Neuroendocrine tumours (NETs) are composed of by a wide range of neoplasm which mostly originate from neuroendocrine cells of gastrointestinal tract, pancreas or lungs. The incidence of NETs in United State was 6.98 cases per 100.000 people in 2012 according to SEER database. Despite its lower prevalence, the incidence rate showed a remarkable increase from 1973 to 2012[1. There are various factors that predict the outcome of disease; including primary tumour site, stage, grade, age, primary tumour size and lymph node involvement2,3. Of those, stage and grade are well established prognostic factors directly correlating with survival4,5.

According to the latest classification of World Health Organization (WHO), NETs are graded as well-differentiated, low-grade (G1); well-differentiated, intermediate-grade (G2); poorly differentiated, high-grade (G3); and a fourth category for pancreatic neuroendocrine tumours: well-differentiated, high grade (G3) on the basis of Ki-67 proliferation index and mitotic count on high-power field (HPF). In 2017, WHO has further divided pancreatic grade 3 neuroendocrine tumours into well-differentiated and poorly-differentiated subgroups6.

Different response rates and survival outcomes of NETs, even within the same group, might be attributed to its heterogeneous nature. There are number of studies in the literature to identify prognostic or predictive markers in NETs7,8,9,10. However, there are still no well-established prognostic factors except for Ki-67 proliferation index and mitotic count which are essential for grading. In the literature, numerous easy-accessible and simple markers, including inflammatory markers such as C-reactive protein, albumin, neutrophil/lymphocyte ratio, platelet/lymphocyte ratio, mean platelet volume, gamma-glutamyltransferase (GGT) were reported to predict outcomes of the patients with NETs. The emerging data suggest that inflammatory response in tumour microenvironment has a role in determining the outcomes of the cancers11. Thereby, recent studies have been focused on the markers that might reflect the inflammatory response in tumour microenvironment8,9,10.

In this study, we aimed to evaluate the prognostic value of inflammation markers specifically for the patients with well differentiated NETs involving only G1 and G2 NETs. We evaluated the inflammatory ratios (i.e neutrophil/lymphocyte ratio, platelet/lymphocyte ratio, monocyte/lymphocyte ratio, GGT/lymphocyte ratio). Besides, we considered to evaluate all of the patients for albumin and hemoglobin concetrations together, since hypoalbuminemia and anemia have been reported to have poor prognostic effect on survival of different cancer types12,13,14,15,16. Therefore, we decided to use a novel parameter including both albumin and hemoglobin.

Patients and data:

We retrospectively reviewed the clinical records of the patients with well differentiated G1 and G2 NETs followed-up in our centre between June 2002 and November 2018. A total of 132 patients' data were reached and pooled from oncologic follow-up files in our centre. All patients were aged over 18 and had histologically confirmed well differentiated G1 and G2 NETs. Patients who had G3 tumours and second primary cancer were excluded from the study.

Various factors were gathered from the follow-up files including age at diagnosis, the Eastern Cooperative Oncology Group (ECOG) performance-status score, primary site of tumour, primary tumour size, presence of metastasis, progression-free survival (PFS), overall survival (OS), treatment modalities, and various parameters from laboratory test results which may have prognostic and/or predictive role (white blood cell count (WBC), neutrophil count, lymphocyte count, platelet count, haemoglobin concentration, albumin concentration, GGT concentration and alkaline phosphatase (ALP) at the time of diagnosis. Progression-free survival (PFS) was determined as the time between the beginning of treatment and first progression or death from any cause, whichever occurred first. Complete blood count analyses were performed from collected blood sample in ethylenediaminetetraacetic acid (EDTA) anticoagulated tubes. White blood cell count, lymphocyte and platelet count was recorded as number in per microliter, albumin and haemoglobin concentration was recorded as gr/dl. Multiplication of hemoglobin and albumin was recorded as a novel parameter and used in survival analyses.

Staging and grading:

Ki-67 proliferation index, mitotic count on high-power fields (HPF) and grade were recorded directly from pathology reports. The staging and grading of the disease were re-performed according to 8th edition of the AJCC Cancer Staging Manual and the latest update of WHO in 20176 17.

Statistical Analyses:

Statistical analysis was performed by using Statistical Package for the Social Sciences Version 22.0 for Windows (SPSS Inc., Chicago, IL, USA). We used Kaplan-Meier test for survival, and outcomes were analysed by Log-rank test. Overall survival was calculated from the date of diagnosis to date of last control for alive patients or death from any cause. Univariate and multivariate Cox proportional hazards regression analysis was used to assess the prognostic significance of the various parameters on OS. Cut-off values are extracted from receiver operating characteristic (ROC) analysis. A p-value <0.05 was considered as statistically significant.

A total of 132 patients with well-differentiated G1 and G2 NETs were included in this study. The number of female patients was 81 (61.4%) and male patients was 51 (38.6%). Median age at the time of diagnosis was 53.0 years (22-80) for all cohort. In the whole group, 68 (51.5%) patients had G1 tumour and 64 (48.5%) patients had G2 tumours. The most common primary site was pancreas with a number of 44 (33.3%); the other locations were gastric, intestinal, appendix, rectum and others with a number of 30 (22.7%), 18 (13.6%), 10 (7.6%), 8 (6.1%) and 22 (16.7%), respectively. At the time of diagnosis, 105 (79.5%) patients had local disease and 23 (17.4%) patients had metastatic disease. Ninety-eight (74.2%) patients had surgery for primary tumour either with or without metastasectomy. Median follow-up period was 49.28 (1.7-186.2) months for all group. Median OS for whole group has not been reached yet. At the time of final analyses, the survival data of 7 patients could not have been reached, and the number of alive patients was 105 (79.5%) within the whole group. Patients characteristics are summarized in Table 1.

Multivariate analyses for OS showed that neutrophil/lymphocyte ratio (p=0.814), platelet/lymphocyte ratio (p=0.090), monocyte/lymphocyte ratio (p=0.115) and GGT/lymphocyte ratio (p=0.131) had no statistically significant effect on OS except for albuminXhemoglobin (AlbXHb) (HR=0.871, 95% CI 0.800-0.947, p=0.001). Additionally, the result of univariate analyses defined AlbXHb as a statistically significant prognostic factor for OS (HR=0.925 95% CI 0.878-0.974, p=0.003).

Median albumin and hemoglobin concentrations were 4.3 gr/dl (2.5-5.2) and 12.4 gr/dl (6.4-17.5) at diagnosis for whole group, respectively. The ROC analysis was performed to set a cut-off value for multiplication of albumin with hemoglobin. The ROC analysis was performed by using survival status as an endpoint and areas under curve (AUC) was 0.771 (95% CI 0.662-0.879) with statistical significance (p=0.001) Figure 1.

Figure 1

The maximal joint point of sensitivity and specificity was 47.94 which had sensitivity and specificity 80% and 66.3%, respectively. All patients with well-differentiated NETs (G1 and G2) were sub-grouped with respect to determined cut-off value. A total of 117 patients could have been categorized according to this scoring system: low risk group (AlbXHb ≥47.94, n:70, 53%) and high risk group (AlbXHb ≤47.94 n:47, 35.6%). Fifteen patients (11.4%) could not have been evaluated for AlbXHb due to absence of data either in albumin or hemoglobin concentration. There was no statistically difference for baseline clinicopathological features between high-risk and low-risk groups Table 2.

Overall survival of these two sub-groups was statistically different, demonstrated by using Kaplan-Meier method with log-rank test. Median OS was 135.5 months ( 95% CI 83.0-188.0) for high-risk group, while it has not been reached in low risk group (p=0.005)Figure 2.

The COX regression analysis showed that the patients in high risk group has increased risk of death compared to low risk patients ( HR:5.16, 95%CI 1.45-18.38 p=0.011).

Well-differentiated NETs are substantially heterogeneous tumours. Furthermore, disease course might differ even within the same grade tumours. The heterogeneity of NETs has not been well-understood until now. Therefore, novel prognostic and predictive factors are needed to explain its heterogeneity. However, recent studies are focused on simple and easy-accessible parameters to make an estimation about treatment response and disease course. Majority of these studies are based on the hypothesis that some plasma proteins and complete blood count parameters might reflect the tumour microenvironment. Various parameters, including inflammatory markers such as C-reactive protein, albumin, neutrophil, lymphocyte, platelet, hemoglobin, mean platelet volume, plasma levels of GGT and AST have been evaluated in those studies. In addition, these parameters have been studied not only in NETs but also in a wide range of tumours. Result of these studies demonstrated a correlation of those parameters with survival (OS, PFS) in spite of some inconsistencies.

In this study, we specifically focused on markers that might be used to figure out tumour microenvironment. Multivariate analyses showed a strong effect of albumin and hemoglobin concentrations on OS. Albumin and hemoglobin are definitely distinct parameters, however multiplication of them seems to correlate with survival. Furthermore, survival data of low-risk and high-risk patients according to AlbXHb confirmed the prognostic value of these novel parameter. Many studies have demonstrated a relation between oxidative stress and many diseases including cancer18. Reactive oxygen species (ROS) have ability to contribute carcinogenesis via induction of DNA damage, epigenetic changes, tumour heterogeneity and key activation factors, including NF-ĸB18,19,20. However, oxidative stress has dual effect on cancer progression with respect to its level. While lower levels of ROS promote carcinogenesis, proliferation and resistance to oxidative stress, higher levels of ROS can limit the proliferation capacity of cancer cells due to the lack of antioxidant metabolic products during metastasis19. There is a substantial association among oxidative stress, inflammation and carcinogenesis. They all promote each other by complex intracellular pathways which in turn cause an inflammatory state both in the tumour microenvironment and in the whole body20. Systemic inflammation results with reduced concentrations of albumin and hemoglobin through secretion of various inflammatory mediators. Additionally, hypoxia is an another state that promote angiogenesis and cell proliferation by hypoxia inducible factor-1 (HIF1), an oxygen-dependant transcription factor. Hypoxia inducible factor-1 (HIF1) is continuously produced and rapidly degraded in normoxic cells; on the contrary, degradation of HIF1 is decreased in hypoxic cells20,21. The relation of plasma hemoglobin concentration with tumour oxygen status was demonstrated in several preclinical animal studies in which evaluation was performed on experimental tumours. These studies showed that tumour hypoxia was more prominent in anemic animals than in non-anemic ones22.

The rationale for using both albumin and hemoglobin concentrations was based on hypothesis that these markers can reflect the tumour microenvironment and aggressiveness indirectly. Since albumin and hemoglobin concentrations are definitely distinct entities, we considered to multiply them and find out a cut-off value for multiplication to increase the statistical meaning. We determined that this novel parameter, AlbXHb, could predict the survival of patients with high sensitivity and specifity. The prognostic value of albumin and hemoglobin was demonstrated separately for various cancer types, despite the fact that they were not evaluated together. In these studies, lower albumin and hemoglobin concentrations were associated with poor prognosis14,16.

However, to best of our knowledge, there are no trials searching the accuracy of prognostic factors specifically for the patients with well-differentiated NETs. Zhou et al showed the prognostic significance of GGT/lymphocyte ratio in a study which included 125 patients with only non-functional pancreatic NETs9. In our study, GGT/lymphocyte ratio was not statistically significant parameter either in univariate analyses or multivariate analyses. This difference might be related to heterogeneity of the study population. Zou J et al aimed to find out inflammation based prognostic factors and revealed a new scoring system in a study which included patients with advanced and metastatic NETs. This scoring system was a combination of Glasgow Prognostic Score, neutrophil/lymphocyte ratio, platelet/lymphocyte ratio, high-sensitivity inflammation-based prognostic index and prognostic nutritional index8. On the contrary of that study, neutrophil/lymphocyte ratio and platelet/lymphocyte ratio failed to predict disease outcome or survival. However, majority (70.3%) of patients had G3 NETs and it might have contributed to this difference.

In conclusion; well-differentiated NETs is an area of deserving more further studies, although they are 'orphan' tumours. Due to its heterogeneity, there are many uncertainty about prognosis, management and classification. There is little data exist about well-differentiated NETs in the literature. Thereby, in this study, we aimed to develop a novel and simple parameter, AlbXHb, which can be used to predict OS of the patients with well-differentiated NETs. This novel parameter, AlbXHb, has statistically significant accuracy for prediction of outcome in this specific patient population. On the other side, major limitation of our study is its retrospective nature. Finally, we considered our results need to be confirmed by further prospective clinical trials with larger number of patients.

Disclosure Statement: The authors have no conflicts of interest to declare.

Xu Y, Shih T, Yao JC: Trends in the Incidence, Prevalence, and Survival Outcomes in Patients With Neuroendocrine Tumors in the United StatesIncidence, Prevalence, and Survival Outcomes in Patients With Neuroendocrine TumorsIncidence, Prevalence, and Survival Outcomes in Patients With Neuroendocrine Tumors. JAMA Oncology 2017;3:1335-1342 2 Man D Shen Z Zhu X: Prognosis of patients with neuroendocrine tumor: a SEER database analysis. Cancer management and research 2018;10:5629-5638 MY 3 Zaidi Smith P. Nathan H. Beems M. Abbott D. Barrett Fields RC Davidson J Cardona K. Lopez-Aguiar AG, Dillhoff M, Beal E, Poultsides G, Makris E, Rocha F, Crown A, Idrees K, Marincola 2018 LB: Global burden of neuroendocrine tumors and changing incidence in Kentucky. Oncotarget 2018;9:19245-19254 5 Yao JC, Hassan M, Phan A, Dagohoy C, Leary C, Mares JE, Abdalla EK, Fleming JB, Vauthey JN, Rashid A, Evans DB: One hundred years after "carcinoid": epidemiology of and prognostic factors for neuroendocrine tumors in 35,825 cases in the United States. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 26 3063 3072 Kim JY, Hong SM, Ro JY: Recent updates on grading and classification of neuroendocrine tumors. Annals of diagnostic pathology 2017 6 Foubert F, Salimon M, Dumars C, Regenet N, Girot P, Venara A, Senellart H, Heymann M-F, Matysiak-Budnik T, Touchefeu Y: Survival and prognostic factors analysis of 151 intestinal and pancreatic neuroendocrine tumors: a single center experience Journal of Gastrointestinal Oncology 7 Zou J. Li Q. Kou F. Zhu Y. Lu M. Li J. Lu Z. Shen L: Prognostic value of inflammation-based markers in advanced or metastatic neuroendocrine tumours. Current oncology 2019 8 Zhou B. Zhan C. Wu J. Liu J. Zhou J. Zheng S: Prognostic significance of preoperative gamma-glutamyltransferase to lymphocyte ratio index in nonfunctional pancreatic neuroendocrine tumors after curative resection Scientific reports 2017 7 Celep B, Ulas M, Dalgic T, Surmelioglu A, Hayran M, Akoglu M: The predictive value of mean platelet volume in differential diagnosis of non-functional pancreatic neuroendocrine tumors from pancreatic adenocarcinomas European Journal of Internal Medicine 2011 22 F 11 Colotta Garlanda C, Allavena P, Sica A, Mantovani A: Cancer-related inflammation, the seventh hallmark of cancer: links to genetic instability. Carcinogenesis 2009;30:1073-1081 12 Danan D, Shonka DC, Jr., Selman Y, Chow Z, Smolkin ME, Jameson MJ: Prognostic value of albumin in patients with head and neck cancer. The Laryngoscope 2016;126:1567-1571 O 13 Bairey Shacham-Abulafia A, Shpilberg O, Gurion R: Serum albumin level at diagnosis of diffuse large B-cell lymphoma: an important simple prognostic factor. Hematological oncology 2016 V 14 Seebacher Grimm C, Reinthaller A, Heinze G, Tempfer C, Hefler L, Polterauer S: The value of serum albumin as a novel independent marker for prognosis in patients with endometrial cancer. European journal of obstetrics, gynecology, and reproductive biology 2013 15 Jiang HH, Li AJ, Tang EJ, Dan X, Chen Y, Zhang Y, Tang M, Xiao YH, Deng XX, Li HG, Lin MB: Prognostic Value of the Combination of Preoperative Hemoglobin, Lymphocyte, Albumin, and Neutrophil in Patients with Locally Advanced Colorectal Cancer. Medical science monitor : international medical journal of experimental and clinical research 2016 A 16 Bottini Berruti A, Brizzi MP, Bersiga A, Generali D, Allevi G, Aguggini S, Bolsi G, Bonardi S, Bertoli G, Alquati P, Dogliotti L: Pretreatment haemoglobin levels significantly predict the tumour response to primary chemotherapy in human breast cancer. British journal of cancer 2003;89:977-982 17 Amin MB ES, Greene FL, et AJCC Cancer Staging al Kohen R, Nyska A: Invited Review: Oxidation of Biological Systems: Oxidative Stress Phenomena, Antioxidants, Redox Reactions, and Methods for Their Quantification Toxicologic Pathology 2002 Gill JG, Piskounova E, Morrison SJ: Cancer, Oxidative Stress, and Metastasis. Cold Spring Harbor symposia on quantitative biology 2016;81:163-175 19 20 Bhattacharyya S Saha J: Tumour, Oxidative Stress and Host T Cell Response: Cementing the Dominance. Scandinavian journal of immunology 2015;82:477-488 Klaunig JE, Kamendulis LM, Hocevar BA: Oxidative stress and oxidative damage in carcinogenesis 21 Toxicol Pathol 2010 22 Vaupel P Mayer A Hockel M: Impact of hemoglobin levels on tumor oxygenation: the higher, the better? Strahlentherapie und Onkologie : Organ der Deutschen Rontgengesellschaft [et 2006