The Correlation Between 18 F-Fluorodeoxyglucose-Positron Emission Tomography/Computed Tomography Semiquantitative Parameters and the Clinical Features and Pathological Biological Indexes of Gastric Cancer

Introduction

The prevalence and mortality of gastric cancer rank third of all malignant tumors in China. ¹ In recent years, the incidence of gastric cancer has reflected an obvious trend, affecting younger populations while showing a high mortality rate and poor prognosis. Surgery is the most effective treatment method for resectable gastric cancer. However, the early diagnosis rate of gastric cancer in China is low. Most of the diagnoses are in the middle and late stages, so these patients have no chance of undergoing surgery. Therefore, a comprehensive and effective treatment plan with accurate prognosis judgment needs to be devised for these patients. The formulation of such depends on the pathological stage, histological type, the status of pathological molecular marker Ki-67, and antihuman epidermal growth factor receptor-2 (HER-2) of gastric cancer. Although a pathological biopsy remains the gold standard for obtaining these parameters, its invasiveness limits the number of sampling sites. Moreover, it is impossible to obtain samples at multiple sites frequently for continuous evaluation. Therefore, the use of noninvasive diagnostic tools to predict oncological behavior is becoming more important in the field of gastric cancer.

¹⁸ F-fluorodeoxyglucose-positron emission tomography/computed tomography ( ¹⁸ F-FDG PET/CT) is a new imaging technology, which combines three-dimensional imaging with metabolic imaging, ² and the most commonly used imaging agent for it is ¹⁸ F-FDG, which is a glucose analogue. The metabolic rate of glucose in cancer cells is higher, so the FDG uptake is higher and the standard uptake value (SUV) reflects this metabolic capacity. This technology is usually used to evaluate the biological invasiveness of malignant tumors. Present research results reveal that maximum standard uptake value (SUV_max) is correlated with clinical features and pathological molecular biological markers in lymphoma, breast cancer, lung cancer, and colorectal cancer. However, studies on its application in gastric cancer are rare. In this study, the correlation of SUV_max with the clinical characteristics and pathological and biological indicators of gastric cancer was analyzed to find noninvasive indicators for predicting the biological behavior of gastric cancer and to provide a basis for developing reasonable and effective treatment programs.

Materials and Methods

Results

The correlation between SUV_max and the different degrees of differentiation

The SUV_max was 4.42 ± 1.58 in the moderately and highly differentiated adenocarcinoma group, 7.06 ± 1.22 in the poorly differentiated adenocarcinoma group, and 2.93 ± 0.83 in the signet-ring cell carcinoma group. The SUV_max of poorly differentiated adenocarcinoma was significantly higher than that of moderately and highly differentiated adenocarcinoma and signet-ring cell carcinoma (both p < 0.01), and the SUV_max of poorly differentiated adenocarcinoma was significantly higher than that of signet-ring cell carcinoma (p < 0.01, Fig. 1). The best critical value of SUV_max for predicting poorly differentiated and moderately and highly differentiated adenocarcinomas by ROC curve was 5.74 (Fig. 2A), and the area under the curve (AUC) was 0.907. When SUV_max was >5.74, the sensitivity of diagnosis of poorly differentiated adenocarcinoma was 89.6%, and the specificity was 77.8%. The best critical value of SUV_max for predicting highly differentiated adenocarcinomas and signet-ring cell carcinoma by ROC curve was 4.25 (Fig. 2B), and the AUC was 0.775. When SUV_max was >4.25, the sensitivity of diagnosis of signet-ring cell carcinoma was 53.1%, and the specificity was 88.6%.

OPEN IN VIEWER

FIG. 1.

The correlation between the different pathological types of gastric cancer and SUV_max. (A) A poorly differentiated adenocarcinoma of the stomach (DAB, × 200). The corresponding figure (D) is the PET/CT image of the same patient, and the SUV_max of the positive lesion is 7.77. (B) A highly differentiated adenocarcinoma of the stomach (DAB, × 200). The corresponding figure (E) is the PET/CT image of the same patient, and the SUV_max of the positive lesion is 4.30. (C) A signet-ring cell carcinoma of the stomach (DAB, × 200). The corresponding figure (F) is the PET/CT image of the same patient, and the SUV_max of the positive lesion is 2.29. SUV_max, maximum standard uptake value.

OPEN IN VIEWER

FIG. 2.

The correlation between SUV_max and the different degrees of differentiation. (A) Diagnosis of poorly differentiated and highly differentiated adenocarcinoma by SUV_max. The AUC is 0.907. (B) Diagnosis of signet-ring cell carcinoma and moderately and highly differentiated adenocarcinoma by SUV_max. The AUC is 0.775. AUC, area under the curve; ROC, receiver operating characteristic.

The correlation between SUV_max and the HER-2 expression of gastric cancer

The SUV_max was 3.65 ± 1.17 in the HER-2 positive group and 5.07 ± 2.07 in HER-2 the negative group; the SUV_max was higher in the HER-2 negative group than in the HER-2 positive group (p < 0.01, Fig. 3). The best critical value of SUV_max for predicting the HER-2 status by ROC curve was 5.25 (Fig. 4A), and the AUC was 0.704. When SUV_max was >5.25, the sensitivity of the diagnosis of HER-2 negative was 49.3%, and the specificity was 95.5%.

OPEN IN VIEWER

FIG. 3.

The correlation between the HER-2 status and SUV_max in gastric cancer immunohistochemistry. (A) A pathological picture of an HER-2 negative patient with gastric cancer (DAB, × 200). The corresponding figure (C) is the PET/CT image of the same patient, and the SUV_max is 5.10. (B) A pathological picture of an HER-2 positive patient with gastric cancer (DAB, × 400). (D) The PET/CT image of the same patient, and the SUV_max is 3.90. HER-2, human epidermal growth factor receptor-2.

OPEN IN VIEWER

FIG. 4.

The correlation between SUV_max with the HER-2 status and the Ki-67 expression level. (A) Diagnosis of the HER-2 state by SUV_max. The AUC is 0.704. (B) Diagnosis of the Ki-67 expression level by SUV_max. The AUC is 0.805.

The correlation between SUV_max and the Ki-67 expression of gastric cancer

The SUV_max was 6.00 ± 2.04 in the Ki-67 high expression group and 3.85 ± 1.37 in the Ki-67 low expression group; the SUV_max was higher in the Ki-67 high expression group than in the Ki-67 low expression group (p < 0.01, Fig. 5). In these 164 patients, the total SUV_max was 4.88 ± 2.03, the value-added index of Ki-67 was 0.50 ± 0.15, and there was a significant positive correlation between these two groups (r = 0.70, p < 0.001). The best critical value of SUV_max for predicting Ki-67 status by ROC curve was 6.05 (Fig. 4B), and the AUC was 0.805. When SUV_max was >6.05, the sensitivity of diagnosis of Ki-67 high expression was 61.5%, and the specificity was 96.5%.

OPEN IN VIEWER

FIG. 5.

The correlation between the Ki-67 status and SUV_max in gastric cancer immunohistochemistry. (A) The expression of Ki-67 is 70% in gastric cancer (DAB, × 200). (C) The PET/CT image of the same patient, and the SUV_max is 6.33. (B) The expression of Ki-67 is 20% in gastric cancer (DAB, × 200). The corresponding figure (D) is the PET/CT image of the same patient, and the SUV_max is 4.20.

Discussion

The results of this study revealed that the difference in SUV_max between patients with stage I–II and stage III–IV gastric cancer was not statistically significant. The SUV_max of poorly differentiated adenocarcinoma was significantly higher than that of moderately and highly differentiated adenocarcinoma and signet-ring cell carcinoma. The best critical value of SUV_max for predicting poorly differentiated and moderately and highly differentiated adenocarcinomas by ROC curve was 5.74. The SUV_max in the highly differentiated adenocarcinoma group was higher than that in the signet-ring cell carcinoma group. The best critical value of SUV_max for predicting moderately and highly differentiated adenocarcinomas and signet-ring cell carcinoma by ROC curve was 4.25. The SUV_max in the HER-2 negative group was higher than that in the HER-2 positive group. The best critical value of SUV_max for predicting the HER-2 status by ROC curve was 5.25. The SUV_max was higher in the Ki-67 high expression group than in the Ki-67 low expression group. In these 164 patients, SUV_max was significantly positively correlated with the Ki-67 value-added index. The best critical value of SUV_max for predicting Ki-67 status by ROC curve was 6.05.

Gastric cancer is one of the most common malignant tumors worldwide, but most patients are in the advanced stages and experiencing complications with malignant proliferation by the time they are first diagnosed. Although surgery and chemotherapy can be effective in the treatment of gastric cancer, a previous study revealed that the median overall survival of patients with advanced gastric cancer was <12 months. ³ This is correlated to the low rate of early diagnosis, the lack of effective predictors of biological behavior, and the limited means of treatment for gastric cancer. At present, the selection of treatment depends on indexes such as clinical stage, pathological differentiation type, and HER-2 status. However, some indexes are not available for patients who cannot undergo a biopsy or who have too few biopsy specimens. Therefore, a noninvasive means of predicting the aforementioned indexes is urgently needed. SUV is the most commonly used semiquantitative index of ¹⁸ F-FDG PET/CT in the diagnosis and treatment of tumors, and SUV_max is the highest intake value in the region of interest in tumor lesions. At present, many studies have proved that SUV_max is associated with the prognosis, stage, and certain biological indexes of a number of tumors. ⁴ However, there are few studies on the correlation between SUV_max and the clinical features and biological indicators of gastric cancer in China.

Clinical stage is the most important index for devising treatments and making prognoses for cancer patients. Gastric cancers are divided into stages I–IV according to the size of the tumor, lymph node, and distant metastasis. At each stage, the treatment plans are different, and the 5-year survival rate can also differ widely. Previous studies have shown that SUV_max increases as the clinical stage of gastric cancer progresses from stage I to stage IV, there being a positive correlation between the two. ^5,6 However, another study indicated that there was no correlation between SUV_max and the clinical stage of gastric cancer, ⁷ and this study also demonstrated that there was no difference in SUV_max between stage I–II and III–IV gastric cancer patients (p > 0.05). Therefore, it would appear that the value of SUV_max cannot be used to predict the clinical stage of gastric cancer.

Gastric cancers can be divided into four types in line with the degree of histological differentiation: highly differentiated, moderately differentiated, poorly differentiated, and undifferentiated (signet-ring cell carcinoma was undifferentiated). ⁸ Differentiation is closely correlated to the malignant degree of the gastric cancer. The malignant degree of poorly differentiated adenocarcinoma is higher than that of moderately and highly differentiated adenocarcinoma, whereas the malignant degree of signet-ring cell carcinoma is the highest. Differentiation needs to be fully considered in formulating treatment plans and making prognoses. Furthermore, SUV_max can reflect the malignant degree of tumors. At present, study results concerning whether there is any correlation between the two are not consistent. Chen et al. revealed that ⁷ the SUV_max of poorly differentiated adenocarcinoma was significantly higher than that of moderately and highly differentiated adenocarcinoma and signet-ring cell carcinoma, and the SUV_max in moderately and highly differentiated adenocarcinoma was higher than that in the signet-ring cell carcinoma. Kim et al. ⁹ also reported that the SUV_max of signet-ring cell carcinoma was lower than that of moderately and highly differentiated and poorly differentiated adenocarcinoma. It has also been shown in other research that the difference in SUV_max between groups of gastric cancer patients with different degrees of differentiation was not statistically significant. ^10,11 The results of this study were that the SUV_max of poorly differentiated adenocarcinoma was significantly higher than that of moderately and highly differentiated adenocarcinoma, and signet-ring cell carcinoma (both p < 0.01), and SUV_max in moderately and highly differentiated adenocarcinoma group was higher than that in the signet-ring cell carcinoma group (p < 0.01). The reasons may be that the lower the degree of differentiation, the higher the malignant degree of the tumor, and the more active the proliferation is, the more active the glucose metabolism is and the higher the SUV_max, as it reflects the glucose metabolism. Signet-ring cell carcinoma has a relatively smaller number of cells, low density, and a scattered distribution; there is a lot of mucus in and outside the cells, and the expression of GLUT1 is very low.

In recent years, the rapid development of molecular targeted therapy has offered a new therapeutic strategy for advanced gastric cancer. Among many molecular targets for gastric cancer, HER 2 is one of the most widely used targets with the most definite clinical significance. Trastuzumab is the first globally approved humanized monoclonal antibody targeting HER2 overexpression. For HER-2 positive patients, chemotherapy combined with trastuzumab has become the standard treatment. Previous studies have revealed that HER-2 was positive in 7.3%–20.2% of patients with gastric cancer, ^{12 –14} and the positive rate of HER-2 in Chinese gastric cancer patients was 12%–13%. ^15,16 A previous retrospective study revealed that the positive expression of HER-2 was correlated to older age, males, intestinal type of histological Lauren classification, and the tumor being in the upper third of the stomach. However, whether it is correlated to SUV_max, which can reflect the tumor's metabolic capacity, has not been confirmed. Previous studies also showed that ^{17 –19} the positive expression rate of HER-2 in moderately and highly differentiated adenocarcinoma was higher than that in poorly differentiated adenocarcinoma and signet-ring cell carcinoma, whereas the difference in the positive rate of HER-2 expression between poorly differentiated adenocarcinoma and signet-ring cell carcinoma was not statistically significant. Chen et al. revealed that ⁷ there was no significant correlation between SUV_max and HER2 expression in gastric cancer. However, when signet-ring cell carcinoma was excluded, the SUV_max was significantly higher in the HER-2 negative group than in the HER-2 positive group. The results of this study indicated that the SUV_max was significantly higher in the HER-2 negative group than in the HER-2 positive group (p < 0.01). The inconsistency in the research results to date suggests that the mechanism remains unclear and needs further study.

Ki-67 is a nuclear antigen found in proliferating cells, expressed in the G1, S, G2, and M phases of the cell cycle. The expression level is highest in the M phase, but not expressed in the G0 phase. The Ki-67 index reflects the proportion of cells entering the proliferative cycle and, to some extent, the proliferative activity of tumor cells. A study concerning breast cancer reported that ²⁰ SUV_max was moderately correlated with the Ki-67 expression but could not yet be used as a surrogate marker for tumor proliferation. The results of a study into ovarian cancer had similar results to those of breast cancer. ²¹ A meta-analysis involving 5600 patients with gastric cancer in 29 studies suggested that ²² a high expression of Ki-67 was significantly associated with Lauren's typing and tumor size but not correlated with lymph node metastasis, tumor stage, and tumor differentiation. However, there are few studies on whether there is a correlation between SUV_max and the Ki-67 index. Deng et al. found that ²³ in gastric cancer, SUV_max was weakly correlated with the Ki-67 index. The results of this study revealed that SUV_max was significantly higher in the Ki-67 high expression group than in the Ki-67 low expression group (p < 0.01), and there was a significant positive correlation between these two groups (r = 0.70). In principle, both expressions can reflect the degree of proliferation of tumor cells, and a positive correlation can be seen.

This study has the following limitations. First, it is a case–control study, not a randomized controlled trial, and it is not blind. Therefore, there is still a certain risk of bias. Second, this study is a single-center clinical trial; the included sample size is small, and multicenter clinical trials with larger sample sizes are still needed. Finally, the correlation between PET/CT SUV_max, the status of HER-2, and the Ki-67 index in gastric cancer patients still require further study.

Conclusion

¹⁸ F-FDG PET/CT SUV_max can, to some extent, predict the degree of pathological type, HER-2 status, and Ki-67 index of gastric cancer patients, and so it can be used as a noninvasive predictor of the biological behavior of patients with gastric cancer, thereby, contributing to clinical diagnosis and treatment.