Clinical Value of Pepsinogen in the Screening, Prevention, and Diagnosis of Gastric Cancer (2024)

,

Abstract

Gastric cancer (GC) has a high incidence in China. Patients with GC tend to be young; the incidence of GC is typically due to stress, diet, and Helicobacter pylori (HP) infection. Reducing the incidence of and mortality from GC is a major public health issue; early diagnosis and treatment remain the best strategies.^1,2 Identification of new diagnostic markers that can rapidly translate into clinical applications and meet the needs for diagnosis and treatment has become a highly dynamic and cutting-edge field in biomedical sciences.

Many studies aim to provide a simple, sensitive, and specific method for the early diagnosis of GC. To date, a range of tumor-related factors have been adopted in clinical practice and have become important assets to tumor diagnosis and treatment. However, their sensitivity and specificity have not reached sufficient reliability.^3–5

Pepsinogen (PG), an aspartic protein enzyme secreted by gastric mucosa cells, is an inactive precursor of pepsin in gastric juice. It is composed of 375 amino acids, has a molecular weight of approximately 42 kDa, and can be divided into 2 subgroups according to differences in biochemical and immune characteristics: PGI (groups 1–5) and PGII (groups 6 and 7). Gene expression sites of PGI and PGII are different, located on chromosomes 11 and 6, respectively.

After synthesis, PG is primarily released into the gastric cavity, with a small proportion entering the blood circulation through the capillary of gastric mucosa.^6,7 PG can reflect the function and state of the gastric mucosa and is closely related to gastric lesions. Xingjie S et al⁸ studied the serum PG content of patients with GC before and after endoscopic treatment and found that PG content after treatment was higher than before the operation. Chunchun C et al⁹ reported that the serum PG content of patients with gastric ulcers was relatively high.

Serological testing of PG has the advantages of being noninvasive and inexpensive, allowing for convenient and dynamic monitoring, and being well accepted by the patients. It also has cost-effectiveness compatible with large-scale screening in areas with high incidence of atrophic gastritis (AG) and GC.¹⁰ To assess the clinical value of tests based on serum PG measurement, we investigated the diagnostic accuracy of these tests for GC diagnosis, the relationship between serum PG levels and the occurrence and development of GC, and the feasibility of implementing a dynamic detection to perform serological monitoring.

Using the pathological examination of gastric mucosal biopsy as a diagnostic reference standard, we compared PG levels in healthy donors and patients with precancerous lesions or GC and explored the relationship between serum PG levels and the occurrence and development of GC. Although previous publications have reported the clinical value of serum PG measurement, most studies were single-centered and lacked validation. Therefore, we conducted a multicenter study to provide a stronger reference for clinical diagnosis and treatment.

Materials and Methods

Inclusion Criteria

Patients with GC were diagnosed by combining clinical symptoms, endoscopy, and histopathology. A specimen of the cancer tissue and its surrounding tissues were taken with forceps and stained with hematoxylin-eosin (H&E), observed under a microscope, and diagnosed according to the morphology and structure of tissues and cells (loss of polarity, disordered arrangement, nuclear atypia, increased nucleoplasmic ratio, and condensed chromatin). The diagnosis of GC, AG, and other related diseases met the consensus for pathological diagnosis of chronic gastritis and epithelial tumor gastric mucosa biopsy.^11–13

No patient experienced serious disease such as liver or kidney failure, other tumors, or had a history of chemotherapy or gastrointestinal surgery, radical treatment of HP, pernicious anemia, intake of drugs acting on the digestive system, or anticancer drugs during the 2 weeks preceding the test.¹⁴ All of the procedures complied with medical ethics requirements. For healthy donors, routine blood tests and liver and kidney function were normal (test results were within the normal ranges: total protein [TP], 65–85 g/L; alanine aminotransferase [ALT], 8–40 u/L; aspartate aminotransferase (AST), 15–35 u/L; total bilirubin [TBIL], 3.5–19 μmol/L; creatinine [CREA], 60–80 μmol/L; uric acid [UA], 202–339 μmol/L; blood urea nitrogen [BUN], 3.0–8.2 mmol/L). There were no space-occupying lesions in any organ, liver, kidney, or digestive tract diseases.

Tumor-node metastasis (TNM) staging of patients with GS was divided into 4 groups according to the latest cancer-staging standard of the American Joint Committee on Cancer (stage I group, stage II group, stage III group, and stage IV group).^15,16 On gastroscopy AG showed a pale mucosa, vascular permeability, small folds or rough surface, erythema, and mucus. Pathological examination showed gland atrophy and hyperplasia. Combined symptoms, such as gastric distension and anemia, completed the diagnosis.^17,18

Cases and Grouping

In total, 117 patients with GC confirmed by gastroscopy and pathological examination between February 2017 and October 2018 were selected from the Anting Hospital of Shanghai, Ruijin Hospital of Shanghai, The First Affiliated Hospital of Suzhou University, and the Kunshan Branch of Shanghai Cancer Hospital. The GC group included 59 men and 58 women aged 32 to 81 years with a median age of 64 years. Amongst these patients, 13 had GC stage I, 47 patients had stage II, 41 patients had stage III, and 16 patients had stage IV. During the same period, 120 donors (including 60 men and 60 women; age, 24–72 years; median age, 47 years) with normal physical examination indices were enrolled in the healthy control group. There were no significant differences in age and sex composition of the groups (P> .05).

Equipment and Methods

The tests were run on an Abbott ARCHITECT i-2000sr automatic chemiluminescence immunoassay system (AbbVie Inc.) that had passed performance verification. The reagents were provided in the kit associated with the instrument. The ARCHITECT Pepsinogen assay is a 2-step immunoassay for the quantitative determination of PG in human serum using chemiluminescent microparticle immunoassay (CMIA) technology with flexible assay protocols, referred to as Chemiflex. In the first step, specimen, assay-specific diluent and antihuman PG antibody-coated paramagnetic microparticles are combined. PG present in the specimen binds to the antihuman PG antibody-coated microparticles. After washing, anti–human PG antibody-acridinium–labeled conjugate is added in the second step. Following another wash step, pretrigger and trigger solutions are added to the reaction mixture. The resulting chemiluminescent reaction is measured as relative light units (RLUs).

A direct relationship exists between the amount of PG in the specimen and the RLUs detected by the ARCHITECT i optical system. The concentration of PG in the specimen is determined using a previously generated ARCHITECT pepsinogen-calibration curve. Interlaboratory quality assessment and indoor quality control were conducted, to ensure the results were controlled. The reportable range was as follows: PGI, 0 to 1000 ng/mL; PGII, 0 to 500 ng/mL. The uncertainty of calibration is 1.99 when PGI is between 0 and 70. When PGII is between 0 and 5, the uncertainty of calibration is 0.13.

The batch numbers of the PGI kits were 73266LI00 and 80439LI00, with the respective expiration dates as April 12, 2018, and January 12, 2019. The batch numbers of the PGII kits were 73371LI00 and 79586LI00, with the respective expiration dates of April 17, 2018 and December 22, 2018.

Fasting venous blood was collected using disposable evacuated blood-collection tubes (BD vacuum red cap [Becton, Dickinson and Company]), and centrifuged at 2664g for 10 minutes. After centrifugation, the serum was collected and stored at −20°C for later use. Hemolysis, lipemic, and other specimens that may affect the results were excluded to eliminate the influence of interfering substances such as bilirubin, hemoglobin, albumin, and triglycerides. Before testing, the specimens were allowed to reach room temperature and centrifuged at 2664g for 10 minutes. Serum PGI and PGII concentrations were determined with an Abbott ARCHITECT i-2000sr automatic chemiluminescence analyzer, and the PGI/PGII ratio (PGR) was calculated. Results were available within 2 hours.

Statistical Analysis

SPSS statistical software, version 21.0, was used to analyze the results. Serum PGI and PGII levels were expressed as mean (SD). The Student t test was used for comparison between 2 groups with normal distributions, and ANOVA was used for comparison between multiple groups. Bivariate correlation analysis was conducted between the 2 variables. The data were compared by X² testing. P <.05 was considered statistically significant. The diagnostic efficacy of PG was evaluated by receiver operating characteristic (ROC) curve.

Results

Serum PGI, PGII, and PGR Levels

Serum PGI and PGII levels in the specimens from 120 healthy donors, 122 patients with AG, and 117 patients with GC were determined, and the PGRs were calculated. Levels from the AG and GC groups were compared with levels from the control group, as well as to each other (Table 1, Figure 1 and Figure 2).

As shown in Table 1, PGI levels between the AG and control groups were significantly different (P = .003). Although the difference in PGII level was not statistically significant between the 2 groups (P = .21), PGR was significantly lower in the AG group (P = .005). Similarly, PGI levels and PGR were significantly lower in the GC, compared with the control group (P <.001) where the PGII levels were not significantly different between the 2 groups (P = .07). Comparison between the GC and AG groups showed significant differences in PGI levels and PGR (P <.001), but no differences were observed in PGII levels (P = .46).

As shown in Figure 1, levels of PGI and PGII and PGR decreased gradually. The highest values were in the control group, intermediate values in the AG group, and lowest values in the GC group. As depicted in Figure 2, PGI levels and PGR showed a decreasing trend from the control group to the AG group to the GC group, with no significant difference in PGII levels across the 3 groups.

ROC Curve Evaluating the Diagnostic Efficacy of PG Measurement in GC

Using pathological diagnoses as the most reliable reference standard, we conducted a comparative analysis between the GC and the control groups and created the corresponding ROC curves (Figure 3). The area under the curve (AUC) was the largest for PGI levels (0.834), followed by PGR (0.752), and PGII levels (0.587), suggesting that serum PGI level and PGR have high diagnostic value for GC detection (Figure 3 and Table 2).

The point of an ROC curve closest to the top left corner represents the optimal cutoff value for the diagnosis of GC based on serum PG level. The cutoff values obtained with this method for the diagnosis of GC are shown in Table 3.

The cutoff values of PGI, PGII, and PGR were 51.2 ng/mL, 13.05 ng/mL, and 5.65, respectively. These corresponded to detection sensitivities and specificities of 81.7% and 65.8% for PGI levels, 54.2% and 68.4% for PGII levels, and 53.8% and 87.2% for PGR, demonstrating the clinical value of PG for the diagnosis of GC.

To improve the sensitivity and specificity of the test, we performed parallel and serial multiple judging. With the parallel method, multiple results are run at the same time. If a specimen scores positive in 1 result, it is considered positive. This method improves the sensitivity of the diagnostic test. Conversely, serial judging consists in performing multiple results successively. Only the cases repeatedly positive are considered true positives. This method improves the specificity of the diagnosis. The results of these combined approaches are shown in Table 4.

When PGI and PGR were combined to detect GC, the sensitivity of parallel judgment was significantly higher than that of serial judgement, and the number of false negatives was reduced. Conversely, the specificity was higher with serial judgements.

Correlation Between GC and Serum PGI and PGR Levels

To further evaluate the value of PG for the diagnosis of GC, we set the control group as the reference. We also used serum PGI and PGR levels as variables for logistic regression analysis (Table 5).

Discussion

In the early stages, GCs are easy to miss because they do not manifest any obvious symptoms or because symptoms are nonspecific (upper abdominal discomfort and fever) and similarly found in other chronic diseases such as gastritis and gastric ulcers. Therefore, the discovery of highly sensitive and specific screening methods is crucial to improving the early diagnosis of and remission in patients with GC.^19,20 Although endoscopic biopsy is regarded as the criterion standard, it is not suitable for screening and early diagnosis because of its low sensitivity and the possible occurrence of trauma and complications such as pneumothorax or cancer-cell proliferation and transplantation. Because of these associated risks, patients tend to avoid this procedure, an attitude that affects the diagnosis and treatment of the disease.^21–24

Serological examination is not only convenient but also noninvasive and could provide early warning of tumor occurrence. Still, there is still not enough positive evidence to prove that PG can be used for mass censuses.¹¹ However, screening for high-risk individuals can not only overcome the shortcomings of pathological diagnosis such as trauma but can also detect lesions early, which is conducive to diagnosis. In 2009, the Chinese Technical Plan for Cancer Screening, Early Diagnosis and Treatment, issued by the Bureau of Disease Control and Prevention of the Ministry of Health of the People’s Republic of China, recommended the use of PG serological tests as a primary screening method in GC.^25,26

By comparing the levels of serum PGI and PGII and PGR from 117 patients with GC, 122 patients with AG, and 120 healthy donors, we found that serum PGI level and PGR in the AG group were significantly lower than in the control group (P <.01), whereas there was no difference in PGII level between these 2 groups (P> .05). Similarly, comparison between the GC and the AG groups showed a significant difference in serum PGI level and PGR (P <.01), but no difference in PGII level (P = .46). The differences in serum PGI level and PGR between the GC and the control groups were also statistically significant (P <.01), contrary to the difference in serum PGII level (P = .07).

The decrease in serum PGI levels in AG and GC may be explained by long-term atrophy or damage of the gastric glands, resulting in decreased number and quality of mucosal cells secreting PGI. Another explanation could be that damaged glandular cells lead to decreased gastric-acid secretion, resulting in decreased serum PGI.

In contrast, PGII is more broadly secreted by different gastric areas, including the main cells of the fundus gland, the mucous neck cells, the mucous cells of the cardiac and pyloric glands, and the upper duodenum. The secretion of PGII is associated with atrophy of the gastric basal-gland duct, metaplasia of the gastric upper dermis or the pseudopyloric gland, and hyperplasia. Local lesions or pathologic magnification caused by single etiologies do not significantly modify the level of secreted PGII. Thus, although changes in PGII levels in GC are minor, the concentration of serum PGI decreases, resulting in decrease of PGR. These results are consistent with those in the study report by Miki et al,²⁷ which showed that changes in the serum levels of PGI and PGII reflect the degree of gastric mucosal atrophy, and that PGR diminishes as gastric mucosal-gland atrophy progresses. Thus, serum PGI level and PGR can be used as indicators of the degree of gastric mucosal-gland atrophy,^27–30 and serum PGI, PGII, and PGR can be used as detectors of AG.

There is an irreversible point in the development of GC (Figure 4), past which it becomes difficult to effectively control the progression of GC.^31–33 The risk of developing cancer in patients with AG is 4 times higher than that of healthy individuals, and in severe cases of AG, this risk factor can be as great as 90 times higher.^34,35 Therefore, early screening of precancerous gastric mucosal lesions and identification of gastric mucosal diseases are important to preventing the occurrence of GC. The most effective way to do this is to make a clear diagnosis that enables early treatment and intervention in patients with AG, thereby preventing the development of GC.

Figure 3 shows that the AUC for PGI was the largest (0.834), followed by the AUC for PGR (0.752), and the AUC for PGII (0.587). The cutoff values of serum PGI, PGII, and PGR were 51.2 ng per mL, 13.05 ng per mL, and 5.65 ng per mL, respectively. The sensitivities of tests based on PGI, PGII, and PGR were 81.7%, 65.8%, 54.2%, respectively, and their specificity was 68.4%, 53.8%, and 87.2%, respectively. In the literature,³⁶ the diagnostic sensitivities of tests based on quantification of CA724, CEA, and CA199 are 60%, 76%, and 68%, respectively, with corresponding specificities of 74%, 53%, and 69%. These data suggest that diagnostic sensitivity based on serum PG is higher than those based on CEA, CA724, and CA199, which already have high predictive values in GC diagnosis.

Shikata et al³⁷ followed up with more than 2000 community residents in Japan for an average period of 10 years and identified cutoff values for GC screening of PGI of 59 ng/mL or less and PGR of 3.9 or less. By this standard, the sensitivity and specificity, compared with the actual incidence of GC, were 71.0% and 69.2%, respectively. In their study of a sample population of Kazakhs in Xinjiang, China, Juan et al³⁸ reported cutoff values for GC screening of PGI of 64 ng/mL or less and PGR of 4.5 or less, corresponding to a sensitivity and specificity of 80.5% and 89.8%, respectively. Fariborz et al³⁹ reported good sensitivity and specificity for cutoff values of 70.95 ng/mL for PGI levels and 2.99 for PGR.

Finally, Li et al⁴⁰ used ROC curves to determine cutoff values for GC screening in the Liaoning Province of China, and found PGI 94.3 ng/mL or less and PGR of 9.3 or less, conferring sensitivity and specificity of 81.4% and 35.0%, respectively, to the initial screening of GC. Overall, these results are in a similar range as our findings. The differences across the different studies may reflect regional and population variability, among other factors. Therefore, for practical implementation, the screening criteria should be adjusted according to the local clinical specifics of each population.

Table 4 shows that for the diagnosis of GC, the combined use of PGI and PGR is more valuable than single indicators for reducing the time before GC diagnosis and improving prognosis. For GC screening, early detection and diagnosis are crucial for treatment success. Therefore, lowering diagnostic criteria to improve the sensitivity would be the best option to reduce the chance of false negatives.

The quality of a diagnostic test also relies on its specificity. Parallel judging improves the sensitivity of the predictive test and reduces the rate of false negatives, but it also reduces the specificity and increases the rate of false positives. Serial judging improves the specificity but decreases the sensitivity, thereby increasing the rate of false negatives and reducing the rate of false positives. Combined approaches help to balance sensitivity and specificity according to different priorities, and to achieve comprehensive testing.

After excluding confounding factors such as sex and age, analysis by logistic regression indicated that the occurrence of GC was negatively correlated with serum PGI level and PGR. PG levels can also be affected by gastric mucosal damage, gland atrophy, ethnicity, cancer cells, and HP, among other factors. We hypothesize that atrophy of the gastric glands leads to a decrease in the production of PG. Gastric ulcer or gastrectomy may lead to gastric mucosal damage and promote the release of more PG into the peripheral blood, thus affecting the content of PG in the peripheral blood. It is unknown whether sex, age, smoking status, alcohol consumption, or dietary habits influence PG levels and how these various factors affect PG levels. Answering these questions requires further study to more accurately and comprehensively assess the relationship between changes in serum PG level and GC progression.^41,42

In conclusion, serum PGI and the PGR levels showed a gradual decrease from the healthy control group to the AG group to the GC group. The differences in PGI level and PGR between the GC and the control groups were significant, suggesting that tests based on PGI level and PGR have good clinical application value and can be used as indicators for the early diagnosis of GC. Moreover, detection of PG level can be indicative of precancerous diseases such as AG. This diagnostic tool is a feasible way to prevent GC and reduce its incidence and mortality rate by detecting PG abnormalities before the irreversible point of GC development is reached.

Conclusion

This study established the clinical value of PG as a biological marker for the prevention, diagnosis, and prediction of GC. This method is convenient, rapid, inexpensive, noninvasive, and easily accepted by patients. It can be used for auxiliary diagnosis, as an effective supplement to endoscopy and X-ray examination, and is worthy of popularization and broader application.

Acknowledgments

We thank the following persons for their guidance and help: Shunchang Sun, from the Hospital of Shanghai, Jiao Tong University Ruijin Hospital; Tao Li, from The First Affiliated Hospital of Anhui Medical University; experts Chen Yang, Jundong Zhou, and Ping Xu, from Suzhou Hospital; and Hong Du, Xuejun Shao, Hong Zhu, Xuefeng Qian, Qingzhen Han, Yang He, and Yimin Zhao, from Suzhou University. Funding was provided by the National Natural Science Foundation of China (81401937) and the Shanghai Science and Technology Project Fund (19ZR1444800, 2020-KY-ZYY-02).

References

Author notes

© The Author(s) 2021. Published by Oxford University Press on behalf of American Society for Clinical Pathology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Issue Section:

Science

Email alerts

Citing articles via

More from Oxford Academic