Home
JournalsCollections
For Authors For Reviewers For Editorial Board Members
Article Processing Charges Open Access
Ethics Advertising Policy
Editorial Policy Resource Center
Company Information Contact Us
OPEN ACCESS

A Nomogram-based Model to Predict Neoplastic Risk for Patients with Gallbladder Polyps

  • Xudong Zhang1,2,# ,
  • Jincheng Wang2,# ,
  • Baoqiang Wu1,
  • Tao Li1,
  • Lei Jin1,
  • Yong Wu1,
  • Peng Gao3,
  • Zhen Zhang3,
  • Xihu Qin1,2,*  and
  • Chunfu Zhu1,* 
Journal of Clinical and Translational Hepatology   2022;10(2):263-272

doi: 10.14218/JCTH.2021.00078

Received:

Revised:

Accepted:

Published online:

 Author information

Citation: Zhang X, Wang J, Wu B, Li T, Jin L, Wu Y, et al. A Nomogram-based Model to Predict Neoplastic Risk for Patients with Gallbladder Polyps. J Clin Transl Hepatol. 2022;10(2):263-272. doi: 10.14218/JCTH.2021.00078.

Abstract

Background and Aims

Gallbladder polyp (GBP) assessment aims to identify the early stages of gallbladder carcinoma. Many studies have analyzed the risk factors for malignant GBPs. In this retrospective study, we aimed to establish a more accurate predictive model for potential neoplastic polyps in patients with GBPs.

Methods

We developed a nomogram-based model in a training cohort of 233 GBP patients. Clinical information, ultrasonographic findings, and blood test findings were analyzed. Mann-Whitney U test and multivariate logistic regression analyses were used to identify independent predictors and establish the nomogram model. An internal validation was conducted in 225 consecutive patients. Performance and clinical benefit of the model were evaluated using receiver operating characteristic curves and decision curve analysis (DCA), respectively.

Results

Age, cholelithiasis, carcinoembryonic antigen, polyp size, and sessile shape were confirmed as independent predictors of GBP neoplastic potential in the training group. Compared with five other proposed prediction methods, the established nomogram model presented better discrimination of neoplastic GBPs in the training cohort (area under the curve [AUC]: 0.846) and the validation cohort (AUC: 0.835). DCA demonstrated that the greatest clinical benefit was provided by the nomogram compared with the other five methods.

Conclusions

Our developed preoperative nomogram model can successfully be used to evaluate the neoplastic potential of GBPs based on simple clinical variables that maybe useful for clinical decision-making.

Keywords

Gallbladder polyps, Neoplastic polyp, Preoperative diagnosis, Nomogram model

Introduction

Gallbladder polyps (GBPs) are elevated lesions that protrude from the gallbladder wall into the lumen, with a prevalence of 5–10% in the general population.1 In recent years, the diagnosis of GBPs has increased because of widespread use of abdominal ultrasonography.2 GBPs are categorized broadly as non-neoplastic (pseudopolyps) and neoplastic (true) polyps. Approximately 70% of GBPs are benign (without malignant tendencies) and are represented by cholesterol, focal adenomyomatosis, and inflammatory pseudopolyps.3 True polyps can present as benign (most commonly adenomas) or malignant adenocarcinomas or metastases. However, an estimated 3% of GBPs are true polyp adenomas that have malignant potential.4

There are various imaging modalities for GBP assessment, such as endoscopic ultrasonography, magnetic resonance imaging, and computed tomography (CT). However, preoperative diagnosis of malignant polyps remains difficult.5 Because of a lack of clinical trials, there are no universally convincing indications for surgery. Considering the rapid progression and poor prognosis of gallbladder carcinoma (GBC), cholecystectomy is generally suggested for GBPs with malignant potential. Current guidelines for management of GBPs mainly focus on polyp size, and cholecystectomy is recommended when polyp diameter is >10 mm.6 However, previous studies have demonstrated that polyp number and shape, patient age, and sessile features are also high-risk factors for GBP malignancy.7,8

A considerable number of patients who underwent cholecystectomy in accordance with GBP management guidelines were shown to have non-neoplastic polyps.9 These patients experienced unnecessary surgical risks and economic burdens. Moreover, incidental GBC in cases with polyps <10 mm have been reported.10,11 Therefore, it is necessary to analyze other preoperative clinical characteristics and ultrasound findings that may be used to integrate variables with greater predictive value for GBP management.12 The aim of this study was to develop a noninvasive, preoperative prediction model for assessing the malignancy risk of GBPs.

Methods

Patients

We reviewed the medical records of 573 patients diagnosed with GBPs by ultrasonography in our hospital between January 2015 and September 2020. Exclusion criteria were: (1) preoperative diagnosis of GBC with liver metastasis (n=31); (2) non-recent examination results (>3 months) (n=8); (3) non-operative treatment (n=32); (4) lack of tumor markers (n=19); (5) lack of polyp characteristics (n=15); (6) lack of lipid tests (n=2); and (7) patients who received emergency surgery (e.g., for acute purulent cholecystitis and severe jaundice) (n=8). After exclusions, 458 cases were included in this study (Fig. 1), and 233 of these patients from between January 2015 and June 2018 were allocated to the training cohort. The remaining 225 patients from July 2018 to September 2020 were included in the validation cohort. This retrospective study was approved by the Institutional Review Board of The Affiliated Changzhou No. 2 People’s Hospital of Nanjing Medical University, Jiangsu, China (approval number [2020]KY222-01). The requirement for written informed consent from the patients was waived because of the retrospective nature of this study. However, at the time of treatment, all patients were informed of the GBP management guidelines and possible surgical risks.

Patient selection flowchart.
Fig. 1  Patient selection flowchart.

GBPs, gallbladder polyps; GBC, gallbladder carcinoma.

Clinical data and pathological diagnoses were collected from medical records. Clinical characteristics included age, sex, body mass index, and the presence of hypertension, diabetes mellitus, fatty liver, or viral hepatitis. Laboratory measurements included white blood cell counts and blood levels of alanine transaminase, total bilirubin, direct bilirubin, triglycerides, total cholesterol, total bile acids, γ-glutamyl transferase, lactic dehydrogenase, and D-dimer. We included other blood measurements that are associated with malignancy, including alpha-fetoprotein (AFP), carcinoembryonic antigen (CEA), and carbohydrate antigen 199 (CA199).13 The pathological diagnoses were categorized as carcinoma (n=41), adenoma with atypical hyperplasia (n=15), adenoma (n=38), inflammatory polyps (n=4), cholesterol polyps (n=313), adenomyomatosis (n=36), or mixed pseudopolyps (n=11). In accordance with the guidelines and malignant risk of adenomas, we classified patients who were diagnosed with neoplastic polyps (carcinoma, adenoma with atypical hyperplasia, and adenoma) into the group with indications for surgery.

Ultrasonography and laboratory analysis of routine blood tests were performed within 1 week before surgery. Clinical symptoms of cholelithiasis were recorded, including abdominal pain, bile reflux gastritis, and jaundice. Polyp characteristics were scored according to the abdominal ultrasonography, which identified gallstones and size and shape of the polyps. The number of polyps was categorized as either single or multiple. For multiple polyps, the size of the largest was recorded. The shape of the polyp was classified as sessile or pedunculated. The threshold for thickening of the gallbladder wall was set at 5 mm.14 Ultrasonic diagnoses were derived from the preoperative diagnosis reports from experienced sonographers.

Validation

The performance of the model was subsequently tested in the independent validation cohort by using the formula and cutoff values derived from the training cohort. Model performance was compared with the ultrasonic report diagnosis (US-reported),15 guidelines from the Japanese Society of Hepato-Biliary-Pancreatic Surgery (JSHBPS),16 European Society of Gastrointestinal and Abdominal Radiology (ESGAR),7 and Chinese Committee of Biliary Surgeons (CCBS),17 and the Korean scoring model.18 These methods, derived from guidelines and previous studies, are summarized in Table 1. Details of these guidelines are described in the Supplemental File 1.

Table 1

Methods derived from guidelines and previous studies

Guideline or modelInstructions for surgical indications
JSHBPSSessile gallbladder polyp and diameter ≥10 mm.
ESGARGallbladder polyps ≥10 mm, polyps <10 mm but patient have symptoms that are attributable to the gallbladder (cholelithiasis or inflammation), polyps 6∼9 mm with risk factors (age >50 years, primary sclerosing cholangitis, Indian ethnicity, or sessile)
CCBSDiameter ≥10 mm, combined gallbladder stones or cholecystitis, single or sessile polyps, with fast growth rate (growth rate >3 mm/ 6 months), adenomatous polyps
US-reportedBased on the size (>10 mm), gallbladder wall thickening (>4 mm), echo intensity (inhomogeneous), procellaneous gallbladder, shape of the polyp and boundary with the surrounding tissues (irregular), diagnosis made by experienced sonologists.
Korean ModelPS (predictive score) = −7.3633 + 0.0374*[age] + 0.6667*[polyp number] + 1.5784*[sessile] + 0.2189*[polyp size]. Probability of neoplastic GBP = ePS / (1 + ePS), where e = 2.7182.

Statistical analysis

Categorical and continuous variables were compared using the χ2 and Mann-Whitney U tests, respectively. We performed the χ2 test or Mann-Whitney U test to determine the variables with significant differences between the training and validation groups, for inclusion in subsequent multivariate logistic regression analysis. Variables with a p-value <0.05 in multivariate logistic regression analysis were identified as independent factors. Based on the β coefficient of each variable, the prediction model was demonstrated by the nomogram. All statistical analyses were performed by R software (version 3.5.1, http://www.r-project.org ). The diagnostic performance of the model was evaluated using receiver-operating characteristic (ROC) curve and area under the curve (AUC) analyses. The Delong test was used to compare AUC values. Decision curve analysis (DCA) was performed to calculate the net benefit from the use of the model at different threshold probabilities.19 A p <0.05 was considered statistically significant.

Results

Baseline characteristics

The baseline characteristics of the patients in the training and validation groups are shown in Table 2. There were no significant differences in pathological markers between these two groups. The incidences of GBPs with neoplastic potential were 19.3% and 21.8% in the training and validation cohorts, respectively. These rates suggest that there was considerable nonessential surgery based on the guidelines for GBP surgical management. No significant differences were found for clinical or ultrasonographic characteristics between the two groups.

Table 2

Baseline characteristics of patients included in this study

Baseline characteristicsTraining, n=233Validation, n=225p
Age in years, mean±SD49.47±13.5349.11±14.110.781
Sex0.658
  Male106107
  Female127118
Physical condition
  BMI (kg/m2)24.14±3.1324.05±3.140.763
  Diabetes, n (%)21 (9)14 (6.2)0.261
  Fatty liver, n (%)50 (21.5)45 (20.0)0.701
  Cholelithiasis, n (%)53 (22.7)56 (24.9)0.591
  Viral hepatitis, n (%)13 (5.6)8 (3.6)0.301
Laboratory findings
  DD (mg/L)0.61±1.780.46±0.840.257
  ALT (U/L)24.71±21.9627.60±27.000.209
  TBil (µmol/L)13.82±9.4914.85±19.370.470
  Triglycerides (mmol/L)1.52±1.051.64±1.020.213
  TCH (mmol/L)4.61±0.984.62±0.980.852
  TBA (µmol/L)5.47±10.725.44±11.890.975
  GGT (U/L)33.47±36.6548.0±115.570.068
Tumor markers
  AFP (ng/mL)2.81±2.122.75±1.630.708
  CEA (ng/mL)2.12±2.142.22±2.650.657
  CA199 (U/mL)20.57±80.8928.1±109.140.401
Ultrasonic diagnosis0.517
  Malignant or suspected, n (%)56 (24.0)60 (26.7)
  Benign, n (%)177 (76.0)165 (73.3)
Polyp characters
  Polyp size (mm)9.60±5.109.83±6.690.679
  Single polyp, n (%)115 (49.4)112 (49.8)0.928
  Sessile polyp, n (%)84 (36.1)83 (36.9)0.852
  GBWT, n (%)87 (37.3)89 (39.6)0.626
  Clinical symptoms, n (%)95 (40.8)76 (33.8)0.122
Neoplastic polyps, n (%)45 (19.3)49 (21.8)0.514

Risk factors for neoplastic GBPs

In the training cohort, age, diabetes, cholelithiasis, CEA, CA199, ultrasonic diagnosis, polyp size, number, sessile shape, and clinical symptoms were predictive clinical and imaging variables for neoplastic GBPs (p<0.05) (Table 3). Multivariate conditional logistic regression analysis identified age, cholelithiasis, CEA levels, polyp size, and sessile shape as independent factors that were associated with neoplastic GBP risk (Table 4). According to ROC curve analysis, we determined that the optimal cut-off values for age and CEA were 58 years and 1.56 ng/mL, respectively. Most of the management guidelines used a polyp diameter of 10 mm as a positive indicator of neoplastic risk. However, our ROC cutoff value for polyp diameter was 15 mm. Therefore, we defined both the 10-mm and 15-mm polyp diameters as cutoff points for a three-way classification in the nomogram as described below.

Table 3

Comparison between neoplastic polyp and pseudopolyps (non-neoplastic) in the training cohort

CharacteristicsNeoplastic, n=45Pseudopolyps, n=188p
Age in years, mean±SD57.49±13.5347.55±12.84*<0.001
Sex0.875
  Male20 (44.4)86 (45.7)
  Female25 (55.6)102 (54.3)
Physical condition
  BMI (kg/m2)24.18±3.3224.13±3.100.829
  Diabetes, n (%)8 (17.8)13 (6.9)*0.023
  Fatty liver, n (%)8 (17.8)42 (22.3)0.504
  Cholelithiasis, n (%)23 (51.1)30 (16.0)*<0.001
  Viral hepatitis, n (%)4 (8.9)9 (4.8)0.283
Laboratory findings
  DD (mg/L)1.06±2.080.52±1.720.071
  ALT (U/L)24.39±18.9724.78±22.660.727
  TBil (µmol/L)15.61±14.9213.38±7.650.842
  Triglyceride(mmol/L)1.64±1.001.49±1.060.333
  TCH (mmol/L)4.44±0.804.65±1.010.193
  TBA (µmol/L)8.97±22.964.63±3.860.626
  GGT (U/L)39.29±47.8332.07±33.430.853
Tumor markers
  AFP (ng/mL)3.25±3.312.70±1.720.511
  CEA (ng/mL)3.61±4.041.76±1.10*<0.001
  CA199 (U/mL)59.82±178.6811.17±12.11*0.001
Ultrasonic diagnosis*<0.001
  Malignant or suspected, n (%)22 (51.1)34 (18.1)
  Benign, n (%)23 (48.9)154 (81.9)
Polyp characters
  Polyp size (mm)13.93±8.308.56±3.24*<0.001
  Single polyp, n (%)29 (64.4)86 (45.7)*0.025
  Sessile polyp, n (%)31 (68.9)53 (28.2)*<0.001
  GBWT, n (%)18 (40.0)69 (36.7)0.681
  Clinical symptoms, n (%)26 (57.8)69 (36.7)*0.010
Table 4

Factors for the prediction of neoplastic risk for patients with gallbladder polyps

VariablesMultivariate analysis
ROC analysis
βORpAUCCutoff
Age in years0.0421.043 (1.010, 1.077)0.0090.685 (0.598, 0.772)58
DiabetesNANA0.39NANA
Cholelithiasis1.062.887 (1.192, 6.993)0.0190.676 (0.581, 0.771)NA
CEA (ng/mL)0.351.420 (1.052, 1.915)0.0220.707 (0.625, 0.789)1.56
CA199 (U/mL)NANA0.573NANA
Ultrasonic diagnosisNANA0.436NANA
Polyp size (mm)0.151.162 (1.047, 1.289)0.0050.707 (0.617, 0.797)15
Single polypNANA0.264NANA
Sessile polyp1.0452.843 (1.209, 6.684)0.0170.703 (0.617, 0.790)NA
Clinical symptomsNANA0.926NANA

Development and validation of the prediction nomogram

Using the results of the univariate and multivariate analyses, we developed a nomogram that incorporated the preoperative predictive variables for neoplastic risk in patients with GBPs. The scoring points for the nomogram are shown in Figure 2A for age (0, ≤58 years; 1, >58 years), cholelithiasis (0, negative; 1, positive), CEA (0, ≤1.56 ng/mL; 1, >1.56 ng/mL), polyp size (0, <10 mm; 1, ≥10 mm and ≤15 mm; 2, >15 mm) and sessile shape (0, pedunculated; 1, sessile). The formula for the weighted value was: Y=1.194 × [age] + 1.177 × [cholelithiasis] + 1.171 × [CEA] + 1.112 × [polyp size] + 1.066 × [sessile] − 3.944.

Developed nomogram presented with ROC.
Fig. 2  Developed nomogram presented with ROC.

(A) The nomogram was established due to the training cohort, with age, cholelithiasis, CEA, polyp size and sessile incorporated. (B) Comparison of ROC curves between our model, US-reported, JSHBPS guideline, ESGAR guideline, CCBS guideline, and Korean model in the training and validation. ROC, receiver operating characteristic; CEA, carcinoembryonic antigen; US-reported, ultrasonic report diagnosis.

The nomogram achieved an overall accuracy rate of 84.1%, with a sensitivity and specificity of 68.1% and 88.2%, respectively. Among the 30 false negative cases, only 1 case was GBC. We plotted the ROC curves to compare the discrimination abilities among our model, the US-reported model, the JSHBPS, ESGAR, and CCBS guidelines, and the Korean scoring model described above. As shown in Figure 2B and summarized in Table 5, the greatest discrimination ability, as demonstrated by the AUC, was observed in our nomogram model in both the training (AUC: 0.846) and validation (AUC: 0.835) cohorts compared with the US-reported alone, the JSHBPS, ESGAR, and CCBS guidelines, and the Korean model.

Table 5

Diagnostic performances of all methods and independent factors for GBPs in the training and validation cohort

Training, n=233Validation, n=225Training vs. validation
MethodsAUROC (95% CI)AUROC (95% CI)Delong test
  Nomogram model0.846 (0.779, 0.913)0.835 (0.765, 0.905)p=0.826
  US-reported0.639 (0.561, 0.717)0.659 (0.603, 0.716)p=0.683
  JSHBPS guideline0.613 (0.544, 0.682)0.635 (0.569, 0.702)p=0.642
  ESGAR guideline0.591 (0.513, 0.670)0.617 (0.561, 0.672)p=0.606
  CCBS guideline0.632 (0.565, 0.699)0.658 (0.598, 0.717)p=0.573
  Korean model0.753 (0.670, 0.836)0.746 (0.663, 0.828)p=0.901
  Age0.685 (0.598, 0.772)0.720 (0.636, 0.804)p=0.569
  Cholelithiasis0.676 (0.581, 0.771)0.693 (0.603, 0.784)p=0.755
  CEA0.707 (0.625, 0.789)0.648 (0.560, 0.736)p=0.336
  Polyp size0.707 (0.617, 0.797)0.749 (0.659, 0.839)p=0.519
  Sessile polyp0.703 (0.617, 0.790)0.708 (0.624, 0.792)p=0.937
Delong test (comparison of AUROC)
  Model vs. US-reportedp<0.001p<0.001
  Model vs. JSHBPSp<0.001p<0.001
  Model vs. ESGARp<0.001p<0.001
  Model vs. CCBSp< 0.001p<0.001
  Model vs. Korean modelp=0.010p=0.007
  Model vs. Agep<0.001p=0.004
  Model vs. Cholelithiasisp< 0.001p< 0.001
  Model vs. CEAp=0.001p<0.001
  Model vs. Polyp sizep=0.001p=0.013
  Model vs. Sessile polypp<0.001p=0.003

To further evaluate and compare these prediction models or guidelines, we determined the net benefits of each using DCA (Fig. 3). Across a reasonable threshold of probability ranges for both the training and validation groups, DCA graphically showed that the nomogram provided greater clinical benefit for predicting malignancy in patients with GBPs than the other methods.

DCA for each prediction method in the training (A) and validation (B) dataset.
Fig. 3  DCA for each prediction method in the training (A) and validation (B) dataset.

The y-axis measures the net benefit. DCA, decision curve analysis; US-reported, ultrasonic report diagnosis.

Discussion

This study established and validated a nomogram model for predicting neoplastic polyps in patients with GBPs. Age, cholelithiasis, serum CEA levels, polyp size, and sessile shape were confirmed as independent predictors for neoplastic risk and integrated into the nomogram model. Subsequently, our model achieved significantly better diagnostic performance and provided more clinical benefit, as demonstrated by ROC and DCA curves, compared with the US-reported model, three different management guidelines, and a Korean scoring model.

We discovered that less than 20% of GBP patients actually required surgery. There is a selection bias for cases that are chosen for inpatient surgery because many patients have cholesterol polyps that do not require surgical intervention. Therefore, the incidence of malignant polyps may be lower than that observed in our study. Greater than 50% of patients included in our study presented with indications for surgery following the guidelines. In a retrospective study, Metman et al.20 determined that the prevalence of neoplastic polyps was much lower than reported and questioned the broad recommendations in the guidelines. From these data, it is clear that more accurate preoperative assessments of GBPs are necessary.

Our nomogram model achieved satisfactory accuracy, good reliability, and reproducibility. The factors included in our final model, such as age, cholelithiasis, polyp size, and sessile shape have been reported as risk factors for gallbladder cancer in other studies.21–23 The predictive effects of serum CEA and CA199 levels have also been demonstrated.24 However, we established a prediction system using a nomogram that integrated a combination of ultrasonic signatures and physiological and tumor markers. The effectiveness of the three guidelines (JSHBPS, ESGAR, and CCBS) for predicting GBP malignancy was similar. The Korean model was more effective than these guidelines but slightly less effective than our model.

In recent years, clinical studies of GBPs have surged. For example, Velidedeoğlu et al.25 expressed doubt about the necessity for cholecystectomy in patients with symptomatic GBPs without first conducting extensive preoperative tests. Zhao et al.26 indicated that dyslipidemia was associated with GBP formation and found that the ratio of non-high density lipoprotein cholesterol to high density lipoprotein cholesterol was an independent factor associated with high risk for GBP formation in Chinese men. Furthermore, fatty liver was found to be an independent risk factor for GBPs.27 Onda et al.28 developed a preoperative scoring system for GBC based on age, the presence of gallstones, polyp size, and solitary and sessile polyps based on ultrasonography and CT scans. In comparison with serum biomarkers, enhanced CT is more sensitive for detecting tumors; however, it is a more expensive method and exposes the patient to radiation. However, this latter study only found two independent risk factors (age and polyp size) for predicting malignant GBPs.

We developed a non-invasive and user-friendly model for predicting malignant GBPs based on easily available data. Not only diagnostic performance but also cost and applicability should be considered. Each of the indicators included in our model can be obtained through an outpatient physical examination. Nomogram modelling has been used effectively in a number of studies.29,30 We recommend that patients who are judged to be at high risk using our diagnostic model should have supplemental CT scans before surgery to confirm the diagnosis and rule-out abdominal metastasis of GBC. Additionally, compared with artificially assigning risk factors, assigning corresponding weights to variables through statistical methods may result in more objective extraction of information from clinical data.

Unlike the adenoma-carcinoma sequence that is well described for colonic polyps, the adenoma-carcinoma sequence for GBPs is not well understood. One study has shown a link between the presence of proximal colon polyps and higher rates of GBPs.31 The evidence suggests that at least some gallbladder adenocarcinomas may have arisen from pre-existing adenomas and atypical hyperplasia of gallbladder adenoma may be a precancerous lesion.32 If GBC is confined to the connective tissue of the gallbladder wall (stage I and II), the 5-year survival rates are more favorable at 57–92%.33 Therefore, early detection and management of GBC is critical. Considering the malignant tendency of gallbladder adenoma and the recommendations of the guidelines, we included adenomas in the recommended cholecystectomy group in our model. Consequently, most of the false-negative cases detected by our model were adenomas. Prior to malignant transformation of gallbladder adenoma, their growth characteristics are different from that of malignant polyps. Currently, the diagnosis of malignancy can only be confirmed by postoperative pathology. If the number of cases is further expanded, attempts can be made to distinguish the polyps that are early malignant adenomas, and the accuracy of GBP management and study of malignant transformation of GBPs may be improved.

In the process of data collection and analysis, we noted certain risk factors for the development of malignant GBPs that have been less recognized. For example, Spearman correlation analysis indicated that diabetes and CA199 were risk factors for malignant GBPs. Systematic reviews have indicated that patients with diabetes had an increased risk of GBC and a higher GBC-mediated mortality compared with non-diabetic individuals.34,35 In addition to CEA and CA199,36 we found that CA724 may be a potential biomarker for GBC. However, in this study, there were too many missing data points, and we are prospectively collecting relevant results to obtain stronger evidence for CA724 as a potential biomarker. Given that our model is simple and easy to understand, we have created a clinical electronic software program for the nomogram to promote it to the public (https://smartglass.nextreal.cn/web/h5/psfnrogp/dev/ ), so that GBP patients can follow-up by themselves and receive accurate and detailed clinical recommendations.

Several limitations in this study should be noted. First, inherent selection biases could not be avoided due to the retrospective nature of this study. The enrolled patients underwent cholecystectomy because of the possibility of malignancy; thus, many patients who were thought to have benign polyps did not undergo surgery and were excluded from this study. Moreover, due to the low incidence of GBC, the total number of positive cases included in this study was low. Second, the accuracy of an ultrasound diagnosis is highly dependent on the experience level of the operator. Incorporating novel specific tumor indicators in a prospective study, such as CA724 or texture analysis of ultrasound signals, may further improve the accuracy of the model. Furthermore, this nomogram was established and validated on the basis of data obtained from a single center. Recognized risk factors, such as Indian ethnicity and primary sclerosing cholangitis, were not examined in this study. We shared this model to increase recognition of risk factors for GBP malignancy and promote cooperation in multi-center prospective research to externally validate our model.

Conclusions

We present an accurate and user-friendly prediction model based on simple clinical variables to improve diagnosis of neoplastic polyps in patients with GBPs. Furthermore, the model facilitates greater accuracy of surgical decisions by both surgeons and patients and may aid in the early diagnosis and treatment of GBC.

Supporting information

Supplemental File 1

Supplemental materials and methods.

(DOCX)

Abbreviations

AFP: 

alpha-fetoprotein

ALT: 

alanine transaminase

AUC: 

area under the curve

BMI: 

body mass index

CA199: 

carbohydrate antigen 199

CCBS: 

Chinese Committee of Biliary Surgeons

CEA: 

carcinoembryonic antigen

CI: 

confidence interval

CT: 

computed tomography

DBil: 

direct bilirubin

DCA: 

decision curve analysis

DD: 

D-dimer

ESGAR: 

European Society of Gastrointestinal and Abdominal Radiology

GBC: 

gallbladder carcinoma

GBPs: 

gallbladder polyps

GBWT: 

gallbladder wall thickening

GGT: 

γ-glutamyl transferase

JSHBPS: 

Japanese Society of Hepato-Biliary-Pancreatic Surgery

LDH: 

lactic dehydrogenase

ROC: 

receiver-operating characteristic curve

TBA: 

total bile acid

TBil: 

total bilirubin

TCH: 

total cholesterol

US-reported: 

ultrasonic report diagnosis

Declarations

Acknowledgement

We thank Xuecun Huang (Department of Ultrasonography, The Affiliated Changzhou No. 2 People’s Hospital of Nanjing Medical University) for his technical consulting service, and Susan Zunino, PhD, from Edanz for editing the language.

Ethical statement

This retrospective study was approved by the Hospital Research Ethics Committee ([2020]KY222-01). The requirement for written informed consent was waived due to its retrospective nature.

Data sharing statement

The datasets analyzed during the current study are available from the corresponding authors on reasonable request.

Funding

This study was supported by the National Natural Science Foundation of China (81702323 and 81672469), the Changzhou Medical Innovation Team (CCX201807), and the Changzhou Sci & Tech Program (CE20165020).

Conflict of interest

The authors have no conflict of interests related to this publication.

Authors’ contributions

Collected the data (BW, XZ, YW, ZZ, PG), analyzed the data (JW, LT), participated in research design (LJ, XQ), wrote the manuscript (XZ, JW), supervised the study (XQ), and revised the paper (CZ). All authors read and approved the final manuscript.

References

  1. Xu A, Hu H. The gallbladder polypoid-lesions conundrum: moving forward with controversy by looking back. Expert Rev Gastroenterol Hepatol 2017;11(11):1071-1080 View Article PubMed/NCBI
  2. McCain RS, Diamond A, Jones C, Coleman HG. Current practices and future prospects for the management of gallbladder polyps: a topical review. World J Gastroenterol 2018;24(26):2844-2852 View Article PubMed/NCBI
  3. Taskin OC, Bellolio E, Dursun N, Seven IE, Roa JC, Araya JC, et al. Non-neoplastic polyps of the gallbladder: a clinicopathologic analysis of 447 cases. Am J Surg Pathol 2020;44(4):467-476 View Article PubMed/NCBI
  4. Patel K, Dajani K, Iype S, Chatzizacharias NA, Vickramarajah S, Singh P, et al. Incidental non-benign gallbladder histopathology after cholecystectomy in an United Kingdom population: Need for routine histological analysis?. World J Gastrointest Surg 2016;8(10):685-692 View Article PubMed/NCBI
  5. Mellnick VM, Menias CO, Sandrasegaran K, Hara AK, Kielar AZ, Brunt EM, et al. Polypoid lesions of the gallbladder: disease spectrum with pathologic correlation. Radiographics 2015;35(2):387-399 View Article PubMed/NCBI
  6. Cha BH, Hwang JH, Lee SH, Kim JE, Cho JY, Kim H, et al. Pre-operative factors that can predict neoplastic polypoid lesions of the gallbladder. World J Gastroenterol 2011;17(17):2216-2222 View Article PubMed/NCBI
  7. Wiles R, Thoeni RF, Barbu ST, Vashist YK, Rafaelsen SR, Dewhurst C, et al. Management and follow-up of gallbladder polyps: joint guidelines between the European Society of Gastrointestinal and Abdominal Radiology (ESGAR), European Association for Endoscopic Surgery and other Interventional Techniques (EAES), International Society of Digestive Surgery - European Federation (EFISDS) and European Society of Gastrointestinal Endoscopy (ESGE). Eur Radiol 2017;27(9):3856-3866 View Article PubMed/NCBI
  8. Bhatt NR, Gillis A, Smoothey CO, Awan FN, Ridgway PF. Evidence based management of polyps of the gall bladder: a systematic review of the risk factors of malignancy. Surgeon 2016;14(5):278-286 View Article PubMed/NCBI
  9. Patel K, Dajani K, Vickramarajah S, Huguet E. Five year experience of gallbladder polyp surveillance and cost effective analysis against new European consensus guidelines. HPB (Oxford) 2019;21(5):636-642 View Article PubMed/NCBI
  10. Pyo JS, Son BK, Lee HY, Oh IW, Chung KH. Incidental carcinoma after cholecystectomy for benign disease of the gallbladder: a meta-analysis. J Clin Med 2020;9(5):1484 View Article PubMed/NCBI
  11. Cavallaro A, Piccolo G, Panebianco V, Lo Menzo E, Berretta M, Zanghì A, et al. Incidental gallbladder cancer during laparoscopic cholecystectomy: managing an unexpected finding. World J Gastroenterol 2012;18(30):4019-4027 View Article PubMed/NCBI
  12. Park JK, Yoon YB, Kim YT, Ryu JK, Yoon WJ, Lee SH, et al. Management strategies for gallbladder polyps: is it possible to predict malignant gallbladder polyps?. Gut Liver 2008;2(2):88-94 View Article PubMed/NCBI
  13. Spadaro A, Tortorella V, Morace C, Fortiguerra A, Composto P, Bonfiglio C, et al. High circulating D-dimers are associated with ascites and hepatocellular carcinoma in liver cirrhosis. World J Gastroenterol 2008;14(10):1549-1552 View Article PubMed/NCBI
  14. Ganeshan D, Kambadakone A, Nikolaidis P, Subbiah V, Subbiah IM, Devine C. Current update on gallbladder carcinoma. Abdom Radiol (NY) 2021;46(6):2474-2489 View Article PubMed/NCBI
  15. Zevallos Maldonado C, Ruiz Lopez MJ, Gonzalez Valverde FM, Alarcon Soldevilla F, Pastor Quirante F, Garcia Medina V. Ultrasound findings associated to gallbladder carcinoma. Cir Esp 2014;92(5):348-355 View Article PubMed/NCBI
  16. Miyazaki M, Yoshitomi H, Miyakawa S, Uesaka K, Unno M, Endo I, et al. Clinical practice guidelines for the management of biliary tract cancers 2015: the 2nd English edition. J Hepatobiliary Pancreat Sci 2015;22(4):249-273 View Article PubMed/NCBI
  17. Quan Z, Hong D. Guideline for the diagnosis and treatment of gallbladder carcinoma (2019 edition). Zhonghua Wai Ke Za Zhi 2020;58(4):243-251 View Article PubMed/NCBI
  18. Yang JI, Lee JK, Ahn DG, Park JK, Lee KH, Lee KT, et al. Predictive model for neoplastic potential of gallbladder polyp. J Clin Gastroenterol 2018;52(3):273-276 View Article PubMed/NCBI
  19. Vickers AJ, Cronin AM, Elkin EB, Gonen M. Extensions to decision curve analysis, a novel method for evaluating diagnostic tests, prediction models and molecular markers. BMC Med Inform Decis Mak 2008;8:53 View Article PubMed/NCBI
  20. Metman MJH, Olthof PB, van der Wal JBC, van Gulik TM, Roos D, Dekker JWT. Clinical relevance of gallbladder polyps; is cholecystectomy always necessary?. HPB (Oxford) 2020;22(4):506-510 View Article PubMed/NCBI
  21. Fujiwara K, Abe A, Masatsugu T, Hirano T, Sada M. Effect of gallbladder polyp size on the prediction and detection of gallbladder cancer. Surg Endosc 2020 View Article PubMed/NCBI
  22. Valibouze C, El Amrani M, Truant S, Leroy C, Millet G, Pruvot FR, et al. The management of gallbladder polyps. J Visc Surg 2020;157(5):410-417 View Article PubMed/NCBI
  23. Shin SR, Lee JK, Lee KH, Lee KT, Rhee JC, Jang KT, et al. Can the growth rate of a gallbladder polyp predict a neoplastic polyp?. J Clin Gastroenterol 2009;43(9):865-868 View Article PubMed/NCBI
  24. Liu XF, Zhou LY, Wei ZH, Liu JX, Li A, Wang XZ, et al. The diagnostic role of circulating inflammation-based biomarker in gallbladder carcinoma. Biomark Med 2018;12(10):1095-1103 View Article PubMed/NCBI
  25. Velidedeoğlu M, Çitgez B, Arıkan AE, Ayan F. Is it necessary to perform prophylactic cholecystectomy for all symptomatic gallbladder polyps diagnosed with ultrasound?. Turk J Surg 2017;33(1):25-28 View Article PubMed/NCBI
  26. Zhao X, Zheng H, Shan S, Wang K, Zhang M, Xie S, et al. Association between the non-HDL-cholesterol-to-HDL-cholesterol ratio and the risk of gallbladder polyp formation among men: a retrospective cohort study. Lipids Health Dis 2020;19(1):146 View Article PubMed/NCBI
  27. Ahn DW, Jeong JB, Kang J, Kim SH, Kim JW, Kim BG, et al. Fatty liver is an independent risk factor for gallbladder polyps. World J Gastroenterol 2020;26(44):6979-6992 View Article PubMed/NCBI
  28. Onda S, Futagawa Y, Gocho T, Shiba H, Ishida Y, Okamoto T, et al. A preoperative scoring system to predict carcinoma in patients with gallbladder polyps. Dig Surg 2020;37(4):275-281 View Article PubMed/NCBI
  29. Wang JC, Fu R, Tao XW, Mao YF, Wang F, Zhang ZC, et al. A radiomics-based model on non-contrast CT for predicting cirrhosis: make the most of image data. Biomark Res 2020;8:47 View Article PubMed/NCBI
  30. Liu D, Wu J, Lin C, Andriani L, Ding S, Shen K, et al. Breast subtypes and prognosis of breast cancer patients with initial bone metastasis: a population-based study. Front Oncol 2020;10:580112 View Article PubMed/NCBI
  31. Lee KC, Jeng WJ, Hsu CM, Kuo CJ, Su MY, Chiu CT. Gallbladder polyps are associated with proximal colon polyps. Gastroenterol Res Pract 2019;2019:9832482 View Article PubMed/NCBI
  32. Albores-Saavedra J, Chablé-Montero F, Méndez-Sánchez N, Mercado MÁ, Vilatoba-Chapa M, Henson DE. Adenocarcinoma with pyloric gland phenotype of the extrahepatic bile ducts: a previously unrecognized and distinctive morphologic variant of extrahepatic bile duct carcinoma. Hum Pathol 2012;43(12):2292-2298 View Article PubMed/NCBI
  33. Misra MC, Guleria S. Management of cancer gallbladder found as a surprise on a resected gallbladder specimen. J Surg Oncol 2006;93(8):690-698 View Article PubMed/NCBI
  34. Gu J, Yan S, Wang B, Shen F, Cao H, Fan J, et al. Type 2 diabetes mellitus and risk of gallbladder cancer: a systematic review and meta-analysis of observational studies. Diabetes Metab Res Rev 2016;32(1):63-72 View Article PubMed/NCBI
  35. Jing C, Wang Z, Fu X. Effect of diabetes mellitus on survival in patients with gallbladder Cancer: a systematic review and meta-analysis. BMC Cancer 2020;20(1):689 View Article PubMed/NCBI
  36. Sachan A, Saluja SS, Nekarakanti PK, Nimisha, Mahajan B, Nag HH, et al. Raised CA19-9 and CEA have prognostic relevance in gallbladder carcinoma. BMC Cancer 2020;20(1):826 View Article PubMed/NCBI