Research Article | Open Access
Hu Ren, Chao-Rui Wu, Guo-Tong Qiu, Li-Peng Zhang, Saderbieke Aimaiti, Cheng-Feng Wang, "Equipping the American Joint Committee on Cancer Staging for Resectable Pancreatic Ductal Adenocarcinoma with Tumor Grade: A Novel Staging System", Journal of Oncology, vol. 2020, Article ID 9093729, 9 pages, 2020. https://doi.org/10.1155/2020/9093729
Equipping the American Joint Committee on Cancer Staging for Resectable Pancreatic Ductal Adenocarcinoma with Tumor Grade: A Novel Staging System
Background. The 8th American Joint Committee on Cancer (AJCC) staging system for pancreatic ductal adenocarcinoma (PDAC) outperforms its previous version in reproducibility but not in survival discrimination. Tumor grade, an indicator of the aggressive biology of PDAC, has been suggested as a reliable prognostic factor. This study aimed to construct a novel staging system with greater prognostication for resectable PDAC by incorporating tumor grade into the 8th AJCC system. Methods. A total of 9966 patients with resectable PDAC from the Surveillance Epidemiology and End Results (SEER) database were randomly separated into training and interval validation sets. Another 324 patients from our center were included as an external validation set. We proposed a novel staging system by sorting the substages yielded by a combination of T, N, and tumor grade based on their overall survival (OS) and grouping them into several stages. Prognostic homogeneity and discrimination were determined using the likelihood ratio χ2 and the linear trend χ2 test, respectively. Prognostic accuracies were evaluated by the area under the receiver operating characteristics curve (AUC). Results. Using the 8th AJCC system, the prognosis of patients within the same stage was quite heterogeneous among different substages. The multivariate Cox model identified the tumor grade (hazard ratio 1.333, 95% confidence interval 1.250–1.423, ) was an independent prognostic factor of the OS. In the training set, the AUC, homogeneity, and discriminatory ability were superior for the novel staging system than for the 8th AJCC system (0.642 vs. 0.615, 403.4 vs. 248.6, and 335.1 vs. 218.0, respectively). Similar results were observed in the internal and external validation sets. Conclusions. The novel staging system incorporating tumor grade into the 8th AJCC system was associated with better prognostic accuracy, homogeneity, and discriminatory ability among resectable PDAC patients. Moreover, the novel staging system also allowed possibly adjuvant chemotherapy decisions.
Pancreatic ductal adenocarcinoma (PDAC) remains one of the most lethal human malignancies, with little improvement in survival over the past decades . The five-year survival rate of PDAC is extremely low (approximately 8%), and PDAC is projected to be the second leading cause of cancer mortality in the next decade [2, 3].
The American Joint Committee on Cancer (AJCC) staging system, which includes the depth of tumor invasion (T) and nodal status (N), was the most frequently used prognosticator for PDAC patients. The 8th AJCC staging system has introduced major revisions to its previous version, including changes to T and N definitions . The new T1–T3 definitions completely depend on tumor size, and T4 refers to tumors that involved the celiac axis, the superior mesenteric artery, and/or common hepatic artery. Additionally, node-positive disease (previous N1) was further categorized into N1 (1–3 positive nodes) and N2 (≥4 positive nodes). However, although the reproducibility of the 8th AJCC staging system was superior, it did not outperform previous versions in terms of survival discrimination [5, 6]. Survival curves of resectable stages IB, IIA, or IIB disease are not well-separated.
With the evolving understanding of PDAC biology, the spectrum of biologic factors that are predictive of prognosis and treatment response was rapidly expanding [7, 8]. In addition to tumor size and lymph node status, those emerging biologic factors should also be incorporated into the existing staging system for more accurate survival prediction [9–12]. Tumor grade, which represents the aggressive biology of the tumor itself, has been widely confirmed as a prognostic factor in PDAC . Wasif et al.  proposed a new staging system by combining the tumor grade and the 7th AJCC stage, which significantly optimized survival discrimination. However, this proposed staging scheme arbitrarily upstaged patients with high tumor grade and downstaged patients with low tumor grade, which was methodologically questionable. Chen et al.  have combined tumor grade and the 8th AJCC staging system into a novel staging system using the SEER database. Liu et al.  also proposed a novel one by integrating the tumor grade and postoperative CA19-9 levels with the 8th AJCC staging system in their single-institutional dataset. However, neither of them had attempted to cross-validate their staging schemes.
This study aimed to assess the performance of the 8th AJCC staging system and the impact of tumor grade on overall survival (OS). We sought to incorporate tumor grade into the 8th AJCC system to form a novel staging system and validate it both internally and externally.
2. Materials and Methods
2.1. Study Cohort
The National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) database, which collects cancer incidence, treatment, and survival data from 18 population-based cancer registries in the US, was used to establish the novel staging system for resectable PDAC. We identified patients with locally resectable PDAC (ICD-O-3 codes 8500/3 and 8140/3, respectively) from 2004 to 2015. Cases with locally unresectable tumor (PDAC with unreconstructable SMV or PV occlusion, and classified as T4 by the 8th AJCC staging system), distant metastasis, history of prior malignancy, age at diagnosis younger than 18 years, and missing information regarding tumor grade, tumor size, tumor extent, and nodal status or OS were excluded. Finally, a total of 9966 patients with resectable PDAC were eligible, which were randomly partitioned into disjoint datasets for model training (n = 4983) and internal validation (n = 4983). To validate the novel staging system, 324 patients with resectable PDAC that matched the inclusion criteria at the China National Cancer Center between January 2010 and October 2017 were included as an external validation set.
2.2. Data Collection
Demographic and clinicopathological variables extracted for each patient were as follows: age at diagnosis, year of diagnosis, gender, tumor size, tumor extent, tumor location (head, body, tail, and others), tumor grade (well-differentiated, moderately differentiated, poorly differentiated, and undifferentiated), nodal status (number of examined lymph nodes and number of positive lymph nodes), the 7th AJCC stage, and survival months. With staging information inferred from the 7th AJCC stage, patients enrolled in the current study were restaged in compliance with new T, N definitions introduced to the 8th AJCC staging system. Tumor differentiation refers to the extent to which a tumor cell morphologically and functionally resembles a normal cell from the same tissue. The extent of glandular differentiation dictates the histologic grade of PDAC. A tumor with unrecognizable tissue of origin was graded as undifferentiated, < 50% of its components being gland as poorly differentiated, 50%–95% as moderately differentiated, and > 95% as well-differentiated. In this study, to minimize interinstitution and interobserver variance in the pathological assessment of tumor grade, tumor grade was dichotomously subdivided into the low grade (well to moderately differentiated, G1) and the high grade (poorly to undifferentiated, G2).
2.3. Statistical Analysis
OS, the primary endpoint in this study, was defined as the duration from the date of initial diagnosis to the date of death or last follow-up (updated on June 1, 2019). The univariate and multivariate Cox proportional hazard regression analysis was performed to identify the independent prognostic predictors for the OS. The OS was analyzed by the Kaplan–Meier survival curves, and log-rank tests were utilized to evaluate the staging systems. The performance of these staging systems was graded based on the area under the curve (AUC), homogeneity, and discriminatory ability. The AUC was compared between the two staging systems using DeLong’s method. Homogeneity, a measure of differences in OS among patients with the same stage within each staging system, was calculated using the likelihood ratio χ2 through the Cox regression model . Discriminatory ability, a measure of differences in OS among patients in different stages within each staging system, was calculated using the linear trend χ2 test . All tests were 2-sided, and was considered statistically significant. Statistical analyses were performed using SPSS software version 25.0 (SPSS Inc., Chicago, IL, USA).
3. Results and Discussion
3.1. Baseline Characteristics
A total of 9966 patients with resectable PADC from the SEER database (2004–2015) were enrolled, which was split equally into a training set (n = 4983) and an internal validation set (n = 4983). In addition, a total of 324 patients with resectable PDAC diagnosed from 2010 to 2017 at the China National Cancer Center were enrolled as an external validation set. The detailed patients’ characteristics of the three cohorts are shown in Table 1. All clinicopathological characteristics were comparable between the training set and the internal validation set. Compared with the training set and internal validation set, the external validation set was associated with a higher proportion of female patients, patients with age at diagnosis <65 years, tumors located at the body and tail of pancreas, higher tumor grade, less than 15 harvested lymph nodes, tumor located within pancreas, less advanced pathological N stage, less advanced 7th AJCC, and less advanced 8th AJCC stage. The median OS of the training set, internal validation set, and external validation set were 19.0, 19.0, and 23.6 months, respectively.
Values in parentheses are percentages.
3.2. Independent Prognostic Factors of Survival in the Training Set
The univariate and multivariate analysis results are shown in Table 2. In the training set, age at diagnosis (), year of diagnosis (), tumor grade (), number of examined lymph nodes (), extrapancreatic invasion (), tumor size (), and number of metastatic lymph nodes () were identified as independent prognostic factors for OS.
Abbreviations: HR, hazard ratio; CI, confidence interval.
3.3. Development of a Novel Staging System in the Training Set
To evaluate the accuracy and appropriateness of the 8th AJCC staging system, the median OS of each substage from the training set was calculated. Even within the same tumor stage, the median OS was quite heterogeneous across different substages (Figure 1(a)). For example, within stage ?, the median OS for T1N2M0, T2N2M0, and T3N2M0 was 18.5, 14.0, and 11.3 months, respectively (). Moreover, patients with stage IIB T1N1M0 tumors survived longer than patients with stage IIA T3N0M0 tumors (median OS: 20.8 vs 17.8 months; ). Similarly, patients with stage III T1N2M0 tumors survived longer than patients with stage IIB T3N1M0 tumors (median OS: 18.5 vs 13.9 months; ). Based on the results of multivariate analysis, the tumor grade was incorporated into the 8th AJCC system to form a novel staging system, which resulted in 18 substages. According to the median OS of each substage, these substages were regrouped into five stage groups, including novel IA (G1T1N0), novel IB (G1T1N1, G1T2N0, and G2T1N0), novel IIA (G1T1N2, G1T2N1, G1T3N0, G2T1N1, and G2T2N0), novel IIB (G1T2N2, G1T3N1, G1T3N2, G2T1N2, G2T2N1, and G2T3N0), and novel III (G2T2N2, G2T3N1, and G2T3N2) (Figure 1(b)).
3.4. Comparison between the 8th AJCC and the Novel Staging System
A histogram showing the distribution of patients across different substages within in the three cohorts are shown (Figure 2). In the training set, 7.9%, 17.8%, 6.6%, 41.8%, and 25.9% of the patients were assigned to stage IA, IB, IIA, IIB, and III (T1-3N2M0) tumors by the 8th AJCC staging system, respectively, while 5.9%, 29.8%, 48.8%, 31.9%, and 15.1% of the patients were assigned to stages IA, IB, IIA, IIB, and III (T1-3N2M0) tumors by the novel staging system, respectively. As to the external validation set, 7.4%, 29.8%, 18.1%, 34.4%, and 10.4% of the patients were assigned to stages IA, IB, IIA, IIB, and III (T1-3N2M0) tumors by the 8th AJCC system, respectively, while 6.1%, 21.2%, 32.8%, 26.4%, and 13.5% of the patients were assigned to stages IA, IB, IIA, IIB, and III (T1-3N2M0) tumors by the novel system, respectively. The proportion of stage IIA disease by the novel stage was much higher than that of stage IIA disease by the 8th AJCC system in the three cohorts.
Suggested by the well-separated survival curves for the three cohorts (Figure 1(d), Figures 3(b) and 3(d)), the novel staging system showed better prognostic discrimination compared with the 8th AJCC staging system (Figure 1(c) and Figures 3(a) and 3(c)). A good separation of survival curves between stages IIA and IIB disease was also noticed. The results of performance evaluation for the three staging systems are shown in Table 3. In the training set, the respective 3-year AUCs of the 8th AJCC and the novel staging system were 0.615, and 0.642, respectively (). The likelihood ratios χ2, which was a measure of homogeneity, of the 8th AJCC and the novel staging system was 248.6 and 403.4, respectively. The discriminatory ability of each system from the linear trend χ2 test was 218.0 and 335.1, respectively. Similar results were observed in the internal and external validation sets.
Higher AUC, discriminatory ability, and homogeneity indicate better performance of the staging system. Abbreviations: AUC, area under the curve. CI, confidence interval.
3.5. This Novel Staging System Predicted Survival Benefits from Adjuvant Chemotherapy
Analysis of the external validation set showed that the median OS was prolonged by adjuvant chemotherapy (31.2 versus 16.8 months, ). The adjuvant chemotherapy benefit was subsequently determined in each stage within the 8th AJCC and the novel staging system. In the subgroup of patients with novel stage IA- IIA disease, there was no significant difference in median OS between patients who received adjuvant chemotherapy and patients who did not (35.4 versus 30.4 months, , Figure 4(a)). Patients with 8th AJCC stage IA- IIA disease had a similar result (31.2 versus 31.4, , Figure 4(c)). Whereas in the subgroup of patients with the novel or 8th AJCC stage IIB-III disease, there was a survival benefit in patients who received adjuvant chemotherapy (novel stage IIB-III, median OS, 20.7 versus 9.3 months, , Figure 4(b); 8th AJCC stage IIB-III, median OS, 24.9 versus 11.1 months, , Figure 4(d)).
Supported by the previous studies [5, 6, 19, 20], a moderate improvement in the accuracy of prediction was yielded by the 8th AJCC staging system. Nevertheless, the survival curves for stages IIA and IIB disease were not clearly separated in the 8th AJCC system, and the overall discrimination was only marginally enhanced compared with the 7th AJCC system. Thus, the 8th AJCC system for resectable PDAC is far from accurate, and improvement should be considered. So far, revisions of the AJCC staging system for PDAC largely relied on changes to T and N definitions , leading us to consider that a staging system that merely revised the existing parameters might fail to gain a satisfying performance of survival discrimination.
In the novel staging system, a moderate improvement in survival discrimination was demonstrated by the elevated AUCs of survival prediction. A good separation of survival curves between stages IIA and IIB disease was also noticed. Prognostic homogeneity and discriminatory ability were also significantly optimized by the novel staging system. Furthermore, we found that this novel staging system had another role in addition to the prediction of survival for resectable PDAC. Although chemotherapy remained the only effective option of systemic therapy, the survival benefit it offered was quite limited, and only a few patients would respond to chemotherapy . So far, there is a paucity of tools to select the subset of patients who would benefit from adjuvant chemotherapy . In the current study, only patients with the 8th AJCC and novel stages IIB or III disease could benefit from adjuvant chemotherapy, which suggested that the tumor grade as well as the tumor stage were reliable indicators of the tumor burden. As shown in Figure 1, the novel staging system can provide more accurate survival prediction for resectable PDAC by incorporating tumor grade into the 8th AJCC system. The information on tumor grade could inform the efficacy of adjuvant chemotherapy for patients within the same 8th AJCC staging system. We think the novel staging system may be a simple and feasible supplement for the 8th AJCC system to identify patients who would benefit from adjuvant chemotherapy.
Chen et al.  have combined tumor grade and the 8th AJCC staging system into a novel staging system using the SEER database. Liu et al.  also proposed a novel one by integrating the tumor grade and postoperative CA19-9 levels with the 8th AJCC staging system in their single-institutional dataset. However, the external validations of their staging schemes were lacking in their studies. In contrast, the newly proposed staging system described in this study has been cross-validated in both the SEER database and our center. Furthermore, lots of nomograms in which diverse prognostic factors were included have been developed to yield absolute prediction [10–12]. However, both patients and clinicians were reluctant to use these nomograms, mostly because of their complexity and the difficulty of using them. In contrast, our proposed staging system only included three components and was subdivided into five stage groups, which was simpler than nomograms. The tumor grade was easy to obtain, which was routinely reported by pathologists. Although the incorporation of the tumor grade complicated the staging system, it significantly optimized the prediction accuracy of the existing AJCC staging system.
Most researchers performed an initial analysis in their single-institution dataset before seeking validation within a larger multi-institution dataset, hoping the analysis can be equally effective in other centers [6, 19, 23], while in the current study, we utilized the SEER database as the training set rather than our institution’s patients. The sample of our single-institution cohort was not adequate enough to establish a reliable staging system. The sample size of the SEER database was adequate while it only included patients from the USA. Moreover, the characteristics were significantly different between the SEER database and the external cohort, which was possibly reflective of the difference between the western and eastern PDAC patients. Therefore, we decided to establish a novel staging system from the SEER database and validate it in an external cohort to evaluate its applicability in eastern patients.
In the current study, we observed that the nodal status was less advanced in the external dataset, but there was no significant difference in tumor size between the SEER data and the external dataset. A smaller proportion of patients with a tumor located at the head of pancreas in the external dataset may account for the smaller number of harvested and metastatic lymph nodes. Furthermore, the interinstitutional variance also accounts for the discrepancy in part.
This novel staging system tends to downstage node-positive disease (T1N1, T2N1, T1N2, and T2N2) regardless of grade classification. A recent study by Shi et al.  demonstrated that the median OS of patients in the same 8th AJCC stage varied widely among the different substages, and they proposed a modified staging system by regrouping the substages. In their proposed staging system, T1N1, T2N1, T1N2, and T2N2 were downstaged as a result of longer survival duration than other substages within the same stage, which was in accordance with our results.
Some limitations should not be ignored. First, despite a superior performance of the novel staging system, the addition of tumor grade created a more complicated system than the AJCC 8th system. Second, the period of the current study was more limited as the preoperative radiographic assessment and postoperative adjuvant therapies of the included patients would be more uniform. Third, we cannot eliminate interinstitution and interobserver variance in protocols for pathological assessment of tumor grade. The two-tiered grading system adopted in this study which divided the grade into high and low could partially minimize that variance. Fourthly, the external validation set has a significantly higher OS compared with the two other sets from the SEER database, which may undermine the efficacy of validation. Finally, the survival discrimination was only moderately improved by the proposed staging system, which owns to insufficient inclusion of predictive factors needed for absolute prediction. For example, although the prognostic value of various clinicopathological factors, such as preoperative CA19-9 level, resection margin status, lymphovascular invasion, nerve invasion, and adjuvant chemotherapy have already been validated by previous studies [16, 21], they were not included as a lack of these information in SEER database.
In conclusion, our newly proposed staging system was associated with better prognostic accuracy, homogeneity, and discriminatory ability among PDAC patients after pancreatectomy compared with the 8th AJCC staging system. Moreover, the novel staging system also allows possible adjuvant chemotherapy decisions. We believe that the superiority of this staging system stems from new revisions to the 8th AJCC T and N definitions and the inclusion of tumor grade that reflects the aggressive behavior of PDAC. However, external validation from other populations is necessary.
The data used to support the findings of this study are available from the corresponding author upon request.
Conflicts of Interest
The authors declare that there are no conflicts of interest regarding the publication of this paper.
Hu Ren Chao-Rui Wu contributed equally to this work.
This work was supported by the CAMS Innovation Fund for Medical Sciences (CIFMS) (grant number: 2016-I2M-1-001). The authors thank the SEER program for providing open access to the database.
- A. McGuigan, P. Kelly, R. C. Turkington, C. Jones, H. G. Coleman, and R. S. McCain, “Pancreatic cancer: a review of clinical diagnosis, epidemiology, treatment and outcomes,” World Journal of Gastroenterology, vol. 24, no. 43, pp. 4846–4861, 2018.
- L. Rahib, B. D. Smith, R. Aizenberg, A. B. Rosenzweig, J. M. Fleshman, and L. M. Matrisian, “Projecting cancer incidence and deaths to 2030: the unexpected burden of thyroid, liver, and pancreas cancers in the United States,” Cancer Research, vol. 74, no. 11, pp. 2913–2921, 2014.
- V. L. Gordon-Dseagu, S. S. Devesa, M. Goggins, and R. Stolzenberg-Solomon, “Pancreatic cancer incidence trends: evidence from the Surveillance, Epidemiology and End Results (SEER) population-based data,” International Journal of Epidemiology, vol. 47, no. 2, pp. 427–439, 2018.
- M. B. Amin, S. Edge, F. Greene et al., AJCC Cancer Staging Manual, Springer, Berlin, Germany, 8th edition, 2017.
- A. M. Schlitter, M. Jesinghaus, C. Jäger et al., “pT but not pN stage of the 8th TNM classification significantly improves prognostication in pancreatic ductal adenocarcinoma,” European Journal of Cancer, vol. 84, pp. 121–129, 2017.
- S. van Roessel, G. G. Kasumova, J. Verheij et al., “International validation of the eighth edition of the American Joint committee on cancer (AJCC) TNM staging system in patients with resected pancreatic cancer,” JAMA Surgery, vol. 153, no. 12, Article ID e183617, 2018.
- Y. R. Lawrence, J. Moughan, A. M. Magliocco et al., “Expression of the DNA repair gene MLH1 correlates with survival in patients who have resected pancreatic cancer and have received adjuvant chemoradiation: NRG Oncology RTOG Study 9704,” Cancer, vol. 124, no. 3, pp. 491–498, 2018.
- D. Liang, S. Shi, C. Liang et al., “Mismatch repair status as a beneficial predictor of fluorouracil-based adjuvant chemotherapy for pancreatic cancer,” Surgery, vol. 163, no. 5, pp. 1080–1089, 2018.
- M. M. Rochefort, J. S. Ankeny, B. E. Kadera et al., “Impact of tumor grade on pancreatic cancer prognosis: validation of a novel TNMG staging system,” Annals of Surgical Oncology, vol. 20, no. 13, pp. 4322–4329, 2013.
- Y.-N. Shen, X.-L. Bai, G. Jin et al., “A preoperative nomogram predicts prognosis of up front resectable patients with pancreatic head cancer and suspected venous invasion,” Hpb, vol. 20, no. 11, pp. 1034–1043, 2018.
- J. Hang, L. Wu, L. Zhu et al., “Prediction of overall survival for metastatic pancreatic cancer: development and validation of a prognostic nomogram with data from open clinical trial and real-world study,” Cancer Medicine, vol. 7, no. 7, pp. 2974–2984, 2018.
- S. H. Choi, S. W. Park, and J. Seong, “A nomogram for predicting survival of patients with locally advanced pancreatic cancer treated with chemoradiotherapy,” Radiotherapy and Oncology, vol. 129, no. 2, pp. 340–346, 2018.
- M. Strijker, J. W. Chen, T. H. Mungroop et al., “Systematic review of clinical prediction models for survival after surgery for resectable pancreatic cancer,” British Journal of Surgery, vol. 106, no. 4, pp. 342–354, 2019.
- N. Wasif, C. Y. Ko, J. Farrell et al., “Impact of tumor grade on prognosis in pancreatic cancer: should we include grade in AJCC staging?” Annals of Surgical Oncology, vol. 17, no. 9, pp. 2312–2320, 2010.
- Y.-T. Chen, Z.-P. Huang, Z.-W. Zhou, and M.-M. He, “Equipping the American Joint Committee on Cancer staging for resectable pancreatic ductal adenocarcinoma with tumor grade: a recursive partitioning analysis,” Medical Oncology, vol. 33, no. 11, pp. 1–8, 2016.
- L. Liu, H.-X. Xu, M. He et al., “A novel scoring system predicts postsurgical survival and adjuvant chemotherapeutic benefits in patients with pancreatic adenocarcinoma: implications for AJCC-TNM staging,” Surgery, vol. 163, no. 6, pp. 1280–1294, 2018.
- D. W. Hosmer, T. Hosmer, S. Le Cessie, and S. Lemeshow, “A comparison of goodness-of-fit tests for the logistic regression model,” Statistics in Medicine, vol. 16, no. 9, pp. 965–980, 1997.
- A. R. Feinstein, “XVI. The process of prognostic stratification (Part 2),” Clinical Pharmacology & Therapeutics, vol. 13, no. 4, pp. 609–624, 1972.
- S. Shi, J. Hua, C. Liang et al., “Proposed modification of the 8th edition of the AJCC staging system for pancreatic ductal adenocarcinoma,” Annals of Surgery, vol. 269, no. 5, pp. 944–950, 2019.
- P. J. Allen, D. Kuk, C. F.-d. Castillo et al., “Multi-institutional validation study of the American Joint commission on cancer (8th edition) changes for T and N staging in patients with pancreatic adenocarcinoma,” Annals of Surgery, vol. 265, no. 1, pp. 185–191, 2017.
- H. Oettle, P. Neuhaus, A. Hochhaus et al., “Adjuvant chemotherapy with gemcitabine and long-term outcomes among patients with resected pancreatic cancer,” JAMA, vol. 310, no. 14, pp. 1473–1481, 2013.
- P. Martinelli, E. Carrillo-de Santa Pau, T. Cox et al., “GATA6 regulates EMT and tumour dissemination, and is a marker of response to adjuvant chemotherapy in pancreatic cancer,” Gut, vol. 66, no. 9, pp. 1665–1676, 2017.
- M. Song, S. B. Yoon, I. S. Lee et al., “Evaluation of the prognostic value of the new AJCC 8th edition staging system for patients with pancreatic adenocarcinoma; a need to subclassify stage III?” European Journal of Cancer, vol. 104, pp. 62–69, 2018.
Copyright © 2020 Hu Ren et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.