用于分期CT扫描中骨病灶全面自动检测与分类的深度学习模型
本研究基于402例患者的分期CT开发nnUNet深度学习模型,用于检测和分类骨病灶,并在69例独立测试集(32例骨转移)中评估;模型检出70%的真实病灶(恶性67%、良性72%),良恶性分类准确率85%(敏感度91%、特异度81%),患者水平AI预测疾病负荷比例的AUC为0.82。
要点
- 模型在独立测试集检出70%真实骨病灶,恶性病灶检出率67%低于良性病灶72%。
- 良恶性分类准确率85%、敏感度91%、特异度81%,部分良性患者出现假阳性。
- 以AI预测疾病负荷比例做患者水平评估,AUC为0.82。
核心结论
- 作者的病灶检测与分类AI模型表现准确,并具有纠正医师错误的潜力。
- 作者认为需进一步研究该模型能否在检出率、分类准确性和阅片时间方面影响医师阅片。
研究设计与样本
- 本研究为采用nnUNet模型的深度学习研究,使用402例患者的CT开发模型。
- 开发数据包括前列腺癌伴良性或恶性成骨性(硬化性)骨病灶患者,以及多种原发癌伴良性或恶性溶骨性骨病灶患者。
- 由一名专家放射科医师勾画真实病灶,模型在病灶水平评估检测性能。
- 分类性能计算准确率、敏感度、特异度及其他指标。
- 保留的测试集为69例患者,其中32例有骨转移;AI预测疾病负荷的AUC在患者水平计算。
诊断标准与关键数值
- 独立测试集中检出70%的真实病灶,其中恶性病灶检出67%、良性病灶检出72%。
- 模型将病灶分类为恶性或良性的准确率为85%,敏感度91%、特异度81%。
- 以AI预测疾病负荷比例进行患者水平评估,AUC为0.82。
临床应用与阅片要点
- 原文未据此给出推荐。
局限性
- AI在若干良性患者中识别出假阳性。
- 作者在结论中提出需进一步研究模型能否影响医师阅片的检出率、分类准确性和阅片时间。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
RATIONALE AND OBJECTIVES: A common site of metastases for a variety of cancers is the bone, which is challenging and time consuming to review and important for cancer staging. Here, we developed a deep learning approach for detection and classification of bone lesions on staging CTs.
MATERIALS AND METHODS: This study developed an nnUNet model using 402 patients' CTs, including prostate cancer patients with benign or malignant osteoblastic (blastic) bone lesions, and patients with benign or malignant osteolytic (lytic) bone lesions from various primary cancers. An expert radiologist contoured ground truth lesions, and the model was evaluated for detection on a lesion level. For classification performance, accuracy, sensitivity, specificity, and other metrics were calculated. The held-out test set consisted of 69 patients (32 with bone metastases). The AUC of AI-predicted burden of disease was calculated on a patient level.
RESULTS: In the independent test set, 70% of ground truth lesions were detected (67% of malignant lesions and 72% of benign lesions). The model achieved accuracy of 85% in classifying lesions as malignant or benign (91% sensitivity and 81% specificity). Although AI identified false positives in several benign patients, the patient-level AUC was 0.82 using predicted disease burden proportion.
CONCLUSION: Our lesion detection and classification AI model performs accurately and has the potential to correct physician errors. Further studies should investigate if the model can impact physician review in terms of detection rate, classification accuracy, and review time.
引用
- Simon BD, Harmon SA, Yang D, et al. A Deep Learning Model for Comprehensive Automated Bone Lesion Detection and Classification on Staging Computed Tomography Scans. Acad Radiol. 2025. PMID: 40634223. doi:10.1016/j.acra.2025.06.021
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。