Although CA242 has been identified and used in clinical practice for some time, this is the first study to profile the levels of this biomarker in 35 different tumors in a large number of patients. Among the 35 tumor types, the serum CA242 level changes in 23 tumor types had never been reported before. We also found that serum CA242 levels in patients with metastatic tumors might be related to the target organs containing metastases for the first time.
We found that CA242 levels were elevated in a variety of digestive system cancers, including pancreatic cancer, gallbladder cancer, cholangiocarcinoma, colorectal cancer, and gastric cancer, which was consistent with previous reports. The AUC values of the ROC curves for pancreatic cancer, gallbladder carcinoma, and cholangiocarcinoma were all greater than 0.8, but only pancreatic cancer has been studied in detail[4, 10, 13, 17, 22, 25–27]. Previous studies have shown that CA242 alone or in combination with other biomarkers can predict the diagnosis, progression, and prognosis of pancreatic cancer[4, 7, 10, 12–16, 23, 24]. Pancreatic cancer has been the leading clinical application for CA242 because of its excellent performance in this cancer. Some small-scale studies have investigated CA242 in gallbladder carcinoma and cholangiocarcinoma[7, 14, 25, 28], but due to the sample size or quality of these studies, the credible AUC values for these two carcinomas are still unknown. Therefore, our findings indicate that CA242 is also suitable for use as a biomarker for cholangiocarcinoma and gallbladder cancer, whose AUCs were 0.8145 (0.7737, 0.8553) and 0.8349 (0.8157, 0.8542), respectively (Table 1).
The study of serum CA242 levels in the ampulla of Vater carcinomas, including ampullary carcinoma and duodenal papilla carcinoma, is shown for the first time. A previous study reported that the mean level in the ampulla of Vater carcinoma is 21.1 ± 26.1 [29], but the level was 47.87 ± 93.24 for the ampulla of Vater carcinoma in our research. Furthermore, the median level was 2.70 times that of healthy controls, and the AUC value reached 0.7268 (0.6581, 0.7955) (Table 1). The discrepancies between the previous study and our study may be due to the sample size in the other study being too small to produce a statistical error or the difference between the ampulla of Vater carcinoma and duodenal papilla carcinoma not being negligible.
CA242 has been reported to be more sensitive and specific than CA50 and CEA in Dukes' A-D colorectal cancer[22], in addition to its considerable value in predicting recurrence and metastasis in clinical stage II and III colorectal cancer[18]. Moreover, many studies have demonstrated that CA242 works well in predicting the diagnosis, progression, and prognosis of colorectal cancer in combination with other tumor markers[18, 19, 26, 30]. Our results showed that the mean value in colorectal cancer was 26.50 ± 59.58 and that the AUC was 0.7241 (0.7073, 0.7408) (Table 1), which was consistent with previous studies[22, 26, 27].
CA242 levels in patients with gastric cancer differed between stage IV and non-stage IV[20], but alone, CA242 was not a good predictor of LNM[21]. However, in a study by Fangxuan Li et al.[31], the mean value of 29.84 ± 83.54 in gastric cancer was not much different from our data, but their AUC for gastric cancer relative to healthy people was 0.809, which is much higher than the value of 0.6839 (0.604, 0.8202) found in this study (Table 1). We suggest that a possible explanation may be that their sample size is too small and that their mean lower than the control is 3.85 ± 3.13. A surprising finding was that the CA242 level in patients with gastric cardia adenocarcinoma was not significantly different from that in gastric cancer patients.
In this study, the mean value in lung cancer was 18.49 ± 45.97 (Table 1), but Huaiqian Dou et al.[10] reported that the mean in lung cancer was 9.34 ± 13.67. In contrast, Xiaochuan Wang et al. [32] reported that it was 28.39 ± 51.91. These differences might be caused by a sampling error, detection kit choice, and instrument differences. However, the results we reported have the largest sample size and therefore higher credibility. The mean value in ovarian cancer was 19.52 ± 42.27 (Table 1), which was slightly higher than the previously reported value of 11.40 ± 16.88[10], and the mean in cervical cancer was 11.54 ± 23.97, which was very similar to the previously reported value[10]. The levels in nasopharyngeal carcinoma, craniofacial malignant tumor, endometrial cancer, breast cancer, renal cell cancer, liver cancer, esophageal cancer, and prostate cancer patients were significantly higher than those in healthy controls (Table 1), indicating that CA242 has potential clinical value in these diseases. Although mediastinal cancer, testicular cancer, ureteral cancer, and neuroendocrine cancer patients had higher mean and median values than health controls (Table 1), the sample size was too small to identify significant differences.
The CA242 levels in colorectal cancer and gastric cancer patients increased not only with the appearance of LNM or DM but also with advancement from stage I to IV[18, 20–22, 26, 30]. The results mentioned above are consistent with the trends in a previous study, and we also found that CA242 was able to distinguish colorectal cancer and gastric cancer patients from healthy controls (Table 1). The AUC values for diagnosing stage IV disease reached 0.85 or higher and, for DM, reached 0.8299 (0.8117, 0.8481) and 0.7359 (0.6955, 0.7764) in colorectal and gastric cancer patients, respectively (Table 2 and Table 3). In patients with pancreatic cancer, cholangiocarcinoma, gallbladder carcinoma, lung cancer, cervical cancer, ovarian cancer, periampullary carcinoma, bladder cancer, breast cancer, or endometrial cancer, serum CA242 levels were positively correlated with LNM and DM (Table 2). Some diseases, such as pancreatic cancer, have been reported, but the rest are reported here for the first time. There were no significant differences in CA242 levels between patients with laryngocarcinoma, esophageal cancer, nasopharyngeal cancer, or renal cancer metastasis and those with NM nor were there different among patients with various stages of disease for these cancers (Table 2).
Among thyroid cancer, prostate cancer, and craniofacial malignant tumor patients, the level of CA242 increased with metastasis, but the differences were not significant (Table 2). While endometrial cancer and bladder cancer showed significant decreases in the sample, although the CA242 level showed an increasing trend with advancing clinical stage, the data showed no significant differences (Table 3). The remaining tumor sample sizes were too small to draw valid conclusions. According to whether the CA242 level increased with metastasis or stage advancement in the tumors listed in Tables 2 and 3, we identified tumors as "CA242 progression-positive tumors". In these tumors, the combination of other tumor markers and CA242 could be further studied to improve the accuracy of the prediction of tumor progression. We speculate that some commonalities exist among these tumors, such as mutations in similar tumor driver genes[33], similar embryonic developmental origins, and microenvironments. Further research is needed to confirm these findings, which may have the potential to facilitate the diagnosis and treatment of tumors.
Furthermore, we also found that in all of the cancers listed in Table 4, the increase in the serum CA242 value for multiple metastases compared to NM was the largest, followed by the increases for liver metastasis and bone metastasis. Some studies have shown that the prognosis of gastric cancer with liver and bone metastasis is worse than that of any other subtype of gastric cancer, and the two organs even account for most metastatic sites [34, 35]. Interestingly, we find that similar situation also exist in colorectal cancer, lung cancer, pancreatic cancer, breast cancer and other tumors [36–40].We deem that the liver and bones may have microenvironments conducive to the synthesis and release of CA242 molecules in the five "CA242 progression-positive tumors" or that the liver and bones enhance metastatic tumor cell growth so that more tumor cells produce more CA242, resulting in elevated levels. Therefore, CA242 may be valuable in predicting the progression of "CA242 progression-positive tumors". Moreover, studying the molecular mechanisms underlying the elevated CA242 levels may produce an understanding of the mechanisms of tumor metastasis.
However, not all malignant tumor patients’ serum CA242 levels were greater than those of healthy controls. For example, the sample sizes for leukemia, laryngocarcinoma, skin cancer, and penile cancer were not small, but there were no significant differences between these patients and the healthy controls in regard to the median levels (Table 1). This finding indicated that serum CA242 levels are not suitable for diagnosing all diseases, but CA242 may be used to identify nonprimary tumors at the same site, that is, metastatic tumors with increased CA242 expression. Moreover, the median serum CA242 values of the two precancerous lesions in this study, benign colorectal neoplasms and cervical intraepithelial neoplasia, were similar to the median value of the healthy controls, and there were no significant differences. However, the median CA242 level in benign colorectal neoplasms was less than half of that in colorectal cancer, but the difference was not significant, which is probably caused by the small sample size for benign colorectal neoplasms. By increasing the number of samples, we might find a difference between these two precancerous lesions and a new marker to identify benign colorectal neoplasms and colorectal cancer, as is done for cervical cancer and cervical intraepithelial neoplasia.
This study has several limitations. First, the sample size is relatively insufficient. Although the total sample size was more than 35,000, there were 35 different kinds of diseases, so each disease included only 1,000 patients on average. Thus, some tumors, such as colorectal cancer and liver cancer, were numerous, with more than 4,000 cases. However, some had many fewer cases, such as mediastinal tumor, ureteral cancer and fallopian tube cancer, which had only 10 cases. The reasons for this situation are not only the disparity in the incidence rates of different diseases but also the previous understanding of the diagnostic role of CA242. CA242 is considered to be a diagnostic marker for digestive system tumors[10, 13, 22, 26], especially pancreatic cancer, so measuring CA242 levels is not an appropriate choice when a doctor suspects a patient may develop other tumors. Second, in the analysis of individual diseases, the subpopulations are not sufficiently accurate. In this study, we stratified subpopulations according to the metastatic status, clinical stage, and metastatic site of the tumors. Although some differences and factors that may affect the level of CA242 were revealed, we still did not find an excellent stratified method that could make the distribution concentrated and failed to find further accurate factors that influenced CA242 levels. Third, the collection of patient clinical information is relatively incomplete, and only some patients had accurate metastasis and clinical staging information. However, we gathered no staging classification information for some tumor patients, such as those with pancreatic cancer, cholangiocarcinoma or gallbladder cancer. Therefore, we missed the opportunity for further analysis of the diseases that most commonly use CA242 in the clinic. Fourth, this was a single-center, retrospective study. Therefore, the numbers of patients in each stage and metastatic status were very different, which may lead to biased conclusions. In addition, the CA242 values for different stages and metastatic states belonged to different patients. In other words, this study had neither a paired design nor a longitudinal design so some confounding factors may have influenced the conclusion of the study.
In this study, we analyzed serum CA242 levels in 35 types of neoplastic disease patients and healthy controls and found that these levels remained unchanged in benign tumors and increased most in malignant tumors, especially pancreatic cancer, gallbladder cancer, cholangiocarcinoma, and periampullary carcinoma. The CA242 levels of these malignant digestive system tumors were significantly higher than those of other tumors and were valuable for patient diagnosis. Moreover, the level of CA242 was also related to metastasis and clinical stage and was able to predict the progression of some tumors.