Application of convolutional neural networks in ultrasound classification of hepatic cystic echinococcosis

doi:10.12140/j.issn.1000-7423.2024.04.005

Abstract

Abstract:

Objective To evaluate the application value of convolutional neural network (VGG19) model in the ultrasound diagnosis of hepatic cystic echinococcosis (HCE). Methods The ultrasound images of patients with hepatic cystic echinococcosis (HCE, including CE1-CE5 types) and patients with non-echinococcosis focal liver lesions (NHFLL) in the First Affiliated Hospital of Xinjiang Medical University from January 2012 to December 2022 were retrospectively collected. The VGG19 model was used to determine the image diagnosis of 6 types of hepatic focal space occupying lesions (CE1-CE5, NHFLL), and the percentage of each type determined was compared. When VGG19 misclassified HCE and NHFLL, the general demographic information and relevant clinical data of patients were compared. The ultrasound images were randomly divided into two groups based on the principle of essentially consistent proportion of different types, of which manual classification summary were performed by 2 junior ultrasound physicians and 2 senior ultrasound physicians on one group each selected randomly. The diagnostic accuracy of the model made by junior and senior ultrasound physicians were compared. The diagnostic performance of VGG19 was evaluated using confusion matrix, precision, recall, specificity, and F1 score. Descriptive analysis was conducted using contingency tables to describe the count data, and chi-square test， Fisher’s exact probability method and paired card square test were used for comparative analysis of differences. Results Among the 871 HCE cases, there were 203 cases of CE1, 227 of CE2, 110 of CE3, 159 of CE4, and 172 of CE5. The 600 NHFLL cases include, 300 cases of hepatic cysts, 150 of hepatic calcified lesions, and 150 solid hepatic occupying lesions (100 cases of hepatic hemangioma, 25 of hepatoma, and 25 of liver abscess). The overall accuracy of the VGG19 model was 82.0%, the recall rate was 87.9%, and the F1 score was 84.3%. The overall accuracy rate of VGG19 model was 86.2% (1 268/1 471), and the accuracy rates of each type from high to low were CE5 (95.3%, 164/172), CE4 (91.2%, 145/159), CE3 (89.1%, 98/110), CE1 (84.7%, 135/159), CE2 (84.6%, 192/227) and NHFLL (82.8%, 497/600), respectively. A total of 203 cases were misdiagnosed, and the misdiagnosis rate was 13.8% (203/1 471). Among them, 100 cases were misdiagnosed between HCE types, including 31 cases of CE1, 35 cases of CE2, 12 cases of CE3, 15 cases of CE4, and 8 cases of CE5. 103 cases of NHFLL were misdiagnosed as HCE, including 68 cases of hepatic cyst, 17 cases of liver calcification, 17 cases of liver hemangioma, and 1 case of liver abscess. No case of liver cancer was misdiagnosed. There were statistically significant differences in age, residential area, history of contact with dogs, education level and province between HCE patients and NHFLL patients misdiagnosed by VGG19 (χ² = 55.116, 24.197, 35.834, 14.069, 11.918, all P < 0.05), but there was no significant difference in gender (χ² = 0.047, P > 0.05). The overall diagnostic accuracy of VGG19 model (86.2%, 1 268/1 471) was higher than that by junior doctors (81.2%, 1 195/1 471) (P < 0.05), lower than that by senior physicians (92.3%, 1 358/1 471) (P < 0.05). Conclusion The VGG19 model can identify the 5 types of HCE and non echinococcal NHFLL, its diagnostic accuracy is lower than that by senior sonographers, but higher than that by junior sonographers. The model is expected to be promoted to primary hospitals to assist correction of ultrasound diagnosis in combination with clinical information.

Key words: Hepatic cystic echinococcosis, Ultrasound, Convolutional neural network, Focal liver lesions

CLC Number:

R532.32

REN Yan, SONG Tao, SHANG Feng, WU Miao, WANG Zhengye, WANG Xiaorong. Application of convolutional neural networks in ultrasound classification of hepatic cystic echinococcosis[J]. CHINESE JOURNAL OF PARASITOLOGY AND PARASITIC DISEASES, 2024, 42(4): 454-460.

Figures/Tables 8

Table 1

Fig. 1

Fig. 2

Table 2

Fig. 3

Fig. 4

Table 3

Table 4

References 22

[1]	Keong B, Wilkie B, Sutherland T, et al. Hepatic cystic echinococcosis in Australia: an update on diagnosis and management[J]. ANZ J Surg, 2018, 88(1/2): 26-31.
[2]	Brunetti E, Kern P, Vuitton DA. Expert consensus for the diagnosis and treatment of cystic and alveolar echinococcosis in humans[J]. Acta Trop, 2010, 114(1): 1-16. doi: 10.1016/j.actatropica.2009.11.001 pmid: 19931502
[3]	Deplazes P, Rinaldi L, Alvarez Rojas CA, et al. Global distribution of alveolar and cystic echinococcosis[J]. Adv Parasitol, 2017, 95: 315-493.
[4]	WHO Informal Working Group, 2003. International classification of ultrasound images in cystic echinococcosis for application in clinical and field epidemiological settings[J]. Acta Trop, 2003, 85(2): 253-261.
[5]	Kern P, Menezes da Silva A, Akhan O, et al. The echinococcoses: diagnosis, clinical management and burden of disease[J]. Adv Parasitol, 2017, 96: 259-369.
[6]	Brunetti E, Tamarozzi F, MacPherson C, et al. Ultrasound and cystic echinococcosis[J]. Ultrasound Int Open, 2018, 4(3): E70-E78.
[7]	Hou J, Wen H, Wang MK, et al. Analysis of the influencing factors of lesion activity in hepatic cystic echinococcosis patients[J]. Chin J Parasitol Parasit Dis, 2022, 40(3): 309-315. (in Chinese)
	(侯娇, 温浩, 王明坤, 等. 肝细粒棘球蚴病手术患者病灶活性状态的影响因素分析[J]. 中国寄生虫学与寄生虫病杂志, 2022, 40(3): 309-315.) doi: 10.12140/j.issn.1000-7423.2022.03.005
[8]	Ferrer Inaebnit E, Molina Romero FX, Segura Sampedro JJ, et al. A review of the diagnosis and management of liver hydatid cyst[J]. Rev Esp Enferm Dig, 2022, 114(1): 35-41.
[9]	Wen H, Vuitton L, Tuxun T, et al. Echinococcosis: advances in the 21st century[J]. Clin Microbiol Rev, 2019, 32(2): e00075-e00018.
[10]	Zhou LQ, Wang JY, Yu SY, et al. Artificial intelligence in medical imaging of the liver[J]. World J Gastroenterol, 2019, 25(6): 672-682.
[11]	Ahmed I, Chehri A, Jeon G. Artificial intelligence and blockchain enabled smart healthcare system for monitoring and detection of COVID-19 in biomedical images[J]. IEEE/ACM Trans Comput Biol Bioinform, 2024, 21(4): 814-822.
[12]	Morley J, DeVito NJ, Zhang J. Generative AI for medical research[J]. BMJ, 2023, 382: 1551.
[13]	Wu M, Yan CB, Wang XR, et al. Automatic classification of hepatic cystic echinococcosis using ultrasound images and deep learning[J]. J Ultrasound Med, 2022, 41(1): 163-174.
[14]	Ali Alghofaily K, Bin Saeedan M, Aljohani IM, et al. Hepatic hydatid disease complications: review of imaging findings and clinical implications[J]. Abdom Radiol (NY), 2017, 42(1): 199-210. doi: 10.1007/s00261-016-0860-2 pmid: 27487777
[15]	Nan JGL, Li R, Wang HX, et al. Ultrasound image classification of hepatic echinococcosis using deep learning[J]. J Shenzhen Univ Sci Eng, 2019, 36(6): 702-708. (in Chinese)
	(南嘉格列, 李锐, 王海霞, 等. 基于深度学习的肝包虫病超声图像分型研究[J]. 深圳大学学报(理工版), 2019, 36(6): 702-708.)
[16]	Ma GX, Yang LF, Yan CB, et al. Ultrasonic image diagnosis of hepatic echinococcosis based on deep Dense Netnetwork[J]. Comput Mod, 2023(9): 100-104. (in Chinese)
	(马国祥, 杨凌菲, 严传波, 等. 基于深度DenseNet网络的肝包虫病超声影像诊断方法[J]. 计算机与现代化, 2023(9): 100-104.)
[17]	Yang YF, Cairang YD, Jiang TA, et al. Ultrasound identification of hepatic echinococcosis using a deep convolutional neural network model in China: a retrospective, large-scale, multicentre, diagnostic accuracy study[J]. Lancet Digit Health, 2023, 5(8): e503-e514.
[18]	Bai JH, Zhang YH, Han XM, et al. Imaging analysis of 79 cases of hepatic echinococcosis[J]. Chin J Schisto Control, 2018, 30(6): 674-677. (in Chinese)
	(白峻虎, 张永海, 韩秀敏, 等. 79例肝棘球蚴病患者影像学分析[J]. 中国血吸虫病防治杂志, 2018, 30(6): 674-677.)
[19]	Surgical Committee for Hydatidosis of the Surgical Branch of the Chinese Medical Association. Expert consensus on diagnosis and treatment of hepatic cystic and alveolar echinococcosis (2019 edition)[J]. Chin J Dig Surg, 2019, 18(8): 711-721. (in Chinese)
	(中国医师协会外科医师分会包虫病外科专业委员会. 肝两型包虫病诊断与治疗专家共识(2019版)[J]. 中华消化外科杂志, 2019, 18(8): 711-721.)
[20]	Liu YY, Zhang TT, Ma X, et al. Awareness and influencing factors of knowledge on echinococcosis prevention and control among adults in Qinghai Province[J]. Chin J Parasitol Parasit Dis, 2022, 40(6): 723-729, 736. (in Chinese)
	(刘玉英, 张甜甜, 马霄, 等. 青海省成年人棘球蚴病防治知识知晓情况及影响因素分析[J]. 中国寄生虫学与寄生虫病杂志, 2022, 40(6): 723-729, 736.) doi: 10.12140/j.issn.1000-7423.2022.06.005
[21]	Li BF, Yang J, Yang JY, et al. Epidemiological investigation and case retrospective analysis of echinococcosis in Dali Prefecture, Yunnan Province[J]. Chin J Parasitol Parasit Dis, 2023, 41(5): 573-578. (in Chinese)
	(李奔福, 杨敬, 杨金玉, 等. 云南省大理州棘球蚴病流行病学调查和病例回顾性分析[J]. 中国寄生虫学与寄生虫病杂志, 2023, 41(5): 573-578.) doi: 10.12140/j.issn.1000-7423.2023.05.008
[22]	Faraj W, Abi Faraj C, Kanso M, et al. Hydatid disease of the liver in the middle east: a single center experience[J]. Surg Infect (Larchmt), 2022, 23(1): 29-34.

模型诊断 Model diagnosis	实际病种 Actual disease type
模型诊断 Model diagnosis	阳性 Positive	阴性 Negative
阳性 Positive	真阳性 True positive	假阳性 False positive
阴性 Negative	假阴性 False negative	真阴性 True negative

类型 Type	CE1	CE2	CE3	CE4	CE5	NHFLL
精确率/% Precision/%	77.8	88.5	68.5	78.0	79.2	100
召回率/% Recall/%	84.7	84.6	89.1	91.2	95.3	82.8
特异度/% Specificity/%	96.1	98.2	96.7	93.1	98.8	100
F1分数/% F1 score/%	81.1	86.5	77.5	84.1	85.6	90.6

特征 Feature	病例数（占比/%) No. case (Percentage/%)		χ²值 χ² value	P值 P value
特征 Feature	HCE （n = 100）	NHFLL （n = 103）	χ²值 χ² value	P值 P value
年龄/岁 Age/Year			55.116	< 0.05^a
0~29	26（26.0）	0（0.0）
30~59	64（64.0）	59（57.3）
≥ 60	10（10.0）	44（42.7）
性别 Gender			0.047	> 0.05
男 Male	52（52.0）	52（50.5）
女 Female	48（48.0）	51（49.5）
住址类别 Address type			24.197	< 0.05
城市 City	44（44.0）	80（77.7）
农村 Countryside	56（56.0）	23（22.3）
犬类接触史 Dog contact history			35.834	< 0.05
有 Yes	65（65.0）	24（23.3）
无 No	35（35.0）	79（76.7）
文化程度 Degree of education			14.069	< 0.05
文盲 Illiteracy	8（8.0）	2（1.9）
小学 Primary school	43（43.0）	26（25.3）
中学 Middle school	36（36.0）	49（47.6）
大学 University	13（13.0）	26（25.2）
省份 Province			11.918	< 0.05^a
新疆 Xinjiang	89（89.0）	103（100）
西藏 Xizang	6（6.0）	0（0）
其他 Other	5（5.0）	0（0）

类别（n） Category (n)	病例数（占比/%) No. case (Percentage/%)				P	病例数（占比/%) No. case (Percentage/%)		P
类别（n） Category (n)	VGG19模型 VGG model		低年资医师 Junior physician		P	高年资医师 Senior physician		P
	正确 True	错误 Wrong	正确 True	错误 Wrong		正确 True	错误 Wrong
CE1（203）	172（84.7）	31（15.3）	191（94.1）	12（5.9）	< 0.05	193（95.1）	10（4.9）	< 0.05
CE2（227）	192（84.6）	35（15.4）	178（78.4）	49（21.6）	< 0.05	208（91.6）	19（8.4）	< 0.05
CE3（110）	98（89.1）	12（10.9）	83（75.5）	27（24.5）	< 0.05	106（96.4）	4（3.6）	< 0.05
CE4（159）	145（91.2）	14（8.8）	130（81.8）	29（18.2）	< 0.05	154（96.9）	5（3.1）	< 0.05
CE5（172）	164（95.3）	8（4.7）	147（85.5）	25（14.5）	< 0.05	166（96.5）	6（3.5）	> 0.05
NHFLL（600）	497（82.8）	103（17.2）	466（77.7）	134（22.3）	< 0.05	531（88.5）	69（11.5）	< 0.05
合计（1 471）	1 268（86.2）	203（13.8）	1 195（81.2）	276（18.8）	< 0.05	1 358（92.3）	113（7.7）	< 0.05