Abstract
Background/Aim: The proportion of patients with liver metastases in patients with appendiceal versus colorectal adenocarcinomas was 3.1 percent and 24 percent, respectively, in our peritonectomy centre. From our internal analyses, carcinoembryonic antigen (CEA) was potentially involved. A hypothesis was proposed regarding the natural progression of appendiceal adenocarcinoma. To support this, a systematic review and meta-analysis were performed to examine whether there was a difference in the proportion of patients with an elevated CEA in appendiceal versus colorectal adenocarcinoma patients in the current literature. Materials and Methods: Medline (PubMed), EMBASE (Ovid), Cochrane Central Register of Controlled Trials (CENTRAL), Cumulative Index to Nursing and Allied Health Literature, Clinicaltrials.gov, Web of Science, and Google Scholar were searched. All studies involving patients with appendiceal and/or colorectal adenocarcinoma were eligible. Data were analysed by grouping appendiceal and colorectal adenocarcinoma in separate meta-analyses, and then comparing their weighted proportions of elevated CEA. Principal summary measures were weighted proportions of patients with elevated CEA. Results: From the initial identification of 1,928 articles, 136 articles were included in the final synthesis. Ninety-two articles were included in the meta-analysis. Proportions of appendiceal and colorectal adenocarcinoma with elevated CEA were 56% (95%CI=47-65%) and 42% (95%CI=38-46%), respectively (p=0.0001). Conclusion: Patients with appendiceal adenocarcinoma had a higher proportion of CEA than those with colorectal adenocarcinoma. Future studies should focus on the several aspects of CEA presented in patients with appendiceal adenocarcinoma. This could provide treatments for patients with colorectal adenocarcinoma by preventing the development of liver metastases.
Differences between appendiceal and colorectal adenocarcinoma have been published extensively, with the most recent regarding gene expression between ‘high-risk’ appendiceal cancer and colorectal cancer (1). Mucin 2 and 5AC, and trefoil factors 1 and 2 are just a few of the many and different genes expressed in the ‘high-risk’ appendiceal cancer patients; whereas homeobox A9 and trinucleotide repeat containing 9 genes are more predominantly expressed in patients with colorectal cancer. The roles of these genes and how they make appendiceal cancer different from colorectal cancer are thought-provoking.
There was a significant difference in the proportion of liver metastases between appendiceal and colorectal adenocarcinomas in our peritonectomy unit, 3.1 percent (n=9/289) and 24 percent (n=95/395, p=0.0001). The reasons behind this have been explored internally. Our results pointed to carcinoembryonic antigen (CEA), a commonly used tumour marker, being potentially involved. CEA is a glycoprotein that belongs to the supergene family of immunoglobulins. The serum level of CEA is used clinically for diagnosis and recurrence surveillance, particularly in patients with colorectal cancer.
A recent paper by Lee and Lee (2) described the role of CEA in the development of liver metastases in patients with colorectal cancer. CEA released by colon cancer cells travels through the portal vein and interacts with a membrane-anchored homolog of heterogeneous nuclear protein M4 (hnRNP M4) on hepatic Kupffer cells (either cytoplasmatic or membranous). This interaction creates a pro-metastatic cascade, which can lead to the development of liver metastases in patients with colorectal cancer (where CEA is positively expressed).
Objectives. A hypothesis has been proposed regarding the natural progression of appendiceal adenocarcinoma (3). To support this, we performed a systematic review and meta-analysis to examine whether there is evidence that the proportion of patients with appendiceal cancer that have an elevated CEA differs from that in patients with colorectal adenocarcinoma.
Materials and Methods
In accordance with the PRISMA guidelines, the systematic review was conducted and reported (4). The protocol has been registered with PROSPERO (CRD42021283615) and is available online (5).
Eligibility criteria. All studies that included hospitalised patients with appendiceal and colorectal adenocarcinoma, aged 18 years or older, were eligible for inclusion. All published and unpublished studies were considered. Reviews, editorials, case reports, paediatric case series, and non-peer-reviewed articles were excluded.
Information sources. Using the following electronic databases: Medline (PubMed), EMBASE (Ovid), Cochrane Central Register of Controlled Trials (CENTRAL), CINAHL, Clinicaltrials.gov, Web of Science, and Google Scholar, a systematic literature search was conducted. Dates of publication were not restricted, with dates of coverage including from January 1950, up until a final search performed on the 26th of September, 2021. Additional relevant articles were manually scanned through the reference lists of all included studies and relevant review papers. By letter or e-mail, the corresponding authors were also requested to provide unpublished data from relevant trials.
Search strategy. The search strategy was constructed in consultation with a senior staff librarian (Table I). Authors (A Cristaudo and S Jennings) independently searched the above databases using keywords related to appendiceal and colorectal adenocarcinoma and carcinoembryonic antigen. A manual search for electronic links to relevant articles and references to selected articles was also performed. No restrictions were placed regarding language, however, only those in English or translated from Turkish, Chinese, Bulgarian, Polish, and Japanese were included in this systematic review.
Search terms used in systematic review.
Selection process. All studies that included patients with either appendiceal or colorectal adenocarcinoma in patients 18 years and above that mentioned CEA were eligible for inclusion from the systematic review literature search. Screening, eligibility, inclusion in the systematic review, and subsequent meta-analysis of studies were performed as per the PRISMA statement by two authors independently (A Cristaudo and S Jennings) (4). Titles, followed by abstracts, and then full-text articles were retrieved and read by both authors to identify those to be included in the systematic review. Data extraction disagreements between two authors were primarily resolved by discussion and consensus. A consensus meeting with a third author (D Morris) resolved disagreements if this failed.
Data collection process. Data were collected independently as per the PRISMA statement by two authors (A Cristaudo and S Jennings) using an electronic database (4). To confirm unclear data and to obtain additional data not available in the original article, investigators of included studies were contacted.
Data items. The collected data included study characteristics (first author’s surname, publication year, and study design), patient characteristics (number of patients, mean or median age), overall cancer type, the type/stage/grade of the primary tumour, CEA cut-off value (including units of measure), and the number and proportion of patients with elevated CEA. Missing data were handled as follows: Firstly, if the proportion of patients with an elevated CEA was not mentioned in the article, it was calculated by dividing the number of patients with an elevated CEA by the total number of patients who had their CEA measured. Secondly, if the CEA cut-off value was not specified, the study was excluded from the subsequent meta-analysis, even if the proportion of patients with an elevated CEA was still available. Lastly, if the outcome of interest was not available, it was not described, and hence the study (and patients) was excluded from the subsequent meta-analysis. No imputation methods were used.
Study risk of bias assessment. A methodological quality assessment was performed for included studies using the Newcastle-Ottawa quality assessment scale (6).
Effect measures. An assessment of the proportion of patients with elevated CEA was the primary key summary measure used in the synthesis of data. Appendiceal and colorectal adenocarcinoma proportions were weighted separately using a meta-analysis.
Synthesis methods. Data collected were qualitatively synthesised noting the number of studies for appendiceal and colorectal adenocarcinoma, as well as the proportion of elevated CEA reported in each study. Data were then quantitatively analysed for appendiceal and colorectal adenocarcinoma using MedCalc® Statistical Software version 20.022. to calculate the overall weighted proportion based on a meta-analysis of proportions (7). An evaluation of heterogeneity was conducted using I2 statistics.
Reporting bias assessment. Egger and Begg’s tests were used to evaluate the risk of publication bias. The Egger’s test estimates the relationship between standard error and the standardised effect using linear regression. Begg’s test measures if the rank of effect estimates is significantly correlated with the rank of their variances.
Results
Study selection. The initial database and registry search identified 1,928 studies (Figure 1). This included 1,066 studies from PubMed.gov, 484 studies from Cumulative Index to Nursing and Allied Health Literature, 294 studies from OVID via Medline, 72 studies from the clinicaltrials.gov website, and 12 studies from Cochrane Central Register of Controlled Trials. An additional 296 studies were identified from the reference lists of included studies and the Web of Science and Google Scholar websites. Of these, 614 were identified as duplicate studies and subsequently excluded. Systematic exclusions were then made, leaving a total of 136 studies in the final review. The final stage of the systematic review (full text) excluded 312 studies. This was due to studies with no CEA proportions available (169 studies), with duplicate datasets (66 studies), with metachronous liver and/or lung metastases (39 studies), with no pre-operative CEA (32 studies), or where the groups were purposefully matched for CEA (three studies). Forty-four studies were further excluded due to the studies involving low-grade appendiceal mucinous neoplasms (LAMN) alone and/or having cut-off values for CEA that were not 5 ng/ml or 5 μg/l, leaving a total of 92 studies in the subsequent meta-analysis.
PRISMA 2020 flow diagram for new systematic reviews, which included searches of databases, registers, and other sources (148).
Study characteristics and results of individual studies. The 136 included studies involved a total of 67,113 patients, with a mean age of 54.7 [standard deviation (SD): 1.71] years for appendiceal and a mean age of 62.3 (SD=5.59) years for colorectal adenocarcinoma [mean difference (MD)=7.6 years, t(96)=−4.4, p=0.00002] (Figure 2) (8-143). Of the included studies, 122 studies involved 64,088 patients with colorectal adenocarcinoma (published from 1972 to 2021), and 14 studies involved 3,025 patients with appendiceal adenocarcinoma (published from 2002 to 2021). Of the 122 studies involving patients with colorectal adenocarcinoma, 85 studies included patients with metastatic disease (eight studies specifically with liver metastases) and 17 studies included rectal adenocarcinoma only. Of the 14 studies involving patients with appendiceal adenocarcinoma, seven studies included those with either LAMN or high-grade appendiceal neoplasms, five studies included those with adenocarcinoma, and two studies included those with LAMN alone.
Number of publications for appendiceal versus colorectal adenocarcinoma from 1972 to 2021.
CEA cut-off values used within the included studies ranged from 2 ng/ml to 200 ng/ml. The most common value was 5 ng/ml (μg/l), noted in 71% of the included studies (n=97).
Ninety-four studies had a retrospective design, and 42 studies had a prospective design. Ninety-one studies were cohort studies, while 13 studies were case-control studies. None of the included studies were randomised-controlled trials.
Further study characteristics and outcomes of individual studies are shown in Table II, Table III, Table IV, Table V, Table VI, Table VII, Table VIII, Table IX, Table X and Table XI. Forest plots for included studies for appendiceal and colorectal adenocarcinoma from meta-analyses are shown in Figure 3 and Figure 4.
Summarised characteristics and outcomes of the 136 included studies.
Summarised characteristics and outcomes of the 136 included studies.
Summarised characteristics and outcomes of the 136 included studies.
Summarised characteristics and outcomes of the 136 included studies.
Summarised characteristics and outcomes of the 136 included studies.
Summarised characteristics and outcomes of the 136 included studies.
Summarised characteristics and outcomes of the 136 included studies.
Summarised characteristics and outcomes of the 136 included studies.
Summarised characteristics and outcomes of the 136 included studies.
Summarised characteristics and outcomes of the 136 included studies.
Forest plot for included studies for appendiceal adenocarcinoma from the meta-analysis.
Forest plot for included studies for colorectal adenocarcinoma from the meta-analysis.
Risk of bias in studies. The median score for methodological quality was 5 (IQR=5 to 6), of a possible 9 using the Newcastle-Ottawa quality assessment scale for those included studies that were appendiceal in origin. One study scored 4 points, nine studies scored 5 points, and four studies scored 6 points. For those included studies involving colorectal adenocarcinoma, the median score for methodological quality was 5 (IQR=4 to 6). Four studies scored 3 points, 35 studies scored 4 points, 46 studies scored 5 points, 26 studies scored 6 points, 9 studies scored 7 points, and two studies scored 8 points (out of a possible 9).
Case-control studies specifically report on control selection, comparability of cases and controls based on design or analysis, exposure assessment, and the same assessment methods for cases and controls. In cohort studies, exposure assessments, demonstration of the absence of endpoints of interest at baseline, and endpoint assessments were reported most frequently. The selection of the unexposed cohort or control group and the adequacy of follow-up were particularly poor in both study designs.
Please refer to Figure 5 and Figure 6 for funnel plots for included studies for appendiceal and colorectal adenocarcinoma, respectively, from the meta-analysis.
Funnel plot of included studies for appendiceal adenocarcinoma from the meta-analysis.
Funnel plot for included studies for colorectal adenocarcinoma from the meta-analysis.
Results of syntheses. Following further exclusions, weighted percentages of elevated CEA (> 5 ng/ml or 5 μg/l) were 56 (95%CI=47-65) for appendiceal and 42 (95%CI=39-46) for colorectal adenocarcinoma (MD: 14; 95%CI=12-16; p<0.0001) (Table XII).
Comparison between appendiceal and colorectal adenocarcinoma using the test of proportions for elevated carcinoembryonic antigen (CEA) (excluding studies involving low-grade appendiceal mucinous neoplasm and/or where the cut-off for CEA was not 5 ng/ml or 5 μg/l).
Reporting biases. The Egger bias test for the included studies was significant for those involving colorectal adenocarcinoma (p=0.0054); however, not significant for those that were appendiceal in origin (p=0.90). The Begg’s test, however, yielded results for both that were not significant (p=0.22 and p=0.93, respectively).
Discussion
Summary of evidence. This systematic review and meta-analysis provided an extensive overview of patients with appendiceal and colorectal adenocarcinoma, comparing the proportion of them with an elevated CEA.
Weighted percentages of elevated CEA were significantly different between patients with appendiceal and those with colorectal adenocarcinoma (MD: 14; 95%CI=12-16; p<0.0001). Reasons behind this finding need further exploration as there are currently several theories surrounding the workings of CEA in patients with colorectal adenocarcinoma that can potentially cover those with appendiceal adenocarcinoma. These are discussed in the sections below.
Strengths and limitations. This is the first systematic review and meta-analysis comparing proportions of elevated CEA between patients with appendiceal and colorectal adenocarcinoma. This allows for a unique insight into the possible role of CEA not only in colorectal but also in appendiceal adenocarcinoma cases regarding liver metastases.
Limitations exist within this systematic review and meta-analysis chiefly regarding the methodological quality, study design, and a paucity of included studies for appendiceal adenocarcinoma. The overall methodological quality of the included studies was moderate for both appendiceal and colorectal adenocarcinoma. This means that the level of evidence behind conclusions drawn from this systematic review and meta-analysis potentially lacks external validity. In terms of study design, as there were no randomised-control trials available for inclusion, this allows for bias to be introduced and subsequently confound results. Lastly, with only 10 percent of the included studies involving appendiceal adenocarcinoma, it must be noted that the rarity of appendiceal adenocarcinoma cases and hence, the paucity of literature regarding this, makes comparison difficult.
How these results fit in with what is known. With this being the first systematic review and meta-analysis, there is little or no information available for comparison regarding the results of this study. However, it paves the way for future research regarding not only CEA but also appendiceal adenocarcinoma and how it varies from colorectal adenocarcinoma. The goal is to identify why there is such a disparity in the proportion of liver metastases between these two adenocarcinomas from different sites.
What this means for future research and practice. Plausible as the aforementioned hypothesis (3) may be, there needs to be further investigation by repeating the methodology presented by Tabuchi et al. (119) but in patients with appendiceal adenocarcinoma. Portal and peripheral vein samples of appendiceal adenocarcinoma patients are to be obtained and CEA levels compared using the exact methods described by Tabuchi et al. (119). If the hypothesis is correct, these values should be similar (or the portal venous CEA level should be lower than the peripheral vein CEA level). We have recently acquired full ethics approval for this to proceed at our institution and now actively recruiting suitable patients.
Another possibility that needs to be considered to explain the significant difference in the proportion of elevated CEA between appendiceal and colorectal adenocarcinoma is the possibility of mutations in the binding region (Pro-Glu-Leu-Pro-Lys; PELPK) of the CEA glycoprotein in patients with appendiceal adenocarcinoma. This has been explored in one study by Zimmer and Thomas (144) in 2001 in patients with colorectal cancer. They explored patients with elevated levels of CEA and found mutations in their CEA, therefore there was a lower binding affinity of these patients’ CEA to the receptors in the Kupffer cells. This meant that they did not seem to develop liver metastases, despite their highly elevated (peripheral) serum CEA levels. If these mutations were to be found in the CEA of patients with appendiceal adenocarcinoma, but in higher proportions, this may also account for the significantly lower proportion of liver metastases (145, 146).
Lastly, the release of CEA from colon cancer cells is dependent on phosphatidyl-specific phospholipase C as described by Sack et al. (147) in 1988. The resulting change in CEA from a membrane-bound, hydrophobic molecule to a soluble, hydrophilic molecule allows its inherent release from the cancer cells to travel throughout the circulation, as described above by Lee and Lee (2). If this was studied in appendiceal adenocarcinoma cells and phosphatidyl-specific phospholipase C was found to be lacking in vivo, it may explain why although CEA is expressed in the majority of appendiceal adenocarcinomas (96), it may not correlate with serum levels of CEA and lack the flow-on effects that CEA would have as described by Lee and Lee (2) regarding liver metastases.
Conclusion
This systematic review and meta-analysis aimed to review studies involving patients with appendiceal and colorectal adenocarcinoma and compare the proportion of elevated CEA levels. In doing this, we have shown that there is a significantly higher proportion in patients with appendiceal adenocarcinoma. Reasons behind this finding have been presented and postulated.
Future research should focus on several areas. Firstly, a study comparing peripheral and portal venous blood samples of patients with appendiceal adenocarcinoma should be conducted. Regarding the possibility of PELPK region mutations in the CEA glycoprotein in patients with appendiceal adenocarcinoma, genomic sequencing of serum samples of patients with appendiceal and colorectal adenocarcinoma with elevated CEAs should be performed. Finally, studies should also focus on the presence or absence of phosphatidyl-specific phospholipase C in appendiceal adenocarcinoma cells, either in vitro or in vivo.
This will also allow for the potential for drug development regarding colorectal adenocarcinoma patients by blocking the effects of CEA and/or manipulation of the CEA receptor (by down-regulation) to reduce the development and recurrence of liver metastases.
Footnotes
Authors’ Contributions
Study concept and design were carried out by Dr. Adam Cristaudo and his supervisor Prof. David Morris. Dr. Adam Cristaudo was involved in all aspects of the project. Dr. Scott Jennings was involved in independently reviewing articles, as well as performing data extraction and methodological assessment of the included studies. All Authors gave their approval for the article before submission.
Conflicts of Interest
The Authors have no conflicts of interest to declare in relation to this study.
- Received June 13, 2022.
- Revision received June 29, 2022.
- Accepted July 4, 2022.
- Copyright © 2022 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
