Objective: This study aimed to compare the postoperative and pathological outcomes between carboplatin, paclitaxel, radiotherapy (CROSS) and 5-FU, leucovorine, oxaliplatin and docetaxel (FLOT) in esophageal adenocarcinoma (EAC) patients from an international, multicenter cohort. Summary of Background Data: Ongoing debate exists around optimum approach to locally advanced EAC, with proponents for perioperative chemotherapy, such as FLOT, or multimodal therapy, in particular the CROSS regimen. Methods: Patients undergoing CROSS (n = 350) and FLOT (n = 368), followed by curative esophagectomy for EAC were identified from the Oesophagogastric Anastomosis Audit. Results: The 90-day mortality was higher after CROSS than FLOT (5% vs 1%, P = 0.005), even on adjusted analyses [odds ratio (OR): 3.97, confidence interval (CI)(95%): 1.34-13.67]. Postoperative mortality in CROSS were related to higher pulmonary (74% vs 60%) and cardiac complications (42% vs 20%) compared to FLOT. CROSS was associated with higher pathologic complete response (pCR) rates (18% vs 10%, P = 0.004) and margin-negative resections (93% vs 76%, P < 0.001) compared with FLOT. On adjusted analyses, CROSS was associated with higher pCR rates (OR: 2.05, CI95%: 1.26-3.34) and margin-negative resections (OR: 4.55, CI95%: 2.70-7.69) compared to FLOT. Conclusions: This study provides real-world data CROSS was associated with higher 90-day mortality than FLOT, related to cardio-pulmonary complications with CROSS. These warrant a further review into causes and mechanisms in selected patients, and at minimum suggest the need for strict radiation therapy quality assurance. Research into impact of higher pCR rates and R0 resections with CROSS compared to FLOT on long-term survival is needed.

Background: No evidence currently exists characterising global outcomes following major cancer surgery, including esophageal cancer. Therefore, this study aimed to characterise impact of high income countries (HIC) versus low and middle income countries (LMIC) on the outcomes following esophagectomy for esophageal cancer. Method: This international multi-center prospective study across 137 hospitals in 41 countries included patients who underwent an esophagectomy for esophageal cancer, with 90-day follow-up. The main explanatory variable was country income, defined according to the World Bank Data classification. The primary outcome was 90-day postoperative mortality, and secondary outcomes were composite leaks (anastomotic leak or conduit necrosis) and major complications (Clavien-Dindo Grade III-V). Multivariable generalized estimating equation models were used to produce adjusted odds ratios (ORs) and 95% confidence intervals (CI95%). Results: Between April 2018 to December 2018, 2247 patients were included. Patients from HIC were more significantly older, with higher ASA grade, and more advanced tumors. Patients from LMIC had almost three-fold increase in 90-day mortality, compared to HIC (9.4% vs 3.7%, p < 0.001). On adjusted analysis, LMIC were independently associated with higher 90-day mortality (OR: 2.31, CI95%: 1.17-4.55, p = 0.015). However, LMIC were not independently associated with higher rates of anastomotic leaks (OR: 1.06, CI95%: 0.57-1.99, p = 0.9) or major complications (OR: 0.85, CI95%: 0.54-1.32, p = 0.5), compared to HIC. Conclusion: Resections in LMIC were independently associated with higher 90-day postoperative mortality, likely reflecting a failure to rescue of these patients following esophagectomy, despite similar composite anastomotic leaks and major complication rates to HIC. These findings warrant further research, to identify potential issues and solutions to improve global outcomes following esophagectomy for cancer. (C) 2020 Elsevier Ltd, BASO similar to The Association for Cancer Surgery, and the European Society of Surgical Oncology. All rights reserved.

AbstractBackgroundThe complexity of oesophageal surgery and the significant risk of morbidity necessitates that oesophagectomy is predominantly performed by a consultant surgeon, or a senior trainee under their supervision. The aim of this study was to determine the impact of trainee involvement in oesophagectomy on postoperative outcomes in an international multicentre setting.MethodsData from the multicentre Oesophago-Gastric Anastomosis Study Group (OGAA) cohort study were analysed, which comprised prospectively collected data from patients undergoing oesophagectomy for oesophageal cancer between April 2018 and December 2018. Procedures were grouped by the level of trainee involvement, and univariable and multivariable analyses were performed to compare patient outcomes across groups.ResultsOf 2232 oesophagectomies from 137 centres in 41 countries, trainees were involved in 29.1 per cent of them (n = 650), performing only the abdominal phase in 230, only the chest and/or neck phases in 130, and all phases in 315 procedures. For procedures with a chest anastomosis, those with trainee involvement had similar 90-day mortality, complication and reoperation rates to consultant-performed oesophagectomies (P = 0.451, P = 0.318, and P = 0.382, respectively), while anastomotic leak rates were significantly lower in the trainee groups (P = 0.030). Procedures with a neck anastomosis had equivalent complication, anastomotic leak, and reoperation rates (P = 0.150, P = 0.430, and P = 0.632, respectively) in trainee-involved versus consultant-performed oesophagectomies, with significantly lower 90-day mortality in the trainee groups (P = 0.005).ConclusionTrainee involvement was not found to be associated with significantly inferior postoperative outcomes for selected patients undergoing oesophagectomy. The results support continued supervised trainee involvement in oesophageal cancer surgery.

Abstract BACKGROUND: Non-small cell lung cancer (NSCLC) accounts for 85% of lung cancer cases, carries a poor prognosis and remains an area of high unmet need. Despite recent advances in treating NSCLC adenocarcinomas with targeted epidermal growth factor receptor (EGFR) kinase inhibitors, few advances have been made in the treatment of squamous cell carcinoma (SCC), which accounts for 25% of NSCLC. Although the incidence of EGFR mutations in SCC is low, a retrospective clinical study of erlotinib in SCC suggested that a proportion of SCC patients would benefit from treatment with an EGFR inhibitor (Tseng et al., 2012). METHODS: NSCLC samples obtained from untreated patients undergoing surgery were collected with ethical consent, disaggregated and implanted subcutaneously in MF-1 nude mice (Harlan UK) admixed with a human stromal cell component to generate patient-derived xenograft (PDX) models (LION PDX models). Tumour material was diagnosed on the basis of operative histology and immunohistochemistry (IHC) used to confirm histological subtype. IHC was also used to evaluate EGFR expression and tumours characterised for mutations in p53, LKB1, EGFR and K-RAS by direct DNA sequencing. Patient-derived xenografts were maintained in vivo and study cohorts of mice dosed when tumours reached 100-200mm3 with either erlotinib or gefitinib or vehicle placebo. Response to treatment was evaluated by tumour growth inhibition and waterfall plots. RESULTS: We have identified a subset of tumours with squamous histology in our LION panel of patient-derived NSCLC xenografts that respond to treatment with a targeted EGFR inhibitor. Four out of 11 SCC models tested (36%) responded to treatment with either erlotinib or gefitinib with greater than 50% growth inhibition and two SCC models responded with ∼30% growth inhibition. Responses correlated well with EGFR expression. Work to characterise a further 8 LION SCC PDX models is also underway. CONCLUSIONS: These data support the proposition that a significant subset of NSCLC patients with squamous histology may benefit from treatment with targeted EGFR inhibitors. Our LION SCC models could be valuable for biomarker studies to identify the basis of the EGFR inhibitor responder subset and for evaluating new agents in a more clinically relevant setting. Citation Format: Aaron Cranston, Andrew Mckenzie, Phil Mallinder, Simon Jiang, Kerry Moakes, Yinfei Yin, Alex Reece-Smith, Anna Grabowska, Martin Page, Rajendra Kumari. Patient-derived xenograft models reveal a subset of clinically relevant squamous non-small cell lung cancers that respond to targeted EGFR inhibition. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 2796. doi:10.1158/1538-7445.AM2013-2796

BACKGROUND:Clinical evidence on patient-reported outcome measures (PROMS) in breast reconstruction is lacking. The aim of this study was to evaluate PROMs in implant-assisted latissimus dorsi (LDI) or tissue-only autologous latissimus dorsi (ALD) flap reconstruction in relation to complications and adjuvant treatments.METHODS:This was a prospective cohort study involving six UK centres. Eligible patients had primary early-stage breast cancer. The European Organization for Research and Treatment of Cancer quality-of-life questionnaire (QLQ)-C30 and QLQ-BR23, Functional Assessment of Cancer Therapy-Breast Cancer scale (FACT-B), Body Image Scale, and Hospital Anxiety and Depression Scale were completed before operation and at 3, 6 and 12 months after surgery.RESULTS:A total of 182 patients (82 LDI and 100 ALD) were recruited between 2007 and 2010 with symptomatic (59·9 per cent) or screen-detected (39·6 per cent) cancers. Some 64·3 per cent had lymph node-negative disease; 30 per cent of the LDI group had radiotherapy, compared with 53·0 per cent in the ALD group (P = 0·004). Early complications up to 3 months after surgery were reported in 66 and 51·0 per cent of patients in the LDI and ALD groups respectively (P = 0·062) and long-term complications (4-12 months) in 48 and 45·0 per cent (P = 0·845). Role functioning and pain (P = 0·002 for both) were adversely affected in the ALD group compared with results in the LDI group, with no significant effects of radiotherapy on any health-related quality of life (HRQL). Chemotherapy and early complications adversely affected HRQL, which improved between 3 and 12 months after surgery (P < 0·010 for all).CONCLUSION:There is evidence of similar HRQL between types of latissimus dorsi breast reconstruction for up to a year after surgery. There appear to be no overarching effects for radiotherapy after mastectomy on the specific HRQL domains studied in the short term. The identification of variables that affect HRQL is important, including their integration into the analysis of PROMs.

Introduction MAGIC1 regimen of three cycles of ECF chemotherapy pre- and post-op is widely used in gastric cancer but its use in oesophageal cancer remains controversial. We have compared post operative survival outcomes of oesophageal and GOJ cancer patients entered into either OEO22 (2 cycles of CF pre-operatively) or MAGIC regimen to determine if either is advantageous. Methods A database of cancer resections was maintained from 2000 until present. Patients fit for treatment were mainly given OEO2 regimen from 2002 to 2006 and MAGIC regimen from 2006-present. The database was searched for patients receiving pre-operative chemotherapy using either ECF/X or CF/X. Kaplan–Meier survival analysis was undertaken using log-rank test for comparisons. Results Median follow-up was more than 8.5 years for the 97 patients with complete data in the OE02 group and 3.7 years for the 138 complete patients in the MAGIC group. Oesophageal cancer patients that had received MAGIC regimen pre-operatively had significantly longer median survival compared to those that had received OE02 (34.0 months vs 23.4 months, p=0.033). A significant benefit was not shown in GOJ cancer (MAGIC: 32.3 months, OE02: 23.5 months, p=0.095) Conclusion Oesophageal cancer patients attending for curative surgery that commenced treatment with MAGIC style chemotherapy have better survival than patients that started OE02 regimen. Competing interests None declared. References 1. Cunningham, et al.NEJM 2006;355:11–20. 2. MRC Working Party. Lancet 2002;359:1727–33.

Abstract Introduction: Evidence for the clinical effectiveness of breast reconstruction based on Patient Reported Outcome Measures (PROMS) is lacking. Methodology evaluating PROMS after types of breast reconstruction has been poor with respect to study design, statistics, missing data and absence of prospective documentation of pre-defined complication data in a systematic review of all studies since 19781. Furthermore, there is no reliable data on the effects of associated radiotherapy (RT) in this context. As a prelude to a proposed randomised trial in breast reconstruction, our aim was to conduct a ‘robust’ cohort study evaluating the effects of either implant-assisted LD (LDI) or tissue only (ALD) LD flap reconstruction in relation to key determinants including clinico-pathological parameters, complications and treatment schedules over a 36 month period. Methods: An MREC approved prospective longitudinal cohort study involving 6 centres commenced in early 2007. Serial PROMS using the EORTC QLQ-C30, BR-23, FACT-B, BIS and HADS, were completed pre-operatively and at 3, 6, 12, 24 and 36 months after surgery. Data up to 12 months were included in this analysis as data were sparse beyond this point; follow-up is ongoing. Demographic and clinical data were compared between the surgical groups. Generalised estimating equations were used to investigate demographic and clinical predictors of HRQL over time. Results: A total of 189 patients (107-ALD, 82-LDI) were recruited, with a mean age of 50 years (range 22–70). Baseline questionnaires were completed by 149 (79%) women, with 167 (88%) available at 12 months. Patients in the ALD group had a higher BMI (mean 27.3 versus 25.2 in LDI, p=0.001) and a greater likelihood of post-mastectomy RT (52% versus 30% in LDI, p=0.004). Only role functioning (p=0.001) and pain (p=0.003) were significantly adversely affected in the ALD v LDI group, with no statistically significant effects from PMRT on HRQL, although chemotherapy impaired global QoL (p<0.001) and social functioning (p=0.001), and increased fatigue (p=0.007). Early complications (< 3 months) significantly impaired HRQL in general, as did poor baseline scores (p<0.01 for various subscales). Significant improvements over time were noted for global QoL, role and social functioning, fatigue, pain and breast symptoms (p<0.001 for all). Conclusion: There is increasing evidence of clinical equipoise between types of LD breast reconstruction and despite acknowledged cosmetic disadvantages the overall effects of PMRT on HRQL are minimal. The identification of important variables that may affect HRQL is crucial in all studies evaluating the effects of surgery on PROMS. Their integration into study results is essential for correct interpretation of clinically based assessments. This remains a challenging aspect in cohort studies, and emphasises the need for pragmatism in design of trials in the field. 1. Winters ZE, Benson JR, Pusic AL. Annals of Surgery 2010;252(6):929–42 Citation Information: Cancer Res 2011;71(24 Suppl):Abstract nr P2-16-01.

Introduction: Treatment recommendations in breast reconstruction should encompass the assessment of both clinical outcomes as well as patient reported outcomes (PROs) that include body image and satisfaction with cosmetic appearance. The Body Image Scale (BIS) has been used to gauge the effects of mastectomy and breast reconstruction as well as study-specific questionnaires (SSQ) that assess patient satisfaction with aesthetic appearance.

Introduction: NICE recommends that the majority of women should be offered immediate breast reconstruction with its potential to improve health related quality of life (HRQL). There is conflicting evidence with a lack of ‘hard’data to best inform clinicians and their patients. Our aim was to evaluate the effects of implant-assisted LD (LDI) versus autologous LD (ALD) breast reconstruction on HRQL over 12 months. Methods: A prospective longitudinal multicentre study commenced in early 2007. Patient reported outcome measures using the EORTC C30 (general HRQL), BR-23 (breast + arm symptoms), Body Image Scale (BIS) and HADS, were completed pre-operatively and at 3, and 12 months after surgery. Longitudinal analyses tested the effects of treatment variables, baseline HRQL, age and time on QL domains (3–12 months). Significance was set at p = 0.01. Results: One hundred and seventy one patients (93 ALD, 78 LDI) were recruited. There were no significant differences in HRQL domains between LDI and ALD (±RT). Chemotherapy patients reported poorer overall HRQL (p < 0.001), poorer role (p = 0.003) and social (p = 0.01) functioning, and greater fatigue (p = 0.002) and depression (p = 0.01). Older patients had fewer HRQL issues (p = 0.01). Significant improvements over time were seen for overall HRQL and other domains (p < 0.001). There were no significant differences between LDI and ALD for patient satisfaction with surgical outcome. Good satisfaction with overall breast appearance and surgical outcome was significantly associated with fewer body image concerns. Conclusion: There is an important need for cumulative clinical evidence in this field on which to base patient informed consent and clinical recommendations.

BIIB015 is an immunoconjugate created for the treatment of solid tumours and is currently in Phase I of clinical evaluation. BIIB015 consists of a humanised monoclonal antibody against the Cripto protein carrying a payload, via a hindered disulphide linker, of the maytansinoid derivative, DM4. Cripto is a GPI-linked protein required for signal transduction of the TGF-beta ligand, Nodal. Cripto has been previously described as an oncogene and fits the classic pattern of an embryonic gene that is re-expressed in a transformed tumour cell. Cripto expression is highly prevalent on a number of solid tumours, including greater than 75% of breast, lung, and colorectal tumours. Our report documents for the first time that targeting the cell surface Cripto protein with an anti-Cripto antibody–cytotoxic conjugate is an effective means of inhibiting or regressing growth of Cripto positive tumours. BIIB015 which utilises a ‘cleavable’ linker containing a disulphide bond exhibits superior activity when compared to huB3F6 mAb conjugates with different linker systems, including one with a ‘non-cleavable’ linker. BIIB015 displays specificity for Cripto in both in vitro and in vivo experiments. In human xenograft models originating from lung (Calu-6), colon (CT-3), testicular (NCCIT) and breast (MDA-MB-231) tumour samples, BIIB015 shows robust activity with results ranging from >50% tumour inhibition to complete tumour regression. The efficacy seen in the MDA-MB-231 model, a triple negative (-HER2, -ER, and -PR) tumour, is particularly exciting since there is currently no approved therapy for this indication. In addition, BIIB015 can be combined with standard of care chemotherapeutics for enhanced efficacy.

Introduction: NICE recommends that the majority of women undergoing mastectomy should be offered immediate breast reconstruction with its potential benefits to improve health related quality of life (HRQL). A systematic review shows poor and conflicting evidence with a lack of 9hard9 data to best inform both clinicians and their patients. Our aim was to evaluate the effects of implant-assisted LD (LDI) versus autologous LD (ALD) breast reconstruction on HRQL over 36 months. Recently, with the exception of the USA Breast-Q module there is a lack of a validated Breast reconstruction questionnaire. Methods: An MREC approved prospective longitudinal cohort study involving 6 centres commenced in early 2007. Serial patient reported outcome measures using the EORTC C30 (global QoL, physical functioning, fatigue, pain), BR-23 (breast and arm symptoms), 10 item Body Image Scale (BIS) and HADS, were completed pre-operatively and at 3, 6, 12, 24 and 36 months after surgery. Overall cosmetic satisfaction of the breast reconstruction was measured by a study-specific question on a 5 point Likert scale. Generalised estimating equations were used to assess change over time and differences between treatment groups.The Spearman9s rank correlation coefficient was used to assess associations between patient reported cosmetic satisfaction with BIS and breast symptoms. Results: 118 patients (65 – ALD, 53 – LDI) were recruited to the study with a mean age of 50 years (range 22-70). Compliance with questionnaires at all time points was between 85-90%. There were no significant differences in HRQL domains between LDI and ALD, except a tendency for worse pain with ALD (p=0.06). Significant improvements over time were seen for overall HRQL (p Conclusion: There is an important need for cumulative clinical evidence in this field on which to base patient informed consent and clinical recommendations. Further analyses will assess whether there is an independent effect of RT on HRQL. A validated EORTC breast reconstruction module is in development. Citation Information: Cancer Res 2009;69(24 Suppl):Abstract nr 3106.