Gastric cancer (GC) is the fifth most common cancer and the third leading cause of cancer-related mortality worldwide. Inflammation and the nutritional status of patients with GC are important factors affecting the therapeutic effect and prognosis. Inflammatory and nutrition-related markers have been shown to be prognostic factors for patients with GC. However, few studies have investigated the relationship of the prealbumin-to-globulin ratio (PGR) with the prognosis of GC patients. The objective of the present study was to examine whether pretreatment PGR is related to the prognosis and chemotherapy outcomes of in-patients with advanced GC undergoing first-line chemotherapy. We retrospectively reviewed the data of 281 patients with unresectable GC from January 2013 to January 2018. The receiver operating characteristic curve analysis determined the cut-off values for the PGR. The relationship between the PGR and chemotherapy effectiveness was evaluated using the chi-square test. Kaplan-Meier’s method was used to plot progression-free survival (PFS) and overall survival (OS) curves, using multivariable Cox regression analysis to identify promising predictors of mortality. The cut-off value for the PGR was 7.21. The high-PGR (≥7.21) group had a higher disease control rate than that of the low-PGR group (93.66% vs. 78.42%,
Although gastric cancer (GC) is the fifth most common cancer, it is the third leading cause of cancer-related mortality worldwide [
The therapeutic effect and prognosis of patients with cancer may be related to their inflammatory and nutritional status. Proinflammatory chemokines and cytokines participate in the occurrence, development, recurrence, and metastasis of tumors, and reduce the effectiveness of cytotoxic drugs, ultimately affecting the prognosis [
Albumin and globulin alone can develop into indicators of the patients’ prognosis. However, albumin and globulin test results are affected by various factors, including physiological and pathological changes. Studies have combined albumin and globulin as predictors to minimize the impact of the inaccuracy of an isolated value on the result. Numerous studies have reported a relationship between the pretreatment AGR and the treatment response in patients with GC [
The main objective of this study was to evaluate whether the pretreatment PGR could develop into a marker of chemotherapy response and a prognostic marker for the progression-free survival (PFS) and OS in patients with unresectable GC receiving first-line chemotherapy. We also compared the PGR with the AGR to determine which is a better prognostic indicator. Thus, this study investigated two potential prognostic factors, the PGR and AGR.
We retrospectively reviewed the records of patients with nonresectable GC evaluated from January 2013 to January 2018 at the Cancer Hospital of the China Medical University in Shenyang City, Liaoning Province, China.
The following were the enrollment criteria: (1) chemotherapy-naive, stage III-IV unresectable gastric adenocarcinoma diagnosed by biopsy and imaging; (2) an Eastern Cooperative Oncology Group (ECOG) performance status score of 0–2; (3) no severe complications that may hinder patients from receiving chemotherapy; (4) completed at least two cycles of chemotherapy; and (5) completed all required blood tests within 1 week before the initial first-line chemotherapy. We excluded patients with kidney or liver diseases other than tumors and those with severe chronic or acute inflammatory diseases.
Radiographic assessments were performed at baseline and every 6–8 weeks until disease progression. After the failure of first-line chemotherapy, the patients were followed up every 8–12 weeks until they died. Variations of approximately 1 week were regarded as a permissible error.
All enrolled patients volunteered to participate in the study, which was approved by the Ethics Committee of the Cancer Hospital of China Medical University.
We collected the following data from the hospital records: the patients’ characteristics before first-line chemotherapy, the location of the tumor in the stomach as determined by gastroscopy, and computed tomography (CT)-, magnetic resonance imaging (MRI)-, and positron emission tomography (PET)/CT-determined metastatic lymph nodes and organs. Peripheral blood analysis and albumin, prealbumin, globulin, hemoglobin, carcinoembryonic antigen (CEA), carbohydrate antigen 72-4 (CA72-4), and carbohydrate antigen 19-9 (CA19-9) levels were obtained within 1 week of treatment initiation. The reference ranges for plasma albumin, prealbumin, globulin, and hemoglobin were 35–50 g/L, 160–450 mg/L, 20–35 g/L, and 115–155 g/L, respectively. The accepted normal ranges of CEA, CA72-4, and CA19-9 were 0–5 ng/ml, 0–6 U/ml, and 0–37 U/ml, respectively. The PGR and AGR were defined as prealbumin (mg/L)/globulin (g/L) ratio and albumin (g/L)/globulin (g/L) ratio, respectively.
All patients received different systemic first-line chemotherapy regimens depending on their ECOG score. Patients were evaluated for therapy response after every two cycles of chemotherapy. We adopted the Response Evaluation Criteria in Solid Tumors (RECIST 1.1) to assess the patients’ chemotherapy response, as follows: complete remission (CR), partial response (PR), stable disease (SD), and disease progression. The objective response rate (ORR) was calculated as
The Statistical Package for Social Sciences version 25 software (SPSS Inc., Chicago, IL, USA) was used for data analysis. We used the cut-off values of the PGR and AGR determined by a receiver operating characteristic curve analysis to divide all patients with unresectable GC into high-value and low-value groups for each factor. Fisher’s exact test and the chi-square test were used to evaluate the relationship between the PGR or AGR and the patients’ characteristics and therapeutic response. Kaplan-Meier’s method was used to plot the PFS and OS curves, and the PFS and OS were analyzed using the log-rank test. Pearson’s correlation analysis was employed to assess the linear correlation between the prognostic factors and the PFS and OS. Univariate and multivariate analyses were performed using a Cox proportional hazards model to identify potential predictors of mortality. We verified that the proportional hazards assumption was met [
The median PFS of 151 days was the state variable, the PGR and AGR were used as the test variables, and then, the cut-off values of the above two prognostic components were determined. We found that a PGR of 7.205 and an AGR of 1.455 were the strongest prognostic factors for the PFS. The area under the curve (AUC) values for the PGR and AGR were 0.660 (95% confidence interval [CI]: 0.597–0.723,
Receiver operating characteristic curve analysis of the prealbumin-to-globulin ratio (PGR) and the albumin-to-globulin ratio (AGR) in patients with unresectable gastric cancer.
A total of 281 patients with advanced GC receiving first-line chemotherapy were included in the analysis. The majority of patients were men (66.55%), and the median age of the patients was 60 years. Among all, 74.38% and 82.92% of patients had a normal body mass index and an ECOG score less than 2 points, respectively. Patients with poorly cohesive or mucinous adenocarcinoma (74.02%) and stage IV disease (80.43%) represented the highest proportions among all patients. The number of patients with more than one organ metastases was the same as that with peritoneal metastases, accounting for 32.38% of the total. SOX (oxaliplatin + S1)/CapeOX (oxaliplatin + capecitabine), FOLFOX (oxaliplatin + leucovorin +5-fluorouracil), and DCF (docetaxel + cisplatin +5-fluorouracil)/DOF (docetaxel + oxaliplatin +5-fluorouracil) were the primary chemotherapy regimens for patients receiving first-line chemotherapy, accounting for 50.53%, 17.44%, and 10.68% of the total, respectively. Based on the RECIST1.1 criteria, the ORR and DCR were 13.17% and 86.12%, respectively. Table
Baseline characteristics of the patients.
Characteristics | Cases ( | Percentage |
---|---|---|
Sex | ||
Male | 187 | 66.55% |
Female | 94 | 33.45% |
Age (yr) | ||
≤60 | 159 | 56.58% |
>60 | 122 | 43.42% |
Body mass index (kg/m2) | ||
<18.5 or >25 | 72 | 25.62% |
18.5–25 | 209 | 74.38% |
ECOG | ||
0–1 | 233 | 82.92% |
≥2 | 48 | 17.08% |
Histological type | ||
Papillary, tubular, | 73 | 25.98% |
Poorly cohesive, mucinous | 208 | 74.02% |
TNM stage | ||
III | 55 | 19.57% |
IV | 226 | 80.43% |
Number of organs affected by metastasis | ||
0–1 | 190 | 67.62% |
≥2 | 91 | 32.38% |
Peritoneal metastasis | ||
Yes | 91 | 32.38% |
No | 190 | 67.62% |
First-line chemotherapy regimen | ||
SOX/CapeOX | 142 | 50.53% |
FOLFOX | 49 | 17.44% |
DCF/DOF | 30 | 10.68% |
Capecitabine/S-1 | 22 | 7.83% |
Others | 38 | 13.52% |
Best response | ||
Complete remission | 0 | — |
Partial response | 37 | 13.17% |
Stable disease | 205 | 72.95% |
Progression of disease | 39 | 13.88% |
Objective response rate | 37 | 13.17% |
Disease control rate | 242 | 86.12% |
Tumor biomarkers | ||
CEA >5 ng/mL | 121 | 43.06% |
CA72-4 >6 U/mL | 158 | 56.23% |
CA19-9 >37 U/mL | 106 | 37.72% |
Hemoglobin (g/L) | ||
<115 | 105 | 37.37% |
≥115 | 176 | 62.63% |
Albumin (g/L) | ||
<35 | 52 | 18.51% |
≥35 | 229 | 81.49% |
Prealbumin (mg/L) | ||
<160 | 83 | 29.54% |
≥160 | 198 | 70.46% |
Globulin (g/L) | ||
≤35 | 259 | 92.17% |
>35 | 22 | 7.83% |
PGR | ||
<7.21 | 139 | 49.47% |
≥7.21 | 142 | 50.53% |
AGR | ||
<1.46 | 128 | 45.55% |
≥1.46 | 153 | 54.45% |
ECOG: Eastern Cooperative Oncology Group; TNM: tumor-node-metastasis; SOX: oxaliplatin + S1; CapeOX: oxaliplatin + capecitabine; FOLFOX: oxaliplatin + leucovorin + 5-fluorouracil; DCF: docetaxel + cisplatin + 5-fluorouracil; DOF: docetaxel + oxaliplatin + 5-fluorouracil; CEA: carcinoembryonic antigen; CA72-4: carbonhydrate antigen 72-4; CA 19-9: carbonhydrate antigen 19-9; PGR: prealbumin-to-globulin ratio; AGR: albumin-to-globulin ratio.
There were some differences in terms of the histopathological and clinical parameters between the low and high groups. The analysis indicated that a low PGR (
Relationship between the pretreatment PGR and AGR and clinicopathological factors.
PGR | AGR | |||||
---|---|---|---|---|---|---|
Low | High | Low | High | |||
Sex | ||||||
Male | 95 | 92 | 0.528 | 86 | 101 | 0.835 |
Female | 44 | 50 | 42 | 52 | ||
Age (yr) | ||||||
>60 | 63 | 59 | 0.523 | 61 | 61 | 0.190 |
≤60 | 76 | 83 | 67 | 92 | ||
Body mass index (kg/m2) | ||||||
<18.5 or >25 | 40 | 32 | 0.231 | 30 | 42 | 0.443 |
18.5–25 | 99 | 110 | 98 | 111 | ||
ECOG | ||||||
0–1 | 113 | 120 | 0.474 | 104 | 129 | 0.497 |
≥2 | 26 | 22 | 24 | 24 | ||
Histological type | ||||||
Papillary, tubular | 35 | 38 | 0.763 | 27 | 46 | 0.088 |
Poorly cohesive, mucinous | 104 | 104 | 101 | 107 | ||
TNM stage | ||||||
III | 26 | 29 | 0.717 | 21 | 34 | 0.221 |
IV | 113 | 113 | 107 | 119 | ||
Number of organs affected by metastasis | ||||||
0–1 | 90 | 100 | 0.309 | 78 | 112 | 0.029 |
≥2 | 49 | 42 | 50 | 41 | ||
Peritoneal metastasis | ||||||
Yes | 52 | 39 | 0.075 | 50 | 41 | 0.029 |
No | 87 | 103 | 78 | 112 | ||
CEA (ng/mL) | ||||||
>5 | 67 | 54 | 0.085 | 56 | 65 | 0.831 |
≤5 | 72 | 88 | 72 | 88 | ||
CA72-4 (U/mL) | ||||||
>6 | 79 | 79 | 0.839 | 76 | 82 | 0.331 |
≤6 | 60 | 63 | 52 | 71 | ||
CA19-9 (U/mL) | ||||||
>37 | 59 | 47 | 0.106 | 52 | 54 | 0.359 |
≤37 | 80 | 95 | 76 | 99 | ||
Hemoglobin (g/L) | ||||||
<115 | 63 | 42 | 0.006 | 61 | 44 | 0.001 |
≥115 | 76 | 100 | 67 | 109 |
PGR: prealbumin-to-globulin ratio; AGR: albumin-to-globulin ratio; ECOG: Eastern Cooperative Oncology Group; TNM: tumor-node-metastasis; CEA: carcinoembryonic antigen; CA72-4: carbonhydrate antigen 72-4; CA 19-9: carbohydrate antigen 19-9.
We evaluated the relationship between the PGR and AGR levels before receiving first-line chemotherapy and the treatment response. Patients in the high-PGR (93.66% vs. 78.42%,
Treatment response to first-line chemotherapy according to the pretreatment PGR and AGR.
Response | PGR | AGR | ||||
---|---|---|---|---|---|---|
High ( | Low ( | High ( | Low ( | |||
Complete response | 0 | 0 | 0 | 0 | ||
Partial response | 22 | 15 | 26 | 11 | ||
Stable disease | 111 | 94 | 118 | 87 | ||
Progressive disease | 9 | 30 | 9 | 30 | ||
Objective response rate | 15.49% | 10.79% | 0.244 | 16.99% | 8.59% | 0.038 |
Disease control rate | 93.66% | 78.42% | <0.001 | 94.12% | 76.56% | <0.001 |
PGR: prealbumin to globulin ratio; AGR: albumin to globulin ratio.
Figures
Kaplan-Meier’s curves for progression-free survival of the PGR (a) and AGR (b) groups. Log of the negative log of the survival curve for progression-free survival of the PGR (c) and AGR (d) groups. PGR: prealbumin-to-globulin ratio; AGR: albumin-to-globulin ratio.
Kaplan-Meier’s curve for overall survival of the PGR (a) and AGR (b) groups. Log of the negative log of the survival curve for overall survival of the PGR (c) and AGR (d) groups. PGR: prealbumin-to-globulin ratio; AGR: albumin-to-globulin ratio.
Pearson’s correlation analysis indicated that the PGR and AGR values were positively correlated with the PFS and OS in patients with GC (all
Pearson’s correlation s analysis for the linear relationship between the PGR group and progression-free survival (a) and overall survival (b); Pearson’s correlation s analysis for the linear relationship between the AGR group and progression-free survival (c) and overall survival (d). PGR: Prealbumin-to-globulin ratio; AGR: albumin-to-globulin ratio.
The univariate and multivariate analyses estimated the factors affecting the PFS and OS (Tables
Correlations between survival and the PGR and other clinicopathological factors.
Progression-free survival | Overall survival | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Univariate analysis | Multivariate analysis | Univariate analysis | Multivariate analysis | |||||||||
Hazard ratio | 95% CI | Hazard ratio | 95% CI | Hazard ratio | 95% CI | Hazard ratio | 95% CI | |||||
Sex (male) | 1.073 | 0.835-1.379 | 0.582 | 1.145 | 0.891-1.471 | 0.290 | ||||||
Age (>60 years) | 0.991 | 0.781-1.256 | 0.939 | 0.797 | 0.627-1.012 | 0.062 | 0.874 | 0.674-1.133 | 0.309 | |||
Body mass index (<18.5 or >25 kg/m2) | 0.923 | 0.705-1.207 | 0.558 | 1.084 | 0.828-1.420 | 0.557 | ||||||
ECOG (≥2) | 1.124 | 0.823-1.535 | 0.463 | 1.174 | 0.859-1.604 | 0.313 | ||||||
Histological type (poorly cohesive, mucinous) | 1.492 | 1.136-1.959 | 0.004 | 1.495 | 1.137-1.965 | 0.004 | 1.537 | 1.172-2.015 | 0.002 | 1.407 | 1.067-1.855 | 0.015 |
TNM stage (IV) | 1.206 | 0.897-1.622 | 0.214 | 1.368 | 1.017-1.839 | 0.038 | 1.186 | 0.855-1.643 | 0.306 | |||
Number of organs affected by metastasis (≥2) | 1.390 | 1.081-1.786 | 0.010 | 1.400 | 1.077-1.818 | 0.012 | 1.364 | 1.058-1.758 | 0.017 | 1.294 | 0.973-1.721 | 0.077 |
Peritoneal metastasis (yes) | 1.526 | 1.185-1.966 | <0.001 | 1.297 | 0.991-1.696 | 0.058 | 1.710 | 1.325-2.205 | <0.001 | 1.426 | 1.062-1.915 | 0.018 |
CEA (>5 ng/mL) | 1.222 | 0.959-1.557 | 0.105 | 1.138 | 0.878-1.474 | 0.328 | 1.278 | 1.004-1.626 | 0.046 | 1.299 | 1.004-1.680 | 0.046 |
CA72-4 (>6 U/mL) | 1.425 | 1.118-1.816 | 0.004 | 1.411 | 1.091-1.824 | 0.009 | 1.254 | 0.989-1.590 | 0.061 | 1.164 | 0.903-1.501 | 0.240 |
CA19-9 (>37 U/mL) | 1.051 | 0.825-1.339 | 0.686 | 1.046 | 0.821-1.332 | 0.717 | ||||||
Hemoglobin (<115 g/L) | 0.905 | 0.708-1.157 | 0.424 | 0.890 | 0.697-1.136 | 0.349 | ||||||
PGR ≥7.21 | 0.644 | 0.508-0.817 | <0.001 | 0.672 | 0.527-0.857 | 0.001 | 0.644 | 0.507-0.817 | <0.001 | 0.675 | 0.530-0.861 | 0.002 |
ECOG: Eastern Cooperative Oncology Group; TNM: tumor-node-metastasis; CEA: carcinoembryonic antigen; CA72-4: carbonhydrate antigen 72-4; CA 19-9: carbohydrate antigen 19-9; PGR: prealbumin-to-globulin ratio.
Correlations between survival and AGR and other clinicopathological factors.
Progression-free survival | Overall survival | |||||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Univariate analysis | Multivariate analysis | Univariate analysis | Multivariate analysis | |||||||||
Hazard ratio | 95% CI | Hazard ratio | 95% CI | Hazard ratio | 95% CI | Hazard ratio | 95% CI | |||||
Sex (male) | 1.073 | 0.835-1.379 | 0.582 | 1.145 | 0.891-1.471 | 0.290 | ||||||
Age (>60 years) | 0.991 | 0.781-1.256 | 0.939 | 0.797 | 0.627-1.012 | 0.062 | 0.877 | 0.678-1.135 | 0.320 | |||
Body mass index (<18.5 or >25 kg/m2) | 0.923 | 0.705-1.207 | 0.558 | 1.084 | 0.828-1.420 | 0.557 | ||||||
ECOG (≥2) | 1.124 | 0.823-1.535 | 0.463 | 1.174 | 0.859-1.604 | 0.313 | ||||||
Histological type (poorly cohesive, mucinous) | 1.492 | 1.136-1.959 | 0.004 | 1.355 | 1.028-1.787 | 0.031 | 1.537 | 1.172-2.015 | 0.002 | 1.352 | 1.023-1.787 | 0.034 |
TNM stage (IV) | 1.206 | 0.897-1.622 | 0.214 | 1.368 | 1.017-1.839 | 0.038 | 1.182 | 0.852-1.640 | 0.318 | |||
Number of organs affected by metastasis (≥2) | 1.390 | 1.081-1.786 | 0.010 | 1.341 | 1.030-1.746 | 0.029 | 1.364 | 1.058-1.758 | 0.017 | 1.299 | 0.977-1.728 | 0.072 |
Peritoneal metastasis (yes) | 1.526 | 1.185-1.966 | <0.001 | 1.239 | 0.947-1.621 | 0.118 | 1.710 | 1.325-2.205 | <0.001 | 1.419 | 1.056-1.907 | 0.020 |
CEA (>5 ng/mL) | 1.222 | 0.959-1.557 | 0.105 | 1.146 | 0.883-1.487 | 0.305 | 1.278 | 1.004-1.626 | 0.046 | 1.381 | 1.066-1.787 | 0.014 |
CA72-4 (>6 U/mL) | 1.425 | 1.118-1.816 | 0.004 | 1.418 | 1.093-1.840 | 0.009 | 1.254 | 0.989-1.590 | 0.061 | 1.150 | 0.891-1.485 | 0.284 |
CA19-9 (>37 U/mL) | 1.051 | 0.825-1.339 | 0.686 | 1.046 | 0.821-1.332 | 0.717 | ||||||
Hemoglobin (<115 g/L) | 0.905 | 0.708-1.157 | 0.424 | 0.890 | 0.697-1.136 | 0.349 | ||||||
AGR ≥1.46 | 0.419 | 0.328-0.535 | <0.001 | 0.454 | 0.353-0.583 | <0.001 | 0.584 | 0.460-0.742 | <0.001 | 0.602 | 0.472-0.769 | <0.001 |
ECOG: Eastern Cooperative Oncology Group; TNM: tumor-node-metastasis; CEA: carcinoembryonic antigen; CA72-4: carbonhydrate antigen 72-4; CA 19-9: carbohydrate antigen 19-9; AGR: albumin-to-globulin ratio.
Systemic inflammation and malnutrition are prevailing in patients with cancer. These factors have a significant impact on the patients’ quality of life, treatment outcomes, diagnosis, and durability, and may increase the risk of longer duration of hospitalization, infection, treatment toxicity, and treatment costs [
Multiple studies have investigated a series of inflammatory and nutritional markers available to doctors, including the PGR, AGR, FAR, PLR, NLR, and LMR, for the prognosis of patients with different types of cancer [
In this study, we explored the prognostic and predictive role of the PGR in patients with GC receiving first-line chemotherapy. First, we found that a low pretreatment PGR was strongly associated with a hemoglobin level of <115 g/L. A previous study has shown that patients with GC who have anemia have a poor prognosis [
First, inflammatory cytokines released by inflammatory cells can cultivate an inflammatory microenvironment, which provides a hotbed for tumor growth. Therefore, chronic inflammation is a contributor to tumor cell growth, angiogenesis, development, progression, recurrence, and metastasis [
Second, malnutrition relates intimately to a decline in the quality of life, reducing the adherence to treatment and response to therapy. As malnutrition is one of the leading causes for immunodeficiency, the nutritional status can be used to comparatively quickly assess the host’s immune status [
Finally, globulin and prealbumin are two potentially valued elements related to the prognosis of patients with cancer. However, these parameters alone are not sufficient to predict the patients’ prognosis. The globulin level is an inversely related prognostic index, while the prealbumin level is a positively related prognostic index. Integrating indicators into the PGR has improved accuracy. By calculating the ratio of globulin to prealbumin, lower values are accompanied by inflammation or/and malnutrition, which may be related to a poor prognosis. On the contrary, a higher value indicates no inflammation or/and great nutritional status, which can be associated with a favorable prognosis. Therefore, compared with globulin and prealbumin alone, the PGR probably has a better ability to distinguish patients with favorable prognosis from those with unfavorable prognosis.
This study has several strengths. We demonstrated for the first time that the PGR is a significant indicator that can independently predict the PFS and OS in patients with GC receiving first-line chemotherapy. This study was the first to use the PGR as a forecasting element of chemotherapeutic response for patients with GC.
There are also limitations to this study. First, the PGR was found not to be superior to the AGR as a prognostic indicator in this study. This might be because patients with advanced GC have gone through protracted disease history, prolonging chronic malnutrition and inflammation in these patients. Therefore, we speculate that the PGR with a shorter half-life of 2–3 days is superior to the AGR with a half-life of 21 days as a prognostic marker in early-stage patients, but further research is needed to validate this hypothesis. Second, the present was a single-center retrospective study. Hence, the usefulness of the PGR still needs to be verified by multicenter large-scale prospective studies.
The prechemotherapy PGR level can predict the chemotherapy effect in patients with advanced GC and is an independent predictor of the PFS and OS. It may be a reliable prognostic tool for medical practitioners.
We declared that materials described in the manuscript, including all relevant raw data, will be freely available to any scientist wishing to use them for noncommercial purposes, without breaching participant confidentiality.
The authors declare that there is no conflict of interest regarding the publication of this paper.
Zhuo Wang and Liqun Zhang are cofirst authors of the article.
This study was supported by the National Key R&D Program of China (grant# 2018YFC1311600), the Scientific Research Foundation For The Introduction Of Talents, Liaoning Cancer Hospital & Institute (grant# Z1702), the Science and Technology Planning Project of Liaoning Province of China (grant# 201800449), the Science and Technology Planning Project of Shenyang (grant# 191124088), and the Science and Technology Planning Project of Shenyang (grant# F15-139-9-27).