Abstract
Background
Recent retrospective subgroup analyses of patients with unresectable colon cancer (CC) receiving systemic chemotherapy have demonstrated that there is a significant difference in treatment outcome between patients with right‐sided CC (RSCC) and those with left‐sided CC (LSCC). However, it is impossible to divide patients with CC randomly into RSCC and LSCC groups before surgery. Therefore, the aim of this study is to explore the impact of primary tumor location (PTL) on survival after curative surgery for patients with CC using propensity score‐matching (PSM) studies instead of randomization.
Materials and Methods
We performed a comprehensive electronic search of the literature up to January 2019 to identify studies that had used databases allowing comparison of postoperative survival between patients with RSCC and those with LSCC. To integrate the impact of PTL on 5‐year overall survival (OS) after curative surgery, a meta‐analysis was performed using random‐effects models to calculate the risk ratio (RR) and 95% confidence interval (CI) for the selected PSM studies.
Results
Five studies involving a total of 398,687 patients with CC were included in this meta‐analysis. Among 205,641 patients with RSCC, 69,091 (33.6%) died during the observation period, whereas among 193,046 patients with LSCC, 63,380 (32.8%) died during the same period. These results revealed that patients with RSCC and those with LSCC had almost the same 5‐year OS (RR, 0.98; 95% CI, 0.89–1.07; p = .64; I 2 = 97%).
Conclusion
This meta‐analysis has demonstrated that there was no significant difference in 5‐year OS between patients with RSCC and those with LSCC after curative resection.
Implications for Practice
To integrate the impact of primary tumor location (PTL) on 5‐year overall survival (OS) after curative surgery, five propensity score‐matching (PSM) studies involving a total of 398,687 patients with colon cancer (CC) were included in this meta‐analysis. Among 205,641 patients with right‐sided CC (RSCC), 69,091 (33.6%) died during the observation period, whereas among 193,046 patients with left‐sided CC (LSCC), 63,380 (32.8%) died during the same period. These results revealed that patients with RSCC and those with LSCC had almost the same 5‐year OS (risk ratio, 0.98; 95% confidence interval, 0.89–1.07; p = .64; I2 = 97%).
Keywords: Left‐sided colon cancer, Meta‐analysis, Primary tumor location, Propensity score, Matching study, Right‐sided colon cancer
Short abstract
There is a significant difference in treatment outcomes for patients with right‐sided versus left‐sided colon cancer. This study explored the effect of primary tumor location on survival after curative surgery using propensity score‐matching studies instead of randomized studies.
Introduction
Recent studies investigating the influence of primary tumor location (PTL) on the outcome of the latest systemic chemotherapies for patients with far advanced or recurrent unresectable colon cancer (CC) have suggested that the outcome for patients with right‐sided CC (RSCC) is worse than that of patients with left‐sided CC (LSCC) [1, 2, 3], in view of several differences between RSCC and LSCC in terms of genetic or epigenetic background. As a consequence, recent consensus guidelines for the management of patients with metastatic colorectal cancer (CRC) have been revised on the basis of evidence‐based or precision medicine [4].
In fact, several systematic reviews and meta‐analyses of retrospective or prospective cohort studies designed to investigate the prognostic differences between these two groups have concluded that RSCC has a significantly poorer prognosis than LSCC in terms of overall patient survival [5, 6, 7, 8].
However, for appropriate comparison of these two groups of patients, their background factors need to be homogeneous. Generally speaking, randomization is an ideal method for dividing patients into two groups. However, it is impossible to divide patients with CC randomly into RSCC and LSCC groups before surgery, because PTL, as well as gender, is decided before surgery.
Recently, in cases where randomization is impossible, propensity score matching (PSM) has been broadly recommended for ensuring that the background of two groups is homogeneous, instead of a randomized controlled trial (RCT) [9, 10]. In fact, our preliminary study to compare patients with RSCC with those with LSCC using PSM revealed that there was no significant intergroup difference in overall survival (OS) after curative surgery. This would suggest that if patients with CC underwent curative surgery, there would be no significant difference in OS between those with RSCC and those with LSCC. The only way to test this hypothesis is to perform a meta‐analysis of PSM studies, as RCT is unable to randomize the background of the two patient groups before surgery.
Materials and Methods
Preliminary Study to Investigate the Influence of PTL on OS After Curative Resection in Patients with CC Using PSM Analysis
We performed a retrospective review of a database of patients who had undergone elective surgery for CRC. All of the procedures had been performed by the same surgical team at the Department of Gastroenterological Surgery, Dokkyo Medical University Hospital, between February 2004 and September 2017. We excluded patients with unresectable CC or patients who had undergone combined resection of other organs. During this period, 1,293 patients with CRC who underwent surgery were enrolled. Among them, 379 patients with CC who underwent curative surgery were selected for PSM study. Using PSM study, patients with CC were divided into two groups (RSCC and LSCC) with almost the same clinical background factors.
The clinical features selected as suitable for PSM study were age, serum albumin level, body mass index, tumor markers such as carbohydrate antigen 19‐9 and carcinoembryonic antigen, American Society of Anesthesiologists (ASA) grade (0, I/ II, III, IV), tumor depth (T1, T2/T3, T4), gender (male/female), lymphatic invasion (absence/presence), lymph node dissection (D1, D2/D3), lymph node metastasis (absence/presence), maximum tumor size (cm), pathology (well/mod or others), tumor‐node‐metastasis stage (I/II, III), and venous invasion (absence/presence).
For the purposes of this preliminary study, RSCC was defined as from the cecum including the appendix to the transverse colon, whereas LSCC was defined as from the descending colon to the recto‐sigmoid colon.
Definition of Lymph Node Dissection
On the basis of the Japanese Classification of Colorectal Carcinoma (Japanese Society for Cancer of the Colon and Rectum, Second English Edition), lymph node dissection is classified as follows: DX, The extent of lymph node dissection cannot be assessed; D0, Incomplete dissection of the pericolic/perirectal lymph nodes; D1, Complete dissection of the pericolic/perirectal lymph nodes; D2, Complete dissection of the pericolic/perirectal and intermediate lymph nodes; D3, Complete dissection of all regional lymph nodes.
Definition of Operative Curability
On the basis of the Japanese Classification of Colorectal Carcinoma (Japanese Society for Cancer of the Colon and Rectum, Second English Edition), residual tumor is diagnosed as follows: R0, no residual tumor; R1, no residual tumor, but tumor at the resection margin suspected; R2, macroscopically evident residual tumor. On the basis of this definition, operative curability is defined as follows: curability A (Cur A), R0 at stage I, II or III; curability B (Cur B), R0 at stage IV or R1 in any stage; curability C (Cur C), R2 at any stage. Therefore, surgery is considered curative when patients with CC have undergone Cur A surgery.
Search Strategy
A systematic literature search was conducted using the Web of Science, MEDLINE, PubMed, and the Cochrane Library covering papers published up to January 2019. The search terms used were “colon cancer,” “right,” “left,” and “propensity.” Of those identified as potentially relevant, complete articles were retrieved and evaluated for inclusion. References from all of the relevant articles were hand‐searched for additional studies.
The meta‐analysis and search strategy complied with the guidelines of Preferred Reporting Items for Systematic Reviews and Meta‐Analyses 2010 [11]. Therefore, the PICO criteria for this study were as follows: Patients (P): patients with CC; Intervention (I): patients with RSCC; Comparison (C): patients with LSCC; Outcome (O): 5‐year OS.
Inclusion and Exclusion Criteria
The inclusion criteria were as follows: (a) Studies that had been analyzed using PSM study. (b) Studies that had provided data suitable for evaluation of 5‐year OS. (c) Studies that had provided data allowing calculation of the risk ratio (RR) or standardized incidence ratios with the 95% confidence interval (CI). (d) Studies that had provided the sample size and other appropriate data. (e) Articles had to be written in English.
The exclusion criteria were as follows: (a) Nonreporting of predefined outcomes for two groups, such as patients with RSCC and with LSCC, or inability to extract the number of outcome events from the published results. (b) Articles that were letters, comments, correspondences, editorials, or reviews. (c) Studies for which the published articles had considerable overlap between authors and centers.
Study Selection and Data Extraction
Full text reviews were performed independently by two of the authors (M.I. and N.S.) on the basis of the inclusion and exclusion criteria and PICO. Any disagreements were resolved by discussion and consensus. The same two authors also independently extracted the following information from each eligible article: the first author's name, the year of publication, the nation in which the study was performed, the study design, the number of patients with CC undergoing surgery, and sample size. If the necessary data could not be extracted from the publication, we contacted the original authors directly whenever possible.
Data Synthesis and Statistical Analysis
Data are presented as median and interquartile range. Intergroup differences were analyzed using the chi‐squared test or the Mann‐Whitney U test, as appropriate.
Kaplan‐Meier analysis and log‐rank test were used to compare the survival curves for the RSCC and LSCC patient groups. Deaths prior to September 30, 2017, were included in this preliminary study.
Statistical analyses were performed using the IBM SPSS statistics version 22.0 software package for Windows (IBM Co., New York, NY) at a significance level of p < .05.
Review Manager (ver. 5.3) for Windows (downloaded from http://ims.cochrane.org/revman/download) was used for this meta‐analysis. Because there were five PSM studies, a random‐effect model was used rather than a fixed‐effect model.
Dichotomous variables were analyzed by assessing the RR of 5‐year OS in patients with RSCC compared with those with LSCC as a control group, along with the 95% CI. An RR of <1 favored patients with RSCC.
Statistical heterogeneity was complemented with the I 2 statistic, which qualified the proportion of total variation across studies due to heterogeneity rather than to chance. The presence of publication bias was assessed by funnel plot. Forest plots were demonstrated in order by weight of each study. A p value of <.05 was considered to indicate statistical significance.
Ethical approval was not required because this was a meta‐analysis of previously published literature.
Results
Results of the Preliminary Study
We enrolled 379 patients (176 male, 203 female), including 193 patients with RSCC and 186 patients with LSCC. Table 1 shows the clinicolaboratory characteristics of the patients in the two groups. Chi‐squared test revealed significant intergroup differences in age (p < .001) and serum albumin level (p = .029), respectively (Table 1).
Table 1.
Relationships between clinicolaboratory characteristics and tumor location in patients with colon cancer who underwent curative resection (before propensity score matching)
| Variable | LSCC (n = 193) | RSCC (n = 186) | p value |
|---|---|---|---|
| Age, yr | 67 (60–75) | 73 (63–80) | <.001 |
| Albumin, g/dL | 4.0 (3.6–4.3) | 3.9 (3.5–4.2) | .029 |
| BMI, kg/m2 | 22.5 (20.6–24.9) | 22.4 (19.9–25.0) | .523 |
| CA19‐9, U/mL | 9.0 (4.0–19.0) | 7.0 (3.0–15.3) | .121 |
| CEA, ng/mL | 3.1 (2.0–5.6) | 2.8 (1.8–6.1) | .590 |
| Comorbidity (ASA grade 0, I/ II, III, IV) | 175/18 | 162/24 | .267 |
| Depth of tumor (T1, T2/T3, T4) | 65/128 | 68/118 | .557 |
| Gender, male/female | 86/107 | 90/96 | .455 |
| Lymphatic invasion, absence/presence | 67/126 | 81/105 | .078 |
| Lymph node dissection (D1, D2/D3) | 54/139 | 50/136 | .811 |
| Lymph node metastasis, absence/presence | 129/64 | 133/53 | .326 |
| Maximum tumor size, cm | 3.8 (2.5–4.9) | 4.0 (2.6–5.5) | .203 |
| Pathology (well/mod or others) | 57/136 | 52/134 | .735 |
| TNM stage (I/II, III) | 61/132 | 60/126 | .892 |
| Venous invasion, absence/presence | 64/129 | 77/109 | .097 |
Chi‐squared test, median (interquartile range), Mann‐Whitney U test.
Bold p values are statistically significant.
Abbreviations: ASA, American Society of Anesthesiologists; BMI, body mass index; CA19‐9, carbohydrate antigen 19‐9; CEA, carcinoembryonic antigen; LSCC, left‐sided colon cancer; mod, moderately differentiated adenocarcinoma; RSCC, right‐sided colon cancer; TNM, tumor‐node‐metastasis; well, well‐differentiated adenocarcinoma.
During the observation period, 27 patients died, 16 of them due to cancer‐related disease. The median and maximum follow‐up periods for the survivors were 884 days and 3,656 days, respectively, with a mean survival period of 1,138 ± 957 days.
Kaplan‐Meier analysis and log‐rank test revealed no significant difference in OS between the two groups (p = .421; Fig. 1A).
Figure 1.
Overall survival. Patients with RSCC and those with LSCC before (A) and after (B) propensity score‐matching analysis. Abbreviations: LSCC, left‐sided colon cancer; RSCC, right‐sided colon cancer.
After propensity score matching, we enrolled 292 patients (156 male, 136 female), including 146 with RSCC and 146 with LSCC. Table 2 shows the clinicolaboratory characteristics of the patients in the two groups. Chi‐squared test revealed no significant background differences between the two groups (Table 2).
Table 2.
Relationships between clinicolaboratory characteristics and tumor location in patients with colon cancer who underwent curative resection (after propensity score matching)
| Variable | LSCC (n = 146) | RSCC (n = 146) | p value |
|---|---|---|---|
| Age, yr | 69 (62–77) | 69 (62–77) | .868 |
| Albumin, g/dL | 3.9 (3.5–4.3) | 4.0 (3.6–4.3) | .428 |
| BMI, kg/m2 | 22.6 (20.1–24.9) | 22.4 (19.7–25.3) | .623 |
| CA19‐9, U/mL | 8.0 (4.0–18.0) | 6.0 (3.0–15.3) | .185 |
| CEA, ng/mL | 2.9 (1.9–5.6) | 2.7 (1.7–6.0) | .569 |
| Comorbidity (ASA grade 0, I/ II, III, IV) | 131/15 | 130/16 | .849 |
| Depth of tumor (T1, T2/T3, T4) | 50/96 | 58/88 | .332 |
| Gender, male/female | 79/67 | 77/69 | .814 |
| Lymphatic invasion, absence/presence | 51/95 | 65/81 | .094 |
| Lymph node dissection (D1, D2/D3) | 43/103 | 39/107 | .602 |
| Lymph node metastasis, absence/presence | 104/42 | 100/46 | .610 |
| Maximum tumor size, cm | 3.9 (2.7–5.0) | 3.8 (2.3–5.0) | .522 |
| Pathology (well/mod or others) | 42/104 | 44/102 | .797 |
| TNM stage (I/II, III) | 47/99 | 51/95 | .620 |
| Venous invasion, absence/presence | 52/94 | 56/90 | .628 |
Chi‐squared test, median (interquartile range), Mann‐Whitney U test.
Abbreviations: ASA, American Society of Anesthesiologists; BMI, body mass index; CA19‐9, carbohydrate antigen 19‐9; CEA, carcinoembryonic antigen; LSCC, left‐sided colon cancer; mod, moderately differentiated adenocarcinoma; RSCC, right‐sided colon cancer; TNM, tumor‐node‐metastasis; well, well‐differentiated adenocarcinoma.
During the observation period, 24 patients died, 14 due to cancer‐related disease. The median and maximum follow‐up periods for the survivors were 952 days and 3,656 days, respectively, with a mean survival period of 1,221 ± 981 days. Kaplan‐Meier analysis and log‐rank test demonstrated no significant difference in OS between the two groups (p = .613; Fig. 1B).
On the basis of this preliminary study, our working hypothesis was that meta‐analysis using PSM studies would demonstrate that if patients with CC underwent curative surgery, there would be no significant difference in 5‐year OS between those with RSCC and those with LSCC.
We did not add our preliminary data to the following meta‐analysis.
Study Identification and Eligibility
An electronic search yielded 52 articles, of which 10 were considered duplicated articles based on a title search. Among the remaining articles, 33 were excluded by title/abstract review on the basis of the selection criteria and PICO. The remaining nine articles were screened by full text review, after which five studies including a total of 398,687 patients with CC were regarded as suitable for inclusion in the data synthesis [12, 13, 14, 15, 16]. The selection process for exclusion is shown in Figure 2.
Figure 2.
Flow diagram detailing the search strategy and identification of studies used in data synthesis.
Characteristics of the Included Studies
All of the five studies were analyzed using PSM study. Among them, one study [12] had collected data from the Surveillance, Epidemiology, and End Results (SEER) database, one [14] had collected data from the National Cancer Database (NCDB), and two had collected data from multicenters [15, 16]. One study had collected data from a single center [13]. The basic characteristics of the five included studies are shown in Table 3.
Table 3.
Summary of the included studies
| Author | Warschkow R [12] | Lim DR [13] | Lee L [14] | Ishihara S [15] | Kishiki T [16] |
|---|---|---|---|---|---|
| Year | 2016 | 2017 | 2018 | 2018 | 2019 |
| Nation | Germany | Korea | U.S | Japan | Japan |
| Study design | Retrospective | Retrospective | Retrospective | Retrospective | Retrospective |
| Data source | SEER database | Single‐center | NCDB | Multicenter | Multicenter |
| Type of analysis | PSM | PSM | PSM | PSM | PSM |
| Definition of right‐sided colon | C & A | C & A & 2/3p‐T | C & A | C & A & T | C & A |
| Definition of left‐sided colon | D & S | D & S&1/3d‐T | D & S& Rs | D & S | D & S |
| Stage | I, II, III | 0, I, II, III | I, II, III | II, III | I, II, III |
| Curability | Cur A | Cur A | Cur A | Cur A | Cur A |
Abbreviations: A, ascending colon; C, cecum; D, descending colon; d‐T, distal site of transverse colon; m‐T, middle site of transverse colon; NCDB, National Cancer Database; PSM, propensity score matching; p‐T, proximal site of transverse colon; Rs, recto‐sigmoid colon; S, sigmoid colon; SEER, Surveillance, Epidemiology, and End Results; T, transverse colon.
All of the studies considered the cecum and ascending colon to be the right‐sided colon (RSC). However, three studies excluded the transverse colon from the RSC [12, 14, 16]. Among the remaining two studies, one included the transverse colon in the RSC [15]. However, one study included the proximal two thirds of the transverse colon [13] in the RSC (Table 3).
All of the studies included the descending colon and sigmoid colon in the left‐sided colon (LSC). However, one study included the recto‐sigmoid colon [14] and one study included the distal one third of the transverse colon [13] in the LSC (Table 3).
With regard to curability, all of the studies regarded Cur A resection as curative. Therefore, in all studies, curative resection was performed for patients with CC on the basis of the Japanese Classification of Colorectal Carcinoma (Japanese Society for Cancer of the Colon and Rectum, Second English Edition; Table 3).
Association Between Primary Tumor Location and Overall Survival
Data on 5‐year OS were available for all five studies analyzed by PSM study. Among the 205,641 patients with RSCC undergoing curative resection, 69,091 (33.6%) died, whereas among 193,046 patients with LSCC, 63,380 (32.8%) died (Table 4).
Table 4.
Summary of 5‐year overall survival of patients with stage I–III right‐sided and left‐sided colon cancer in the included studies
| Author | Stage I–III right‐sided colon cancer5‐year OS | Stage I–III left‐sided colon cancer5‐year OS | ||||
|---|---|---|---|---|---|---|
| Total | Alive | Dead | Total | Alive | Dead | |
| Warschkow R [12] 2016 |
51,331 (100%) |
33,622 (65.5%) |
17,709 (34.5%) |
38,736 (100%) |
24,404 (63.0%) |
14,332 (37.0%) |
| Lim DR [13] 2017 |
207 (100%) |
170 (82.1%) |
37 (17.9%) |
207 (100%) |
184 (88.9%) |
23 (11.1%) |
| Lee L [14] 2018 |
148,540 (100%) |
97,888 (65.9%) |
50,652 (34.1%) |
148,540 (100%) |
100,265 (67.5%) |
48,275 (32.5%) |
| Ishihara S [15] 2018 |
2,168 (100%) |
1,886 (87.0%) |
282 (13.0%) |
2,168 (100%) |
1,849 (85.3%) |
319 (14.7%) |
| Kishiki T [16] 2019 |
3,395 (100%) |
2,984 (87.9%) |
411 (12.1%) |
3,395 (100%) |
2,964 (87.3%) |
431 (12.7%) |
| Total |
205,641 (100%) |
136,550 (66.4%) |
69,091 (33.6%) |
193,046 (100%) |
129,666 (67.2%) |
63,380 (32.8%) |
Abbreviation: OS, overall survival.
There was no significant difference in 5‐year OS between patients with RSCC and those with LSCC (RR, 0.98; 95% CI, 0.89–1.07; p = .64; I 2 = 97%; Fig. 3A). The basic funnel plot of the studies included in this meta‐analysis indicated no evidence of publication bias, in view of its symmetry (Fig. 3A).
Figure 3.
Forest plots. Five‐year overall survival of patients with stage I–III RSCC and stage I–III LSCC (A), patients with stage I RSCC and stage I LSCC (B), patients with stage II RSCC and stage II LSCC (C), and patients with stage III RSCC and stage III LSCC (D) undergoing curative surgery, and funnel plot analyses of these patients using integrated data. Abbreviations: CI, confidence interval; LSCC, left‐sided colon cancer; RR, risk ratio; RSCC, right‐sided colon cancer.
Subclass Analyses for Association Between Primary Tumor Location and Overall Survival
Stage I Patients with CC
Data on 5‐year OS were available for four studies analyzed by PSM study. Among the 55,105 patients with RSCC undergoing curative resection, 11,190 (20.3%) died, whereas among 52,554 patients with LSCC, 10,714 (20.4%) died (Table 5).
Table 5.
Summary of 5‐year overall survival of patients with stage I right‐sided and left‐sided colon cancer in the included studies
| Author | Stage I right‐sided colon cancer5‐year OS | Stage I left‐sided colon cancer5‐year OS | ||||
|---|---|---|---|---|---|---|
| Total | Alive | Dead | Total | Alive | Dead | |
| Warschkow R [12] 2016 |
13,622 (100%) |
10,543 (77.4%) |
3,079 (22.6%) |
11,302 (100%) |
8,465 (74.9%) |
2,837 (25.1%) |
| Lim DR [13] 2017 |
39 (100%) |
37 (94.9%) |
2 (5.1%) |
53 (100%) |
52 (98.1%) |
1 (1.9%) |
| Lee L [14] 2018 |
40,625 (100%) |
32,541 (80.1%) |
8,084 (19.9%) |
40,362 (100%) |
32,532 (80.6%) |
7,830 (19.4%) |
| Ishihara S [15] 2018 | NA | NA | NA | NA | NA | NA |
| Kishiki T [16] 2019 |
819 (100%) |
794 (96.9%) |
25 (3.1%) |
837 (100%) |
791 (94.5%) |
46 (5.5%) |
| Total |
55,105 (100%) |
43,915 (79.7%) |
11,190 (20.3%) |
52,554 (100%) |
41,840 (79.6%) |
10,714 (20.4%) |
Abbreviations: NA, not available; OS, overall survival.
There was no significant difference in 5‐year OS between patients with RSCC and those with LSCC (RR, 0.93; 95% CI, 0.81–1.06; p = .29; I 2 = 90%; Fig. 3B). The basic funnel plot of the studies included in this meta‐analysis indicated no evidence of publication bias, in view of its symmetry (Fig. 3B).
Stage II Patients with CC
Data on 5‐year OS were available for five studies analyzed by PSM study. Among the 75,475 patients with RSCC undergoing curative resection, 22,666 (30.0%) died, whereas among 69,290 patients with LSCC, 20,276 (29.3%) died (Table 6).
Table 6.
Summary of 5‐year overall survival of patients with stage II right‐sided and left‐sided colon cancer in the included studies
| Author | Stage II right‐sided colon cancer5‐year OS | Stage II left‐sided colon cancer5‐year OS | ||||
|---|---|---|---|---|---|---|
| Total | Alive | Dead | Total | Alive | Dead | |
| Warschkow R [12] 2016 |
19,790 (100%) |
13,517 (68.3%) |
6,273 (31.7%) |
13,212 (100%) |
8,442 (63.9%) |
4,770 (36.1%) |
| Lim DR [13] 2017 |
93 (100%) |
81 (87.1%) |
12 (12.9%) |
86 (100%) |
76 (88.4%) |
10 (11.6%) |
| Lee L [14] 2018 |
53,137 (100%) |
36,983 (69.6%) |
16,154 (30.4%) |
53,537 (100%) |
38,332 (71.6%) |
15,205 (28.4%) |
| Ishihara S [15] 2018 |
1,140 (100%) |
1,033 (90.6%) |
107 (9.4%) |
1,133 (100%) |
997 (88.0%) |
136 (12.0%) |
| Kishiki T [16] 2019 |
1,315 (100%) |
1,195 (90.9%) |
120 (9.1%) |
1,322 (100%) |
1,167 (88.3%) |
155 (11.7%) |
| Total |
75,475 (100%) |
52,809 (70.0%) |
22,666 (30.0%) |
69,290 (100%) |
49,014 (70.7%) |
20,276 (29.3%) |
Abbreviation: OS, overall survival.
There was no significant difference in 5‐year OS between patients with RSCC and those with LSCC (RR, 0.90; 95% CI, 0.77–1.05; p = .18; I 2 = 97%; Fig. 3C). The basic funnel plot of the studies included in this meta‐analysis indicated no evidence of publication bias, in view of its symmetry (Fig. 3C).
Stage III Patients with CC
Data on 5‐year OS were available for four studies analyzed by PSM study. Among the 75,057 patients with RSCC undergoing curative resection, 31,074 (41.4%) died, whereas among 71,197 patients with LSCC, 27,527 (38.7%) died (Table 7).
Table 7.
Summary of 5‐year overall survival of patients with stage III right‐sided and left‐sided colon cancer in the included studies
| Author | Stage III right‐sided colon cancer5‐year OS | Stage III left‐sided colon cancer5‐year OS | ||||
|---|---|---|---|---|---|---|
| Total | Alive | Dead | Total | Alive | Dead | |
| Warschkow R [12] 2016 |
17,919 (100%) |
9,479 (52.9%) |
8,440 (47.1%) |
14,222 (100%) |
7,580 (53.3%) |
6,642 (46.7%) |
| Lim DR [13] 2017 |
71 (100%) |
47 (66.2%) |
24 (33.8%) |
63 (100%) |
50 (79.4%) |
13 (20.6%) |
| Lee L [14] 2018 |
54,778 (100%) |
32,648 (59.6%) |
22,130 (40.4%) |
54,641 (100%) |
34,205 (62.6%) |
20,436 (37.4%) |
| Ishihara S [15] 2018 |
1,028 (100%) |
814 (79.2%) |
214 (20.8%) |
1,035 (100%) |
829 (80.1%) |
206 (19.9%) |
| Kishiki T [16] 2019 |
1,261 (100%) |
995 (78.9%) |
266 (21.1%) |
1,236 (100%) |
1,006 (81.4%) |
230 (18.6%) |
| Total |
75,057 (100%) |
43,983 (58.6%) |
31,074 (41.4%) |
71,197 (100%) |
43,670 (61.3%) |
27,527 (38.7%) |
Abbreviation: OS, overall survival.
There was no significant difference in 5‐year OS between patients with RSCC and those with LSCC (RR, 1.06; 95% CI, 1.00–1.12; p = .06; I 2 = 85%; Fig. 3D). The basic funnel plot of the studies included in this meta‐analysis indicated no evidence of publication bias, in view of its symmetry (Fig. 3D).
The Quality of Surgery
In order to investigate the quality of surgery, the data of surgical results of lymph node dissection and results of histological findings were added to the study.
There were no significant differences in number of positive regional lymph nodes between patients with RSCC and those with LSCC, except for one study [13] (Table 8).
Table 8.
Summary of lymph node metastasis in the included studies
| Author | Number of positive regional lymph nodes | Number of retrieved regional lymph nodes | ||||
|---|---|---|---|---|---|---|
| Right sided | Left sided | p value | Right sided | Left sided | p value | |
| Warschkow R [12] 2016 | 0 65.1% | 0 63.3% | .733 | <12 19.9% | <12 20.9% | .931 |
| 1 10.9% | 1 12.2% | |||||
| 12–16 28.9% | 12–16 28.1% | |||||
| ≥17 51.2% | ≥17 51.0% | |||||
| 2–3 11.0% | 2–3 11.0% | |||||
| 4–6 7.1% | 4–6 7.2% | |||||
| ≥7 6.0% | ≥7 6.3% | |||||
| Lim DR [13] 2017 | 0 65.7% | 0 69.6% | .019 | NA | NA | NA |
| 1 9.7% | 1 10.6% | |||||
| 2–3 9.2% | 2–3 11.6% | |||||
| 4–6 4.8% | 4–6 5.8% | |||||
| ≥7 10.6% | ≥7 2.4% | |||||
| Lee L [14] 2018 | 0 67% | 0 67% | .748 | <12 21% | <12 22% | .832 |
| 12–21 49% | 12–21 48% | |||||
| ≥22 30% | ≥22 30% | |||||
| 1–3 24% | 1–3 24% | |||||
| ≥4 9% | ≥4 9% | |||||
| Ishihara S [15] 2018 | 0 52.6% | 0 52.3% | .92 | NA | NA | NA |
| 1–3 46.4% | 1–3 46.8% | |||||
| ≥4 1.0% | ≥4 0.9% | |||||
| Kishiki T [16] 2019 | 0 62.9% | 0 63.6% | .816 | NA | NA | NA |
| 1–3 27.1% | 1–3 26.5% | |||||
| ≥4 10.0% | ≥4 9.9% | |||||
Abbreviation: NA, not available.
There were no significant differences in histological grading between patients with RSCC and those with LSCC, except for one study [13] (Table 9).
Table 9.
Summary of histological results in the included studies
| Author | Histological grading | T stage | ||||
|---|---|---|---|---|---|---|
| Right sided | Left sided | p value | Right sided | Left sided | p value | |
| Warschkow R [12] 2016 | G1 9.1% | G1 9.3% | .954 | T1 12.4% | T1 12.9% | .908 |
| T2 18.2% | T2 20.3% | |||||
| T3 57.9% | T3 54.9% | |||||
| T4 11.5% | T4 11.9% | |||||
| G2 67.5% | G2 67.7% | |||||
| G3 18.7% | G3 17.9% | |||||
| G4 2.0% | G4 1.8% | |||||
| NA 2.7% | NA 3.3% | |||||
| Lim DR [13] 2017 | G1 19.3% | G1 21.3% | <.05 | T1 13.3% | T1 20.3% | .354 |
| T2 9.9% | T2 11.4% | |||||
| T3 69.0% | T3 63.4% | |||||
| T4 7.9% | T4 4.9% | |||||
| G2 57.0% | G2 73.4% | |||||
| G3 13.5% | G3 4.3% | |||||
| G4 10.2% | G4 1.0% | |||||
| Lee L [14] 2018 | G1, 2 80.9% | G1, 2 81.1% | .214 | T1 15.3% | T1 15.6% | .245 |
| T2 17.6% | T2 17.6% | |||||
| T3 56.4% | T3 56.0% | |||||
| T4 10.7% | T4 10.8% | |||||
| G3 15.3% | G3 15.3% | |||||
| NA 3.8% | NA 3.6% | |||||
| Ishihara S [15] 2018 | G1, 2 96.0% | G1, 2 96.1% | .87 | T1 1.6% | T1 1.7% | .93 |
| T2 3.6% | T2 4.0% | |||||
| G3, 4 4.0% | G3, 4 3.9% | T3 71.9% | T3 71.5% | |||
| T4 22.9% | T4 22.8% | |||||
| Kishiki T [16] 2019 | G1, 2 48.6% | G1, 2 49.5% | .7662 | T1 14.5% | T1 14.4% | .8163 |
| T2 14.8% | T2 15.6% | |||||
| T3 52.7% | T3 52.2% | |||||
| T4 18.0% | T4 17.8% | |||||
|
G3 46.0% |
G3 45.2% |
|||||
|
G4 5.4% |
G4 5.3% |
|||||
Bold p values are statistically significant.
Abbreviations: G, grade; NA, not available.
Discussion
Surprisingly, although all four previous systematic reviews [5, 6, 7, 8] that had included meta‐analyses of retrospective and prospective studies had demonstrated that the prognosis of patients with RSCC was worse than that of patients with LSCC, the results of the present meta‐analysis using PSM studies revealed no significant difference between patients with RSCC and those with LSCC after curative resection.
It is well known that the RSC and LSC have different embryologic origins [17]. In fact, a recent study revealed that RSC develops from the midgut and LSC from the hindgut [17]. In view of this, there are a number of biological, molecular, and pathological differences between patients with RSCC and those with LSCC [18]. RSCCs are more likely to be diploid and to have a mucinous histology [19], high microsatellite instability [20], CpG island methylation [21], and BRAF mutations [22]. On the other hand, LSCCs are frequently infiltrating, constricting lesions, with a phenotype characterized by chromosomal instability and aneuploidy [23, 24, 25].
This means that a simple comparison of patients with RSCC with those with LSCC reveals significant differences in clinical background between the two. In fact, a recent nationwide cohort study challenging the categorization of RSCC and LSCC revealed that patients with RSCC were older and had higher proportions of individuals with a high ASA score (>3) and a high Union for International Cancer Control stage (III + IV) than those with LSCC [8].
Therefore, in order to assess whether patients with CC undergoing curative surgery show no significant difference in OS between RSCC and LSCC, it is not appropriate to perform meta‐analysis of retrospective or prospective studies that have investigated the significance of PTL on postoperative survival because of background heterogeneity. In fact, it has been reported that patients with RSCC have a worse prognosis than those with LSCC, because the former have a worse clinical background than the latter.
In order to resolve this issue, the background of the two groups should be made more homogeneous by using PSM study, as it is impossible to categorize patients with CC as having RSCC or LSCC randomly before surgery, or by gender. Although the results of a meta‐analysis using several RCTs might be expected to yield the highest level of evidence, in order to obtain a valid conclusion, a meta‐analysis using PSM studies would be required when RCTs cannot be performed.
In fact, although only one study [13] among the five included studies had background heterogeneities between patients with RSCC and LSCC in the number of positive lymph nodes and histological grading, the population of the study was about 0.01% in all included studies. Therefore, it was supposed that the quality of the surgery was guaranteed, because the backgrounds of surgical and histological factors were regarded as almost the same between patients with RSCC and those with LSCC in this study.
Among the five studies included in the present meta‐analysis, the latest two Japanese studies [15, 16] demonstrated that the 5‐year OS rate for patients with CC (stage II, III or stage I, II, III) undergoing curative surgery was almost 90%. Also, according to the latest Japanese guidelines for CRC, the cumulative 5‐year survival of patients with CC ranged from 63.8% to 93.0% (stage I, 92.3%; stage II, 85.4%; stage IIIa, 80.4%; stage IIIb, 63.8%) [26, 27]. Three reasons for this cumulative 5‐year survival of patients with CC in Japan can be suggested. First, Japanese surgeons undergo surgery on the basis of not only R0 resection but also D3 lymph node dissection, especially for patients with stage III CC [28, 29]. Second, effective postoperative chemotherapies are performed not only for patients with stage III CC but also for high‐risk patients with stage II CC [30]. Third, surveillance after surgery for CRC [31] is recommended in the guidelines [26, 27].
On the other hand, the authors [14] concluded that patients with RSCC show a worse oncologic outcome than patients with LSCC, although a higher lymph node yield led to improved survival in the study using PSM study from the NCDB for the period 2004–2014 [14].
Although there is a possibility that the results of this meta‐analysis might have been affected by the findings of two large‐scale data studies [12, 14] (the former SEER database had a 25.9% weighting [12] and the latter NCDB had a 26.1% weighting [14]; Fig. 3), there was no significant difference in the 5‐year OS rates between patients with RSCC and those with LSCC, the difference in the rates being very tight (0.8%).
Therefore, the most important factor appears to be the quality of surgery. If sufficient lymph node dissection had been performed, there would have been no significant difference in OS between patients with RSCC and those with LSCC, even if RSCC has a higher malignant potential than LSCC. In fact, a previous study had concluded that patients with RSCC tend to have a poorer oncologic outcome than patients with LSCC owing to insufficient lymph node harvest [14].
Because several studies have used large‐scale data from the National Clinical Database in Japan [32, 33], a PSM study using the National Clinical Database should be performed not only to examine the hypothesis but also to increase the volume of Japanese data for meta‐analysis in the near future.
Thus, we have been able to verify the hypothesis that even if patients with RSCC have higher malignant potential than those with LSCC, there would be no significant difference in OS between patients with RSCC and those with LSCC if curative resection is performed. The addition of further large‐scale data from studies involving high‐quality surgery for patients with CC would undoubtedly strengthen our findings.
Conclusion
In conclusion, this meta‐analysis has demonstrated that there was no significant difference in 5‐year OS between patients with RSCC and those with LSCC after curative resection.
On the other hand, although all recent studies including subclass analyses that had investigated the effect of PTL on outcomes of the latest systemic chemotherapies revealed that PTL is an important prognostic factor in patients with metastatic CRC, no studies have compared RSCC with LSCC using PSM study to ensure background homogeneity. Thus, similar meta‐analysis using PSM studies will be needed in order to address the issue of whether PTL is really important when performing systemic chemotherapy for patients with far advanced and recurrent unresectable CRC.
Author Contributions
Conception/design: Mitsuru Ishizuka
Provision of study material or patients: Norisuke Shibuya
Collection and/or assembly of data: Kazutoshi Takagi, Hiroyuki Hachiya, Yusuke Nishi, Kotaro Suda, Taku Aoki
Data analysis and interpretation: Takayuki Shimizu
Manuscript writing: Mitsuru Ishizuka
Final approval of manuscript: Mitsuru Ishizuka, Takayuki Shimizu, Norisuke Shibuya, Kazutoshi Takagi, Hiroyuki Hachiya, Yusuke Nishi, Kotaro Suda, Taku Aoki, Keiichi Kubota
Disclosures
The authors indicated no financial relationships.
Acknowledgments
We thank Dr. Akihisa Matsuda (Department of Gastrointestinal and Hepato‐Biliary‐Pancreatic Surgery, Nippon Medical School, Tokyo, Japan) for useful advice and suggestions regarding meta‐analysis. We also thank Dr. Soichiro Ishihara (Department of Surgical Oncology, Faculty of Medicine, The University of Tokyo, Tokyo, Japan) for data on 5‐year overall survival of patients with stage II and III colon cancer.
Disclosures of potential conflicts of interest may be found at the end of this article.
No part of this article may be reproduced, stored, or transmitted in any form or for any means without the prior permission in writing from the copyright holder. For information on purchasing reprints contact Commercialreprints@wiley.com. For permission information contact permissions@wiley.com.
References
- 1. Loupakis F, Yang D, Yau L et al. Primary tumor location as a prognostic factor in metastatic colorectal cancer. J Natl Cancer Inst 2015;107:1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Holch JW, Ricard I, Stintzing S et al. The relevance of primary tumour location in patients with metastatic colorectal cancer: A meta‐analysis of first‐line clinical trials. Eur J Cancer 2017;70:87–98. [DOI] [PubMed] [Google Scholar]
- 3. Arnold D, Lueza B, Douillard JY et al. Prognostic and predictive value of primary tumour side in patients with RAS wild‐type metastatic colorectal cancer treated with chemotherapy and EGFR directed antibodies in six randomized trials. Ann Oncol 2017;28:1713–1729. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Van Cutsem E, Cervantes A, Adam R et al. ESMO consensus guidelines for the management of patients with metastatic colorectal cancer. Ann Oncol 2016;27:1386–1422. [DOI] [PubMed] [Google Scholar]
- 5. Hansen IO, Jess P. Possible better long‐term survival in left versus right‐sided colon cancer ‐ A systematic review. Dan Med J 2012;59:1–6. [PubMed] [Google Scholar]
- 6. Yahagi M, Okabayashi K, Hasegawa H et al. The worse prognosis of right‐sided compared with left‐sided colon cancers: A systematic review and meta‐analysis. J Gastrointest Surg 2016;20:648–655. [DOI] [PubMed] [Google Scholar]
- 7. Lee GH, Malietzis G, Askari A et al. Is right‐sided colon cancer different to left‐sided colorectal cancer? ‐ A systematic review. Eur J Surg Oncol 2015;41:300–308. [DOI] [PubMed] [Google Scholar]
- 8. Petrelli F, Tomasello G, Borgonovo K et al. Prognostic survival associated with left‐sided vs right‐sided colon cancer: A systematic review and meta‐analysis. JAMA Oncol 2017;3:211–219. [DOI] [PubMed] [Google Scholar]
- 9. Jupiter DC. Propensity score matching: Retrospective randomization? J Foot Ankle Surg 2017;56:417–420. [DOI] [PubMed] [Google Scholar]
- 10. Kuss O, Blettner M, Borgermann J. Propensity score: An alternative method of analyzing treatment effects. Dtsch Arztebl Int 2016;113:597–603. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Moher D, Liberati A, Tetzlaff J et al.; PRISMA Group. Preferred reporting items for systematic reviews and meta‐analyses: The PRISMA Statement. Open Med 2009;3:e123–e130. [PMC free article] [PubMed] [Google Scholar]
- 12. Warschkow R, Sulz MC, Marti L et al. Better survival in right‐sided versus left‐sided stage I ‐ III colon cancer patients. BMC Cancer 2016;16:554. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Lim DR, Kuk JK, Kim T et al. Comparison of oncological outcomes of right‐sided colon cancer versus left‐sided colon cancer after curative resection: Which side is better outcome? Medicine (Baltimore) 2017;96:e8241. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Lee L, Erkan A, Alhassan N et al. Lower survival after right‐sided versus left‐sided colon cancers: Is an extended lymphadenectomy the answer? Surg Oncol 2018;27:449–455. [DOI] [PubMed] [Google Scholar]
- 15. Ishihara S, Murono K, Sasaki K et al. Impact of primary tumor location on postoperative recurrence and subsequent prognosis in nonmetastatic colon cancers: A multicenter retrospective study using a propensity score analysis. Ann Surg 2018;26:917–921. [DOI] [PubMed] [Google Scholar]
- 16. Kishiki T, Kuchta K, Matsuoka H et al. The impact of tumor location on the biological and oncological differences of colon cancer: Multi‐institutional propensity score‐matched study. Am J Surg 2019;217:46–52. [DOI] [PubMed] [Google Scholar]
- 17. Bufill JA. Colorectal cancer: Evidence for distinct genetic categories based on proximal or distal tumor location. Ann Intern Med 1990;113:779–788. [DOI] [PubMed] [Google Scholar]
- 18. Benedix F, Meyer F, Kube R et al. Influence of anatomical subsite on the incidence of microsatellite instability, and KRAS and BRAF mutation rates in patients with colon carcinoma. Pathol Res Pract 2012;208:592–597. [DOI] [PubMed] [Google Scholar]
- 19. Li P, Xiao Z, Braciak TA et al. A relationship to survival is seen by combining the factors of mismatch repair status, tumor location and age of onset in colorectal cancer patients. PLoS One 2017;12:e0172799. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Kim ST, Kim HK, Lee J et al. The impact of microsatellite instability status and sidedness of the primary tumor on the effect of bevacizumab‐containing chemotherapy in patients with metastatic colorectal cancer. J Cancer 2018;9:1791–1796. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Lee MS, Menter DG, Kopetz S. Right versus left colon cancer biology: Integrating the consensus molecular subtypes. J Natl Compr Canc Netw 2017;15:411–419. [DOI] [PubMed] [Google Scholar]
- 22. Gao XH, Yu GY, Gong HF et al. Differences of protein expression profiles, KRAS and BRAF mutation, and prognosis in right‐sided colon, left‐sided colon and rectal cancer. Sci Rep 2017;7:7882. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Iacopetta B. Are there two sides to colorectal cancer? Int J Cancer 2002;101:403–408. [DOI] [PubMed] [Google Scholar]
- 24. Iacopetta B, Grieu F, Amanuel B. Microsatellite instability in colorectal cancer. Asia Pac J Clin Oncol 2010;6:260–269. [DOI] [PubMed] [Google Scholar]
- 25. Baran B, Mert Ozupek N, Tetik NY et al. Difference between left‐sided and right‐sided colorectal cancer: A focused review of literature. Gastroenterology Res 2018;11:264–273. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Watanabe T, Itabashi M, Shimada Y et al. Japanese Society for Cancer of the Colon and Rectum (JSCCR) Guidelines 2014 for treatment of colorectal cancer. Int J Clin Oncol 2015;20:207–239. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Watanabe T, Muro K, Ajioka Y et al. Japanese Society for Cancer of the Colon and Rectum (JSCCR) guidelines 2016 for the treatment of colorectal cancer. Int J Clin Oncol 2018;23:1–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Xie D, Yu C, Gao C et al. An optimal approach for laparoscopic D3 lymphadenectomy plus complete mesocolic excision (D3+CME) for right‐sided colon cancer. Ann Surg Oncol 2017;24:1312–1313. [DOI] [PubMed] [Google Scholar]
- 29. Kitano S, Inomata M, Mizusawa J et al. Survival outcomes following laparoscopic versus open D3 dissection for stage II or III colon cancer (JCOG0404): A phase 3, randomised controlled trial. Lancet Gastroenterol Hepatol 2017;2:261–268. [DOI] [PubMed] [Google Scholar]
- 30. Meyers BM, Cosby R, Quereshy F et al. Adjuvant chemotherapy for stage II and III colon cancer following complete resection: A Cancer Care Ontario systematic review. Clin Oncol 2017;29:459–465. [DOI] [PubMed] [Google Scholar]
- 31. Fuccio L, Rex D, Ponchon T et al. New and recurrent colorectal cancers after resection: A systematic review and meta‐analysis of endoscopic surveillance studies. Gastroenterology 2019;156:1309–1323.e3. [DOI] [PubMed] [Google Scholar]
- 32. Gotoh M, Miyata H, Hashimoto H et al. National Clinical Database feedback implementation for quality improvement of cancer treatment in Japan: From good to great through transparency. Surg Today 2016;46:38–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Takahara T, Wakabayashi G, Konno H et al. Comparison of laparoscopic major hepatectomy with propensity score matched open cases from the National Clinical Database in Japan. J Hepatobiliary Pancreat Sci 2016;23:721–734. [DOI] [PubMed] [Google Scholar]