Assessment of patient-reported outcome measures in the surgical treatment of patients with gastric cancer

Background Gastric cancer is responsible for 10 % of all cancer-related deaths worldwide. With improved operative techniques and neo-adjuvant therapy, survival rates are increasing. Outcomes of interest are shifting to quality of life (QOL), with many different tools available. The aim of this study was to assess which patient-reported outcome measures (PROMs) are used to measure QOL after a gastrectomy for cancer. Methods A comprehensive search was conducted for original articles investigating QOL after gastrectomy. Two authors independently selected relevant articles, conducted clinical appraisal and extracted data (P.J. and J.S.). Results Out of 3414 articles, 26 studies were included, including a total of 4690 patients. These studies included ten different PROMs, which could be divided into generic, symptom-specific and disease-specific questionnaires. The EORTC and the FACT questionnaires use an oncological overall QOL module and an organ-specific module. Only one validation study regarding the use of the EORTC after surgery for gastric cancer was available, demonstrating good psychometric properties and clinical validity. Conclusions A great variety of PROMs are being used in the measurement of QOL after surgery for gastric cancer. A questionnaire with a general module along with a disease-specific module for the assessment of QOL seems most desirable, such as the EORTC and the FACT with their specific modules. Both are developed in different treatment modalities, such as in surgical patients. EORTC is the most widely used questionnaire and therefore allows for comparison of new studies to existing data. Future studies are needed to assess content validity in surgical gastric cancer patients.

questionnaire with a general module along with a diseasespecific module for the assessment of QOL seems most desirable, such as the EORTC and the FACT with their specific modules. Both are developed in different treatment modalities, such as in surgical patients. EORTC is the most widely used questionnaire and therefore allows for comparison of new studies to existing data. Future studies are needed to assess content validity in surgical gastric cancer patients.
Keywords Quality of life Á Gastric cancer Á Gastrectomy Á PROMs Gastric cancer is responsible for 10 % of all cancer-related deaths worldwide, with the highest incidences in Eastern Asia, Eastern Europe and South America [1]. Although multiple treatment modalities exist, surgical resection of the primary tumour and regional lymph nodes is still the only curative treatment available for gastric cancer [2]. Currently, the 5-year survival rate after oesophageal resection is approximately 20 % [3]. With the implementation of minimally invasive techniques and additional treatments such as neo-adjuvant chemotherapy, survival rates have improved and an according number of long-term survivors exists [4][5][6]. Laparoscopic techniques have been shown to improve quality of life sooner after surgery [7].
With increasing survival and decreased morbidity, a shift in interest of outcome parameters is seen from survival and morbidity rates to the impact of radical gastrectomy and chemoradiotherapy on patient-reported outcomes, such as quality of life (QOL) [8]. Information about QOL outcomes should be an important outcome parameter in research regarding the optimal treatment for gastric cancer.
The World Health Organization (WHO) defined QOL as an individuals' perception of their position in life in the cultural context and in the value system in which they live and in relation to their goals, expectations, standards and concerns [9]. QOL data provide direct measures of benefit as perceived by the patient and may be useful in clarifying treatment preferences. Many different questionnaires are available, both validated and non-validated, to assess the quality of life [7]. Although the different instruments focus on different aspects of QOL, no consensus exists as to which instrument is optimal in the assessment of QOL after gastrectomy for gastric cancer [10]. The aim of this systematic review was to assess which PROMs are used in the assessment of QOL after surgery for gastric cancer.

Literature search
To identify all relevant publications, a systematic search in the bibliographic databases PubMed, EMBASE and The Cochrane Library (via Wiley) from inception to 14 October 2014 was performed. Search terms included controlled terms from MeSH in PubMed, Emtree in EMBASE.com as well as free text terms. Free text terms were only used in The Cochrane library. Search terms expressing ''stomach neoplasm'' were used in combination with search terms comprising ''surgery''. Moreover, an extensive search filter for finding patient-reported outcome measures was used, developed by the University of Oxford (''Appendix''). The reference list of included articles was hand-searched for relevant publications.

Selection criteria and definitions
Two authors (P.J. and J.S.) independently evaluated the search findings for potential eligibility for systematic review using the MEDLINE, EMBASE and Cochrane databases. The inclusion criteria were: (1) article published in English language; (2) only full-text articles, no abstracts or case reports were included and (3) the study had to investigate  30:1920-1929 1921 QOL after gastric resection using questionnaires (i.e. nonstructured interviews were not included). (4) Only patients with gastric carcinoma were included. Studies that described gastrointestinal stromal tumours (GIST) and benign tumours were excluded. Distal, proximal, subtotal and total gastrectomies were included. Wedge resections and local resections were excluded. Regarding surgical techniques, both open and minimally invasive procedures were included, and various reconstructive methods were included (i.e. Roux-en Y or Billroth reconstruction).

Data extraction and quality assessment
The reviewers (P.J. and J.S.) extracted the following data from each study: first author, title of the article, year of publication, type of study, type of gastrectomy, type of reconstruction, number of patients included and the PROMs used to assess QOL. All articles that were deemed suitable after fulltext analysis were assessed for quality of the performed study.

Study selection
Initially, the literature search of MEDLINE, EMBASE and Cochrane resulted in 4529 hits, after removal of duplicates 3414 hits remained. The articles were screened based on title and abstract by two different authors (P.J. and J.S.) independently, and this resulted in a selection of 141 articles for full-text analysis. Of these 141 articles, another 115 were excluded since they did not meet the predefined criteria as described in the methods section; 28 articles were published in another language than English; 45 references consisted only of conference abstracts; 39 articles included a different subject; a final three articles were excluded because they did not use questionnaires but self-reported interviews for QOL assessment. Twenty-six articles remained for further analysis. A flow chart of the article selection is depicted in Fig. 1.

Study characteristics
Twenty-six articles were included for full-text analysis, of which twelve articles were prospective cohort studies, six of which were randomized controlled trials, and fourteen were retrospective cohort studies with prospective QOL assessment, including a total of 4690 patients. One study was a development and validation study [11]. There was great dispersion in follow-up data, ranging from 6 months to 5 years. An overview of the included articles is given in Table 1 for prospective articles and Table 2 for retrospective studies.
Secondly, symptom-specific questionnaires were used, namely the Gastrontestinal Quality of Life Index (GIQLI) and the Gastrointestinal Symptom Rating Scale (GSRS). The GIQLI is developed in patients with benign and malignant disorders [14]. The GRSR was initially developed in patients with irritable bowel disease and not specifically designed for oncological or postoperative patients [15]. Only one study assessed QOL with the GIQLI score [16]. The GSRS score was used in two studies and allowed for overall assessment and of assessment of the individual items [17,18]. GIQLI and GSRS are specifically designed for gastrointestinal symptoms, not for overall QOL.
A third group consists of disease-specific questionnaires. The Functional Assessment of Cancer Therapy (FACT) questionnaires consist of a general health module (FACT-G), and disease-specific modules can be added, such as FACT-Ga for gastric cancer [19,20], thus allowing for the assessment of overall QOL and assessment of diseasespecific symptoms by adding the appropriate module. The FACT-Ga is developed in patients with gastric cancer who underwent different treatment modalities, such as gastrectomy, chemotherapy and radiotherapy [20]. One study has used the FACT questionnaire [21].
The European Organisation for Research and Treatment of Cancer (EORTC) questionnaires work in a similar fashion, consisting of a general health questionnaire, the EORTC QLQ-C30, which is aimed specifically at cancer patients [22]. Disease-specific modules can be added, such as the EORTC QLQ-STO22 for gastric cancer. The EORTC QLQ-STO22 is developed in patients with gastric cancer who underwent different treatment modalities, such as surgery, chemo-or chemoradiotherapy in curative or palliative setting [23,24]. The EORTC QLQ-STO22 and the FACT-Ga are site-specific questionnaires that are related to gastric cancer [20,23]. Fifteen out of twenty-six studies have used the EORTC QLQ-C30 of which twelve studies also included the EORTC QLQ-STO22 module.
Only one validation study was identified, which assessed the use of the STO22 module in patients who were operated in curative or palliative setting. The module was found to have a good internal consistency (Crohnbach's alpha's[0.7) and was deemed reliable and sensitive to changes in both individual patient status and differences between patient groups [23].
Postoperative patients are considered a different entity in the DAUGS20 and Korenaga's score, and these questionnaires focus specifically on patients following gastrectomy for cancer [11]. The questionnaires measure treatment-specific symptoms, such as appetite, swallowing, heartburn and diarrhoea [25,26]. The Dysfunction After Upper Gastrointestinal Surgery (DAUGS20) questionnaire was originally designed in gastric and oesophageal cancer patients who had undergone surgery. The DAUGS is designed to measure QOL postoperative, and no baseline measurement is included [26]. An overview of the different PROMs is provided in Table 3.

Discussion
The here-presented systematic review aimed to review what PROMs are available in assessing the QOL in patients with gastric cancer who undergo gastric resection. Ten PROMs were identified in 26 studies regarding different surgical techniques or comparison of different treatment modalities.
Gastrectomy with radical resection margins of 5 cm around the tumour along with adequate lymfadenectomy is currently the only curative therapy available in gastric cancer [27]. Overall QOL and even separate domains of QOL may differ between different treatment modalities. Question remains whether surgical patients should be considered a separate entity, and whether questionnaires should be developed or adapted for patients undergoing gastrectomy. In an optimal setting, the PROMs should allow for overall assessment of QOL, along with specific modules to assess specific effects associated with the disease and treatment [28].
The DAUGS20 and Korenaga's score consider surgical patients to be a different entity. These questionnaires are specifically aimed at the postoperative patient who had surgery for gastric cancer [25,29]. No validation studies regarding these questionnaires were available. DAUGS20 and Korenaga's score are not developed for overall QOL assessment and are preferably to be used alongside a general QOL PROM [26,30]. Since the questionnaires aim specifically at the postoperative patient, they do not allow for comparison of QOL among different treatment modalities such as chemotherapy and radiotherapy. They do allow for comparison of QOL among different surgical techniques.
The EORTC and FACT questionnaires consider gastric cancer patients as a whole. Both the EORTC and FACT questionnaires consist of a general cancer QOL module to which organ-specific module can be added (EORTC QLQ-STO22 and FACT-Ga), allowing for general and diseasespecific QOL assessment between different treatment modalities. Both questionnaires were developed in patients with gastric cancer undergoing different treatment modalities, including surgery. With regard to comparability and reproducibility, the EORTC was used more often and might therefore allow for comparison to conducted studies, taking into account the heterogeneity in research questions, time points of QOL measurement and follow-up.
Fourteen (54 %) of the included studies consisted of retrospective cohort studies. Only six randomized studies were available. Differences in study design, endpoints, patient groups, surgical techniques and time points in the studies further limited assessment and pooling of data. No validation studies were available for the use of these PROMs in patients undergoing surgery for gastric cancer; hence, comparison of the performance of the different PROMs with regard to validity, internal consistency and discriminative ability was not possible.
Future research should focus on content validity of the used questionnaires in postgastrectomy patients in order to assess whether all the important domains are truly assessed and no items are missing. In order to further assess the use of PROMs in treatment of individual patients, our project group is currently aiming to develop a core outcome set of patient-reported outcomes in gastric cancer patients.
In conclusion, in the assessment of QOL in surgical gastric cancer patients, a great variety of PROMs are being used. A questionnaire with a general module to assess overall QOL, which can be supplemented with disease-specific modules allowing for the assessment or QOL of different treatment modalities, seems to be most desirable. With regard to current practice, the EORTC QLQ-C30 with STO22 module was developed in gastric cancer patients with different treatments, and it is used most widely, allowing for comparison of new data to studies that were already conducted. Future research should assess the need for treatment-specific modules.