DOI: https://doi.org/10.21203/rs.3.rs-2000787/v1
We aimed to assess the prognostic value of the combination of post-operative CEA and CA199 in patients with locally advanced rectal cancer (LARC) undergoing neoadjuvant chemoradiotherapy followed by TME.
Combined serum tumor biomarkers (CSTB) score were defined: Score 0: post-operative CEA < 2.550 and post-operative CA199 < 16.36; Score 2: post-operative CEA > 2.550 and post-operative CA199 > 16.36. Score 1: Other situations. The clinical outcome were overall survival (OS) and disease-free survival (DFS).
According to multivariate analysis, for OS, only post-operative CA199 score were prognostic predictors, while post-operative CEA was not. For DFS, only CSTB score and perineural invasion were prognosis predictors.
This research revealed the novel scoring system combining postoperative CEA and CA199 had better prognostic value than those two index alone.
Previous researches indicated that pathological factors had strongest prognostic value for LARC patients, including lymphatic invasion, CRM invasion, perineural invasion, vascular invasion, T staging and lymph node metastasis[1, 2]. Nonetheless, pathological factors were frequently affected by the quality of operation and specimen processing, were highly subjective and were difficult to quantify. EphA4 [3], special proteins[4, 5], and microRNA[6, 7] were novel molecular indicators significantly associated with patient outcomes. Regrettably, measuring these marks was costly and time consuming. Therefore, it is essential to find easily available and valuable prognostic markers to accurately predict the survival of LARC patients undergoing NCRT.
Clinical and laboratory components were significantly related to the clinical outcome of RC by numerous previous research, among which post-operative serum CEA and CA199 value were the strongest predictors [8, 9]. However, the researches on these factors almost always focused on one index, and the results were not consistent. Moreover, CEA and CA199 levels could be interfered with by many factors, which limit their application and credibility in the clinic, and also makes the conclusion of the article inconsistent.
The combined serum tumour biomarker (CSTB) score combined the two strong prognostic factors. We hypothesized that the CSTB score might also have a high prognostic value. Thus, for the first time, we aimed to evaluate the prognostic value of CSTB score in LARC patient who underwent NCRT.
This is a retrospective study from our single centre. LARC patients diagnosed from January 1, 2011, to February 28, 2020, were examed. The study was approved by our institution’s medical ethics committee (Approval NO.2020-18). Individual consent for this retrospective analysis was waived, since this was a retrospective study The diagnoses of LARC were confirmed by pre-NCRT MRI, CT, Transrectal ultrasound and pathologically biopsies. The inclusion criteria were: 1) patients were histologically diagnosed as RC; 2) locally advanced stage diagnosed by pre-NCRT MRI, CT and transrectal ultrasound. 3) patients received 5-fu based pre-operative chemoradiotherapy; 4) patients received the biochemistry test after surgery; 5) patients had a normal liver function. The exclusion criteria were: 1) patients with detectable haematology diseases, liver diseases, inflammatory or infectious diseases; 2) recurrent tumours; 3) combined with other tumours; 4) patients with incomplete or unavailable data. The patient screening flow chart was shown as Fig. 1
The optimal cut-off value of post-operative CEA and CA199 level for prognosis were determined by the ROC curve. Post-operative CEA > 2.550 and CA199 > 16.36 were associated with poor prognosis
The end-points of this research were overall survival (OS) and disease-free survival (DFS). OS was defined as the time interval from the operation date to the death date caused by any reason or the last follow-up date. DFS was defined as the time interval from the date of operation to recurrence, metastasis or the last follow-up. Follow-up information was obtained from telephone contacts or out-patients clinic records.
Statistical analyses were processed by the IBM SPSS Statistics for Windows, version 22.0 (IBM Corp, USA) [10]. The best cut-off values were achieved by receiver operating characteristics (ROC) curves [11]. Continuous and categorical variables were presented as the median (IQR) and frequencies, respectively. Univariate Cox regression analyses were adopted to detect possible risk factors associated with a poor prognosis for DFS and OS [12], and factors with a P-value < 0.05 were included in multivariate analyses [13]. Kaplan-Meier analysis was applied for survival curve [14], and the log-rank test was applied for survival difference [15].
Two Cox regression models excluding and including the CSTB score were constructed to avoid the effect of post-operative CEA and CA199 level on the CSTB score. P-values < 0.05 were considered statistically significant in all statistical tests and all p values were two-sided.
According to the inclusion and exclusion criteria, we included a total of 234 LARC patients undergoing NCRT. The primary characteristics of patients were presented in Table 1.
Characteristics |
N / Median (IQR) |
---|---|
Sex |
|
Female |
95 |
Male |
139 |
Age |
56.0 (50.0–66.0) |
CEA |
1.93 (1.19–2.96) |
CA199 |
10.7 (6.7–16.8) |
pT stage1 |
|
T0 |
49 |
T1 |
11 |
T2 |
57 |
T3 |
106 |
T4 |
11 |
pN stage1 |
|
N0 |
169 |
N1 |
51 |
N2 |
14 |
pTNM stageA |
|
0 |
46 |
I |
51 |
II |
72 |
III |
65 |
Vascular invasionA |
|
Yes |
13 |
No |
221 |
Lymphatic invasionA |
|
Yes |
16 |
No |
218 |
Perineural invasionA |
|
Yes |
45 |
No |
189 |
CRMA |
|
Positive |
9 |
Negative |
225 |
A: All the above were pathological status. TNM: tumor node metastasis |
Among these 234 patients, 139 patients were male, the median age of all patients was 56.0 years, with interquateral range (IQR) 50.0–66.0. The median post-operative CEA level was 1.93 (with IQR 1.19–2.96), and the CA199 level was 10.7 (with IQR 6.7–16.8). A total of 46 patients had complete tumour response after NCRT. Sixty-five patients presented pathologic LN metastasis, 13 patients presented pathologic vascular invasion, 16 patients with pathologic lymphatic invasion, 45 patients with pathologic perineural invasion and nine patients with positive CRM.
The median follow-up period was 33 months (range 2–122 months), with eight patients (3.4%) lost to follow up. During this period, 34 patients died, and 50 patients developed tumour recurrence. Among these 50 patients, no patients presented local recurrence. All were distal metastasis.
The cut-off value of the post-operative CEA for predicting prognosis was 2.550 (sensitivity of 68.1% and specificity of 75.5%, area under the curve = 0.677). The cut-off value of the post-operative CA199 for predicting prognosis was 16.36 (sensitivity of 44.1% and specificity of 75.5%, AUC = 0.630)
According to K-M curve analysis, patients with higher post-operative CEA or CA199 level were significantly associated with a decreased OS rate and DFS rate (Fig. 2 for CEA and Fig. 3 for CA199). Patients with CSTB score 2 had significantly poorer OS (Fig. 4A, p = 0.002) and DFS (Fig. 4B, p = 0.007) than patients with score 1–2.
Univariate analyses demonstrated that the post-operative CEA level < 2.550 (P = 0.007), post-operative CA199 level < 16.36 (P = 0.004), CSTB score < 2 (P = 0.002), lower pathological TNM stage (P = 0.029), negative pathologic lymphatic invasion (P = 0.012), and negative pathological perineural invasion (p = 0.007) were potential predict factors for a better OS. Multivariate analyses indicated that post-operative CEA level < 2.550 was not significantly correlated with better OS (hazard ratio (HR) = 0.497, 95% confidence interval (95% CI): 0.247-1.000, P = 0.050). Only post-operative CA199 level < 16.36 (HR = 0.481; 95% CI: 0.235–0.987; p = 0.046) and CSTB score < 2 (HR = 0.636; 95% CI: 0.168–0.787; p = 0.010)were. CSTB score was a more significant predictor for OS. Detailed information is shown in Table 2.
Univariate analysis |
Multivariate analysis |
||||||
---|---|---|---|---|---|---|---|
Model 1 |
Model 2 |
||||||
Characteristics |
HR (95% CI) |
P- value |
HR (95% CI) |
p- value |
HR (95% CI) |
p- value |
|
Sex |
Male vs female |
1.991 (0.929–4.268) |
0.077 |
||||
Post-operative CEA |
< 2.550 vs > 2.550 |
0.392 (0.199–0.772) |
0.007 |
0.497 (0.247-1.000) |
0.050 |
||
Post-operative CA199 |
< 16.36 vs > 16.36 |
0.373 (0.190–0.732) |
0.004 |
0.481 (0.235–0.987) |
0.046 |
||
CSTB score |
0–1 vs 2 |
0.302 (0.143–0.635) |
0.002 |
0.636 (0.168–0.787) |
0.010 |
||
Pathological TNM stage |
(0-I / II-III) |
0.396 (0.172–0.909) |
0.029 |
0.541 (0.225–1.301) |
0.170 |
0.548 (0.227–1.324) |
0.182 |
Vascular invasionA |
Absent vs Present |
0.376 (0.132–1.072) |
0.067 |
||||
Lymphatic invasion A |
Absent vs Present |
0.320 (0.132–0.775) |
0.012 |
0.711 (0.260–1.942) |
0.506 |
0.677 (0.253–1.809) |
0.436 |
Perineural invasion A |
Absent vs Present |
0.375 (0.185–0.761) |
0.007 |
0.546 (0.247–1.210) |
0.136 |
0.527 (0.241–1.153) |
0.109 |
CRM A |
Absent vs Present |
0.854 (0.203–3.588) |
0.829 |
||||
Model 1: Post-operative CEA and CA199 were included in multivariate analysis, CSTB score was not. Model 2: CSTB score was in multivariate analysis, Post-operative CEA and CA199 were not; CSTB score: Combined serum tumor biomarker score; A: All the above were pathological status. TNM: tumor node metastasis |
Univariate analyses demonstrated that the post-operative CEA level < 2.550 (P = 0.029), CSTB score < 2 (P = 0.007), lower pathological TNM stage (P = 0.009), negative pathologic vascular invasion (P = 0.003), negative pathologic lymphatic invasion (P = 0.003), and negative pathological perineural invasion (P = 0.007) and perineural invasion (P = 0.001) were potential factors for a better DFS, while post-operative CA199 level < 16.36 was not. Multivariate analyses indicated that only post-operative CSTB score < 2(HR = 0.482; 95% CI: 0.248–0.936; p = 0.031)and negative pathological perineural invasion (HR = 0.457; 95% CI: 0.243–0.857; p = 0.015)were. Detailed data is presented in Table 3.
Univariate analysis |
Multivariate analysis |
||||||
---|---|---|---|---|---|---|---|
Model 1 |
Model 2 |
||||||
Characteristics |
HR (95% CI) |
P- value |
HR (95% CI) |
p- value |
HR (95% CI) |
p- value |
|
Sex |
Male vs female |
0.597 (0.326–1.095) |
0.095 |
||||
Post-operative CEA |
< 2.550 vs > 2.550 |
0.535 (0.306–0.937 |
0.029 |
0.573 (0.326–1.007) |
0.053 |
||
Post-operative CA199 |
< 16.36 vs > 16.36 |
0.574 (0.323–1.020) |
0.059 |
||||
CSTB score |
0–1 vs 2 |
0.410 (0.214–0.788) |
0.007 |
0.482 (0.248–0.936) |
0.031 |
||
Pathological TNM stage |
(0-I / II-III) |
0.408 (0.209–0.797) |
0.009 |
0.560 (0.276–1.136) |
0.108 |
0.592 (0.289–1.210) |
0.151 |
Vascular invasionA |
Absent vs Present |
0.302 (0.135–0.672) |
0.003 |
1.014 (0.309–3.335) |
0.981 |
1.004 (0.328–3.072) |
0.994 |
Lymphatic invasion A |
Absent vs Present |
0.281 (0.136–0.580) |
0.001 |
0.446 (0.159–1.366) |
0.164 |
0.522 (0.189–1.441) |
0.210 |
Perineural invasion A |
Absent vs Present |
0.339 (0.191–0.601) |
0.001 |
0.486 (0.260–0.911) |
0.024 |
0.457 (0.243–0.857) |
0.015 |
CRM A |
Absent vs Present |
0.789 (0.237–2.627) |
0.789 |
||||
Model 1: Post-operative CEA and CA199 were included in multivariate analysis, CSTB score was not. Model 2: CSTB score was in multivariate analysis, Post-operative CEA and CA199 were not; CSTB score: Combined serum tumor biomarker score; A: All the above were pathological status. TNM: tumor node metastasis |
The prognostic value of the combination of post-operative CEA and CA199 for LARC patients undergoing NCRT was tested in this study. Patients with post-operative CEA < 2.550 had better OS and DFS, while post-operative CA199 < 16.36 only had better OS. The results of multivariate analysis showed that CSTB score was an independent factor associated with OS and was more significant than post-operative CEA level. However, for DFS, only CSTB score and negative pathological perineural invasion were independent predictors; post-operative CEA and CA199 level were not.
Serum carcinoembryonic antigen (CEA) is a glycoprotein anchored on the surface of glycosyl phosphatidylinositol (GPI) cells [16], which is the key to metastasis and dissemination of colon cancer cells [17]. It was usually measured in the pre-treatment examination of RC patients. The value of serum CEA level on prognosis has been widely discussed in related literature [18, 19] [17, 20]: Patients with elevated CEA levels before or after NCRT have poor tumour response and increased risk of recurrence [20, 21]. Researches also stated that the reduction ratio of pre- to post-CRT serum CEA levels might be a prognostic factor for DFS in RC patients with a pre-NCRT CEA of more than six ng/ml [22]. However, it was rarely reported prognostic value for post-operative CEA level in LARC patients. Our study revealed that lower post-operative CEA level associated with better prognosis, which was in consistence with many previous studies.
CA19-9 is an antigen expressed by the glycosylated extracellular MUC1 protein and plays an essential role in cancer invasion by indirectly enhancing cell adhesion and promoting angiogenesis [23]. Many studies showed that CA199 could be a prognosis predictor for RC patients [24–26]. Unfortunately, CA19-9 can be raised by several types of gastrointestinal cancer, such as esophageal cancer, colorectal cancer, and hepatocellular carcinoma [27–29]. Apart from cancer, elevated levels could be present in benign diseases, including pancreatitis, cirrhosis, and diseases of the bile ducts [30, 31]. Therefore, it failed to be widely used in the clinic to forecast the prognosis for RC patients, especially in LARC patients who underwent NCRT.
Instead of the single serum tumour biomarker, CSTB was a score that combined these two. We assumed that the CSTB score could predict prognosis in RC patients mainly for the following two reasons. First, the CSTB score was a combination of two significant serum tumour biomarker predictive factors with a physiological mechanism. Second, the application of the ROC curve, not just whether the tumour index exceeds the standard value, contributed to finding the best cut-off value of prognosis prediction. Therefore, we believed the CSTB score was capable of evaluating outcomes of LARC patients.
The prognostic value of CEA and CA199 were inspected by a few studies. Yang et al. manifested a significant connection between pre-operative CEA level and RC patients’ prognosis .In the study of Yang et al., low pre-operative CEA level was proved to be a valuable predictor in RC [32]. Zhang et al. reported that the elevated CA199 was an independent risk factor of worse prognosis in LARC patients [33]. However, none of them focused on combining post-operative CEA and CA199 level and specific on LARC patients undergoing NCRT. Consequently, we designed the present research to investigate the correlation between the CSTB score and prognosis in LARC patients undergoing NCRT, and the results suggested that the CSTB score < 2 was associated with significantly better OS and DFS.
Our study's most important clinical significance was as follows: Firstly, we focused on post-operative CEA and CA199 level instead of pre-treatment since post-operative indicators have a more robust predictive effect on prognosis. Moreover, we focused specifically on LARC patients undergoing NCRT, a research hot spot. Finally, we combined these two scores and adopted the ROC curve to determine the optimal cut-off point. CSTB score may not only assess the risk of LARC patients but also contributed to making treatment decisions. In detail, compared with LARC patients with low CSTB score, patients with high CSTB scores were indicated to be treated more aggressively, and post-operative adjuvant therapy may be more beneficial and adjustable to patients in the same circumstances.
The research had significant drawbacks. First, this was a retrospective study. Second, we determined 2.550 as the optimal cut-off value of post-operative CEA and 16.36 for CA199. The cut-off values may not be the most suitable. However, the ideas and methods in this research can be widely used the ideas and methods in this research can be widely used. Even the optimal cut-off may vary among different people, irrespective of the specific cut-off point, this study could still show the prognosis of patients with higher CEA CA199 level was poor.
In conclusion, our study first reported the correlation between the CSTB (combination of post-operative CEA and CA199 level) score and prognosis of LARC patients undergoing NCRT. CSTB score was significantly associated with OS and DFS and had better value for predicting prognosis than those two.
LARC, locally advanced rectal cancer; NCRT, neoadjuvant chemoradiotherapy; CSTB, Combined serum tumor biomarkers; OS, overall survival; DFS, disease-free survival; ROC, receiver operating characteristic;
Data Availability Statement
Inquiries can be directed to the corresponding author [email protected]
Conflicts of interests
All authors have no conflicts of interest to declare.
Fundings
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Author Contributions
Runzhuo Sun: Data curation, Formal analysis, Writing- Original draft preparation. Yimiao Zeng: Conceptualization, Methodology, Writing- Original draft preparation. Yuanyuan Fan: Visualization. Dan Lin: Investigation. Siyu Du: Software, Validation. Xiaoyu Wang: Supervision, Writing- Reviewing and Editing.
All authors reviewed the manuscript.
Reporting Checklist
The authors have completed the STROBE reporting checklist
Ethical statement
The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by our institution’s medical ethics committee (Approval NO.2020-18),and individual consent for this retrospective analysis was waived, since this was a retrospective study.