1
|
Chen T, Zhang D, Chen S, Lu J, Guo Q, Cai S, Yang H, Wang R, Hu Z, Chen Y. Machine learning for differentiating between pancreatobiliary-type and intestinal-type periampullary carcinomas based on CT imaging and clinical findings. Abdom Radiol (NY) 2024; 49:748-761. [PMID: 38236405 PMCID: PMC10909762 DOI: 10.1007/s00261-023-04151-1] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/16/2023] [Revised: 12/03/2023] [Accepted: 12/05/2023] [Indexed: 01/19/2024]
Abstract
PURPOSE To develop a diagnostic model for distinguishing pancreatobiliary-type and intestinal-type periampullary adenocarcinomas using preoperative contrast-enhanced computed tomography (CT) findings combined with clinical characteristics. METHODS This retrospective study included 140 patients with periampullary adenocarcinoma who underwent preoperative enhanced CT, including pancreaticobiliary (N = 100) and intestinal (N = 40) types. They were randomly assigned to the training or internal validation set in an 8:2 ratio. Additionally, an independent external cohort of 28 patients was enrolled. Various CT features of the periampullary region were evaluated and data from clinical and laboratory tests were collected. Five machine learning classifiers were developed to identify the histologic type of periampullary adenocarcinoma, including logistic regression, random forest, multi-layer perceptron, light gradient boosting, and eXtreme gradient boosting (XGBoost). RESULTS All machine learning classifiers except multi-layer perceptron used achieved good performance in distinguishing pancreatobiliary-type and intestinal-type adenocarcinomas, with the area under the curve (AUC) ranging from 0.75 to 0.98. The AUC values of the XGBoost classifier in the training set, internal validation set and external validation set are 0.98, 0.89 and 0.84 respectively. The enhancement degree of tumor, the growth pattern of tumor, and carbohydrate antigen 19-9 were the most important factors in the model. CONCLUSION Machine learning models combining CT with clinical features can serve as a noninvasive tool to differentiate the histological subtypes of periampullary adenocarcinoma, in particular using the XGBoost classifier.
Collapse
Affiliation(s)
- Tao Chen
- Department of Radiology, The First Affiliated Hospital, Zhejiang University School of Medicine, 79 Qingchun Road, Hangzhou, 310003, Zhejiang, China
| | - Danbin Zhang
- Department of Radiology, The First Affiliated Hospital, Zhejiang University School of Medicine, 79 Qingchun Road, Hangzhou, 310003, Zhejiang, China
| | - Shaoqing Chen
- Department of Radiology, The Second Affiliated Hospital and Yuying Children's Hospital of Wenzhou Medical University, 109 Xueyuan West Road, Wenzhou, 325027, Zhejiang, China
| | - Juan Lu
- Department of Computer Science and Software Engineering, The University of Western Australia, Crawley, WA, 6009, Australia
- School of Medicine, The University of Western Australia, Crawley, WA, 6009, Australia
- Harry Perkins Institute of Medical Research, Murdoch, WA, 6150, Australia
| | - Qinger Guo
- Department of Radiology, The First Affiliated Hospital, Zhejiang University School of Medicine, 79 Qingchun Road, Hangzhou, 310003, Zhejiang, China
| | - Shuyang Cai
- Department of Radiology, The First Affiliated Hospital, Zhejiang University School of Medicine, 79 Qingchun Road, Hangzhou, 310003, Zhejiang, China
| | - Hong Yang
- Department of Radiology, The First Affiliated Hospital, Zhejiang University School of Medicine, 79 Qingchun Road, Hangzhou, 310003, Zhejiang, China
| | - Ruixuan Wang
- School of Electronics and Computer Science, University of Liverpool, Brownlow Hill, Liverpool, Merseyside, L69 3BX, UK
| | - Ziyao Hu
- School of Electronics and Computer Science, University of Liverpool, Brownlow Hill, Liverpool, Merseyside, L69 3BX, UK
| | - Yang Chen
- Liverpool Centre for Cardiovascular Science at University of Liverpool, Liverpool John Moores University and Liverpool Heart and Chest Hospital, William Henry Duncan Building, 6 West Derby St, Liverpool, Merseyside, L7 8TX, UK.
| |
Collapse
|
2
|
Nalbant MO, Oner O, Akinci O, Hocaoglu E, Inci E. Analysis of Pancreatobiliary and Intestinal Type Periampullary Carcinomas Using Volumetric Apparent Diffusion Coefficient Histograms. Acad Radiol 2023; 30 Suppl 1:S238-S245. [PMID: 37211479 DOI: 10.1016/j.acra.2023.04.031] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 04/02/2023] [Revised: 04/21/2023] [Accepted: 04/22/2023] [Indexed: 05/23/2023]
Abstract
RATIONALE AND OBJECTIVES Magnetic resonance imaging plays an important role in the evaluation of patients with known or suspected periampullary masses. The utilization of volumetric apparent diffusion coefficient (ADC) histogram evaluation for the entire lesion eradicates the potential for subjectivity in the region of interest placement, thus guaranteeing the accuracy of computation and repeatability. PURPOSE To investigate the value of volumetric ADC histogram analysis in the differentiation of intestinal-type (IPAC) and pancreatobiliary-type periampullary adenocarcinomas (PPAC). MATERIALS AND METHODS This retrospective study included 69 patients with histopathologically confirmed periampullary adenocarcinoma (54 PPAC and 15 IPAC). Diffusion-weighted imaging was obtained at b values of 1000 mm²/s. The histogram parameters of ADC values, comprising the mean, minimum, maximum, 5th, 10th, 25th, 50th, 75th, 90th, and 95th percentiles, as well as skewness, kurtosis, and variance, were calculated independently by two radiologists. Using the interclass correlation coefficient, the interobserver agreement was evaluated. RESULTS The ADC parameters for the PPAC group were all lower than those of the IPAC group. The PPAC group had higher variance, skewness, and kurtosis than the IPAC group. However, the difference between the kurtosis (P = .003), the 5th (P = .032), 10th (P = .043), and 25th (P = .037) percentiles of ADC values was statistically significant. The area under the curve (AUC) of the kurtosis was the highest (AUC=0.752; cut-off value=-0.235; sensitivity=61.1%; specificity=80.0%). CONCLUSION Volumetric ADC histogram analysis with b values of 1000 mm²/s can discriminate subtypes noninvasively before surgery.
Collapse
Affiliation(s)
- Mustafa Orhan Nalbant
- University of Health Sciences, Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Radiology Department, Tevfik Saglam Cad. No: 11, Zuhuratbaba, 34147 Bakırkoy, Istanbul, Turkey.
| | - Ozkan Oner
- University of Health Sciences, Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Radiology Department, Tevfik Saglam Cad. No: 11, Zuhuratbaba, 34147 Bakırkoy, Istanbul, Turkey
| | - Ozlem Akinci
- University of Health Sciences, Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Radiology Department, Tevfik Saglam Cad. No: 11, Zuhuratbaba, 34147 Bakırkoy, Istanbul, Turkey
| | - Elif Hocaoglu
- University of Health Sciences, Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Radiology Department, Tevfik Saglam Cad. No: 11, Zuhuratbaba, 34147 Bakırkoy, Istanbul, Turkey
| | - Ercan Inci
- University of Health Sciences, Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Radiology Department, Tevfik Saglam Cad. No: 11, Zuhuratbaba, 34147 Bakırkoy, Istanbul, Turkey
| |
Collapse
|
3
|
Nalbant MO, Inci E, Akinci O, Aygan S, Gulturk U, Boluk Gulsever A. Evaluation of Imaging Findings of Pancreatobiliary and Intestinal Type Periampullary Carcinomas with 3.0T MRI. Acad Radiol 2023; 30:1846-1855. [PMID: 36585328 DOI: 10.1016/j.acra.2022.12.028] [Citation(s) in RCA: 2] [Impact Index Per Article: 2.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/23/2022] [Revised: 12/13/2022] [Accepted: 12/16/2022] [Indexed: 12/29/2022]
Abstract
RATIONALE AND OBJECTIVES The aim of this study was to differentiate pancreatobiliary and intestinal type periampullary carcinomas using dynamic contrast MRI and MRCP (Magnetic Resonance Cholangiopancreatography) with diffusion-weighted imaging (DWI) MATERIALS AND METHODS: MRI and MRCP images of 70 patients with pathologically proven periampullary adenocarcinoma were included. MRCP image features, extra-ampullary features, enhancement patterns, and apparent diffusion coefficient (ADC) values derived from b-values of 1000 s/mm² were evaluated by two radiologists independently. The interclass correlation coefficient (ICC) or Cohen's kappa statistic was used to evaluate the interobserver agreement. RESULTS 51 patients were diagnosed with pancreatobiliary type carcinomas, and 19 with intestinal type. In the pancreatobiliary subtype, the distal wall of the common bile duct was usually irregular (p = 0.047). Although the progressive enhancement pattern was evident in the pancreatobiliary type, an oval filling defect in the distal common bile duct was found to be more common in the intestinal type (p<0.001). The pancreatic duct cut-off sign (p<0.001), gastroduodenal artery involvement (p <0,001), and lymphadenopathy (p<0.05) were mostly observed in pancreatobiliary carcinomas. The ADCmin, ADCmean, and ADCmax values of the pancreatobiliary type carcinomas were all lower compared to the intestinal type carcinomas (p <0.05). CONCLUSION The oval filling defect seen in MRI and MRCP examinations suggests intestinal type, whereas the progressive contrasting pattern of the masses with irregular narrowing in the distal margin of the common bile duct, the pancreatic duct cut-off sign, gastroduodenal artery involvement, lymphadenopathy, and low ADC values indicate pancreatobiliary type carcinomas.
Collapse
Affiliation(s)
- Mustafa Orhan Nalbant
- Radiology Department, University of Health Sciences, Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Istanbul, Turkey.
| | - Ercan Inci
- Radiology Department, University of Health Sciences, Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Istanbul, Turkey
| | - Ozlem Akinci
- Radiology Department, University of Health Sciences, Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Istanbul, Turkey
| | - Sinan Aygan
- Radiology Department, University of Health Sciences, Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Istanbul, Turkey
| | - Ulas Gulturk
- Radiology Department, University of Health Sciences, Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Istanbul, Turkey
| | - Aycan Boluk Gulsever
- Radiology Department, University of Health Sciences, Bakirkoy Dr. Sadi Konuk Training and Research Hospital, Istanbul, Turkey
| |
Collapse
|
4
|
Dynamic contract-enhanced CT-based radiomics for differentiation of pancreatobiliary-type and intestinal-type periampullary carcinomas. Clin Radiol 2021; 77:e75-e83. [PMID: 34753589 DOI: 10.1016/j.crad.2021.09.010] [Citation(s) in RCA: 2] [Impact Index Per Article: 0.7] [Reference Citation Analysis] [Abstract] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 01/29/2021] [Accepted: 09/09/2021] [Indexed: 11/21/2022]
Abstract
AIM To investigate whether computed tomography (CT) radiomics can differentiate pancreatobiliary-type from intestinal-type periampullary carcinomas. MATERIALS AND METHODS CT radiomics of 96 patients (54 pancreatobiliary type and 42 intestinal type) with surgically confirmed periampullary carcinoma were assessed retrospectively. Volumes of interest (VOIs) were delineated manually. Radiomic features were extracted from preoperative CT images. A single-phase model and combined-phase model were constructed. Five-fold cross-validation and five machine-learning algorithms were utilised for model construction. The diagnostic performance of the models was evaluated by receiver operating characteristic (ROC) curves, and indicators included area under the curve (AUC), accuracy, sensitivity, specificity, and precision. ROC curves were compared using DeLong's test. RESULTS A total of 788 features were extracted on each phase. After feature selection using least absolute shrinkage and selection operator (LASSO) algorithm, the number of selected optimal feature was 18 (plain scan), nine (arterial phase), two (venous phase), 23 (delayed phase), 15 (three enhanced phases), and 29 (all phases), respectively. For the single-phase model, the delayed-phase model using the logistic regression (LR) algorithm showed the best prediction performance with AUC, accuracy, sensitivity, specificity, and precision of 0.89, 0.83, 0.80, 0.88, and 0.93, respectively. Two combined-phase models showed better results than the single-phase models. The model of all phases using the LR algorithm showed the best prediction performance with AUC, accuracy, sensitivity, specificity, and precision of 0.96, 0.88, 0.90, 0.93, and 0.92, respectively. CONCLUSION Radiomic models based on preoperative CT images can differentiate pancreatobiliary-type from intestinal-type periampullary carcinomas, in particular, the model of all phases using the LR algorithm.
Collapse
|
5
|
Lu JY, Yu H, Zou XL, Li Z, Hu XM, Shen YQ, Hu DY. Apparent diffusion coefficient-based histogram analysis differentiates histological subtypes of periampullary adenocarcinoma. World J Gastroenterol 2019; 25:6116-6128. [PMID: 31686767 PMCID: PMC6824280 DOI: 10.3748/wjg.v25.i40.6116] [Citation(s) in RCA: 6] [Impact Index Per Article: 1.2] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Download PDF] [Journal Information] [Submit a Manuscript] [Subscribe] [Scholar Register] [Received: 07/23/2019] [Revised: 09/17/2019] [Accepted: 09/28/2019] [Indexed: 02/06/2023] Open
Abstract
BACKGROUND For periampullary adenocarcinoma, the histological subtype is a better prognostic predictor than the site of tumor origin. Intestinal-type periampullary adenocarcinoma (IPAC) is reported to have a better prognosis than the pan-creatobiliary-type periampullary adenocarcinoma (PPAC). However, the classification of histological subtypes is difficult to determine before surgery. Apparent diffusion coefficient (ADC) histogram analysis is a noninvasive, non-enhanced method with high reproducibility that could help differentiate the two subtypes.
AIM To investigate whether volumetric ADC histogram analysis is helpful for distinguishing IPAC from PPAC.
METHODS Between January 2015 and October 2018, 476 consecutive patients who were suspected of having a periampullary tumor and underwent magnetic resonance imaging (MRI) were reviewed in this retrospective study. Only patients who underwent MRI at 3.0 T with different diffusion-weighted images (b-values = 800 and 1000 s/mm2) and who were confirmed with a periampullary adenocarcinoma were further analyzed. Then, the mean, 5th, 10th, 25th, 50th, 75th, 90th, and 95th percentiles of ADC values and ADCmin, ADCmax, kurtosis, skewness, and entropy were obtained from the volumetric histogram analysis. Comparisons were made by an independent Student's t-test or Mann-Whitney U test. Multiple-class receiver operating characteristic curve analysis was performed to determine and compare the diagnostic value of each significant parameter.
RESULTS In total, 40 patients with histopathologically confirmed IPAC (n = 17) or PPAC (n = 23) were enrolled. The mean, 5th, 25th, 50th, 75th, 90th, and 95th percentiles and ADCmax derived from ADC1000 were significantly lower in the PPAC group than in the IPAC group (P < 0.05). However, values derived from ADC800 showed no significant difference between the two groups. The 75th percentile of ADC1000 values achieved the highest area under the curve (AUC) for differentiating IPAC from PPAC (AUC = 0.781; sensitivity, 91%; specificity, 59%; cut-off value, 1.50 × 10-3 mm2/s).
CONCLUSION Volumetric ADC histogram analysis at a b-value of 1000 s/mm2 might be helpful for differentiating the histological subtypes of periampullary adenocarcinoma before surgery.
Collapse
Affiliation(s)
- Jing-Yu Lu
- Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430030, Hubei Province, China
- Department of Radiology, the First Affiliated Hospital of Soochow University, Suzhou 215006, Jiangsu Province, China
| | - Hao Yu
- Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430030, Hubei Province, China
| | - Xian-Lun Zou
- Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430030, Hubei Province, China
| | - Zhen Li
- Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430030, Hubei Province, China
| | - Xue-Mei Hu
- Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430030, Hubei Province, China
| | - Ya-Qi Shen
- Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430030, Hubei Province, China
| | - Dao-Yu Hu
- Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan 430030, Hubei Province, China
| |
Collapse
|
6
|
Lu J, Hu D, Tang H, Hu X, Shen Y, Li Z, Peng Y, Kamel I. Assessment of tumor heterogeneity: Differentiation of periampullary neoplasms based on CT whole-lesion histogram analysis. Eur J Radiol 2019; 115:1-9. [PMID: 31084752 DOI: 10.1016/j.ejrad.2019.03.021] [Citation(s) in RCA: 20] [Impact Index Per Article: 4.0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 10/18/2018] [Revised: 03/25/2019] [Accepted: 03/26/2019] [Indexed: 12/19/2022]
Abstract
PURPOSE To investigate the utility of whole-lesion histogram analysis from multidetector computed tomography (MDCT) for discrimination of duodenal adenocarcinoma (DAC), pancreatic ductal adenocarcinoma (PDAC) and gastrointestinal stromal tumor (GIST) around the periampullary area. MATERIALS AND METHODS 171 patients suspicious of periampullary tumors were examined by MDCT (arterial and venous phases) and treated with surgery. A total of 74 patients were finally included in this retrospective study (26 DACs, 20 PDACs, and 28 GISTs). The interobserver agreement was evaluated by intra-class correlation coefficient (ICC) test between two radiologists. Volumetric histogram analysis based on CT Kinetics software was performed on enhanced MDCT images that recorded different histogram parameters of arterial and venous phases, including mean, median, 10th, 25th, 75th, and 90th percentiles, as well as skewness, kurtosis and entropy. The extracted histogram parameters were compared between DAC, PDAC and GIST respectively by Mann-Whitney U tests with Bonferroni corrections. Receiver operating characteristic (ROC) curve analysis was used to determine the diagnostic ability of each significant parameter and the area under the curve (AUC) was calculated. RESULTS The whole-lesion CT histogram analysis demonstrated significant differences between DAC, PDAC, and GIST with different histogram features on both arterial and venous phase scans (all P < 0.05). In the ROC analysis, the 90th percentile of venous phase demonstrated the highest AUC of 0.854 (P < 0.001) for discriminating DAC from PDAC. Excellent discriminators of periampullary tumors were noted among the histogram features, namely the 90th percentile of arterial phase, which demonstrated AUCs of 0.809 and 0.936 (P < 0.001) respectively for distinguishing DAC and PDAC from GIST. CONCLUSION The whole-lesion CT histogram analysis could be useful for differential diagnosis of DAC, PDAC and GIST arising from the periampullary area. Further assessment is warranted to investigate the clinical role of histogram analysis based on MDCT.
Collapse
Affiliation(s)
- Jingyu Lu
- Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1095 Jiefang Avenue, Wuhan, Hubei, 430030, PR China.
| | - Daoyu Hu
- Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1095 Jiefang Avenue, Wuhan, Hubei, 430030, PR China.
| | - Hao Tang
- Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1095 Jiefang Avenue, Wuhan, Hubei, 430030, PR China.
| | - Xuemei Hu
- Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1095 Jiefang Avenue, Wuhan, Hubei, 430030, PR China.
| | - Yaqi Shen
- Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1095 Jiefang Avenue, Wuhan, Hubei, 430030, PR China.
| | - Zhen Li
- Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1095 Jiefang Avenue, Wuhan, Hubei, 430030, PR China.
| | - Yang Peng
- Department of Radiology, Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology, 1095 Jiefang Avenue, Wuhan, Hubei, 430030, PR China.
| | - Ihab Kamel
- Russell H. Morgan Department of Radiology and Radiological Science, The Johns Hopkins Medical Institutions, Baltimore, Maryland, USA.
| |
Collapse
|
7
|
Albayrak E, Sahin S. Evaluation of upper abdominal organs with DWI in patients with familial Mediterranean fever. Abdom Radiol (NY) 2017; 42:1393-1399. [PMID: 27909774 DOI: 10.1007/s00261-016-1005-3] [Citation(s) in RCA: 1] [Impact Index Per Article: 0.1] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Indexed: 01/02/2023]
Abstract
PURPOSE To investigate the diagnostic efficiency of diffusion-weighted magnetic resonance imaging (DWI) for the evaluation of functional changes that can occur in upper abdominal organs in patients with familial Mediterranean fever (FMF). METHODS The study included 50 controls, 45 patients with FMF, and 14 patients with FMF who had accompanying proteinuria. Measurement of apparent diffusion coefficient (ADC) was performed using DWI sections obtained from liver, spleen, kidney, and pancreas parenchyma with 1.5T MRI using b = 500 and b = 1000 s/mm2 values both in patients and control groups. Mean ADC values were compared between patient and control groups. RESULTS Renal ADC values were lower in the patient groups compared to the control group. Additionally, renal ADC values showed further decrease in the patient group in the presence of accompanying proteinuria, when compared to the FMF group without proteinuria (p < 0.001). Based on the ROC analysis, calculated cutoff values for the determination of FMF and FMF accompanied by proteinuria were 2.26 × 10-3 and 2.04 × 10-3 mm2/s, respectively. Liver, spleen, and pancreas ADC values did not show remarkable change between patient and control groups. CONCLUSION Present findings indicate that the presence of FMF and its clinical progression expressed by proteinuria can be differentially determined with renal DWI.
Collapse
|