Abstract
Background/Aim: To evaluate whether the serum levels of CEA or CA19-9 concentration is a useful predictor of survival in patients with metastatic colon cancer (mCC). Patients and Methods: Between 2012 and 2015, 113 patients with mCC who underwent chemotherapy according to the Japanese Colorectal Cancer Treatment Guidelines at four Jikei University Hospitals were enrolled in this study. The two serum tumor makers, CEA and CA19-9 were measured before first-line chemotherapy and at four months thereafter. Results: Serum CA19-9 concentration at four months after first-line chemotherapy (p=0.003, HR=3.761) and first-line chemotherapy including oxaliplatin (p=0.038, HR=0.312) were independent predictors of survival in patients with mCC. By excluding the transverse colon, only serum CA19-9 concentration at four months after first-line chemotherapy (p=0.005, HR=3.660) was identified as the predictor of survival. Conclusion: Serum CA19-9 concentration after first-line chemotherapy seems to be a useful predictor of survival in patients with mCC.
Until the 1980s, colorectal cancer had been regarded as a cancer marginally affected by chemotherapy. In the early 1990s, two new anticancer drugs, irinotecan (CPT-11) and oxaliplatin (1-15) attracted attention, and new regimens with strong anticancer effects, such as infusion of fluorouracil and leucovorin combined with either irinotecan (FOLFIRI) (1-3, 5, 8, 14, 15) or oxaliplatin (FOLFOX) (1, 4-8, 12, 13) were introduced for patients with metastatic colorectal cancer (mCRC). During the last decade, the clinical outcomes of patients with mCRC have improved due to combination of the aforementioned regimens with antiangiogenetic drug or anti-epidermal growth factor receptor agents. The median overall survival (OS) for patients with mCRC is currently estimated to exceed 30 months (6-8, 12-15).
RAS, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA), and BRAF have gained attention as predictors of the effect of chemotherapy and have been used to determine the indications for combined treatment with anti-epidermal growth factor receptor agents (9-11). Generally, predictors used in colorectal cancer include serum levels of the tumor markers, carcinoembryonic antigen (CEA) and carbohydrate antigen 19-9 (CA19-9); however, whether serum CEA and CA19-9 concentrations are useful predictors of survival in patients with metastatic colon cancer (mCC) remains unclear. The aim of this retrospective study was to evaluate whether serum CEA and CA19-9 concentration are useful predictors of survival of patients with mCC.
Patients and Methods
The Ethics Committee for Biomedical Research of the Jikei Institutional Review Board approved the protocol of this retrospective study (30-431(9452)). Between 2012 and 2015, 113 patients with mCC whose primary tumor lesion was resected, and were 80 years of age or younger, underwent chemotherapy according to the Japanese colorectal cancer treatment guidelines at four Jikei university hospitals, who were enrolled into this study. Serum CEA and CA19-9 levels were measured twice: once before first-line chemotherapy and then at four months after starting chemotherapy, the latter being performed to evaluate the resectability of metastatic lesions. The cutoff values for serum CEA and CA19-9 concentrations were 50 ng/ml and 37 U/ml, respectively.
This study consisted of 64 male and 49 female patients with a mean age of 66.1 years (range: 38-80 years). Eighty-seven patients (77%) had metastatic recurrence in only one organ, and the other 26 patients (23%) had metastatic recurrence in two or more organs. The liver was the most common organ with metastatic recurrence, occurring in 77 patients (68%), followed by lung in 24 patients (24%) and peritoneal dissemination in 35 patients (31%). One hundred and two patients (90%) were administered oxaliplatin-based regimen (FOLFOX, S-1 plus oxaliplatin (SOX) (6,12), capecitabine plus oxaliplatin (XELOX) (13)) as the first-line chemotherapy, while 11 patients (10%) were given CPT-11-based regimens (FOLFIRI, irinotecan plus S-1 (IRIS) (14), irinotecan plus capecitabine (XELIRI) (15)) as the first-line chemotherapy. Only two patients (2%) underwent conversion surgery (Table I).
Statistical analysis. Continuous variables were expressed as a mean and range. χ2 test was used for comparison of categorical data, and t-tests were used for comparison of continuous variables. The Kaplan–Meier method was used to calculate survival rates, and the Wilcoxon test was used for comparison of survival rates. Cox proportional hazard regression model was used to analyze the predictive factors for the effect of chemotherapy. A p-value of less than 0.05 was considered to indicate significance. All data were analyzed with the IBM SPSS Statistics software, version 22.0 (IBM Japan, Ltd, Tokyo, Japan).
Results
Correlation between the effect of chemotherapy and each factor (Table II). To determine the variables affecting survival after starting chemotherapy, 10 parameters (age, gender, pathological type of the primary tumor, number of metastatic sites, CEA and CA19-9 levels before and at four months after starting chemotherapy, and first-line regimen) were analyzed using Cox proportional hazard regression. Only two factors, serum CA19-9 levels (p=0.003; HR=3.761) at four months after starting chemotherapy and the use of an oxaliplatin-based regimen as first line chemotherapy (p=0.04; HR=0.312) were independent contributing factors to predict survival after chemotherapy.
Comparison of primary cancer sites in the right- and left-side colons (Table III). In subgroup analysis for primary tumor located in right or left side colon, 15 patients with transverse colon cancer were excluded and the other 98 patients, 47 right-sided and 51 left-sided were selected. The number of patients with the serum CA19-9 levels above 37.0 U/ml at four months after starting chemotherapy was significantly higher in patients with right-side colon cancer as compared to those with left-side colon cancer (p=0.05). A larger percentage of patients with left-sided colon cancer were treated with bevacizumab as compared to patients with right-sided colon cancer (84% vs. 53%, p=0.002).
Correlations between the effect of chemotherapy and various factors depending on location of the primary cancer site (Table IV). In order to determine the variables affecting survival after starting chemotherapy, 12 parameters (age, gender, pathological type of the primary tumor, number of metastatic sites, CEA and CA19-9 levels before and four months after starting chemotherapy, and first line regimen including Bmab or Cmab) were analyzed using the Cox proportional hazard regression in the subgroup. Only one factor, serum CA19-9 levels (p=0.005; HR=3.660) at four months after starting chemotherapy were the independent contributing factor to predict survival after starting chemotherapy. However, no significant difference was observed regarding the site of the primary cancer (left or right) (p=0.65).
Characteristics of the patients with colon cancer
Comparison of cumulative survival rates depending on the location of the primary cancer (Figure 1). We compared the prognosis based on the site of the primary cancer. We found no significant difference in overall survival between left-sided and right-sided colon cancer (p=0.50) and demonstrated no significant superiority of prognosis in patients whose primary cancer site were left-side colon, over those whose primary cancer site were right-side colon.
Differences in tumor marker changes before and after chemotherapy (Table V). Evaluation of the changes in serum CEA levels before and at four months after starting chemotherapy revealed that 50 patients (44%) had serum CEA levels of 50 ng/ml or more before starting chemotherapy; this number decreased significantly to 31 patients (27%) at four months after starting chemotherapy (p=0.008). In contrast, 59 patients (52%) and 51 patients (45%) had serum CA19-9 levels of 37 U/ml or more before and at four months after starting chemotherapy, with no statistically significant difference (p=0.29).
Correlation between the effect of chemotherapy and each factor
Discussion
With the recent advent of CPT-11, oxaliplatin and molecularly targeted agents, chemotherapy for unresectable recurrent colon cancer has made remarkable progress (6-8, 12-15). Genetic variants of RAS, PIK3CA, BRAF, and UGT1A1 are regarded as prognostic factors for therapeutic agents (9-11, 16). Circulating tumor cells (CTCs) are also considered as a prognostic factor (17, 18) but determining CTC levels is not currently commonly used due to its high cost and complicated protocol. The most general and inexpensive chemotherapy predictors are serum levels of the tumor markers CEA and CA19-9.
Morita et al. (19) and Louhimo et al. (20) have reported that preoperative serum CEA but not CA19-9 level is a prognostic factor in patients with resectable colon cancer. Conversely, Park et al. (21) have measured preoperative and postoperative levels of serum CEA and CA19-9 in patients with resectable colon cancer and reported that high preoperative levels of serum CA19-9 were an independent risk factor for recurrence, while the postoperative elevation of serum CA19-9 was a factor associated with an increased risk of peritoneal recurrence.
Adam et al. (22) have reported that elevated serum CA19-9 levels after chemotherapy in unresectable recurrent colon cancer correlated with shortened recurrence-free and overall survival. Similarly, Sakamoto et al. (23) have reported that elevated serum CA19-9 levels after chemotherapy shortened recurrence-free and overall survival, while elevated serum CEA level after chemotherapy was a risk factor for recurrence. Mitsuyama et al. (24) have reported that serum CEA levels of ≥100 ng/ml and CA19-9 levels of ≥100 U/ml before chemotherapy in patients with colon cancer and unresectable liver metastasis were prognostic factors for poor prognosis. Thus, given that there is no consensus on this matter, we evaluated serum CEA and CA19-9 before implementing first-line chemotherapy for unresectable colon cancer and again at four months after starting chemotherapy to determine whether serum CEA and CA19-9 were prognostic factors.
Comparison of primary cancer sites in the right- and left-side colons.
We reported a correlation between serum CEA and CTC levels in which CTC appear when the serum CEA levels were at least 10-fold higher than the normal levels (50 ng/ml), and any subsequent increase in serum CEA was associated with an increase in CTC. Accordingly, the serum CEA cutoff levels in the current study were 50 ng/ml (17, 18). The CA19-9 cut-off levels were set at 37 U/ml which was the upper normal value.
Analysis of the Cox proportional hazard model revealed significant differences in two factors: the serum CA19-9 levels at four months after starting chemotherapy and the use of an oxaliplatin-based regimen (FOLFOX, SOX, and XELOX) as the first-line chemotherapy regimen. The pre-chemotherapy serum CEA and CA19-9 levels were not prognostic factors, which was an unexpected finding. The use of potent oxaliplatin-based chemotherapies, such as FOLFOX, SOX, and XELOX, as first-line chemotherapy could explain why these levels were not prognostic. Also, the life-prolonging effect of second and third-line therapies might have influenced the outcome.
Analysis of the Cox proportional hazard model regarding differences in primary cancer site (right-side or left-side colon), which has recently become a topic of discussion (25-27), revealed a significant difference only in serum CA19-9 levels at four months after starting chemotherapy, and the serum CA19-9 levels had a higher hazard ratio than the difference between the primary cancer being in the left or the right colon. It has been reported that little effect can be expected from anti-epidermal growth factor receptor (EGFR) antibodies as a concomitant drug with chemotherapy in patients with primary cancer in the right-side colon (25-27). In the current study, there were more cases in which angiogenesis inhibitors were used as a concomitant drug with chemotherapy in patients with primary cancer in the right-side colon as compared to anti-EGFR antibodies. In fact, we found no significant difference in overall survival between left-side and right-side primary cancers. Using angiogenesis inhibitors as a concomitant drug with chemotherapy may be effective for patients with primary cancer in the right side of the colon.
Correlations between chemotherapy effect in patients with the primary cancer site in the left or right side of the colon and various factors.
Assessment of the difference in changes in tumor marker levels before and after chemotherapy revealed a significant reduction of serum CEA levels after chemotherapy and no significant difference in CA19-9 levels. Perhaps, CTC disappeared due to the potent chemotherapy, resulting in the serum CEA levels below the cut-off levels, which disqualified serum CEA as an effective predictor. Patients with a serum CA19-9 levels of ≥37 U/ml at four months after starting chemotherapy had a poor prognosis. Therefore, early switching to a second-line therapy and the selection of a concomitant agent may be worthwhile in this category of patients.
Differences in tumor marker changes before and after chemotherapy.
In conclusion, serum CA19-9 concentration at four months after first-line chemotherapy seems to be a useful predictor of survival in patients with mCC.
Comparison of overall survival rate after starting chemotherapy between the two groups. No significant difference was demonstrated.
Footnotes
Authors' Contributions
All Authors performed operations, analyzed the data of patients regarding their clinical features, and have been involved in drafting the manuscript. KY had given final approval of the version to be published. All Authors read and approved the final manuscript.
This article is freely accessible online.
Conflicts of Interest
The Authors have no conflicts of interest to declare regarding this study.
- Received July 31, 2019.
- Revision received August 14, 2019.
- Accepted August 22, 2019.
- Copyright © 2019 The Author(s). Published by the International Institute of Anticancer Research.
