S232 (n ¼ 75) and selected a number of published clinically relevant prognostic/predictive transcriptional gene signatures in order to assess the potential for classification of patients into multiple clusters based on the tumoural region profiled. Results: Consensus annotated gene lists for each of the gene signatures were used, followed by semi-supervised hierarchical clustering within our multi-regional clinical dataset. Combined clustering metrics were used to generated individual “gene signature stability scores” based on the extent to which signatures could cluster samples by patient, regardless of tumoural region (inter-patient heterogeneity (IPH)), tumoural region regardless of patient (intra-tumoural heterogeneity (ITH)) or which displayed little clustering by either tumoural region or patient. Conclusion: Our analyses indicate that there are variations in sensitivity and specificity of clinically relevant gene expression signatures, due to ITH in CRC. These underlying issues are particularly problematic for signatures based on non-epithelial derived gene expression. There was a wide variation in the ability of clinically relevant gene signatures to consistently cluster samples based on IPH signalling, indicating that careful consideration of confounding ITH effects must be taken during selection of gene transcripts in the signature development stage. http://dx.doi.org/10.1016/j.ejso.2016.07.065
154. Short-course r adiotherapy foll owed by neo-adjuvan t chemotherapy in rectal cancer with synchronous liver metastases: A single institution experience Mark Teo1, Fei Sun2, Nathalie Casanova2, Rachel Cooper2, David Sebag-Montefiore1 1 University of Leeds, UK 2 Leeds Cancer Centre, UK Background: Neo-adjuvant treatment strategies vary widely for rectal cancer patients with potentially resectable synchronous liver metastases. Since 2012, our institution has offered neo-adjuvant short-course radiotherapy (SCRT) to the primary tumour followed by neoadjuvant chemotherapy prior to surgery. We report the early and late outcomes of this strategy. Method: From January 2012 to October 2014, 16 rectal cancer patients with potentially resectable synchronous liver metastases received SCRT (25Gy in 5 fractions) at our institution followed by neoadjuvant oxaliplatin-fluoropyrimidine chemotherapy. Clinical data and outcomes were collected retrospectively from electronic records. Pelvic failure was defined as no curative primary resection attempted (including due to distant progression) or pelvic progression/recurrence. Distant failure was defined as curative metastatectomy not attempted (including due to pelvic progression) or distant metastatic progression/recurrence. Results: Fifteen of 16 patients had T3/4 disease. Liver metastases were deemed borderline resectable or resectable at presentation in 12 patients. Five patients also had synchronous potentially resectable non-liver metastases. Mesorectal fascia was involved or threatened in 15 patients. All patients completed SCRT followed by a median of 6 (range: 2e6) chemotherapy cycles. Median follow-up was 20.6 months. At MRI response assessment, 13 patients had partial or complete response of the primary tumour with no cases of primary tumour progression. Seven patients had partial responses of their liver metastases but 7 patients developed systemic progression. Ten patients had primary tumour resections (90% resection margin negative) of which, 7 also had metastatectomies. Two patients had pelvic failures only, 10 distant failures only, and 2 both pelvic and distant failure. Four patients had salvaged pelvic surgery or metastatectomy following recurrence. Two-year pelvic and distant failure free survival and overall survival were 37.7%, 18.8% and 65.8% respectively, with 5(31.3%) patients disease-free. Conclusion: SCRT followed by neoadjuvant chemotherapy is a feasible strategy achieving good pelvic responses, survival and potential cure in early metastatic rectal cancer. http://dx.doi.org/10.1016/j.ejso.2016.07.066
ABSTRACTS 155. Is there a role for palliative pelvic radiotherapy in pelvic symptom control in metastatic/unresectable rectal cancer? Mark Teo1, Yanos Reynolds-Khan2, Victor Chew2, Daniel Swinson2, David Sebag-Montefiore1 1 University of Leeds, UK 2 Leeds Cancer Centre, UK Background: Rectal symptoms of bowel obstruction are reported in 10e15% of patients with primary rectal cancer in situ often needing surgical intervention (colostomy/stenting). Palliative radiotherapy can be offered for rectal symptoms. This retrospective study investigates uptake of palliative radiotherapy and surgical intervention for rectal symptoms. Method: Patients diagnosed with rectal adenocarcinoma whose primary tumours were not resected between 2011 and 2012 were identified from electronic records at the Leeds Cancer Centre. Results: Thirty-nine patients had primary tumours left in situ due to metastatic disease, being unsuitable for surgery, or progression following neoadjuvant treatment. Median survival was 7.8 months. Radiotherapy was not considered in 17 patients. 10/17 had early surgical interventions prior to seeing an oncologist e 6 for rectal symptoms, 4 needed emergency surgery (fistula (N ¼ 1), bowel obstruction (N ¼ 1), perforation (N ¼ 2)). 1/ 17 developed bowel obstruction while receiving primary chemotherapy. 6/ 17 were considered unfit for any oncological treatment by the multi-disciplinary team. Radiotherapy was considered in 22 patients. 6/22 were asymptomatic, had no radiotherapy with none developing bowel obstruction. 1/22 had severe symptoms requiring a colostomy prior to neoadjuvant radiotherapy. 15/22 received radiotherapy (3 neoadjuvant, 1 consolidation following chemotherapy, 11 for rectal symptoms) with only one requiring a colostomy for recurrent symptoms. The colostomy rate following palliative radiotherapy for rectal symptoms was 9.1% (1/11). Conclusion: In incurable rectal cancer patients, colostomies are occurring early in the treatment pathway with the majority for symptom control. Recurrence rates of bowel obstruction symptoms are low following palliative radiotherapy, hence avoiding a stoma. Palliative radiotherapy could be considered earlier for symptomatic patients. http://dx.doi.org/10.1016/j.ejso.2016.07.067
157. Impact of dementia on treatment of older patients with breast cancer. An interim analysis of the bridging the age gap in Breast Cancer Study Osama Zaman1, Sue Ward2, Karen Collins3, Kwok-Leung Cheung4, Thompson Robinson5, Riccardo Audisio6, Oscar Bortolami7, Charlene Martin7, Malcolm Reed8, Lynda Wyld7 1 University of Sheffield Medical School, UK 2 ScHARR, University of Sheffield, UK 3 Sheffield Hallam University, UK 4 The University of Nottingham, UK 5 University of Leicester, UK 6 University of Liverpool, UK 7 The University of Sheffield, UK 8 Brighton and Sussex Medical School, UK Background: One third of all breast cancer cases occur in women over 70 in whom co-morbidity rates may be higher than younger patients. Seven percent of women aged over 65 have dementia and so the two conditions often co-exist. Dementia is associated with a reduced life expectancy and increases the risk of acute or chronic post-operative cognitive dysfunction after surgery. This study has examined the impact of dementia on women with early breast cancer in terms of treatment decisions and post-operative complications. Method: The Age Gap study is a large, multicentre, prospective cohort study of older women with early breast cancer. Interim comparison of treatment allocation and surgical complications in women with and without a clinical or MMSE confirmed diagnosis of dementia, was performed. Comparison of rates of surgery in women with and without dementia was assessed via simple logistic regression. Surgically treated
ABSTRACTS patients with or without dementia were compared using Chi2 to detect any association between the presence or absence of dementia and incidence of local and systemic complications. Results: Data was available for 1965 patients recruited between April 2013 and December 2015, from 51 UK hospitals. Of these 207 (10.5%) had dementia. Median age of the patient cohort was 77 years (70e101). Patients with dementia were 66% less likely to receive surgery than those without dementia (OR 0.344, 95% CI 0.249e0.473, p < 0.001). In patients undergoing surgery, dementia was not a significant predictor of local (p < 0.16) or systemic (p < 0.84) complications. Conclusion: This study suggests that whilst older patients with dementia seem to tolerate surgery well, they are significantly less likely to undergo breast cancer surgery. Whether this will translate into higher breast cancer specific mortality in dementia patients is unknown, although overall survival in dementia patients is known to be reduced. http://dx.doi.org/10.1016/j.ejso.2016.07.068
162. Why do I choose to get screened? An examination of factors influencing participation in the Scottish Bowel Screening Programme Kirsty Bisset1, Paul G. Horgan1, Emilia M. Crighton2, David Mansouri1 1 University of Glasgow, UK 2 Department of Public Health Screening Unit, NHS Greater Glasgow & Clyde, UK Background: Uptake of screening for colorectal cancer using the faecal occult blood test is influenced by demographics such as age, sex and socioeconomic deprivation. In addition, an individuals health beliefs and behavioural factors may be of importance. The aim of the present study was to further examine the health beliefs of individuals participating in the Scottish Bowel Screening Programme (SBoSP). Method: A prospective study of all patients undergoing a colonoscopy following a positive bowel screening test via the SBoSP in our geographical area was carried out. Ultimate (demographics), proximal (health beliefs and social support) and distal (background) factors that could potentially influence participation behaviour were examined using a previously validated questionnaire. Results: Of the 96 individuals invited, 94(98%) participated. There were 45 (48%) males, 50 (53%) under the age of 65 and 55 (57%) patients were in the two most deprived quintiles of deprivation. Examining proximal variables, participants reported low perceived susceptibility to bowel cancer (mean score 9/20), and high levels of response-efficacy (mean score 8/10), self-efficacy (mean score 20/24) and social support (mean score 3/ 4). Examining distal variables, a low fear of bowel cancer (mean score 13/ 32) and high levels of general self-efficacy (mean score 19/25) were noted. There were no significant differences in proximal factors according to age, sex or deprivation status, however there was weak trend for females (mean score 12 vs 14, p ¼ 0.055) and the more deprived (mean score 14 vs 12, p ¼ 0.097) to have a higher fear of bowel cancer. Conclusion: High levels of social support, response-efficacy and selfefficacy are present in screening participants across all demographics. Promotion of response-efficacy, which is an individual’s perception of the efficacy of the screening programme and treatment, in particular, may represent a potential modifiable target to improve uptake. Further work examining these factors in non-participants is required. http://dx.doi.org/10.1016/j.ejso.2016.07.069
168. General practitioner attitudes towards prescribing aspirin to carriers of Lynch Syndrome: Findings from a national survey Samuel Smith1, Robbie Foy2, Jennifer McGowan3, Lindsay Kobayashi3, John Burn4, Karen Brown5, Lucy Side3, Jack Cuzick1 1 Queen Mary University of London, UK 2 University of Leeds, UK 3 University College London, UK
S233 4 5
Newcastle University, UK University of Leicester, UK
Background: The CaPP2 trial provided evidence for the effectiveness of aspirin (600 mg) for cancer prevention among Lynch Syndrome carriers. A dose non-inferiority study comparing 100mg, 300mg and 600mg of aspirin is currently underway (CaPP3 trial). To encourage rapid implementation of the CaPP3 findings, we investigated general practitioner (GP) attitudes towards aspirin prescribing for Lynch Syndrome carriers. Method: We conducted an online survey of UK GPs in 2016 (N ¼ 1007). Using a within-subjects design, GPs read a statement on the harms and benefits of aspirin and indicated their willingness to prescribe aspirin at three doses (100mg, 300mg, 600mg). Doses were presented randomly to prevent order effects. Additional data collected included awareness of Lynch Syndrome (and its associated names) and the preventive effects of aspirin among carriers, and comfort discussing harms and benefits of aspirin. Results: Almost one third (29.2%) of GPs had not heard of Lynch Syndrome, and 62.0% of GPs were unaware of the preventive effects of aspirin among carriers of the syndrome. GPs were most willing to prescribe at the lowest dose (100mg, 91.3%), and there was a significant graded decline for prescribing at 300mg (81.8%) and 600mg (62.3%) (p < 0.001 for trend). In multivariable analyses adjusted for country, practice partnership, gender, and special interests, willingness to prescribe aspirin (600mg) was higher among GPs older than 50 years (OR ¼ 1.46 [95% CI, 1.03e2.07], p ¼ 0.033), and those with more than 10 years’ experience (OR ¼ 1.50 [95% CI, 1.10e2.04], p ¼ 0.010). Two-thirds (68.1%) of GPs indicated they would be comfortable discussing the harms and benefits of aspirin with a Lynch Syndrome patient. Conclusion: There is low awareness among GPs of Lynch Syndrome and the preventive effects of aspirin among carriers. To ensure the optimal dose identified in the CaPP3 trial is readily available to patients, strategies to educate GPs and establish prescribing pathways should be developed. http://dx.doi.org/10.1016/j.ejso.2016.07.070
169. General practitioner attitudes towards prescribing tamoxifen for the primary prevention of breast cancer: Results of a vignette study Samuel Smith1, Robbie Foy2, Jennifer McGowan3, Lindsay Kobayashi3, Karen Brown4, Lucy Side3, Jack Cuzick1 1 Queen Mary University of London, UK 2 University of Leeds, UK 3 University College London, UK 4 University of Leicester, UK Background: A key recommendation of the Independent Cancer Taskforce was to ensure general practitioners (GPs) are appropriately prescribing tamoxifen for breast cancer primary prevention. We investigated the barriers to prescribing tamoxifen for chemoprevention among GPs in the NHS. Method: We conducted an online survey of GPs in England, Northern Ireland and Wales in 2016 (N ¼ 928). Respondents were randomised to read one of four vignettes (1:1:1:1) describing a healthy patient seeking a tamoxifen prescription. Using a 2 2 between-subjects design, the patient’s breast cancer risk (‘moderate risk’ vs. ‘high risk’) and the clinician responsible for initiating the prescription (‘GP prescriber’ vs. ‘secondary care clinician [SCC] prescriber’) were manipulated. Outcomes were willingness to prescribe, comfort discussing harms and benefits, comfort managing the patient, factors affecting the prescribing decision, and awareness of tamoxifen and NICE guidelines (CG164). Results: Half (51.7%) of the GPs knew that tamoxifen can reduce breast cancer risk; a quarter (24.1%) were aware of relevant NICE guidelines. Respondents who were asked to initiate prescribing (‘GP prescriber’) were less willing to prescribe tamoxifen than those who were asked to take over the prescribing role from secondary care (‘SCC prescriber’) (68.9% vs. 84.6%, p < 0.001). The GP prescribers also reported