基于深度学习的印度北部超声检测胆囊癌：一项前瞻性诊断研究。
Deep-learning enabled ultrasound based detection of gallbladder cancer in northern India: a prospective diagnostic study.
作者信息
Gupta Pankaj, Basu Soumen, Rana Pratyaksha, Dutta Usha, Soundararajan Raghuraman, Kalage Daneshwari, Chhabra Manika, Singh Shravya, Yadav Thakur Deen, Gupta Vikas, Kaman Lileswar, Das Chandan Krushna, Gupta Parikshaa, Saikia Uma Nahar, Srinivasan Radhika, Sandhu Manavjit Singh, Arora Chetan
机构信息
Department of Radiodiagnosis and Imaging, Postgraduate Institute of Medical Education and Research, Chandigarh, 160012, India.
Department of Computer Science and Engineering, Indian Institute of Technology, New Delhi, 110016, India.
出版信息
Lancet Reg Health Southeast Asia. 2023 Sep 11;24:100279. doi: 10.1016/j.lansea.2023.100279. eCollection 2024 May.
BACKGROUND
Gallbladder cancer (GBC) is highly aggressive. Diagnosis of GBC is challenging as benign gallbladder lesions can have similar imaging features. We aim to develop and validate a deep learning (DL) model for the automatic detection of GBC at abdominal ultrasound (US) and compare its diagnostic performance with that of radiologists.
METHODS
In this prospective study, a multiscale, second-order pooling-based DL classifier model was trained (training and validation cohorts) using the US data of patients with gallbladder lesions acquired between August 2019 and June 2021 at the Postgraduate Institute of Medical Education and research, a tertiary care hospital in North India. The performance of the DL model to detect GBC was evaluated in a temporally independent test cohort (July 2021-September 2022) and was compared with that of two radiologists.
FINDINGS
The study included 233 patients in the training set (mean age, 48 ± (2SD) 23 years; 142 women), 59 patients in the validation set (mean age, 51.4 ± 19.2 years; 38 women), and 273 patients in the test set (mean age, 50.4 ± 22.1 years; 177 women). In the test set, the DL model had sensitivity, specificity, and area under the receiver operating characteristic curve (AUC) of 92.3% (95% CI, 88.1-95.6), 74.4% (95% CI, 65.3-79.9), and 0.887 (95% CI, 0.844-0.930), respectively for detecting GBC which was comparable to both the radiologists. The DL-based approach showed high sensitivity (89.8-93%) and AUC (0.810-0.890) for detecting GBC in the presence of stones, contracted gallbladders, lesion size <10 mm, and neck lesions, which was comparable to both the radiologists (p = 0.052-0.738 for sensitivity and p = 0.061-0.745 for AUC). The sensitivity for DL-based detection of mural thickening type of GBC was significantly greater than one of the radiologists (87.8% vs. 72.8%, p = 0.012), despite a reduced specificity.
INTERPRETATION
The DL-based approach demonstrated diagnostic performance comparable to experienced radiologists in detecting GBC using US. However, multicentre studies are warranted to explore the potential of DL-based diagnosis of GBC fully.
FUNDING
None.
背景
胆囊癌（GBC）具有高度侵袭性。由于良性胆囊病变可能具有相似的影像学特征，GBC的诊断具有挑战性。我们旨在开发并验证一种深度学习（DL）模型，用于在腹部超声（US）检查时自动检测GBC，并将其诊断性能与放射科医生的诊断性能进行比较。
方法
在这项前瞻性研究中，使用2019年8月至2021年6月期间在印度北部一家三级护理医院研究生医学教育与研究机构获取的胆囊病变患者的US数据，训练了一个基于多尺度、二阶池化的DL分类器模型（训练和验证队列）。在一个时间上独立的测试队列（2021年7月至2022年9月）中评估DL模型检测GBC的性能，并与两名放射科医生的性能进行比较。
结果
该研究包括训练集中的233例患者（平均年龄48±（2标准差）23岁；142名女性）、验证集中的59例患者（平均年龄51.4±19.2岁；38名女性）和测试集中的273例患者（平均年龄50.4±22.1岁；177名女性）。在测试集中，DL模型检测GBC的灵敏度、特异度和受试者操作特征曲线下面积（AUC）分别为92.3%（95%CI，88.1 - 95.6）、74.4%（95%CI，65.3 - 79.9）和0.887（95%CI，0.844 - 0.930），与两名放射科医生的表现相当。基于DL的方法在存在结石、胆囊收缩、病变大小<10 mm和颈部病变的情况下检测GBC显示出高灵敏度（89.8 - 93%）和AUC（0.810 - 0.890），与两名放射科医生的表现相当（灵敏度p = 0.052 - 0.738，AUC p = 0.061 - 0.745）。基于DL检测壁增厚型GBC的灵敏度显著高于其中一名放射科医生（87.8%对72.8%，p = 0.012），尽管特异度有所降低。
解读
基于DL的方法在使用US检测GBC方面显示出与经验丰富的放射科医生相当的诊断性能。然而，需要进行多中心研究以充分探索基于DL诊断GBC的潜力。
资金
无。
Zotero 插件