Abstract
Purpose
To verify whether radiomics techniques based on dual-modality ultrasound consisting of B-mode and superb microvascular imaging (SMI) can improve the accuracy of the differentiation between gallbladder neoplastic polyps and cholesterol polyps.
Methods
A total of 100 patients with 100 pathologically proven gallbladder polypoid lesions were enrolled in this retrospective study. Radiomics features on B-mode ultrasound and SMI of each lesion were extracted. Support vector machine was used to classify adenomas and cholesterol polyps of gallbladder for B-mode, SMI and dual-modality ultrasound, respectively, and the classification results were compared among the three groups.
Results
Six, eight and nine features were extracted for each lesion at B-mode ultrasound, SMI and dual-modality ultrasound, respectively. In dual-modality ultrasound model, the area under the receiver operating characteristic curve (AUC), classification accuracy, sensitivity, specificity, and Youden’s index were 0.850 ± 0.090, 0.828 ± 0.097, 0.892 ± 0.144, 0.803 ± 0.149 and 0.695 ± 0.157, respectively. The AUC and Youden’s index of the dual-modality model were higher than those of the B-mode model (p < 0.05). The AUC, accuracy, specificity and Youden’s index of the dual-modality model were higher than those of the SMI model (p < 0.05).
Conclusions
Radiomics analysis of the dual-modality ultrasound composed of B-mode and SMI can improve the accuracy of classification between gallbladder neoplastic polyps and cholesterol polyps.
Similar content being viewed by others
Introduction
Gallbladder polyps are defined as a general term for a group of diseases that originate in the wall of the gallbladder and protrude restrictively into the lumen [1], also known as polypoid lesions of gallbladder (PLGs). With the popularity of ultrasound, the detection rate of PLGs has increased. Epidemiological studies show that the prevalence of PLGs is about 0.3–12.3% in adults, where gallbladder neoplastic polyps account for only about 5% of PLGs [1, 2]. As the most common gallbladder polyps, cholesterol polyps are foamy cell clusters formed by cholesterol crystals deposited in the gallbladder wall and phagocytosed by macrophages, with the surface covering the mucosal layer of the gallbladder, and no malignant tendency has been reported in the literature [3]. According to the 2019 WHO Classification of Tumors [4], gallbladder neoplastic polyps include pyloric gland adenomas and intracholecystic papillary neoplasms. Gallbladder neoplastic polyps are prone to atypical hyperplasia and may progress to gallbladder cancer, and they are considered as precancerous lesions [5, 6]. Therefore, the accurate differentiation between cholesterol polyps and precancerous gallbladder neoplastic polyps by preoperative imaging is a pressing issue in clinical practice.
Ultrasound is the preferred imaging method for evaluating PLGs, but it is difficult to accurately distinguish cholesterol polyps from gallbladder neoplastic polyps by conventional ultrasound [7, 8]. How to improve the accuracy of ultrasonic identification of gallbladder polyp-like lesions is an urgent clinical problem. Superb microvascular imaging (SMI), as a new modality of ultrasound imaging, has unique advantages in showing the morphology of microvasculature with low flow velocity in the lesion, thereby significantly improving the resolution, sensitivity and specificity of ultrasound diagnosis [9]. The combination of the two ultrasonic modalities, namely conventional ultrasound and SMI, could contribute to more accurate diagnosis of PLG.
Recently, novel imaging technologies based on radiomics (AI) have made rapid advances, where algorithms process medical imaging data sets through hierarchical mathematical models that can learn to use biometrics to detect diagnostic patterns. Zhang et al. [10] established a neoplastic predictive model and evaluated the effectiveness of radiomics in predicting malignancy in patients with gallbladder polyps. A single-center study by ** to 8% when there is lymph node metastasis, and the rate of stage 4b gallbladder cancer is only 2% [24]. Therefore, early detection of gallbladder cancer and precancerous lesions and early intervention are important to improve the survival rate of patients.
Ultrasound has been recognized as the first choice of imaging examination for the screening and follow-up of gallbladder polyps. However, it is difficult to identify gallbladder neoplastic polyps or cholesterol polyps solely from the echogenicity, morphology, or vascularity characteristics of the lesion [25, 26]. In Table 1 of this study, although the polyp diameter, pedicle and blood flow are statistically different between the gallbladder neoplastic polyp group and cholesterol polyp group, the diagnostic accuracy is low based on these indicators. Accordingly, Domestic and international scientific guidelines over the years recommend cholecystectomy for gallbladder polyps larger than 1 cm in diameter [27]. However, the guidelines for gallbladder polyps have resulted in a large number of unnecessary cholecystectomy, which has been questioned by many scholars [28, 29]. Therefore, there is an urgent need for a new method with high accuracy and reproducibility to accurately identify gallbladder neoplastic polyps and cholesterol polyps.
The application of artificial intelligence in medical imaging is one of the hot spots in medical research. Based on the big data analysis of computer, we can obtain numerous objective image feature data with a resolution far beyond the human eye [16, 30,31,32]. In our previous study, we extracted spatial and morphological features of single-modality gray-scale ultrasound. Our study reported that adenomas polyps have a more uniform pixel distribution, with a relatively smaller proportion of hyperechoic areas inside the polyps, and adenomas polyps are larger and more irregular in morphology. These are closely related to the pathophysiological features of gallbladder neoplastic polyps and cholesterol polyps [3, 6].
In this study, we extracted dual-modality ultrasound image datasets of gray-scale ultrasound and SMI. SMI technique is a method to evaluate tissue microvascular perfusion, and it can detect low speed flow signal without contrast agent, thus giving us more diagnostic information. In the process of radiomics analysis in this paper, the SVM model using dual-modality images’ features had the best discriminative ability, and its AUC, ACC, SEN, SPE and YI all reached the best level. There were statistical differences in the AUC, YI between the dual-modality images and the B-mode images. And there were statistical differences in the AUC, ACC, SPE and YI between the dual-modality images and SMI images. It could be seen that the calibration curve (Fig. 5) of the dual-modality model was closer to the diagonal dotted line corresponding to the perfect prediction model than the single-modality models, which indicated that the dual-modality model had better prediction performance. In conclusion, the dual-modality images combined with B-mode images and SMI images have the potential to improve the accuracy of classification of PLGs. In future clinical practice, the dual-modality radiomics features will be extracted, and the classification model based on the dual-modality radiomics features will help to accurately and early identify the gallbladder neoplastic polyps from cholesterol polyps.
Our study also had some limitations. This study was our preliminary attempt to obtain dual-modality ultrasound images’ parameters as well as apply radiomics technology to identify gallbladder neoplastic polyps and cholesterol polyps of gallbladder. The sample size of this study was small, it will be expanded for deep learning in the follow-up [33]. Moreover, this study was a single-center study, and a prospective multicenter study with a large sample size need to be conducted for further validation in the future. The results in this paper are preliminary and guarantee further robust studies in the future.
Conclusions
In conclusion, with the analysis of radiomics, the dual-modality ultrasound combining B-mode ultrasound and SMI showed an excellent classification accuracy for gallbladder neoplastic polyps and cholesterol polyps of gallbladder. Our model has high sensitivity and specificity at differentiating gallbladder polyps, which means that it could be potentially used in clinical practice to avoid unnecessary cholecystectomies and missing diagnosis of gallbladder neoplastic polyps.
Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Abbreviations
- SMI:
-
Superb microvascular imaging
- PLG:
-
Polypoid lesion of gallbladder
- LASSO:
-
Least absolute shrinkage and selection operator
- GLCM:
-
Gray level co-occurrence matrix
- AUC:
-
Area under the receiver operating characteristic curve
- ACC:
-
Accuracy
- SEN:
-
Sensitive
- SPE:
-
Specificity
- YI:
-
Youden’s index
References
McCain RS, Diamond A, Jones C, et al. Current practices and future prospects for the management of gallbladder polyps: a topical review. World J Gastroenterol. 2018;24:2844–52. https://doi.org/10.3748/wjg.v24.i26.2844.
Şahiner İT, Dolapçı M. When should gallbladder polyps be treated surgically? Adv Clin Exp Med. 2018;27:1697–700. https://doi.org/10.17219/acem/75678.
Oestmann A. Polyps of the gallbladder. Praxis. 2012;101:581–4. https://doi.org/10.1024/1661-8157/a000919.
Nagtegaal ID, Odze RD, Klimstra D, et al. The 2019 WHO classification of tumours of the digestive system. Histopathology. 2020;76:182–8. https://doi.org/10.1111/his.13975.
Guettier C. Pathology of gallbladder and extrahepatic bile ducts. Case 7. Biliary-type tubulopapillary adenoma without dysplasia. Ann Pathol. 2014;34:315–23. https://doi.org/10.1016/j.annpat.2014.06.001.
Kai K, Aishima S, Miyazaki K. Gallbladder cancer: clinical and pathological approach. World J Clin Cases. 2014;2:515–21. https://doi.org/10.12998/wjcc.v2.i10.515.
Wang X, Zhu J-A, Liu Y-J, et al. Conventional ultrasound combined with contrast-enhanced ultrasound in differential diagnosis of gallbladder cholesterol and adenomatous polyps (1–2 cm). J Ultrasound Med. 2022;41:617–26. https://doi.org/10.1002/jum.15740.
Yuan H-X, Cao J-Y, Kong W-T, et al. Contrast-enhanced ultrasound in diagnosis of gallbladder adenoma. Hepatobiliary Pancreat Dis Int. 2015;14:201–7. https://doi.org/10.1016/s1499-3872(15)60351-4.
Miller DD, Brown EW. Artificial intelligence in medical practice: The question to the answer? Am J Med. 2018;131:129–33. https://doi.org/10.1016/j.amjmed.2017.10.035.
Zhang X, Wang J, Wu B, et al. A nomogram-based model and ultrasonic radiomic features for gallbladder polyp classification. J Gastroenterol Hepatol. 2022;37:1380–8. https://doi.org/10.1111/jgh.15841.
**ang F, Liang X, Yang L, et al. Contrast-enhanced CT radiomics for prediction of recurrence-free survival in gallbladder carcinoma after surgical resection. Eur Radiol. 2022;32:7087–97. https://doi.org/10.1007/s00330-022-08858-5.
Ohno Y, Fujimoto T, Shibata Y. A new era in diagnostic ultrasound, superb microvascular imaging: preliminary results in pediatric hepato-gastrointestinal disorders. Eur J Pediatr Surg. 2017;27:20–5. https://doi.org/10.1055/s-0036-1593381.
Yang X, Liu Y, Guo Y, et al. Utility of radiomics based on contrast-enhanced CT and clinical data in the differentiation of benign and malignant gallbladder polypoid lesions. Abdom Radiol (NY). 2020;45:2449–58. https://doi.org/10.1007/s00261-020-02461-2.
Peng JB, Peng YT, Lin P, et al. Differentiating infected focal liver lesions from malignant mimickers: value of ultrasound-based radiomics. Clin Radiol. 2022;77:104–13. https://doi.org/10.1016/j.crad.2021.10.009.
Yu B, Li Y, Yu X, et al. Differentiate thyroid follicular adenoma from carcinoma with combined ultrasound radiomics features and clinical ultrasound features. J Digit Imaging. 2022. https://doi.org/10.1007/s10278-022-00639-2.
Colling R, Pitman H, Oien K, et al. Artificial intelligence in digital pathology: a roadmap to routine use in clinical practice. J Pathol. 2019;249:143–50. https://doi.org/10.1002/path.5310.
Zhou K, Liu Z, He W, et al. Application of 3D whole-brain texture analysis and the feature selection method based on within-class scatter in the classification and diagnosis of Alzheimer’s disease. Ther Innov Regul Sci. 2022. https://doi.org/10.1007/s43441-021-00373-x.
Liu Z, Zhu G, Jiang X, et al. Survival prediction in gallbladder cancer using CT based machine learning. Front Oncol. 2020;10:604288. https://doi.org/10.3389/fonc.2020.604288.
Yu F, Hang J, Deng J, et al. Radiomics features on ultrasound imaging for the prediction of disease-free survival in triple negative breast cancer: a multi-institutional study. Br J Radiol. 2021;94:20210188. https://doi.org/10.1259/bjr.20210188.
Wu Y-Q, Gao R-Z, Lin P, et al. An endorectal ultrasound-based radiomics signature for preoperative prediction of lymphovascular invasion of rectal cancer. BMC Med Imaging. 2022;22:84–84. https://doi.org/10.1186/s12880-022-00813-6.
Yang M, Liu H, Dai Q, et al. Treatment response prediction using ultrasound-based pre-, post-early, and delta radiomics in neoadjuvant chemotherapy in breast cancer. Front Oncol. 2022. https://doi.org/10.3389/fonc.2022.748008.
Hearst MA. Support vector machines. Ieee Intell Syst Appl. 1998;13:18–21. https://doi.org/10.1109/5254.708428.
Li Q, Jiang T, Zhang C, et al. A nomogram based on clinical information, conventional ultrasound and radiomics improves prediction of malignant parotid gland lesions. Cancer Lett. 2022;527:107–14. https://doi.org/10.1016/j.canlet.2021.12.015.
Amin MB, Greene FL, Edge SB, et al. The eighth edition AJCC cancer staging manual: continuing to build a bridge from a population-based to a more “personalized” approach to cancer staging. CA Cancer J Clin. 2017;67:93–9. https://doi.org/10.3322/caac.21388.
Yu MH, Kim YJ, Park HS, et al. Benign gallbladder diseases: imaging techniques and tips for differentiating with malignant gallbladder diseases. World J Gastroenterol. 2020;26:2967. https://doi.org/10.3748/wjg.v26.i22.2967.
Okaniwa S. Role of conventional ultrasonography in the diagnosis of gallbladder polypoid lesions. J Med Ultrason. 2021;48:149–57. https://doi.org/10.1007/s10396-019-00989-5.
Babu BI, Dennison AR, Garcea G. Management and diagnosis of gallbladder polyps: a systematic review. Langenbecks Arch Surg. 2015;400:455–62. https://doi.org/10.1007/s00423-015-1302-2.
Lu D, Radin R, Yung E, et al. Malignant transformation of a 5-mm gallbladder polyp over 2 years a case report and review of current literature. Ultrasound Q. 2015;31:66–8. https://doi.org/10.1097/ruq.0000000000000094.
Chae H-D, Lee JY, Jang J-Y, et al. Photoacoustic imaging for differential diagnosis of benign polyps versus malignant polyps of the gallbladder: a preliminary study. Korean J Radiol. 2017;18:821–7. https://doi.org/10.3348/kjr.2017.18.5.821.
Dalmis MU, Gubern-Merida A, Vreemann S, et al. Artificial intelligence–based classification of breast lesions imaged with a multiparametric breast MRI protocol with ultrafast DCE-MRI, T2, and DWI. Invest Radiol. 2019;54:325–32. https://doi.org/10.1097/rli.0000000000000544.
O’Sullivan S, Heinsen H, Grinberg LT, et al. The role of artificial intelligence and machine learning in harmonization of high-resolution post-mortem MRI (virtopsy) with respect to brain microstructure. Brain Inform. 2019;6:3–3. https://doi.org/10.1186/s40708-019-0096-3.
Weisberg EM, Chu LC, Park S, et al. Deep lessons learned: radiology, oncology, pathology, and computer science experts unite around artificial intelligence to strive for earlier pancreatic cancer diagnosis. Diagn Interv Imaging. 2020;101:111–5. https://doi.org/10.1016/j.diii.2019.09.007.
Huang Q, Luo H, Yang C, et al. Anatomical prior based vertebra modelling for reappearance of human spines. Neurocomputing. 2022. https://doi.org/10.1016/j.neucom.2022.05.033.
Acknowledgements
Not applicable.
Funding
This work was supported by the National Natural Science Foundation of China [62071285]; Shanghai Science and Technology Innovation Action Plan [21Y11911200]; Shanghai Municipal Key Clinical Specialty [shslczdzk03501]; and Fujian Province for Health and Science Research Project [2019-ZQNB-39].
Author information
Authors and Affiliations
Contributions
YHX collected and organized the patient data regarding the gallbladder polyps and was a major contributor in writing the manuscript. WCY analyzed and interpreted the patient data regarding the gallbladder polyps and was a major contributor in writing the manuscript. TCY collected and organized the patient data regarding the gallbladder polyps and assisted in revising the manuscript. YQQ collected and organized the patient data regarding the gallbladder polyps and assisted in revising the manuscript. ZQ analyzed and interpreted the patient data regarding the gallbladder polyps and designed the experimental process. WWP collected and organized the patient data regarding the gallbladder polyps and assisted in revising the manuscript. All authors read and approved the final manuscript.
Corresponding authors
Ethics declarations
Ethics approval and consent to participate
The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013) and approved by the Ethics Committee of Zhongshan Hospital (B2022-187R) for the retrospective review of images and patients’ medical records. The need for informed consent was waived by the Ethics Committee of Zhongshan Hospital for the retrospective design.
Consent for publication
Not applicable.
Competing interests
The authors declare that they have no competing interests.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
About this article
Cite this article
Yuan, Hx., Wang, C., Tang, Cy. et al. Differential diagnosis of gallbladder neoplastic polyps and cholesterol polyps with radiomics of dual modal ultrasound: a pilot study. BMC Med Imaging 23, 26 (2023). https://doi.org/10.1186/s12880-023-00982-y
Received:
Accepted:
Published:
DOI: https://doi.org/10.1186/s12880-023-00982-y