Evaluation of post-operative surveillance strategies for esophageal and gastric cancers: a systematic review and meta-analysis

Summary Background There is no consensus or guidelines internationally to inform clinicians of how patients should be monitored for recurrence after esophagogastric resections. Aim This systematic review and meta-analysis summarizes the latest evidence investigating the usefulness of surveillance protocols in patients who underwent esophagectomy or gastrectomy. Methods A systematic review of the literature was performed using MEDLINE, EMBASE, the Cochrane Review and Scopus databases. Articles were evaluated for the use of surveillance strategies including history-taking, physical examination, imaging modalities and endoscopy for monitoring patients post-gastrectomy or esophagectomy. Studies that compared surveillance strategies and reported detection of recurrence and post-recurrence survival were also included in the meta-analysis. Results Fifteen studies that described a surveillance protocol for post-operative patients were included in the review. Seven studies were used in the meta-analysis. Random-effects analysis demonstrated a statistically significant higher post-recurrence survival (standardized mean difference [SMD] 14.15, 95% CI 1.40–27.26, p = 0.03) with imaging-based planned surveillance post-esophagectomy. However, the detection of recurrence (OR 1.76, 95% CI 0.78–3.97, p = 0.17) for esophageal cancers as well as detection of recurrence (OR 0.73, 95% CI 0.11–5.12, p = 0.76) and post-recurrence survival (SMD 6.42, 95% CI –2.16–18.42, p = 0.14) for gastric cancers were not significantly different with planned surveillance. Conclusion There is no consensus on whether surveillance carries prognostic survival benefit or how surveillance should be carried out. Surveillance may carry prognostic benefit for patients who underwent surgery for esophageal cancer. Randomized controlled trials are required to evaluate the survival benefits of intensive surveillance strategies, determine the ideal surveillance protocol and tailor it to the appropriate population.


INTRODUCTION
Gastric and esophageal cancers are aggressive and associated with a 5-year overall survival rate of approximately 25%. 1,2 This poor prognosis can be primarily attributed to a recurrence rate as high as 80% even after multi-modal treatment with curative intent. [3][4][5][6] Routine post-operative surveillance of patients can help to detect recurrence and initiate intervention early. However, there is no international consensus on the best strategy for follow-up of patients who underwent surgery with curative intent for esophago-gastric cancers. 7 Surveillance has been beneficial in detecting recurrence in patients who underwent surgical resection for colorectal cancer, leading to improved survival outcomes. There is well-established evidence that have demonstrated an overall survival advantage from intensive surveillance in patients with colorectal cancer. 8,9 Subsequently, surveillance protocols are now standard in various international guidelines. 10,11 In contrast, the evidence for routine intensive surveillance for esophago-gastric cancers is ambiguous. Some studies have reported success in detecting asymptomatic recurrences earlier than symptomatic recurrences when using an intense post-operative surveillance program. Other studies noted that the overall survival was not increased in these studies, especially when factoring in lead-time biases. 12 Currently, there is no randomized controlled trial that has investigated the impact of intensive routine surveillance against symptom-triggered surveillance. As a result, there is subsequently no internationally accepted surveillance strategy or regime for patients 1 after surgical resection. Messager et al. 13 reported a wide geographical variation in surveillance strategies and a lack of consensus between the guidelines even within a relatively homogeneous group of countries in western and central Europe. The most recent review on this topic was in 2012, and since then, more work has been carried out to determine if intensive surveillance has a prognostic benefit, and more importantly, the structure of the ideal regimen. 14 Furthermore, routine surveillance was traditionally not advocated given that the treatments for recurrent esophago-gastric cancers were limited. The recent advances in the radical management of recurrent cancers with multimodal therapy (surgery and/or chemoradiotherapy) provides further justification for adopting a more intensive approach towards post-operative surveillance. This systematic review and meta-analysis aims to summarize the latest evidence investigating the usefulness of various surveillance protocols in patients who underwent surgical intervention for esophago-gastric cancers.

METHODS
A systematic review of studies evaluating recurrence and survival in patients who were followed-up after surgery for esophageal and gastric cancers was conducted. This systematic review follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. 15 Studies that compared planned and unplanned surveillance strategies were subsequently included in the meta-analysis.

Search strategy
A literature search was performed in MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, and CINAHL databases on the 17 July 2021 to identify relevant studies describing surveillance strategies in patients who have undergone surgery for esophageal and gastric cancers. The search included the following keywords 'esophageal cancer,' 'esophageal adenocarcinoma,' 'esophageal squamous cell carcinoma,' 'esophagectomy,' 'Ivor Lewis,' 'McKeown,' 'trans-hiatal esophagectomy,' 'gastric cancer' OR 'gastric adenocarcinoma' OR 'gastric squamous cell carcinoma'; 'gastrectomy' OR 'total gastrectomy' OR 'partial gastrectomy' OR 'subtotal gastrectomy' OR 'distal gastrectomy'; and 'surveillance' OR 'monitoring' OR 'followup.' The individual strings were combined using the AND modifier. References of included articles were screened and a hand-search was performed to identify missing articles. Two reviewers (SC and VS) independently assessed the titles and abstracts for inclusion of relevant references. In cases where there was disagreement for inclusion, a third author (SRM) was consulted.

Study selection
Randomized controlled trials (RCTs), quasi-randomized trials, cohort studies and case-control studies were included if they investigated the use of a surveillance protocol in patients who had undergone surgery for gastric cancer or oesophageal cancer, or compared the factors between symptomatic and asymptomatic recurrence. Any form of surveillance such as regular clinical follow-up (history and examination), blood tests, radiological investigations and endoscopy were considered as part of a surveillance strategy. Any type of surgery carried out with a curative intent regardless of operative technique, approach and additional procedures was included. Owing to the heterogeneity of cases included, it was not feasible to control for a specific grade, stage or histological sub-type of gastric cancer, hence all cancers of the stomach were included. Studies involving patients who underwent surgery alongside adjuvant or neoadjuvant chemotherapy were included as well.
Review articles and case reports were excluded. Studies were also excluded if there was no comparative group present. Comparative studies were excluded if no outcome data was provided for the control or the intervention group. Studies were excluded if surveillance post-surgery was performed in a cohort of patients who would have required monitoring regardless, such as patients with hereditary syndromes (CHD1 mutation leading to hereditary diffuse gastric cancer). Finally, articles were excluded if patients did not undergo surgery but underwent only other curative therapies and if other cancer types were included.

Outcome measures and data extraction
Our main aim was to assess the impact of postoperative surveillance on survival outcomes, hence primary outcome measures were length of survival and detection of recurrence. Secondary aims included readmissions as well as trigger of other investigations or interventions (e.g. imaging shows evidence of recurrence hence endoscopy was carried out); and the incidence of postoperative morbidity assessed by a Clavien-Dindo Classification (CDC) of 2 or higher. In addition, the following data were extracted from each study: first author, year of publication, study design, sample size, demographic data (age and gender), oncological details (stage, grade and histological subtype of cancer); surgical intervention (operation, approach, intervention, details of lymphadenectomy); other forms of treatment undergone (adjuvant and/or neoadjuvant chemotherapy and/or radiotherapy) and details of surveillance (frequency, setting, persons involved, clinical evaluation carried out, investigations undergone and why these were performed). The definition of recurrence was based on the primary studies, which used a combination of clinical symptoms, CT or PET evidence, and direct visualization on endoscopy. However, given that the included studies investigated the use of CT or PET, detection of recurrence was based primarily on evidence from imaging.

Quality assessment of selected studies
Two reviewers (SC and VS) assessed quality of each included study by independently evaluating the risk of bias using the Newcastle-Ottawa Scale (NOS) for the assessment of non-randomized studies. 16,17 The NOS scores ranging from 0 to 9, with a higher score indicating a lower risk of bias. In this review, we considered a score of 0-3, 4-6 and 7-9 as low, moderate and high quality of studies, respectively.

Statistical methods
Review Manager 5.3 (Cochrane Collaboration, Oxford, United Kingdom) was used for statistical analysis of the data. Two types of modeling were used to assess the heterogeneity of the data: fixedeffects and random-effects. The random-effects model was chosen for all analysis due to the significant heterogeneity between studies. Data are given as odds ratio and 95% confidence intervals (CI) for all non-continuous data, and as standardized mean difference and 95% CI for all continuous data. In all cases, statistical heterogeneity was assessed by using I 2 statistic and was categorized as low, moderate and high for an I 2 statistic of above 25%, 50% and 75%, respectively. Results above 60% were considered as substantial heterogeneity. All data given as medians were converted to means and standard deviations, as outlined by Hozo et al. 18

Quality appraisal
Assessment of studies using the Newcastle-Ottawa tool showed that studies were of a moderatehigh quality ( Table 1). The non-randomized studies included were evaluated for sources of bias using the NOS. All studies achieved an excellent score of 7/8 Bilici * * * * * * * Hosokawa * * * * * * * * Lee * * * on the NOS. They attained maximum points for the 'selection' category. Some risk of bias was present due to heterogeneity of the population that reduced the comparability of the study cohorts. Two studies did not report an accurate follow-up time.

Study and patient characteristics
Fifteen different studies were included in this paper. In total, this accounted for 6271 patients. Patients underwent a combination of partial gastrectomy, subtotal gastrectomy or total gastrectomy for gastric cancers; one study did not explicit state the type of gastrectomy carried out. 27 One study included only patients with early gastric cancer, while other studies included patients with tumors of stages T1-T4 (Table 2). Accordingly, except 1 study, all other patients underwent chemotherapy and/or radiotherapy. Histological subtypes as reported in six of the studies included poor to well-differentiated carcinoma of adeno, squamous or signet ring morphologies. Studies on esophageal cancer included T1-T4 cancers in most studies of both adenocarcinoma and squamous cell histological subtypes. A range of procedures including transthoracic, trans-hiatal, vagal-sparing or minimally invasive thoracoscopic/laparoscopic approaches were used. Some studies also carried out gastrectomy due to tumor location. All of the studies involved a combination of surgery and chemoradiotherapy as per standardized guidelines ( Table 3).

Detection of recurrence rate
In total, 11 studies included a comparison of detection rate of recurrence rate between planned and unplanned surveillance strategies. The total sample size was 3032 patients and 3239 patients in the planned and unplanned surveillance cohorts, respectively. The prevalence of recurrence detected was 1145 and 1086 events in the respective groups. Of these, 674 and 578 patients were detected in    (Fig. 2).

Post-recurrence survival
In total, seven studies included a comparison of postrecurrence survival between planned and unplanned surveillance strategies.  (Figure 3). Eight studies also reported on the overall survival between planned and unplanned surveillance. Three of the studies reported a better overall survival with planned surveillance. 23,28,30 Five studies reported no advantage in overall survival. For example, Antonowicz et al. reported no overall survival advantage with planned surveillance. In contrast, Abate et al. demonstrated that the grouped overall survival was significantly longer in patients treated for the recurrence, although it is difficult to fully attribute this to surveillance strategy, as they had combined both cohorts in survival analysis. For these reasons, a meta-analysis could not be performed as not all the required data for pooled measures was available. 6,12,19,26,27

DISCUSSION
Our study is an updated review of the surveillance programs available for patients who have undergone surgery for gastric and esophageal cancers. Currently, there are no published or ongoing randomized trials comparing intensive surveillance for asymptomatic recurrence with symptom-driven followup for esophago-gastric cancers. Thus, we have only identified observational evidence to answer this question. Furthermore, the latest study on this topic was published in 2016, indicating that while there has been much debate around how surveillance should be performed, there has been minimal work dedicated towards it. Based on the studies included in the quantitative analysis, there is no concrete evidence to show that planned surveillance improved the detection of recurrence post-esophagectomy, although post-recurrence survival was improved. In contrast, these effects were not observed for planned surveillance post-gastrectomy. These results, whilst no different to earlier summaries on this topic, emphasize the need for more rigorous evidence in this area.
The main argument for intensive surveillance is due to the high recurrence rate of the cancers. Detection of recurrence at an earlier potentially treatable stage before symptoms arise provides a longer time frame for intervention and may improve survival outcomes. The included studies that carried out routine surveillance reported a high detection rate of recurrence amongst asymptomatic patients, the highest being 75% by Eom et al. 27 However, the impact on survival outcomes is mixed. Jiang et al. reported that patients who underwent intensive surveillance had similar pathological characteristics and peri-operative treatments but had similar postrecurrence and overall survival. 23 In another study, Sisic et al. reported that survival analysis of 242 propensity-matched patients showed significantly improved overall survival by 100% for patients with standardized follow-up, and further multivariate analysis showed intensive surveillance to be a positive prognosticator. 21 In contrast, other studies carried out in Asia, Canada and Europe have reported no clear survival benefit. 24,26 It is unclear why detection of recurrence did not translate to a survival advantage in these studies, although this may partly be due to a lack of effective and potentially curative treatment for recurrence, or due to poor study design. Since these studies, there has been a significant improvement in surgical and non-surgical treatments for recurrence, and it is therefore appropriate to revisit whether there may be an improvement in survival outcomes. 32 Traditionally, the aggressive nature of these cancers deemed even primary treatment not feasible in most patients. Consequently, in the subset of patients who undergo treatment but experience recurrence, further curative treatments were scarcely undertaken. 33 This led to a lack of surveillance post-operatively, as detection of recurrence often did not merit re-intervention. 20 However, in recent years, there has been a shift towards curative management of recurrent cancer, at least in cases with loco-regional recurrence. Studies on salvage esophagectomy and chemoradiotherapy have shown promise in providing acceptable survival benefits in selected patients. 34,35 The presence of salvage treatments justifies an intensive surveillance approach. 36 In our paper, majority of the included studies have shown a better overall and post-recurrence survival in patients who underwent intensive surveillance, despite a similar median time to recurrence to account for any lead-time bias. For example, Bennett et al. reported longer PRS and disease-specific survival at 13.5 and 29.4 months in patients who underwent intensive surveillance compared to 4.8 months and 21.6 months in the routine surveillance cohorts, respectively. 22 However, other studies noted no difference in any survival advantage between these surveillance strategies. 20 One limitation is that not all studies report a clear timeline in time to recurrence and length of follow-up, which may introduce lead-time bias as noted by some studies. 6,12 This further necessitates a randomized controlled trial to evaluate all possible metrics of survival in an unbiased manner between routine and intensive surveillance.
In our quantitative analysis, planned surveillance was beneficial for esophageal cancer but not gastric cancer. This dissonance may be explained by better multi-modal treatments available for esophageal cancer that can be used to manage recurrence once detected. For example, in all the included studies, the majority of the patients with recurrence of esophageal cancer underwent either chemotherapy or radiotherapy, whereas not all studies reported patients with recurrence with gastric cancer undergoing these therapies. Furthermore, the sample sizes for studies involving patients with esophageal cancers are larger and therefore their results have more weight towards estimating the overall effect. Moreover, it is possible that surveillance is more rigorous for esophageal cancer. The preliminary results reported by Elliot et al exemplifies the afore-mentioned reasons, since it is a large study involving 4682 patients; 60% of these patients received chemoradiotherapy; and all patients underwent rigorous planned surveillance. Further work is necessary to identify the reasons driving this difference in outcomes.
There are currently no standardized guidelines for monitoring post-operative patients, and a geographical variation is noted in the extent of surveillance. There is unanimity amongst Western nations in advocating follow-up with symptom-driven recurrence investigations only. In Europe, the National Institute for Health and Care Excellence (NICE) and other joint society guidelines advise against routine followup (for the detection of recurrence) of asymptomatic patients. 37,38 Similarly, the National Comprehensive Cancer Network (NCCN) does not support the use of routine investigations. 39 The Japan Gastric Cancer Association recognizes that follow-up at outpatient clinic can help patients readjust to normal life and cope with post-gastrectomy syndrome, but addresses the lack of high grade evidence for detection of recurrence and calls for further work 'to scientifically verify the prognostic relevance of postoperative follow-up programs'. 40 Hence, while there is consensus that patients should be followed up for a maximum of 5 years and any surveillance should be tailored to the patient's risk, there is a lack of agreement on how this should be carried out or what it entails, and should form the basis of future work.
Organizational factors have to be considered in implementing any form of surveillance programs. For example, in public healthcare systems such as the National Health Service (NHS), routine surveillance for asymptomatic patients may potentially place undue pressure on secondary and tertiary healthcare sectors. This necessitates the inclusion of general practitioners as well as specific staff and pathways dedicated to looking after this subgroup of patients. In one study, primary healthcare practitioners reported that they would be willing to assume exclusive responsibility for the follow-up care of adult cancer survivors if they were appropriately supported by the specialist center with patient-specific letters from the specialist, printed guidelines, clear routes of rereferral and rapid access to investigations for suspected recurrence. 41 This is increasingly required in low income countries where cancer is an increasing burden but has gone unnoticed and under-reported due to a lack of nationwide cancer surveillance networks. 42

Limitations
Our review is a comprehensive update of existing literature on the surveillance protocols that have been trialed for post-gastrectomy patients. However, there are several limitations that need to be addressed. Firstly, there is no standardized intensive surveillance protocol that has been compared to routine symptom-triggered surveillance, which reduces the comparability of studies to each other. Secondly, the included studies are largely observational in design and have small sample sizes that have often rendered them underpowered. Thirdly, surveillance is largely aimed at identifying recurrence, which depends on patient, surgical and pathological factors; however, most included studies have not tailored their surveillance strategy to factor these variables. Some of the studies have different outcome measures over a variable follow-up time period. Lastly, there is an inherent confounding by indication in studies that have analyzed only patients with recurrence and not comparing it to a more representative cohort of patients without recurrence.

Areas of future work
Future work should be in the form of well-powered randomized controlled trials that allocate patients to a standardized intensive surveillance program; a routine surveillance program based on symptoms; or no further surveillance. Patients should be stratified based on intraoperative factors, oncological factors and pathological factors, and need for alternative multimodal treatments to fully account for any other confounding variables. The survival outcome measures attained over an adequately long follow-up time would better inform whether intensive routine surveillance offers any viable advantage. Ultimately, the feasibility of surveillance depends on the infrastructure set up to facilitate it, and this varies with the structure of the healthcare system. [42][43][44] Economic analysis tailored to organizational factors have to be incorporated to conclude if routine intensive surveillance is feasible.

CONCLUSION
Gastric and esophageal cancers are malignancies associated with a poor prognosis, largely due to their high recurrence rates. Yet, there is no standardization of guidelines internationally to inform clinicians of how patients should be monitored after their operation. Although this has been a topic of debate for decades, our study shows that there is still no consensus on this. More importantly, our study highlights that planned surveillance has good survival benefit for patients with esophageal cancer, while there is a lack of high grade evidence for its benefit in gastric cancers. Further work in the form of randomized controlled trials is required to confirm these findings and subsequently establish how surveillance should be carried out and wand what it should comprise of in a practical and clinically safe manner.