Development and Validation of a Nomogram Based on CT Imaging Features for Differentiating Pancreatic Head Cancer in Periampullary Carcinomas

doi:10.21203/rs.3.rs-4694686/v1

Download PDF

Research Article

Development and Validation of a Nomogram Based on CT Imaging Features for Differentiating Pancreatic Head Cancer in Periampullary Carcinomas

https://doi.org/10.21203/rs.3.rs-4694686/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Purpose To construct a predictive nomogram for differentiate pancreatic head cancer from other periampullary cancers based on CT imaging features.

Methods

This is a retrospective analysis, Patients diagnosed with periampullary carcinoma by pathological findings from April 2013 to April 2024 were consecutively collected. The variables evaluated included imaging characteristics (direct and indirect signs) and clinical data. Univariate and multivariate regression analyses were used to find statistically significant variables. A nomogram prediction models based on regression analysis and was internally validated.

Results

Multivariable analysis revealed that the distance from the end of the dilated pancreatic duct to the medial wall of the papilla (P＜0.05), the distance from the end of the dilated bile duct to the medial wall of the papilla (P＜0.01), papilla enlargement（P＜0.01）, and the presence of pancreatic and/or bile ducts between the tumor and the papilla (P＜0.05)were identified as independent risk factors for differentiating pancreatic head cancer from non-pancreatic head cancers, and were used to construct a nomogram. The nomogram demonstrated high accuracy, with an AUC of 0.826 in the development cohort and 0.801 in the validation cohort.

Conclusions

This study is based on CT imaging features to differentiate pancreatic head cancer from non-pancreatic head cancer in periampullary cancer. Multiple imaging signs with differential diagnostic significance were obtained, Development and validation of a nomogram that integrates these imaging features, providing a basis for treatment and comprehensive assessment in the clinic. Keywords Periampullary cancer·Pancreatic head cancer·Differential diagnosis·Computed tomography (CT)·Nomogram

Periampullary carcinomas occur within 2 cm of the duodenal papilla and include ampullary cancer, distal cholangiocarcinoma, pancreatic head cancer, and duodenal cancer [1]. Surgical treatment is recommended for most ampullary tumors [2]. However, the treatment strategies for these four types of cancers differ. In the case of ampullary cancer, even if preoperative imaging shows positive lymph nodes, aggressive resection should be performed, whereas patients with resectable pancreatic head cancer are usually recommended to undergo adjuvant chemotherapy or chemoradiotherapy after surgery [3]. While periampullary cancers present with similar clinical manifestations and symptoms, their long-term postoperative prognosis varies, potentially due to biological and genetic differences [4]. Patients with pancreatic head cancer have the worst prognosis, with a 5-year survival rate of 5–20% after resection. In comparison, patients with non-pancreatic head periampullary cancers have a 5-year survival rate of over 20% [5]. Given the significant differences in surgical strategies and long-term prognosis between pancreatic head cancer and other periampullary cancers, it is crucial to accurately diagnose the specific type of cancer before treatment. This study aims to differentiate pancreatic head cancer from other periampullary cancers based on CT imaging features to better guide clinical treatment.

Patients

Patients diagnosed with periampullary carcinoma by pathological findings at our center from April 2013 to April 2024 were consecutively collected. Classified into pancreatic head cancer(Group A) and non-pancreatic head cancer (Group B) according to pathologic findings. Patients were divided into a development cohort and a validation cohort in a 7:3 ratio based on the order of diagnosis. Inclusion criteria were as follows:(1) Pathologically confirmed pancreatic head cancer, distal cholangiocarcinoma, ampullary cancer, or duodenal (peripapillary) cancer via surgery, biopsy, or ERCP; (2) Underwent non-contrast and dual-phase contrast-enhanced CT scans within two weeks before surgery. Exclusion criteria included: (1) Prior adjuvant tumor treatment, such as radiotherapy or biliary drainage, before CT examination; (2) Pathologically confirmed non-pancreatic ductal carcinoma (e.g. cystadenocarcinoma, neuroendocrine carcinoma); (3) Poor image quality that did not meet diagnostic requirements (Fig. 1). This single-center retrospective study was approved by the Institutional Review Board, with informed consent waived.

Imaging protocol

All CT examinations were performed using a 64-CT system (Aquilion 64, Toshiba Medical Systems). Patients were routinely instructed to fast for 8–10 hours prior to the CT examination. Additionally, each patient was asked to drink 500 mL of water 30 minutes before imaging, and an additional 300 mL of water immediately before getting on the examination bed to dilate the gastrointestinal tract, unless dietary restrictions were present.

Our standard imaging protocol primarily included a single breath-hold scan from the right diaphragmatic dome to the right renal hilum, followed by unenhanced and biphasic contrast-enhanced examinations [4]. For contrast-enhanced images, an auto-triggered imaging mode with a 5-second delay was used, set at a preset CT value of 200 HU based on the region of interest in the descending aorta reference image. In this mode, after initiating the injection of nonionic contrast agent (Optiray 320, Tyco Healthcare) at a rate of 3–4 mL/s, the arterial and portal venous phases were delayed by 19–30 seconds and 58–72 seconds, respectively. The parameters were: 120 kVp ; imaging time per rotation of 0.5 seconds; detector collimation of 1.0 mm × 32 or 0.5 mm × 64, with pitch coefficients of 0.844 and 0.828, respectively; tube currents with automatic dose modulation ranging from 98 to 440 mA; and a field of view (FOV) of 30–35 cm. The total effective dose per patient ranged from 8.40 to 14.62 mSv (mean dose 11.03 mSv). For all raw data, unenhanced and enhanced biphasic CTs were reconstructed using a low spatial resolution algorithm to reduce image noise. Additional parameters included: slice thickness of 0.5-1.0 mm; reconstruction interval of 0.6–0.8 mm; a 512 × 512 matrix; and a field of view of 22–30 cm. The reconstructed source images were then transferred to the workstation (Advantage Workstations, version 3.1; GE Medical Systems).

Radiological assessment

All imaging data were evaluated by two radiologists, each with 10 and 15 years of experience in diagnostic abdominal CT, respectively, using a double-blind method. They were blinded to the clinical data and pathological findings of each case prior to evaluation and were trained in the evaluation method beforehand.

The assessment focused on the following direct and indirect signs. Direct signs include: the distance of the end of the dilated bile duct from the medial wall of the papilla(the distance of the dilated bile duct), the distance of the end of the dilated pancreatic duct from the medial wall of the papilla(the distance of the dilated pancreatic duct), unenhanced CT value, relative enhancement value (i.e. the difference of CT value between the arterial phase/venous phase and the unenhanced CT value), enlargement of the papilla (defined as a diameter > 10 mm) [6], thickened and enhanced duodenal wall, and thickened and enhanced distal common bile duct (CBD) wall.

Indirect signs included: pancreatic morphology (no change/atrophy/enlargement with peripheral exudation), morphology of the main pancreatic duct (MPD) (no change/deformation/proximal obstruction with distal dilatation/full-length dilatation), dilated diameter of the MPD (pancreatic duct diameter > 3 mm) [7], visibility of the pancreatic and/or bile duct between the tumor and papilla, pattern of CBD narrowing (gradual tapering/abrupt narrowing), types of duct dilatation (CBD dilatation/intrahepatic bile duct (IHD) dilatation/MPD dilatation/CBD and MPD dilatation/CBD and intrahepatic duct (IHD) dilatation), extent of IHD dilatation (mild: diameter of dilatation smaller than the portal vein/moderate to severe: diameter of dilatation larger than the portal vein) [8], presence and specific vessels involved in vascular invasion, as well as lymph node metastasis and distant metastasis.

When there was a dispute in the evaluation between the two radiologists, another radiologist with 15 years of diagnostic experience reassessed the content, and a consensus was ultimately reached.

Clinical data analysis

Clinical data included gender, age, CA199, CEA, direct bilirubin (DBIL), and total bilirubin (TBIL). All clinical data were obtained within three weeks prior to imaging.

Statistical analyses

Data were statistically analyzed using SPSS 27.0 (IBM, Armonk,). Normally distributed continuous variables were expressed as mean ± standard deviation, and t-tests were used for comparisons between groups. Non-normally distributed continuous variables were expressed as median (quartiles), and the Mann-Whitney test was used for comparisons between groups. Count data were expressed as frequencies and percentages, with comparisons between groups made using the rank sum test and Fisher's exact test. After performing univariate comparative analyses between groups, the Youden index was analyzed using the receiver operating characteristic (ROC) curve to assess the cut-off value of continuous variables, which were then transformed into ordered categorical variables. Variables with a P-value less than 0.1 were included in a logistic regression analysis to calculate odds ratios (OR) and 95% confidence intervals (CI) [9]. Based on the results of the logistic regression, nomograms were created by R version 4.3.3 (Te R Foundation for Statistical Computing). The concordance index (C-index) was used to assess the performance of the nomograms, and internal validation was performed to reduce overfitting bias. Calibration focused on the agreement between predicted probabilities and actual outcomes, which was assessed in both the development cohort and the internal validation cohort. A P-value of less than 0.05 was considered statistically significant.

Characteristics of Study Patients

The study collected 176 patients who were pathologically confirmed to have periampullary cancer, and finally included 171 patients in this study. The clinical baseline data of the patients in the development cohort and the validation cohort are shown in Table 1. there was no significant difference between the two cohorts in terms of age and gender. However, there was a statistically significant difference between the two cohorts in the indicators of DBIL (P < 0.05).

Table 1

Characteristics of Study Patients
Characteristics	Development Cohort			Validation Cohort
Characteristics	Group A(n = 72)	Group B (n = 63)	P-value	Group A(n = 18)	Group B (n = 18)	P-value
Age^a(y)	68.6 ± 9.0	66.3 ± 12.0	0.21	70.2 ± 7.9	69.3 ± 8.2	0.73
Sex			0.72			1
Male	44(61.1)	41(65.1)		12(66.7)	11(61.1)
Female	28(38.9)	22(34.9)		6(33.3)	7(38.9)
CA199(u/ml)	209.3(84.2-736.3)	125.9(21.8-345.3)	0.07	615.1(102.7-1911.6)	176.9(87.8-712.2)	0.08
CEA(u/ml)	3.3(2.5–5.7)	3.7(2.2–5.3)	0.53	7.8(2.8–59.6)	10.7(2.3–69.7)	0.95
DBIL(umol/l)	90.2(14.1–166.0)	66.6(8.5-115.3)	0.05	108.2(34.1-141.3)	39.4(14.3–90.9)	0.05
TBIL(umol/l)	111.8(22.6-220.7)	87.2(22.4-146.7)	0.08	130.3(16.7-178.8)	79.2(21.1-148.9)	0.45
^aData are means ± SDs
Data in parentheses are percentages or interquartile ranges
CA199 Carbohydrate antigen199, CEA Carcinoembryonic antigen, DBIL Direct bilirubin, TBIL Total bilirubin

Imaging characteristics

In the analysis of imaging features, the distance of the end of the dilated bile duct from the medial wall of the papilla and the distance of the end of the dilated pancreatic duct from the medial wall of the papilla in the pancreatic head cancer group were greater than those in the group B (P < 0.05). The relative enhancement CT value in the arterial phase and venous phase in the group A was smaller than that group B (P < 0.05). The number of patients with thickened and enhanced distal bile duct walls and duodenal walls in the pancreatic head cancer group was lower than that in the group B (P < 0.05). In the non-pancreatic cancer group, 27 cases (43%) had enlarged papillae, compared to 7% in the group A, showing a statistically significant difference (P < 0.05) (Table 2).

Table 2

CT findings of Study Patients
	Development Cohort			Validation Cohort
	Group A(n = 72)	Group B (n = 63)	p-value	Group A(n = 18)	Group B (n = 18)	P-value
Direct signs
CT plain scan value^a(HU)	34.72 ± 8.23	36.74 ± 8.13	0.21	37.2 ± 6.5	39.4 ± 7.4	0.35
Relative enhancement value in the arterial phase^a	22.43 ± 19.68	35.12 ± 20.11	<0.01	16.5(12.8–27.5)	27.5(18.5–42.8)	0.07
Relative enhancement value in the venous phase ^a	34.26 ± 23.95	52.48 ± 24.81	<0.01	27.0(15.8–39.0)	46.036.8–75.3)	<0.01
The distance of the dilated bile duct(mm)	31.5(15.5–40.0)	15.0(7.0–24.0)	<0.01	21.0(0–33.0)	9(0-17.3)	0.07
The distance of the dilated pancreatic duct(mm)	23.0(0–40.0)	0(0–13.0)	<0.01	8.0(0–33.0)	0(0–10.0)	0.09
Enlargement of the papilla	5(6.9)	27(42.9)	<0.01	2(11.1)	8(44.4)	0.06
Thickened duodenal wall	0(0)	8(12.7)	0.02	0(0)	8(44.4)	<0.01
Thickened distal CBD wall	13(18.1)	24(38.1)	<0.01	2(11.1)	3(16.7)	0.63
Indirect signs
Pancreatic morphology			0.15			0.28
Invariably	48(66.7)	51(81.0)		10(55.6)	14(77.8)
Pancreatic atrophy	18(25.0)	10(15.9)		7(38.9)	4(22.2)
Pancreatic enlargement with exudation	6(8.3)	2(3.2)		1(5.6)	0(0)
Dilated diameter of MPD^a(mm)	6.01 ± 3.15	5.55 ± 2.11	0.48	3.0(2.0-9.3)	2.0(2.0-5.8)	0.41
Morphology of MPD			0.76			0.53
Invariably	29(40.3)	29(46.0)		6(33.3)	9(50.0)
Deformation	3(4.2)	1(1.6)		3(16.7%)	1(5.6)
Proximal obstruction and Distal dilatation	12(16.7)	9(14.3)		4(22.2%)	5(27.8)
Full scale dilatation	28(38.9)	24(38.1)		5(27.8)	3(16.7)
Pancreatic duct and/or bile duct showing between tumor and papilla			<0.01			<0.01
Pancreatic duct	17(23.6)	2(3.2)		3(16.7)	1(5.6)
bile duct	11(15.3)	4(6.3)		0(0)	2(11.1)
Pancreatic duct and bile duct	16(22.2)	5(7.9)		12(66.7)	1(5.6)
Pattern of the distal CBD narrowing	11	2	0.02			0.73
Gradual tapering	26(36.1)	35(55.6)		7(38.9)	8(44.4)
Abrupt narrowing	35(48.6)	26(41.3)		6(33.3)	7(38.9)
Dilated MPD and/or bile duct			0.20			0.69
CBD	2(2.8)	2(3.2)		0(0)	1(5.6)
MPD	5(6.9)	1(1.6)		1(5.6)	0(0)
CBD and MPD	30(41.7)	27(42.9)		5(27.8)	6(33.3)
CBD and IHD	29(40.3)	32(50.8)		8(44.4)	8(44.4)
IHD dilatation extent			0.06			0.22
Mild	26(36.1)	32(50.8)		8(44.4)	5(27.8)
Moderate to severe	29(40.3)	25(39.7)		4(22.2)	9(50.0)
Vascular invasion			0.17			0.60
Arterial invasion	3(4.2)	0(0)		15(83.3)	17(0)
Venous invasion	0(0)	0(0)		3(16.7)	1(5.6)
Both	1(1.4)	0(0)		0(0)	0(0)
Lymphatic metastasis	20(27.8)	5(7.9)	<0.01	6(33.3)	1(5.6)	0.09
Distant metastasis	1(1.4)	0(0)	1	2(11.1)	0(0)	0.48
^a Data are means ± SDs
Data in parentheses are percentages or interquartile ranges
MPD Main pancreatic duct, CBD Common bile duct, IHD Intrahepatic duct

Among indirect signs, 61% of group A showed pancreatic duct and/or bile duct visibility between the tumor and the papilla, while only 18% of group B showed this feature, which was statistically significant (P < 0.05). Although vascular invasion and distant metastasis were not statistically different, it is noteworthy that 2 cases of group A had peripheral arterial invasion, with 1 case involving the hepatic artery and 1 case involving the celiac trunk and superior mesenteric artery. Another case had simultaneous invasion of the superior mesenteric artery and superior mesenteric vein. In terms of the incidence of distant metastasis, there was no statistically significant difference between the two groups, with only 1 case of pancreatic head cancer patient developing liver metastasis.

Logistic regression

Factors that were statistically different in the univariate analysis of the development cohort were further analyzed using multivariable logistic regression analysis (Table 3). Ultimately, the distance of the dilated bile duct, the distance of the dilated bile duct, papilla enlargement, and the presence of pancreatic and/or bile ducts between the tumor and the papilla were identified as independent risk factors for differentiating pancreatic head cancer from non-pancreatic head cancers. The cut-off values for the distance of the dilated bile duct and the distance of the dilated bile duct, determined using the Youden index of the ROC curve, were 20.50 mm and 27 mm, respectively.

Table 3

Multivariable logistic regression
	Odds ratio	95%CI	P-value
CA199	1.53	0.39–5.97	0.54
DBIL	2.04	0.19–22.29	0.56
TBIL	1.38	0.12–15.67	0.80
The distance of the dilated bile duct	31.83	4.12-246.25	<0.01
The distance of the dilated pancreatic duct	8.76	1.44–53.16	0.02
Relative enhancement value in the arterial phase	0.19	0.03–1.08	0.06
Relative enhancement value in the venous phase	0.76	0.18–3.16	0.71
Thickened distal CBD wall	0.43	0.09–1.87	0.26
Enlargement of the papilla	0.03	0.003–0.25	<0.01
Thickened duodenal wall	---	---	0.99
Pancreatic duct and/or bile duct showing between tumor and papilla	3.97	1.01–15.64	0.049
Pattern of the distal CBD narrowing
Gradual tapering	0.16	0.01–1.98	0.15
Abrupt narrowing	0.17	0.01–2.15	0.17
Lymphatic metastasis	0.38	0.05–2.74	0.34
CA199 Carbohydrate antigen199, DBIL Direct bilirubin, TBIL Total bilirubin, CBD Common bile duct
Only variables identified as significant (p < 0.1) in the univariable analyses were entered into the multivariable analysis

Establishment of Nomogram prediction models

A clinical outcome prediction model was developed using the nomogram (Fig. 2A). In the development cohort, this model demonstrated an AUC of 0.826 (95% CI, 0.754–0.897), with a sensitivity of 0.889 and a specificity of 0.698. The nomogram also performed well in the validation cohort, with an AUC of 0.801 (95% CI, 0.651–0.950), a sensitivity of 0.611, and a specificity of 0.889 (Fig. 2B). The calibration curves of the development and validation cohorts demonstrated good agreement (Fig. 2C and Fig. 2D). Two classic cases illustrating the use of the nomogram(Fig. 3 and Fig. 4).

The aim of this study was to differentiate pancreatic head cancer from non-pancreatic head cancer in periampullary cancer based on CT imaging features, and to comprehensively analyze the direct and indirect signs of CT imaging features in patients with periampullary cancer. In this retrospective study, we identified independent risk factors for differentiating pancreatic head cancer from non-pancreatic head cancer in periampullary cancer through multifactorial regression analysis. These imaging signs were the distance of the distance of the dilated bile duct, the distance of the dilated bile duct, enlargement of the papilla, and pancreatic duct and/or bile duct visible between the tumor and the papilla. Therefore, in cases where the specific type of periampullary cancer cannot be clinically determined, this study provides an imaging-based predictive model for the differential diagnosis of periampullary cancer. This model can better guide the selection of clinical treatment options.

In recent years, imaging has been applied to several research aspects of periampullary lesions, such as MRI differential diagnosis of benign and malignant periampullary obstructions [10], the diagnostic value of CT for signs of biliary obstruction [7, 11], and the diagnostic accuracy of CT for benign and malignant periampullary stenosis [12]. Additionally, studies have reported on the assessment of CT for benign and malignant differentiation of periampullary tumors, with an AUC of about 81% [13], indicating a high accuracy of CT in identifying benign and malignant tumors around the periampullary area. These studies highlight the role of different imaging methods in the diagnosis of periampullary lesions. In the routine clinical examination of patients with suspected periampullary tumors, the most commonly used examination modality is enhanced spiral CT [14], which offers advantages such as fast imaging speed, short post-processing time, the ability to perform various image reconstructions, and the ability to obtain numerous image features. Therefore, based on previous studies, we chose CT imaging features as the primary focus of this research. Starting from the specific direct and indirect signs, we subdivided the imaging features to collect a comprehensive set of imaging characteristics.

Chang S et al. [15] found that the presence of a measurable papillary mass in the arterial phase and homogeneous enhancement of the papilla/papillary mass suggested an ampullary tumor by thin-layer CT. Their findings showed that a papillary mass greater than 12.3 mm was the critical value for diagnosing ampullary tumors and benign papillary stricture, with a sensitivity of 91.7% and a specificity of 92.3%. Similarly, Chung YE et al. [16] confirmed that, excluding benign obstruction, unexplained papillary enlargement suggests that ampullary obstruction is malignant. Previous studies have mostly used papillary enlargement for diagnosing ampullary tumors or ampullary cancer in ampullary obstructions, also finding that expanded papillae with diameters greater than 10 mm are effective in detecting periampullary lesions [6]. In this study, we used the imaging feature of papillary enlargement to differentiate pancreatic head cancer from non-pancreatic head cancer among periampullary cancers. We found the diagnostic sensitivity and specificity of the sign of papillary enlargement to be 93% and 43%, respectively. The relatively low specificity may be due to the fact that papillary enlargement can be caused by ampullary tumors or other benign conditions such as papillary inflammation and/or papillary stenosis, which aligns with the findings of previous studies.

Kim JH et al. [1] showed that the distance from the lumen of the duodenum to the end of the dilated duct in pancreatic head carcinoma was 14–42 mm (mean 25 mm), whereas in ampullary cancer, it was 2–9 mm (mean 5 mm). In our study, the mean distance from the end of the dilated duct to the papilla in pancreatic head cancer was 25 mm, whereas in non-pancreatic head cancer, the mean distance was 11 mm, demonstrating a significant difference between the two. This difference may be because, in pancreatic head cancer, the infiltrated part of the pancreatic and bile ducts is often located in the distal non-bile duct, and a section of normal pancreaticobiliary duct remains between the papilla and the tumor. Consequently, four segments (two bile duct segments and two pancreatic duct segments) can be observed, a phenomenon termed the "four-segment sign," which is frequently seen in pancreatic head cancer but is rare in other periampullary cancers [1, 16]. Additionally, Kim JH et al. found that the distance from the duodenal papilla to the proximal end of the stenotic segment of the pancreatic duct was longer in pancreatic head cancer than in other periampullary cancers [1]. Our analysis determined cut-off values for the distances of the dilated pancreatic duct and bile duct from the papilla in pancreatic head and non-pancreatic head cancers, which were 20.5 mm and 27 mm, respectively. Our study further quantified these indices, providing high diagnostic specificity, thus offering a quantitative diagnostic basis for clinical practice. In pancreatic head cancer, the "four-segment sign" [1, 17] is often observed, indicating that distal pancreatic duct and bile duct involvement is rare. As a result, the pancreatic duct and/or bile duct are more frequently seen between the tumor and the papilla. In contrast, patients with non-pancreatic head cancer, such as those with distal cholangiocarcinoma, often exhibit distal bile duct involvement [1]. Consequently, the likelihood of finding the pancreatic duct and bile duct between the tumor and the papilla in non-pancreatic head cancer patients is lower than in pancreatic head cancer patients. In this study, within the development cohort, 44 patients with pancreatic head cancer (61.6%) showed the presence of pancreatic ducts and/or bile ducts between the tumor and the papilla. In contrast, only 11 patients with non-pancreatic head cancer (17.4%) exhibited this feature, demonstrating a statistically significant difference between the two groups.

There are several limitations to this study. First, it was a retrospective single-center study with a relatively small sample size, leading to potential selection bias. Second, this study was a unimodal imaging examination and did not include MRI image data for analysis. Third, although an internally validated nomogram was performed, further studies using external and prospective validation in larger cohorts are needed. Therefore, in subsequent studies, more patients with periampullary cancer will be included, and more imaging data for differential diagnosis of Periampullary cancer will be obtained from a multimodal perspective by combining the patients’ MRI imaging data and prognosis.

This study is based on CT imaging features to differentiate pancreatic head cancer from non-pancreatic head cancer in periampullary cancer. Multiple imaging signs with differential diagnostic significance were obtained, and the nomogram constructed on the basis of these factors can be used to differentiate pancreatic head cancer from non-pancreatic head cancer in periampullary cancer, providing a basis for treatment and comprehensive assessment in the clinic.

Author Contribution

XHZ, JQW and LZ contributed to study design. WJW, ZYZ, FMC, and DYZ obtained and analyzed the data. XHZ and JQW drafted the manuscript and fifigures. LZ contributed to the manuscript reviewing. All authors read, edited, and approved the final manuscript.

Kim J H, Kim M J, Chung J J, et al ( 2002) Differential diagnosis of periampullary carcinomas at MR imaging. Radiographics 22(6):1335-1352. doi: https://doi.org/10.1148/rg.226025060
Chang S, Lim J H, Choi D, et al (2008) Differentiation of ampullary tumor from benign papillary stricture by thin-section multidetector CT. Abdominal Imaging 33(4):457-462. doi: https://doi.org/10.1007/s00261-007-9295-0
Wang F B, Ni J M, Zhang Z Y, et al (2014) Differential diagnosis of periampullary carcinomas: comparison of CT with negative-contrast CT cholangiopancreatography versus MRI with MR cholangiopancreatography. Abdominal Imaging 39(3):506-517. doi: https://doi.org/10.1007/s00261-014-0085-1
Li B, Zhang L, Zhang Z Y, et al (2015) Differentiation of noncalculous periampullary obstruction: comparison of CT with negative-contrast CT cholangiopancreatography versus MRI with MR cholangiopancreatography. European Radiology 25(2):391-401. doi: https://doi.org/10.1007/s00330-014-3430-4
Chen S C, Shyr Y M, Wang S E. (2013) Longterm survival after pancreaticoduodenectomy for periampullary adenocarcinomas. HPB 15. doi: https://doi.org/10.1111/hpb.12071
Kim T U, Kim S, Lee J W, et al (2008) Ampulla of Vater: Comprehensive anatomy, MR imaging of pathologic conditions, and correlation with endoscopy. European Journal of Radiology 66(1):48-64. doi: https://doi.org/10.1016/j.ejrad.2007.04.005
Kim SW, Kim SH, Lee DH, et al (2017) Isolated Main Pancreatic Duct Dilatation: CT Differentiation Between Benign and Malignant Causes. AJR Am J Roentgenol 209:1046–1055. doi: https://doi.org/10.2214/AJR.17.17963
Lee J E, Choi S Y, Lee M H, et al (2022) Differentiating between benign and malignant ampullary strictures: a prediction model using a nomogram based on CT imaging and clinical findings. European Radiology 32(11):7566-7577. doi: https://doi.org/10.1007/s00330-022-08856-7
Wang Y, Niu Z R, Tao L Y, et al (2021) Nomogram to predict pregnancy outcomes of emergency oocyte freeze-thaw cycles. Chinese Medical Journal 134. doi: https://doi.org/10.1097/CM9.0000000000001731
Chung Y E, Kim M J, Kim H M, et al (2011) Differentiation of benign and malignant ampullary obstructions on MR imaging. European Journal of Radiology 80(2):198-203. doi: https://doi.org/10.1016/j.ejrad.2010.04.017
Zhang Z Y, Wang D, Ni J M, et al (2012) Comparison of three-dimensional negative-contrast CT cholangiopancreatography with three-dimensional MR cholangiopancreatography for the diagnosis of obstructive biliary diseases. European Journal of Radiology 81(5):830-837. doi: https://doi.org/10.1016/j.ejrad.2011.02.036
Lee J E, Choi S Y, Lee M H, et al (2022) Differentiating between benign and malignant ampullary strictures: a prediction model using a nomogram based on CT imaging and clinical findings. European Radiology 32(11):7566-7577. doi: https://doi.org/10.1007/s00330-022-08856-7
Andersson M, Kostic S, Johansson M, et al (2005) MRI combined with MR cholangiopancreatography versus helical CT in the evaluation of patients with suspected periampullary tumors: a prospective comparative study. Acta Radiol 46(1):16-27. doi: https://doi.org/10.1080/02841850510016018
Freeny, P. C (1999) Computed tomography in the diagnosis and staging of cholangiocarcinoma and pancreatic carcinoma.Annals of Oncology Official Journal of the European Society for Medical Oncology. Ann Oncol 10 Suppl 4:12-7. doi: https://doi.org/10.1093/annonc/10.suppl_4.S12
Chang S, Lim J H, Choi D, et al (2008) Differentiation of ampullary tumor from benign papillary stricture by thin-section multidetector CT. Abdominal Imaging 33(4):457-462. doi: https://doi.org/10.1007/s00261-007-9295-0
Chung Y E, Kim M J, Kim H M, et al (2011) Differentiation of benign and malignant ampullary obstructions on MR imaging. European Journal of Radiology 80(2):198-203. doi: https://doi.org/10.1016/j.ejrad.2010.04.017
Chen FM, Ni JM, Zhang ZY, et al (2016) Presurgical Evaluation of Pancreatic Cancer: A Comprehensive Imaging Comparison of CT Versus MRI. AJR Am J Roentgenol 206(3):526-35. doi: https://doi.org/10.1016/10.2214/AJR.15.15236

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Development and Validation of a Nomogram Based on CT Imaging Features for Differentiating Pancreatic Head Cancer in Periampullary Carcinomas

Status:

Version 1

Abstract

Figures

Introduction

Materials and Methods

Patients

Imaging protocol

Radiological assessment

Clinical data analysis

Statistical analyses

Result

Characteristics of Study Patients

Imaging characteristics

Logistic regression

Establishment of Nomogram prediction models

Discussion

Conclusion

Declarations

Author Contribution

References

Additional Declarations

Status:

Version 1