Prognostic Significance of Deprivation in Upper Gastrointestinal Cancer

Special Article – Gastric Cancer

Gastrointest Cancer Res Ther. 2017; 2(2): 1018.

Prognostic Significance of Deprivation in Upper Gastrointestinal Cancer

Blake PA¹*, Karran AL¹, Chan DSY¹, White C² and Lewis WG¹

¹South East Wales Upper GI Cancer Network, University Hospital of Wales, Cardiff, CF14 4XW, UK

²Welsh Cancer Intelligence and Surveillance Unit, Public Health Wales, Floor 13, Brunel House, Cardiff CF24 0HA, UK

*Corresponding author: Blake PA, Department of Upper GI and General Surgery, University Hospital of Wales, Heath Park, Cardiff, UK CF14 4XW, UK

Received: March 27, 2017; Accepted: April 27, 2017; Published: May 04, 2017


The aim of this study was to determine the influence of the Index of Multiple Deprivation (IMD) and Health Deprivation (HD) on upper gastrointestinal (UGI) cancer outcome.

Consecutive 1185 patients (697 oesophageal, 488 gastric cancer) were studied prospectively. Deprivation scores were calculated using the IMD of the Welsh Government. Mortality data were obtained from the Office for National Statistics (ONS) and this data, as well as survival data, were independently verified by the Welsh Cancer Intelligence and Surveillance Unit. Primary outcome measure was survival from diagnosis.

Median survival for gastric cancer patients was 8 months (0.25 to 64) compared with 10 months (0.25 to 62) for oesophageal cancer patients. Open and close laparotomy for all surgical patients was commoner in patients residing in deprived geographical areas with a 6.5% open and close rate in the least deprived IMD quintile versus 13.5% in the most deprived quintile (P=0.006). On post-operative histopathology, IMD was associated with pT (r=-0.146, P=0.043), pN (r =-0.158, P=0.029), and pM stage (r=-0.189, P=0.016). On univariate analysis survival was associated with oesophageal versus gastric tumour site (P=0.028), histopathological cell type (P<0.0001), age (P<0.0001), radiological (r) TNM stage (P<0.0001), radical treatment intent (P<0.0001), IMD (P<0.0001) and HD (P<0.0001). On multivariate analysis age (HR 1.021, 95% CI, 1.014- 1.028, P<0.0001), rTNM stage (HR 1.559, 95% CI, 1.427-1.704 P<0.0001), radical treatment intent (HR 0.338, 95% CI, 0.274-0.418, P<0.0001), and IMD rank (HR 1.000, 95% CI, 1.000-1.000, P=0.084) were associated with duration of survival.

In conclusion deprivation is an important prognostic indicator in UGI cancer.

Keywords: Gastric cancer; Oesophageal cancer; Deprivation; Surgery; Survival


Deprivation is a broad concept which describes limited access to the opportunities and resources which society might expect such as good health, a clean and safe living environment, and protection from crime [1]. Eight types of deprivation, or domains, have been described, including; employment, income, education, health, community, geographical access to services, housing, and physical environment. Multiple deprivation refers to the different types that might occur, and represents a far more profound notion than poverty alone. Deprivation varies geographically, and Wales is recognised as having relatively high levels when compared with England and several other European countries. Indeed, when compared with the UK as a whole, the general health of the population of Wales is significantly poorer with more emergency hospital admissions per capita, and an overall life expectancy one year shorter when compared with England [2].

Linear relationships between levels of deprivation and survival have been reported for no fewer than 44 of 47 specific anatomical cancer sites, including oesophageal, colon and rectal cancer [3]. Deprivation is also associated with an increased incidence of upper gastrointestinal cancer [4,5], and several reports have highlighted a survival benefit for patients residing in less deprived geographical areas when compared with more deprived areas [6-8]. Discrepancies in cancer related survival cannot be explained entirely by differences in the stage at diagnosis [9,10] or by higher co-morbidity among patients from deprived backgrounds [11]. Moreover, a widening of survival inequality with time has been reported, whereby the improved outcomes experienced by patients living in less deprived geographical areas over the past 25 years have not been shared by patients from the more deprived areas [12-14]. The NHS Cancer Plan of September 2000 [15], and subsequent government targets introduced in 2003, was aimed at reducing such inequalities across the socio-economic divide, and specific and demanding NHS targets were set [16]. It remains to be established whether deprivation per se directly influences outcome in UGI cancer, and if so, whether the effect may be analogue or digital in nature. As prognosis for patients diagnosed with UGI cancer is often poor, the potential benefit from understanding and addressing reversible factors is substantial. The aims of this study were to determine the influence of deprivation on outcomes for patients with UGI cancer, with particular emphasis on survival following potentially curative therapy. The setting was a UK cancer network serving a population of 1.4 million people.

Materials and Methods

Between 1st August 2008 and 31st July 2012, a total of 1185 patients were diagnosed with UGI cancer and managed by the South East Wales UGI multidisciplinary team [median age 72 (22-97) years, 783 male, 402 female, 697 oesophageal, 488 gastric cancer, 903 adenocarcinoma (ACA), 206 squamous cell carcinoma (SCC)]. The details of these patients were collected prospectively and data was cross-referenced with the oncology (CANISC) database. Mortality data were obtained from the Office for National Statistics (ONS) and this data, as well as survival data, were independently verified by the Welsh Cancer Intelligence and Surveillance Unit. Deprivation rankings were designated for each patient using the Welsh Index of Multiple Deprivation (IMD) 2011 [17], as determined by the National Assembly for Wales [1]. This index gives the official measure of multiple deprivation for every postcode in Wales and is based on the eight previously described forms of deprivation. The country is divided into 1,896 areas each having about 1,500 people with the most deprived geographical area ranked 1 and the least deprived area ranked 1,896. The IMD for all areas was sub-classified into equally sized socio-economic quintiles; the most deprived group was labelled quintile 1, and the least deprived quintile 5. These cut-off points allowed subgroup analysis of patients from similarly deprived areas while facilitating comparison across the spectrum. Health deprivation (HD) was also examined, the indicators for which are cancer incidence, all-cause death rate, percentage of live single births <2.5kg, and the number of inhabitants with limiting long-term illness per 100,000 of the population. HD was similarly sub-classified into equally sized quintiles.

Staging investigations

Patients deemed to have potentially curable tumours underwent diagnostic gastroscopy with histopathological confirmation of oesophageal or gastric cancer and computed tomography (CT) of the thorax and upper abdomen. Patients selected for radical treatment also underwent endoluminal ultrasound (EUS), CT Positron Emission Tomography (CT-PET) and laparoscopy, if appropriate. Tumours were staged according to the unified TNM classification of UGI cancer edition 6 [18] until 2010 and edition 7 [19] thereafter.

Multidisciplinary management

Patients were initially discussed at one of three local multidisciplinary team (MDT) meetings and if deemed potentially curative they were then referred to and discussed at the regional South East Wales UGI MDT meeting. The MDT consists of seven specialist upper GI surgeons, oncologists, palliative care physicians, radiologists, pathologists, specialist nurses and dieticians. Patients were selected for appropriate radical treatment based on histopathological stage, co-morbidity, the technical feasibility of surgery and patient choice according to an algorithm described previously [20]. Those not suitable or in favour of radical therapy were offered palliative care by specialist palliative care physicians.

American society of anaesthesiologists grade

The ASA grade was calculated for surgical patients as a measure of co-morbidity. The system has five grades: normal healthy individual; mild systemic disease that does not limit activity; severe systemic disease that limits activity but is not incapacitating; incapacitating systemic disease which is constantly life-threatening; moribund, not expected to survive 24 hours.

Surgical treatment

Surgery was performed by one or a combination of seven upper gastrointestinal surgeons working within the parameters of the MDT. For patients with oesophageal cancer a transhiatal resection as described by Orringer was performed in those with T1-2, N0 tumours [21]. It was also employed selectively for patients with adenocarcinomas of the lower third of the oesophagus which were more advanced (T3 N1) and for patients with associated significant comorbidity (ASA grade III). The remaining oesophageal cancer patients underwent standard subtotal oesophagectomy as described by Lewis or Tanner [22,23]. For those with gastric cancers it was the policy to perform a modified radical D2 resection with extended lymphadenectomy but preserving the pancreas and spleen where possible [24-26]. The definition of a potentially curative resection was that all visible tumours were removed and that both proximal and distal resection margins were free of tumour on histological examination. Morbidity and mortality included all in-hospital complications and deaths. Morbidities were recorded against a specific list agreed by all the surgeons involved and graded using the Clavien-Dindo Classification of surgical complications [27].

Definitive chemoradiotherapy (dCRT)

Patients undergoing dCRT received a treatment protocol which involved four 3-weekly cycles of cisplatin (dose 60mg/m²) and infusional 5-fluorouracil (5-FU, 300mg/m²/day). Cycles three and four were given concurrently with five weeks of radiotherapy (50Gy in 25F), during which time the 5-FU was reduced to 225mg/m²/day. If during the course of treatment the glomerular filtration rate (GFR) was less than 40ml/min or the patients experienced significant neuroor nephrotoxicity, cisplatin was discontinued and replaced with carboplatin.


Patients undergoing surgery were reviewed every three months for the first year and every six months thereafter. Definitive chemoradiation patients were followed up by the oncologists at equivalent periods. Endoscopy and CT were performed if recurrent disease was suspected. Patients treated with palliative intent were followed up by both oncology and palliative care physicians. All patients were followed up for a minimum of 6 months or until death, and no patients were lost to follow-up. Dates of death were obtained from the Office for National Statistics thus ensuring accurate survival times and dates of death for all patients. Nine hundred and eighty five patients (83.1%) were followed up for two years (n=157) or until death (n=828).

Statistical analysis

Statistical analysis appropriate for non-parametric data was used. Grouped data were presented as median (range), and quintiles were grouped to allow accurate Cox regression analysis. Bivariate correlations were calculated using Spearman`s correlation test. Differences were deemed statistically significant when P<0.05. Cumulative overall survival was calculated by the life-table method of Kaplan and Meier [28]. Differences in survival between groups of patients were analysed using the log-rank method [29]. Factors found to be significantly associated with duration of survival on univariate analysis and with P-value <0.10 were entered into a multivariate analysis using Cox’s proportional hazards model. To identify any potential confounding factors, a separate stepwise regression was also performed using the univariate effect of deprivation as the first step. Data analysis was carried out with the Statistical Package for Social Sciences (SPSS) version 20 package (IBM Corporation, New York) (Figure 1).