Nomograms for predicting survival outcomes in patients with neuroendocrine neoplasms of the gallbladder undergoing primary tumor resection: a population-based study

doi:10.21203/rs.3.rs-1832326/v1

Background

Neuroendocrine neoplasms of the gallbladder (GB-NENs) are a rare group of histologically heterogeneous tumors, and surgical resection of the primary tumor is the mainstream treatment at the moment. The current study aimed to establish and validate novel nomograms for patients with GB-NENs undergoing primary tumor resection to predict the 6-, 12-, and 18- months overall survival (OS) and cancer-specific survival (CSS).

Methods

Clinicopathological information of patients with GB-NENs undergoing primary tumor resection between 2004 and 2018 was derived from the Surveillance, Epidemiology, and End Results (SEER) database. The cohort was then randomly assigned to the training and internal validation datasets (7:3). Candidate prognostic factors were selected by Cox regression analyses, and the nomograms were constructed. Finally, concordance index (C-index), calibration plot, and area under the curve from the receiver operating characteristic curve (AUC), were conducted to assess the effective performance of the nomograms. The decision curve analysis (DCA) was employed to further confirm the clinical effectiveness of nomograms. To test the generalizability of the prediction models, we also included 12 patients with GB-NENs undergoing resection from Wuhan Union Hospital.

Results

A total of 221 patients with GB-NENs undergoing resection were enrolled in our study. Using the Cox regression analyses, age, tumor size, and stage were identified as the independent prognostic factors of patients with GB-NENs undergoing resection. We constructed individualized nomograms to visually predict the OS and CSS in patients with GB-NENs undergoing resection. The C-indexes of OS and CSS in training dataset were 0.798 (95%CI: 0.748–0.848) and 0.825 (95%CI:0.771–0.879), while that of internal validation dataset was 0.849 (95%CI: 0.791–0.907) and 0.875 (95%CI: 0.819–0.931), respectively. The AUC of the training and internal validation datasets also indicated that the prediction models performed accurately. The calibration curves demonstrated that the observed survival probability of nomograms was extremely compatible with the predicted probability. The DCA showed that our nomograms were clinically useful.

Conclusions

Taken together, the nomograms are accurate enough to predict the prognostic factors of GB-NENs patients undergoing resection, allowing for treatment decision-making and clinical monitoring for future clinical work.

neuroendocrine neoplasms of the gallbladder

nomogram

Cox regression

prognosis

SEER database

Neuroendocrine neoplasms (NENs) represent an uncommon group of heterogeneous tumors, which originate from the diffuse neuroendocrine system and occur in almost all tissues and organs of the human body [1, 2]. With the improvement of early disease detection and stage migration, the incidence of NENs is steadily growing. As reported by the SEER database, the latest annual age-adjusted incidence rate of NENs has increased by 6.4 times from 1.09 per 100,000 in 1973 to 6.98 per 100,000 in 2012 [3]. In the general population, the incidence of NENs depends on the specific anatomical location, with the highest incidence in the gastrointestinal pancreas and lungs.

However, due to the lack of neuroendocrine cells in the mucosa, NENs are rarely detected in the gallbladder, and neuroendocrine neoplasms of the gallbladder (GB-NENs) currently account for only 0.5% of all NENs cases [4, 5].

Given the rarity of GB-NENs, we lacked sufficient understanding of the possible pathological mechanism driving the malignant transformation of GB-NENs. As a result, the only acceptable therapy for GB-NENs is the surgical removal of the whole gallbladder

[6]. However, because of the diverse biological behaviors and clinical characteristics of GB-NENs, prognostic prediction is a highly challenging task [1]. Lee et al found the complete resection and application of postoperative adjuvant therapy may lead to a better clinical outcome for patients with neuroendocrine carcinomas of the gallbladder [7]. A study of 754 patients with GB-NENs selected from the National Cancer Database (NCDB) showed that advanced age, positive surgical margins, and large cell histology were significantly associated with shorter survival after resection. Besides, the prognosis of primary GB-NENs is worse than NENs of other gastrointestinal sites [8]. Notably, none of these risk factors can well predict the survival probability of each patient. There was an urgent need to seek out a way to estimate the individualized prognosis of patients with GB-NENs undergoing resection through a large-scale cohort study.

Nomograms are generally considered to be a reliable and visualized statistical prediction model to accurately stratify risk by including important prognostic indicators of the diseases [9]. For many cancers, the use of nomograms is superior to the traditional staging systems, thus it has been recommended as an alternative or even as a new standard [10]. In this study, we utilized the data from the SEER database to determine the significant prognostic indicators influencing OS and CSS of patients, and then the individualized nomograms were constructed. Furthermore, we used the validation dataset from the SEER database and our hospital to evaluate the clinical predictive performance of the nomograms.

Ethics Approval and Informed Consent

Since the data from the SEER database are publicly available, and all patients’ personal information is anonymous, informed consent was not required in this study. The patients from our hospital included in this study all provided oral consent. This retrospective study was approved by the Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (IORG number: IORG0003571) and was performed by the Declaration of Helsinki.

Source data and screening criteria

The SEER database provides demographic and clinicopathological information on cancer cases for approximately 30.0% of the United States population [11]. We collected GB-NENs cases from 18 population-based registries (2000–2018) in the SEER database using SEER* Stat 8.3.9 software. The strategy to identify GB-NENs cases is shown in Fig. 1. Data of interest were extracted, including baseline demographics (age, sex, race, marital status, vital survival status, survival time), tumor features (tumor size, grade, stage, lymph node(LN) metastasis, tumor metastasis), treatments methods (other sites surgery, lymph node surgery, radiotherapy, chemotherapy). In this retrospective study, the primary endpoint was OS and CSS. The OS was the time from diagnosis to death of any cause. And CSS was the time from diagnosis to death due to GB-NENs rather than any other cause of death.

The inclusion criteria were defined as follows: (1) year of diagnosis ranged from 2004 to 2018, (2) primary site of the tumor was gallbladder (primary site code: C23.9), (3) histological types were confined to 8013/3, 8041/3, 8240/3, 8244/3, 8246/3, 8249/3 according to International Classification of Disease for Oncology, third edition (ICDO-3). The exclusion criteria were described below: (1) unknown demographic information (2) incomplete pathological diagnosis, (3) not first tumor, (4) no surgery on the primary site, and (5) survival time was zero or unknown.

The clinical records of 12 patients with GB-NENs receiving primary tumor surgery in the Wuhan Union Hospital were collected from 2011 to 2021. The clinical record and follow-up data were complete. The last follow-up was on November 15, 2021. The cause of the patient’s death was specifically caused by GB-NENs. The exclusion criteria were previously mentioned.

Construction and validation of the nomogram models

All cases from the SEER database were randomly assigned into two groups (7:3), which consisted of the training set (n = 154, 70% of total cases) and the internal validation set (n = 67, 30% of total cases). The cases collected from our hospital were used for external validation. The training set was used to construct the nomograms, and the internal and external validation sets were used for validation. The Chi-square test was used to compare categorical variables between the two groups. We performed the univariate Cox regression analysis to select variables for constructing the prediction model. Multivariate Cox proportional hazards (CPH) regression analysis was performed to estimate OS and CSS, then the nomograms were constructed.

Concordance index (C-index), calibration plot, area under the curve (AUC) from the receiver operating characteristic (ROC) curve, and decision curve analysis (DCA), were utilized to evaluate the CPH prediction model. The C-index ranging from 0 to 1 was used to assess the prediction performance of the nomograms. A C-index greater than 0.7 usually indicated that the models had a good discriminatory ability [10]. The value of AUC ranged from 0.5 to 1, which was positively correlated with predictive ability. Calibration plots at 6, 12, and 18 months were drawn to compare the predicted survival probability with the survival probability observed in our study, and the 45-degree calibration curve represents the actual result of the perfect model [12]. The constructed nomograms were formulated using the Cox regression method with the best predictive accuracy to personalize the estimated probability of survival. As an emerging method, DCA was used to evaluate the potential clinical value of the prediction models [13]. All analyses were performed using R software version 4.0.5. P values of less than 0.05 were defined as statistically significant.

1: Patient characteristics

A total of 221 patients with GB-NENs undergoing resection from the SEER database were enrolled. All patients were randomly assigned to two datasets: training (N = 154) and internal validation (N = 67). Of the patients included in this study, 119 (53.8%) were younger than 65 years old,175 (79.2%) were white, 143 (64.7%) were women, and 146 (66.1%) were married. Among all patients, the median survival time was 22 months, 85 (38.5%) tumors were NET, 95 (47.0%) tumors were well-differentiated and the mean tumor size was 27.07 ± 4.31 mm. There were 150 (67.9%) patients in stage N0 and 161 (72.9%) patients in stage M0. There were 147 (66.5%) patients without LN surgery and 191 (86.4%) patients without other site surgery. There were 73 (33.0%) patients who received chemotherapy and 30 (13.6%) patients who received radiotherapy. The clinicopathological information of the patients is shown in Table 1, and there was no significant difference between the training and internal validation datasets. Moreover, the clinicopathological information of the patients collected from tur medical center is shown in Table S1.

2: Identification Of Prognostic Factors

The univariate Cox regression analysis was conducted to identify the prognostic factors in the training set. In the OS analysis, all 14 characteristics were reduced to 8 potential predictors (age, pathological classification, histology, N stage, M stage, stage, tumor size, and chemotherapy) with P < 0.05. Then the selected 8 variables were incorporated into the multivariate Cox regression analysis. The CSS was analyzed in the same way, 5 characteristics (gender, race, marital status, surgery at other sites, and LN surgery) were excluded. The selected 9 variables were also incorporated in the multivariate Cox regression analysis. In the multivariable Cox regression analysis of the training set, younger age, smaller tumor, and localized stage were associated with longer OS (P < 0.05) (Table 2).

3: Construction And Validation Of The Nomograms

Age, tumor size, and stage were identified as independent prognostic factors by Cox regression analysis (Table 2). Then the prediction models that contained the above predictors were constructed using the CPH regression model based on the training set and are shown as the nomograms (Fig. 2). In internal validation, the C-indexes for the nomograms of OS and CSS were 0.798 (95% confidence interval [CI]: 0.748–0.848) and 0.825 (95%CI:0.771–0.879) in the training set, respectively. In the internal validation set, the C-indexes for the models of OS and CSS were 0.849 (95%CI: 0.791–0.907) and 0.875 (95%CI: 0.819–0.931), respectively, presenting a good discriminatory ability of the nomograms. In addition, AUC values at three different time points (6 months, 12 months, and 18 months) were calculated to evaluate the predictive accuracy of the prediction models. In OS analysis, for each of the 3-time points, the AUC values of the training set were 0.790, 0.816, and 0.870; the AUC values of the internal validation set were 0.950, 0.874, and 0.919 (Fig. 3A, C). In CSS analysis, for each of the 3-time points, the AUC values of the training set were 0.822, 0.828, and 0.881; the AUC values of the internal validation set were 0.940, 0.881, and 0.935 (Fig. 3B, D). Furthermore, the calibration plots showed that the nomograms had good performance in predicting the 12-month OS and CSS in the training and internal validation sets (Fig. 4, 5).

4: Clinical Application Of The Nomograms

In the DCA curves, when the threshold probability was greater than 25%, the net benefits of the constructed nomograms and the SEER stage system were comparable in predicting the OS and CSS. Within this range, the clinical performance of the nomograms was significantly better than the SEER stage system in predicting OS and CSS (Fig. 6).

In external validation, we analyzed the data of GB-NENs patients who received primary tumor surgery (n = 12) at Wuhan Union hospital from 2010 to 2021. The last follow-up was on November 15, 2021. Demographics and clinicopathological characteristics was shown in Table S1. Given the rarity of GB-NENs, our data of GB-NENs can only validate CSS at 6 and 12 months. In CSS analysis, the C-indexes were 0.751 (95%CI: 0.652–0.851). In addition, the AUCs of 6- and 12-month (0.917 and 0.861) of the nomogram were greater than the SEER stage system (0.833 and 0.833) (Fig. 7). The results suggested that our prediction model had good applicability and predictive performance in external data.

Neuroendocrine neoplasms are highly heterogeneous diseases, and the prognosis depends on the original site. As shown in the previous study [14], the prognosis of neuroendocrine neoplasms of the gallbladder, bile duct, and ampulla of Vater (AoV) are different. GB-NENs are relatively rare tumors, and the main treatment is surgical resection of the whole gallbladder. However, the prognostic factors of patients undergoing primary tumor resection are still unknown. Thus it is very important and necessary to carry out an independent study on the prognosis of patients with GB-NENs undergoing primary tumor resection. In the current study, we constructed individualized nomograms to predict the prognosis of patients. After bootstrap validation internally, the CPH survival models were evaluated by C-indexes, AUC, calibration plots, and DCAs, showing the nomograms have good discriminatory capabilities and calibration. Fourteen clinical characteristics were collected as potential prognostic factors, and these most valuable variables were selected by COX regression analysis and incorporated into the nomograms. Measured by standard deviation along with nomogram scales[15], the stage was the most important prognostic factor, followed by tumor size and age. These characteristics are easily obtained from patients, making it easy to use the models in real life.

Age as a personal characteristic is widely considered to be a prognostic predictor for gallbladder cancer as well as neuroendocrine neoplasms in other sites, such as gastric NENs, colorectal NENs, and pancreatic NENs[16–20]. With age, the chance of cancer-causing genetic mutations increases. Aging is linked to highly reproducible DNA methylation alterations, which may explain why older people are more likely to get cancer[21]. Older patients with GB-NENs often have a worse prognosis than younger patients [22]. In an NCDB-based study, 300 patients who underwent resection of GB-NENs were analyzed, and the results suggested that elderly patients may have worse survival [8]. The poorer survival rate of elderly GB-NENs patients is due to coexisting diseases and weakened resistance to injury caused by tumor invasion, surgery, or adjuvant therapy [16, 17]. Therefore, age can be considered an important prognostic factor for GB-NENs patients.

In addition to personal characteristics, tumor biological characteristics may play an important role in the prognosis of GB-NENs patients. Tumor size and stage may indicate poorer survival of GB-NENs patients. The high stage means extensive tumor infiltration, reflecting the high malignancy and rapid progression of GB-NENs, which often leads to the involvement of adjacent important organs [23]. Furthermore, the expansion of the tumor infiltration area will make radical surgery difficult, then the probability of tumor recurrence after surgery will be greatly increased [24]. Our study further verified that tumor size and stage were the significant risk factors for GB-NENs patients receiving primary tumor surgery, which represented the inherent characteristics of tumor affecting prognosis. The GB-NENs are currently staged according to the same AJCC staging criteria as gallbladder adenocarcinoma, but the AJCC staging system may not be very suitable for neuroendocrine neoplasms [25, 26]. Therefore, the prediction models seem to be a very practical and effective tool, especially for GB-NEN patients receiving primary tumor surgery.

At present, surgery remains the cornerstone of treatment for localized tumors, and the systemic treatment choices for patients with advanced NENs have expanded considerably[1]. However, due to the lack of sufficient data, the treatment strategy for patients with GB-NENs is still controversial. Iype et al reported chemotherapy drugs, including cisplatin, carboplatin, and etoposide, perhaps lead to partial response and added a marginal advantage for patients with GB-NENs [27]. Similarly, Chorath et al reported a patient with high-grade gallbladder neuroendocrine carcinomas experiencing partial response to carboplatin, etoposide, nivolumab, and ipilimumab[28]. The role of radiotherapy in the treatment of GB-NENs is unclear since NENs are generally insensitive to traditional radiotherapy[6]. In our analysis, chemotherapy and radiotherapy were not identified as potential prognostic factors in patients with GB-NENs. But the results of univariate Cox regression analysis showed that chemotherapy had a negative effect on OS and CSS, while radiotherapy only had a positive effect on CSS. Thus, in the treatment of GB-NENs, chemotherapy may not be recommended for a routine postoperative adjuvant therapy, and radiotherapy was recommended when the condition permitted. But it is still important and necessary to clarify the efficacy of chemotherapy and radiotherapy in larger sample studies.

We developed and validated the individual nomograms for OS and CSS in patients with GB-NENs. This study has some advantages. First, to avoid heterogeneity between different medical institutions, a large sample dataset from the SEER database was combined with the sample data set from a single medical institution. Second, the variables incorporated in the nomograms are available and are often easily obtained in daily clinical practice. We also analyzed the effectiveness of chemotherapy and radiotherapy in the treatment of GB-NENs. The results showed that chemotherapy may negatively affect the prognosis of GB-NENs, and radiotherapy was recommended when necessary. Third, the nomograms had good calibration and discriminatory ability (the C-indexes for OS and CSS are 0.798 and 0.825 respectively). This can help us make better decisions in the actual clinical environment. Meanwhile, we plotted the DCA curves of the clinical impact of the nomograms, and the results indicated that our nomograms had greater clinical prediction performance than the SEER stage system. The limitation of this study lies in the retrospective nature with potential selective bias, which may not reflect problems encountered in actual clinical practice. In addition, our nomograms were constructed using only three clinicopathological characteristics, lacking other important variables, such as Ki-67, chromogranin A, and neuron-specific enolase[29, 30], which may reduce the prediction accuracy of our nomograms. Although the C-indexes, AUC, calibration plots, and DCAs had been applied to validate the nomograms, multi-center validation of large samples is still necessary. More work to enhance the validity of the prediction model is warranted.

This study constructed and validated the individualized nomograms to evaluate OS and CSS in patients with GB-NENs receiving primary tumor surgery. The nomograms used three priority variables to predict the prognosis in the actual clinical scenario and showed good prognostic prediction ability. In population-based analysis, the nomograms provide an individualized estimation of OS and CSS, which can be easily used by medical service providers to provide treatment decisions, prognosis suggestions, and clinical follow-up for patients and their caregivers.

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics statement

This retrospective study was approved by the Ethics Committee of Tongji Medical College, Huazhong University of Science and Technology (IORG number: IORG0003571) and was performed by the Declaration of Helsinki.

Consent for publication

Not applicable.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Authors' contributions

* Y. Z. and G.H. contributed equally to this work. H.S and R.L. designed and supervised the study and data analysis; Y.Z. and G.H. performed most of the data collection, analysis and wrote the manuscript; M.F. provided pathological assessment and analysis; H.Y. and C.J. provided help in data collection and analysis. All authors contributed to the article and approved the submitted version.

Funding

This study was supported by the National Natural Science Foundation of China (No. 81770539 and 81974068), the National key research and development program of China (No. 2017YFC0110003), and the Interdisciplinary Program of Wuhan National High Magnetic Field Center(Grant No.WHMFC202104), Huazhong University of Science and Technology.

Acknowledgements

The authors wish to acknowledge all participants in this study and everybody involved in the set-up and implementation of the study.

AUC, area under the curve

C-index, concordance index

CI, confidence interval

CPH, Cox proportional hazards

CSS, cancer-specific survival

DCA, decision curve analysis

GB-NENs, neuroendocrine neoplasms of the gallbladder

LN, lymph node

NEN, neuroendocrine neoplasm

NCDB, National Cancer Database

OS, overall survival

ROC, receiver operating characteristic

SEER, Surveillance, Epidemiology, and End Results

Cives, M. and J.R. Strosberg, Gastroenteropancreatic Neuroendocrine Tumors. CA Cancer J Clin, 2018. 68(6): p. 471–487.doi:10.3322/caac.21493.
Mafficini, A. and A. Scarpa, Genetics and Epigenetics of Gastroenteropancreatic Neuroendocrine Neoplasms. Endocr Rev, 2019. 40(2): p. 506–536.doi:10.1210/er.2018-00160.
Dasari, A., C. Shen, D. Halperin, et al., Trends in the Incidence, Prevalence, and Survival Outcomes in Patients With Neuroendocrine Tumors in the United States. JAMA Oncol, 2017. 3(10): p. 1335–1342.doi:10.1001/jamaoncol.2017.0589.
Yao, J.C., M. Hassan, A. Phan, et al., One hundred years after "carcinoid": epidemiology of and prognostic factors for neuroendocrine tumors in 35,825 cases in the United States. J Clin Oncol, 2008. 26(18): p. 3063–72.doi:10.1200/JCO.2007.15.4377.
O'Toole, D., R. Kianmanesh, and M. Caplin, ENETS 2016 Consensus Guidelines for the Management of Patients with Digestive Neuroendocrine Tumors: An Update. Neuroendocrinology, 2016. 103(2): p. 117–8.doi:10.1159/000443169.
Eltawil, K.M., B.I. Gustafsson, M. Kidd, et al., Neuroendocrine tumors of the gallbladder: an evaluation and reassessment of management strategy. J Clin Gastroenterol, 2010. 44(10): p. 687–95.doi:10.1097/MCG.0b013e3181d7a6d4.
Lee, S.M. and C.O. Sung, Neuroendocrine Carcinomas of the Gallbladder: A Clinicopathologic and Immunohistochemical Analysis of 34 Resected Cases. Am J Surg Pathol, 2020. 44(10): p. 1308–1321.doi:10.1097/PAS.0000000000001536.
Ayabe, R.I., M. Wach, S. Ruff, et al., Primary Gallbladder Neuroendocrine Tumors: Insights into a Rare Histology Using a Large National Database. Ann Surg Oncol, 2019. 26(11): p. 3577–3585.doi:10.1245/s10434-019-07440-6.
Balachandran, V.P., M. Gonen, J.J. Smith, et al., Nomograms in oncology: more than meets the eye. Lancet Oncol, 2015. 16(4): p. e173-80.doi:10.1016/S1470-2045(14)71116-7.
Wang, Y., J. Li, Y. Xia, et al., Prognostic nomogram for intrahepatic cholangiocarcinoma after partial hepatectomy. J Clin Oncol, 2013. 31(9): p. 1188–95.doi:10.1200/JCO.2012.41.5984.
Duggan, M.A., W.F. Anderson, S. Altekruse, et al., The Surveillance, Epidemiology, and End Results (SEER) Program and Pathology: Toward Strengthening the Critical Relationship. Am J Surg Pathol, 2016. 40(12): p. e94-e102.doi:10.1097/PAS.0000000000000749.
Wang, S., L. Yang, B. Ci, et al., Development and Validation of a Nomogram Prognostic Model for SCLC Patients. J Thorac Oncol, 2018. 13(9): p. 1338–1348.doi:10.1016/j.jtho.2018.05.037.
Vickers, A.J. and E.B. Elkin, Decision curve analysis: a novel method for evaluating prediction models. Med Decis Making, 2006. 26(6): p. 565–74.doi:10.1177/0272989X06295361.
Zhou, S., S. Jiang, W. Chen, et al., Biliary Neuroendocrine Neoplasms: Analysis of Prognostic Factors and Development and Validation of a Nomogram. Front Oncol, 2021. 11: p. 654439.doi:10.3389/fonc.2021.654439.
Wu, J., H. Zhang, L. Li, et al., A nomogram for predicting overall survival in patients with low-grade endometrial stromal sarcoma: A population-based analysis. Cancer Commun (Lond), 2020. 40(7): p. 301–312.doi:10.1002/cac2.12067.
Yadav, S., S.H. Tella, A. Kommalapati, et al., A Novel Clinically Based Staging System for Gallbladder Cancer. J Natl Compr Canc Netw, 2020. 18(2): p. 151–159.doi:10.6004/jnccn.2019.7357.
Hallet, J., C. Law, S. Singh, et al., Risk of Cancer-Specific Death for Patients Diagnosed With Neuroendocrine Tumors: A Population-Based Analysis. J Natl Compr Canc Netw, 2021: p. 1–10.doi:10.6004/jnccn.2020.7666.
Hu, P., J. Bai, M. Liu, et al., Trends of incidence and prognosis of gastric neuroendocrine neoplasms: a study based on SEER and our multicenter research. Gastric Cancer, 2020. 23(4): p. 591–599.doi:10.1007/s10120-020-01046-8.
Wang, Z.J., K. An, R. Li, et al., Analysis of 72 patients with colorectal high-grade neuroendocrine neoplasms from three Chinese hospitals. World J Gastroenterol, 2019. 25(34): p. 5197–5209.doi:10.3748/wjg.v25.i34.5197.
Miao, D.L., W. Song, J. Qian, et al., Development and Validation of a Nomogram for Predicting Overall Survival in Pancreatic NeuroendocrineTumors. Transl Oncol, 2018. 11(5): p. 1097–1103.doi:10.1016/j.tranon.2018.06.012.
Lin, Q. and W. Wagner, Epigenetic Aging Signatures Are Coherently Modified in Cancer. PLoS Genet, 2015. 11(6): p. e1005334.doi:10.1371/journal.pgen.1005334.
Cen, D., H. Liu, Z. Wan, et al., Clinicopathological features and survival for gallbladder NEN: a population-based study. Endocr Connect, 2019. 8(9): p. 1273–1281.doi:10.1530/EC-19-0124.
Shapera, E. and C. Bitting, Survival: a rare outcome in large cell neuroendocrine carcinoma of the gallbladder. Acta Gastroenterol Belg, 2019. 82(3): p. 433–436
Zheng, Z., C. Chen, B. Li, et al., Biliary Neuroendocrine Neoplasms: Clinical Profiles, Management, and Analysis of Prognostic Factors. Front Oncol, 2019. 9: p. 38.doi:10.3389/fonc.2019.00038.
Wang, W.Q., W.H. Zhang, H.L. Gao, et al., A novel risk factor panel predicts early recurrence in resected pancreatic neuroendocrine tumors. J Gastroenterol, 2021. 56(4): p. 395–405.doi:10.1007/s00535-021-01777-0.
Zhang, X.F., F. Xue, Z. Wu, et al., Development and Validation of a Modified Eighth AJCC Staging System for Primary Pancreatic Neuroendocrine Tumors. Ann Surg, 2020.doi:10.1097/SLA.0000000000004039.
Iype, S., T.A. Mirza, D.J. Propper, et al., Neuroendocrine tumours of the gallbladder: three cases and a review of the literature. Postgrad Med J, 2009. 85(1002): p. 213–8.doi:10.1136/pgmj.2008.070649.
Chorath, J., V.R. Placencio-Hickok, M. Guan, et al., Durable Response to Carboplatin, Etoposide, Nivolumab, and Ipilimumab in Metastatic High-Grade Neuroendocrine Carcinoma of the Gallbladder. Pancreas, 2020. 49(2): p. e19-e20.doi:10.1097/MPA.0000000000001480.
Hofland, J., W.T. Zandee, and W.W. de Herder, Role of biomarker tests for diagnosis of neuroendocrine tumours. Nat Rev Endocrinol, 2018. 14(11): p. 656–669.doi:10.1038/s41574-018-0082-5.
Clay, V., G. Papaxoinis, B. Sanderson, et al., Evaluation of diagnostic and prognostic significance of Ki-67 index in pulmonary carcinoid tumours. Clin Transl Oncol, 2017. 19(5): p. 579–586.doi:10.1007/s12094-016-1568-z.

Table 1: Demographic and clinicopathological characteristics of patients with GB-NENs receiving primary tumor surgery.

Variables	ALL (N=221)	Training dataset (N = 154)	Internal validation dataset (N = 67)	P-value
Gender [N (%)]
Female	143 (64.7)	96 (62.3)	47 (70.1)	0.3352
Male	78 (35.3)	58 (37.7)	20 (29.9)
Race [N (%)]
White	175 (79.2)	122 (79.2)	53 (79.1)	0.4525
Black	30 (13.6)	19 (12.3)	11 (16.4)
Other	16 (7.2)	13 (8.4)	3 (4.5)
Age at diagnosis [N (%)]
≤65 years	119 (53.8)	82 (53.2)	37 (55.2)	0.9012
>65 years	102 (46.2)	72 (46.8)	30 (44.8)
Marital status [N (%)]
Married	146 (66.1)	103 (66.9)	43 (64.2)	0.8137
Single	75 (33.9)	51 (33.1)	24 (35.8)
Pathological Classification [N (%)]
NET	85 (38.5)	57 (37.0)	28 (41.8)	0.7961
NEC	90 (40.7)	64 (41.6)	26 (38.8)
Unknown	46 (20.8)	33 (21.4)	13 (19.4)
Histology [N (%)]
Well-differentiated	95(43.0)	64 (41.6)	31 (46.3)	0.6155
Poorly differentiated	126(57.0)	90 (58.4)	36 (53.7)
N stage [N (%)]
N0	150 (67.9)	104 (67.5)	46 (68.7)	0.697
N1	49 (22.2)	36 (23.4)	13 (19.4)
Unknown	22 (10.0)	14 (9.1)	8 (11.9)
M Stage [N (%)]
M0	161 (72.9)	114 (74.0)	47 (70.1)	0.694
M1	42 (19.0)	27 (17.5)	15 (22.4)
Unknown	18 (8.1)	13 (8.4)	5 (7.5)
Tumor size [N (%)]
≤2cm	92 (41.6)	62 (40.3)	30 (44.8)	0.3148
>2cm	82 (37.1)	55 (35.7)	27 (40.3)
Unknown	47 (21.3)	37 (24.0)	10 (14.9)
Stage [N (%)]
Localized	95 (43.0)	67 (43.5)	28 (41.8)	0.4248
Regional	72 (32.6)	53 (34.4)	19 (28.4)
Distant	54 (24.4)	34 (22.1)	20 (29.9)
Surgery at other sites [N (%)]
No	191 (86.4)	134 (87.0)	61 (91.0)	0.5301
Yes	30 (13.6)	20 (13.0)	6 (9.0)
LN surgery [N (%)]
No	147 (66.5)	104 (67.5)	43 (64.2)	0.7411
Yes	74 (33.5)	50 (32.5)	24 (35.8)
Chemotherapy [N (%)]
No	148 (67.0)	104 (67.5)	44 (65.7)	0.9086
Yes	73 (33.0)	50 (32.5)	23 (34.3)
Radiotherapy [N (%)]
No	191 (86.4)	132 (85.7)	59 (88.1)	0.7993
Yes	30 (13.6)	22 (14.3)	8 (11.9)

Table 2: Univariate and multivariate Cox regression analyses of OS and CSS in training set.

Variables	OS				CSS
	Univariate analysis		Multivariate analysis		Univariate analysis		Multivariate analysis
	HR (95%CI)	P	HR (95%CI)	P	HR (95%CI)	P	HR (95%CI)	P
Gender
Female	Ref.		-		Ref.		-
Male	1.03(0.648,1.64)	0.895			0.866(0.509,1.472)	0.594
Race
White	Ref.		-		Ref.		-
Black	0.759(0.347,1.66)	0.489			0.512(0.185,1.418)	0.198
Others	0.985(0.449,2.16)	0.969			0.795(0.317,1.994)	0.625
Age at diagnosis
≤65 years	Ref.				Ref.
>65 years	2.942(1.825,4.742)	<0.001	2.806(1.785-4.412)	<0.001	2.826(1.660,4.811)	<0.001	2.501(1.507-4.149)	<0.001
Marital status
Married	Ref.		-		Ref.
Single	0.568(0.337,0.957)	0.0337			0.426(0.226,0.803)	0.00834
Pathological classification
NET	Ref.				Ref.
NEC	4.176(2.357,7.400)	<0.001	1.638(0.969-2.769)	0.0655	4.667(2.445,8.905)	<0.001	1.696(0.934-3.077)	0.0825
Unknown	1.001(0.469,2.137)	0.998	0.776(0.399-1.510)	0.455	0.907(0.357,2.307)	0.838	0.996(0.446-2.223)	0.992
Histology
Well-differentiated	Ref.				Ref.
Poorly differentiated	5.139(2.848,9.272)	<0.001	0.708(0.355-1.411)	0.326	9.541(4.094,22.235)	<0.001	0.998(0.421-2.368)	0.997
N stage
N0	Ref.				Ref.
N1	2.678(1.641,4.371)	<0.001	1.038(0.645-1.670)	0.879	3.159(1.860,5.364)	<0.001	1.093(0.652-1.835)	0.735
Unknown	0.635(0.196,2.062)	0.45	2.162(0.756-6.183)	0.151	0.770(0.235,2.523)	0.665	2.881(0.974-8.526)	0..0559
M stage
M0	Ref.				Ref.
M1	4.602(2.741,7.727)	<0.001	0.685(0.312-1.504)	0.346	4.723(2.704,8.250)	<0.001	0.616(0.273-1.390)	0.244
Unknown	0.465(0.112,1.930)	0.291	0.708(0.0744-1.193)	0.0871	0.529(0.127,2.209)	0.383	0.238(0.0559-1.016)	0.0526
Tumor size
≤2cm	Ref.				Ref.
>2cm	6.375(3.305,12.298)	<0.001	1.991(1.100-3.602)	0.0228	15.411(5.462,43.479)	<0.001	3.462(1.578-7.595)	0.002
Unknown	5.089(2.535,10.219)	<0.001	3.887(2.105-7.176)	<0.001	12.597(4.312,36.797)	<0.001	6.710(2.964-15.192)	<0.001
SEER Stage
Localized	Ref.				Ref.
Regional	4.621(2.384,8.960)	<0.001	3.531(1.668-7.475)	0.001	6.287(2.724,14.512)	<0.001	4.038(1.539-10.596)	0.004
Distant	13.894(6.910,27.934)	<0.001	13.971(5.005-38.998)	<0.001	19.152(8.137,45.076)	<0.001	18.340(4.434-61.905)	<0.001
Surgery at other sites
No	Ref.				Ref.		-
Yes	1.59(0.854,2.97)	0.144			1.718(0.891,3.312)	0.106
LN surgery
No	Ref.		-		Ref.		-
Yes	0.926(0.570,1.505)	0.756			1.242(0.737,2.091)	0.416
Chemotherapy
No	Ref.				Ref.
Yes	2.827(1.778,4.494)	<0.001	1.029(0.642-1.651)	0.905	3.462(2.067,5.799)	<0.001	1.129(0.657-1.942)	0.660
Radiotherapy
No	Ref.		-		Ref.
Yes	1.62(0.916,2.86)	0.0974			1.867(1.025,3.399)	0.0414	0.804(0.441-1.467)	0.477

No competing interests reported.

TableS1.docx

Nomograms for predicting survival outcomes in patients with neuroendocrine neoplasms of the gallbladder undergoing primary tumor resection: a population-based study

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Figures

Introduction

Methods

Results

1: Patient characteristics

2: Identification Of Prognostic Factors

3: Construction And Validation Of The Nomograms

4: Clinical Application Of The Nomograms

Discussion

Conclusions

Declarations

List Of Abbreviations

References

Tables

Additional Declarations

Supplementary Files

Status:

Version 1