steering committee for the MISSION trial. C. Lathia: Dr. Lathia is an employee of Bayer HealthCare. T.J. Ong: Dr. Ong is an employee of, and owns shares in, Bayer HealthCare. C. Pena: Dr. Pena is an employee of Bayer HealthCare. All other authors have declared no conflicts of interest.

Background High BRCA1 mRNA levels affected progression-free survival (PFS) to erlotinib in EGFR-mutant NSCLC p (Rosell et al. CCR 2011). LMO4 is a negative regulator of BRCA1 function, and CtIP can bind to BRCA1 and LMO4. We have assessed the expression of CtIP, LMO4 and BRCA1 and examined the impact of CtIP and LMO4 levels on outcome. Methods mRNA expression of BRCA1, LMO4 and CtIP was examined by RT-PCR in the original pretreatment tumor biopsies of 72 erlotinb-treated NSCLC p with sensitive EGFR mutations. Results p characteristics: median age, 68; 61.8% female; 98.2% Caucasian; 63.6% never-smokers; 81.8% ECOG PS Conclusions Low BRCA1 and high LMO4 levels were associated with longer PFS in erlotinib-treated advanced NSCLC p with EGFR mutations. We recommend baseline assessment of BRCA1 and LMO4 mRNA expression to predict outcome and provide alternative individualized treatment to patients when indicated. HR 95%CI P BRCA1 ≤4.92 1 ref. 4.92-10.7 5.32 1.45-19.52 0.03 >10.7 5.13 1.25-20.99 0.02 CtIP ≤1.21 1 ref. 1.21-2.1 0.69 0.24-2.02 0.50 >2.1 1.10 0.42-2.89 0.84 LMO4 ≤1.29 6.02 1.51-23.94 0.01 1.29-1.86 2.81 0.85-9.21 0.09 >1.86 1 ref. Disclosure All authors have declared no conflicts of interest.

Background Preclinical and clinical data have shown that low BRCA1 expression increases sensitivity to platinum. However, the complexity of DNA repair pathways suggests that the BRCA1 function could be modulated by several proteins. MDC1, the protein initiating the response to DNA double strand breaks, activates two parallel pathways: 53BP1 and BRCA1. We hypothesized that the mRNA expression of components involved in the 53BP1 pathway could influence the predictive value of BRCA1. Methods mRNA expression levels of BRCA1, MDC1, UBC13, RNF8, 53BP1, PIAS4, MMSET and UBC9 were assessed by real-time PCR analysis in 115 paraffin-embedded tumor samples collected from advanced NSCLC p treated with first-line platinum-based chemotherapy without taxanes. Results Expression levels of BRCA1 were correlated with MDC1 (r = 0.47, P Conclusions Advanced NSCLC p with low levels of both BRCA1 and 53BP1 showed significantly improved outcome to first-line platinum-based chemotherapy. The assessment of BRCA1 and 53BP1 can be useful for customizing chemotherapy. Disclosure All authors have declared no conflicts of interest.

Background At the final cutoff of April 2012, in the randomized phase III EURTAC trial, erlotinib improved progression-free survival (PFS) in comparison with chemotherapy as first-line treatment in European p with EGFR-mutation-positive NSCLC (HR, 0.34; P Methods We have assessed EGFR T790M and TP53 mutations, the EML4-ALK translocation and BIM mRNA expression in pretreatment tumor samples of 95 p from the EURTAC trial and correlated our findings with outcome. Results Concomitant T790M was found in 37.89%, TP53 in 24.21% and EML4-ALK in 15.8% of p. BIM expression was low or intermediate in 55.8% of p and high in 31.6%. 86.7% of the 45 p receiving chemotherapy had crossed over to receive EGFR TKIs at the time of progression. In p treated with erlotinib, overall response rates (ORR) were 87.5% in p with high BIM expression and 34.6% in p with low/intermediate BIM; ORR in the chemotherapy group were 11.1 and 14.2, respectively (P = 0.0002). Among p in the erlotinib group without T790M mutations, ORR was 100% for p high BIM expression vs 35.2% for p low/intermediate BIM levels (P = 0.01). In the multivariate analysis, only erlotinib (HR, 0.36; P Conclusions Our results can lead to the design of studies of treatment based on the presence of the EGFR T790M mutation and BIM expression levels. We are currently designing a clinical trial based on our findings. Disclosure All authors have declared no conflicts of interest.

We report a case of hemangiopericytoma with simultaneous affectation of soft palate and mediastinum in a 46-year-old male. We describe the outstanding histologic features of this uncommon tumor, as well as its treatment and evolution. This case had a 5-year survival in spite of pulmonary metastases. We also reviewed the literature and discussed this rare clinical presentation of hemangiopericytoma.

M. Gracia, M. Majem, M.T. Martinez Aguillo, N. Martinez Banaclocha, M. Provencio, E. Arriola, M. Codes, A. Artal, R. Cajal, B. Massuti Sureda. Hospital Universitario Central de Asturias, Medical Oncology, Oviedo, Spain; Hospital de la Santa Creu i Sant Pau, Medical Oncology, Barcelona, Spain; Hospital de Navarra, Medical Oncology, Pamplona, Spain; Hospital General Universitario de Elche, Medical Oncology, Alicante, Spain; Hospital Universitario Puerta de Hierro de Majadahonda, Medical Oncology, Madrid, Spain; Hospital del Mar, Medical Oncology, Barcelona, Spain; Hospital Virgen de la Macarena, Medical Oncology, Sevilla, Spain; Hospital Universitario Miguel Servet, Medical Oncology, Zaragoza, Spain; AstraZeneca Farmaceutica Spain, Medical, Madrid, Spain; Hospital General Universitario de Alicante, Medical Oncology, Alicante, Spain