Abstract
Background/Aim: The Geriatric Nutritional Risk Index (GNRI) predicts prognosis in various cancers. This study examined the correlation between GNRI, complete adjuvant chemotherapy (AC), and prognosis in patients with resected pancreatic cancer. Patients and Methods: We retrospectively evaluated 123 patients with pancreatic cancer who underwent pancreatectomies at our institute between January 2010 and December 2020. Kaplan–Meier and Cox regression methods were used to assess survival. Factors associated with complete AC were identified using logistic regression analysis. Results: Among the 123 patients with pancreatic cancer, 93 (75.6%) initiated AC, and 55 (44.7%) completed AC. In multivariate analysis, the pre-operative GNRI was an independent prognostic factor for overall survival (OS) [hazard ratio=1.63, 95% confidence interval (CI)=1.01-2.63; p=0.046]. Additionally, pre-operative GNRI was an independent predictor of complete AC (odds ratio=0.38, 95%CI=0.17-0.83; p=0.015). In the high (≥98) and low GNRI (<98) groups, patients who underwent complete AC had significantly longer OS than those who did not (p<0.001, respectively). However, the patients in the low GNRI group who underwent complete AC had no significant difference in OS compared to those in the high GNRI group who did not undergo complete AC (p=0.523). Conclusion: Pre-operative GNRI may predict complete AC and prognosis in resected pancreatic cancer.
- Pancreatic cancer
- geriatric nutritional risk index
- complete adjuvant chemotherapy
- prognosis
- pancreatectomy
Recent advancements have been made in diagnostic and multidisciplinary therapies; nonetheless, pancreatic cancer (PC) prognosis remains extremely poor (1). Radical surgery is the best method for prolonging the survival of patients with PC, with post-operative adjuvant chemotherapy (AC) as the standard treatment to enhance the overall survival (OS) rate (2). However, randomized controlled trials have revealed completion rates of 54-79%, and single-institution data reported lower rates of 40-60% (3). Therefore, increasing the completion rate of AC is essential to improve the prognosis of patients with resected PC. Moreover, the factors predicting AC completion in patients with resected PC and strategies to address these factors urgently need to be explored.
Nutritional status significantly influences patient quality of life, treatment intensity, and disease outcomes early in the treatment course for various cancers (4). Nutritional assessment traditionally relies on body composition, blood biochemical data, imaging techniques, measuring instruments, or a combination of these methods. The geriatric nutritional risk index (GNRI) is a nutritional screening tool used to assess body weight, height, and serum albumin levels. The GNRI was originally designed to evaluate mortality and morbidity risks in older hospitalized patients (5), and has recently been reported as a reliable prognosis predictor in older and younger adult patients with cancer, especially in gastrointestinal cancers, including esophageal (6), gastric (7), colorectal (8), pancreatic (9), and hepatocellular carcinomas (10). Additionally, GNRI has been a useful predictor of post-operative complications in patients with PC (11). Nevertheless, the relationship between GNRI and AC in patients with resected PC remains unclear.
Therefore, this study aimed to elucidate the association between pre-operative GNRI, complete AC, and prognosis in patients with resected PC.
Patients and Methods
Patient selection. This retrospective study included patients who underwent pancreatectomy for PC at a single tertiary care center–Gifu University Hospital, Gifu, Japan—between January 2010 and December 2020. Patients who were lost to follow-up, diagnosed at Stage IV on the final pathology of frozen section specimens, received chemoradiotherapy or gemcitabine (GEM) + nab-paclitaxel post-operatively, and those who died owing to surgery-related complications were excluded (Figure 1). All the patients in this study were of Japanese ethnicity and provided informed consent using an opt-out option before enrollment in the study. This study was conducted per the human and ethical principles of the Declaration of Helsinki, and the study protocol was approved by the Institutional Review Board of Gifu University Hospital (approval number: 2023-140).
Flow chart of patient selection.
Treatment. Enrolled patients were radiologically diagnosed with resectable or borderline resectable PC per the National Comprehensive Cancer Network Guidelines (2). Twenty-one patients with borderline resectable PC received neoadjuvant chemotherapy with GEM plus nab-paclitaxel or chemoradiotherapy. Twelve patients with resectable PC received neoadjuvant chemotherapy with GEM and S-1 in 2019 following the results of the Prep-02/JSAP05 trial (12). Surgical procedures were performed according to institutional policies and institutional cancer board recommendations.
AC was post-operatively administered after the patient had fully recovered from surgery. Regarding the AC regimen, patients received GEM or S-1 based on the results of the CONKO-001 (13) and JASPAC01 trials (14), respectively. During 2010-2013, patients were mainly treated with intravenous GEM infusions (1,000 mg/m2; days 1, 8, and 15, every 4 weeks for 6 cycles). Subsequently, S-1 was administered orally at 80, 100, and 120 mg/day for 4 weeks based on body surface areas of <1.25 m2, between ≥1.25 m2 and <1.5 m2, and ≥1.5 m2, respectively, followed by 2 weeks of rest for 4 cycles. In instances of adverse reactions, adjustments were made by reducing the dose, modifying the dosing interval, or temporarily discontinuing the administration.
Complete AC was defined as the point at which a planned cycle of chemotherapy was reached: GEM, six cycles; S-1, four cycles. However, heterogeneity was observed among the treatment regimens over the study duration. Patients who experienced relapse during AC comprised the incomplete AC group.
Evaluation of the clinicopathological features. We investigated the demographic and clinical variables, including age, sex, body mass index, American Society of Anesthesiologists physical status, tumor location, tumor size, neoadjuvant therapy regimens, carbohydrate antigen 19-9 (CA19-9) level, neutrophil-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte ratio (LMR), and GNRI. These variables were measured pre-operatively. Tumor size was determined using multidetector computed tomography at the initial diagnosis. The treatment factors included information on the treatment or complications after pancreatectomy. Post-operative complications were defined and classified according to the Clavien–Dindo classification (15). Post-operative pancreatic fistulas were graded according to the definition of the International Study Group on Pancreatic Fistula (16). We collected all information regarding the tumor from the pathological examination. The pathological stage of all tumor specimens was determined according to the TNM staging system of the UICC, 8th edition (17).
GNRI definition. GNRI was calculated using body weight, height, and serum albumin level. The pre-operatively data were obtained as follows: GNRI=[14.89 × serum albumin (g/l)] + [41.7 × present/ideal body weight (kg)] (5). The ideal body weight was defined as 22× patient height (m)2. If the current body weight was higher than the ideal body weight, the present/ideal body weight was set to 1.
Statistical analysis. Patient characteristics were expressed as frequencies and percentages for categorical variables and medians with interquartile ranges or means with standard deviations for continuous variables. The chi-square test was performed for categorical variables, and the Student’s t-test or Mann–Whitney U-test was used to compare continuous variables to evaluate differences between the two groups. Receiver operating characteristic (ROC) analysis determined the cutoff values for CA19-9, NLR, LMR, GNRI, and length of hospital stay. In addition, cutoff values were determined using the Youden Index. Univariate and multivariate analyses of patient survival were performed using Cox regression analysis to identify prognostic factors for patient survival. Survival curves were estimated using the Kaplan–Meier method. Survival time was calculated from the surgery date. Logistic regression analysis was used to identify factors independently associated with AC completion. A p-value of <0.05 indicated statistical significance. Statistical analyses were performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan) (18), which is a graphical user interface for R (R Foundation for Statistical Computing, Vienna, Austria), EZR is a modified version of the R commander designed to add statistical functions frequently used in biostatistics.
Results
Patient characteristics. We retrospectively analyzed 123 consecutive patients who underwent pancreatectomy for PC between January 2010 and December 2020. Table I summarizes the patient characteristics; 46 patients (37.4%) were females, with a median age of 70. Eighty (65.0%) patients had pancreatic head cancer. The resectability status was borderline in 26 (21.1%) patients. The median CA19-9 level was 104 U/ml, median GNRI was 98, R1 (microscopic residual tumor) resection was achieved in 25 patients (20.3%), and 29 patients (23.6%) experienced post-operative complications (Clavien–Dindo grade ≥III). Of the cohort, 93 (75.6%) patients initiated AC, and 30 (24.4%) did not. Among the 93 patients who initiated AC, 55 (44.7%) underwent complete AC, and 38 (30.9%) were unable to complete AC. Recurrence was confirmed in 89 (72.4%) patients.
Patient characteristics.
Factors associated with overall survival. Univariate and multivariate Cox regression analyses were used to evaluate factors in the overall cohort (Table II). Multivariate analysis revealed that CA19-9 [hazard ratio (HR)=1.83, 95% confidence interval (CI)=1.11-3.01; p=0.018], LMR (HR=1.95, 95%CI=1.10-3.46; p=0.022), GNRI (HR=1.63, 95%CI=1.01-2.63; p=0.046), and incomplete AC (HR=3.28, 95%CI=2.00-5.39; p<0.001) were independent prognostic factors for OS.
Univariate and multivariate analyses of variables for overall survival in patients with resected pancreatic cancer.
Association with overall survival. We evaluated the association between GNRI or AC and OS. Median survival time (MST) was 56.4 months in the high GNRI (≥98) group and 20.1 months in the low GNRI (<98) group (p<0.001, Figure 2A). MST was significantly higher in the complete AC group (56.4 months) than in the incomplete AC group (19.8 months; p<0.001, Figure 2B). No significant difference in MST was observed between the incomplete and the no-AC groups (14.5 and 13.0 months, respectively) (p=0.999, Figure 2C).
Overall survival (OS) curves of patients with pancreatic cancer. (A) OS curves of patients stratified using pre-operative Geriatric Nutritional Risk Index. (B) OS curves of patients who underwent complete adjuvant chemotherapy (AC) and those who did not. (C) OS curves of patients who underwent complete adjuvant AC, those who underwent incomplete AC, and those who did not undergo AC.
Factors associated with complete adjuvant chemotherapy. We divided the patients into two groups based on complete or incomplete AC and compared their clinicopathological features (Table III). Univariate and multivariate analyses revealed age and low GNRI as independent factors associated with complete AC [odds ratio (OR)=0.26, 95%CI=0.12-0.56; p<0.001 and OR=0.38, 95%CI=0.17-0.83; p=0.015, respectively].
Univariate and multivariate analyses for predictors of complete adjuvant chemotherapy in patients with resected pancreatic cancer.
Association between patients’ clinicopathological factors and GNRI. Low GNRI was significantly associated with tumor location, pre-operative LMR, operation, and incomplete AC (Table IV). We compared the entire cohort stratified by GNRI and AC completion into four groups (Figure 3). In the high- and low-GNRI groups, patients who received complete AC had a significantly longer OS than those who did not (p<0.001). However, the patients in the low GNRI group who underwent complete AC (median OS: 39.3 months) had no significant difference in OS compared to those in the high GNRI group who underwent complete AC (median OS: 73.3 months; p=0.072) and those in the high GNRI group who did not undergo complete AC (median OS: 28.0 months; p=0.523).
Comparison of perioperative characteristics of patients with pancreatic cancer according to Geriatric Nutritional Risk Index (GNRI).
Overall survival curves of patients with pancreatic cancer based on pre-operative Geriatric Nutritional Risk Index (GNRI) and adjuvant chemotherapy.
Discussion
This study revealed that the rates of initiating AC and AC completion were 75.6% and 44.7%, respectively. Pre-operative GNRI and no complete AC were independent predictors of OS in patients with resected PC. Additionally, GNRI and age were independent predictors of complete AC. Patients with complete AC in the high and low GNRI groups had significantly longer OS than those who did not complete AC. However, no significant difference in OS was observed between the low GNRI group with complete AC and the high GNRI group without complete AC.
Previous studies have reported that complete AC was associated with improved survival rates in patients with resected PC (3, 19). This study revealed that the complete AC group had significantly higher OS than the incomplete AC and no initial AC groups. Therefore, treatment completion is essential for maximizing the efficacy of AC. Several factors can influence whether patients with resected PC receive appropriate AC and can be categorized as patient, perioperative, or social factors (3, 19-30). Regarding patient-related factors, older age, female sex, higher performance status, comorbidity severity, body weight loss (BWL), skeletal muscle loss, low post-operative serum albumin levels, low prognostic nutritional index, advanced tumor stage, and poor tumor differentiation grade limit AC administration. In contrast, positive lymph nodes and high Evans’s grade can promote AC induction. Peri-operative factors, such as vascular resection, intra-operative blood transfusion, post-operative complications, and surgical site infections can restrict AC administration. In contrast, the administration of Phellinus linteus promotes AC completion rates. Concerning social factors, non-white race, uninsured status, lower income, low education, community residence, high area deprivation index, distance from the treating facility, low hospital volume for surgical procedures, and administration of AC by local oncologists can result in low rates of initiation or completion of AC. In this study, older age and a low pre-operative GNRI were independent factors associated with AC completion rates. However, sex, performance status, pathological stage, and post-operative complications did not affect AC completion. This suggests that older age and a low pre-operative GNRI are the most important factors that influence AC completion rates.
Among the factors influencing AC administration in patients with resected PC, nutritional management is pivotal for surgeons and oncologists. The GNRI is a practical and readily applicable nutritional assessment tool, based on serum albumin level, height, and the body weight of patients. Lower serum albumin levels are associated with poor survival across various cancers (31). Albumin, beyond its nutritional role, serves various chemotherapy-related biological functions, including maintaining osmolality, delivering bioactive anticancer molecules, suppressing over-activated inflammation and oxidative stress, and modulating immune responses (8). Hypoalbuminemia is associated with increased morbidity and mortality in patients with PC (32). On the other hand, BWL is associated with a lower quality of life, malnutrition, and poor prognosis in various cancers. Many patients with PC suffer from malnutrition owing to insufficient production and secretion of digestive enzymes and hormones and increased caloric requirements, resulting in BWL. Over 80% of patients with PC have BWL at initial diagnosis, with over a third losing over 10% of their initial body weight (32). Post-operative BWL is associated with poor outcomes after pancreatectomy (33). Therefore, the GNRI is a good nutritional tool for patients with PC because it is based on serum albumin levels and body weight.
In this study, the low GNRI group had a higher percentage of pancreatic head cancers than the high GNRI group. Pancreatic head cancers may have inadequate production and secretion of digestive enzymes owing to the wider extent of tumor-related obstructive pancreatitis (34), resulting in a worse nutritional status compared with pancreatic tail cancers. Furthermore, the high GNRI group had a significantly longer OS and higher completion rate of AC than the low GNRI group. Subgroup analyses revealed a trend toward longer OS in patients with a high GNRI who underwent complete AC than in those with a low GNRI. Moreover, no significant difference in OS was observed between patients with a low GNRI who underwent complete AC and those with a high GNRI who did not undergo complete AC. This result suggests that patients with malnutrition may not fully benefit from AC after pancreatectomy, resulting in a poor prognosis. Furthermore, early intervention, including nutritional support, may contribute to improving the prognosis after pancreatectomy in patients with PC who are malnourished. Peri-operative nutritional intervention decreases post-operative infectious and non-infectious complications and length of hospital stay in patients undergoing gastrointestinal cancer surgery (35). The updated Enhanced Recovery After Surgery recommendations for pancreatoduodenectomy reported pre-operative nutritional interventions for patients with severe weight loss (>15% weight loss or BMI <18.5 kg/m2), and post-operative nutritional interventions are strongly recommended (36). Therefore, intervening in intensive nutritional management is essential, starting from the initial treatment of malnourished patients with PC. However, many related issues remain unresolved, including optimal nutrition screening tools, the best indicators of malnutrition status, and the role and methods of intervention. Further analyses are required to clarify these findings.
Study limitations. First, it was a retrospective study conducted at a single institution. Second, the sample size was relatively small and only of East Asian ethnicity; therefore, the findings may be biased with limited generalizability. Third, the GNRI cutoff value was set at 98 using ROC analysis. However, the optimal cutoff value for patients with resected PC remains unclear. Therefore, a more comprehensive large-scale prospective study should be conducted to validate our findings.
Conclusion
Our data suggest that a low pre-operative GNRI score is a crucial risk factor for AC completion and poor prognosis in patients with PC. Therefore, nutritional assessment using the GNRI may be profound for patients with PC, and early intervention with nutritional support for malnourished patients may be critical in promoting the completion of AC and improving their prognosis.
Acknowledgements
The Authors thank Editage (www.editage.com) for English language editing.
Footnotes
Authors’ Contributions
All Authors conceived the study concept and design. TH made the initial proposal for this study, collected, and analyzed the data and statistical analyses, and wrote and edited the manuscript. TH, KM, RY, MK, MF, JYT, SK, YT, and NO monitored the patients. KM and NM revised and supervised this study. The final manuscript was approved by all the Authors who agreed to be accountable for the content of this work.
Conflicts of Interest
NM received grants from Taiho Pharma. The Authors declare no conflicts of interest in relation to this study.
- Received November 29, 2023.
- Revision received December 16, 2023.
- Accepted December 18, 2023.
- Copyright © 2024 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
