从结构化分级到智能分级：人工智能与ACR-RADS融合的研究进展

From structured reporting to intelligent grading：research advances in the integration of artificial intelligence and ACR-RADS

摘要

影像报告与数据系统（reporting and data system，RADS）是美国放射学会（American College of Radiology，ACR）旨在降低影像判读主观性、提升一致性的结构化风险分层框架。虽已广泛应用，但在真实世界实践中仍面临评分一致性受限、特征量化不足及跨中心可重复性差等方法学挑战。近年来，人工智能（artificial intelligence，AI）尤其是大语言模型（large language model，LLM）的发展为解决上述局限提供了新路径。文章系统梳理了主要RADS体系的结构特征及实践局限，综述了AI在病灶识别、风险再分层及流程规范化等方面的实证进展。现有证据显示，AI更适宜作为RADS的“增强层”，以提升其客观性与重复性。文章重点探讨了LLM在语义理解、自动推理及质控中的潜力，并从监管视角展望了人机协作的演进方向。该趋势对推动我国影像诊断同质化、提升基层诊疗水平及构建智能化监管体系具有重要指导意义。

Abstract

The reporting and data systems（RADS），established by the American College of Radiology（ACR），serve as structured risk-stratification frameworks designed to mitigate subjectivity and enhance consistency in radiological interpretation. Despite their widespread adoption，RADS still face methodological challenges in real-world practice，including limited inter-observer consistency，insufficient feature quantification，and suboptimal cross-center reproducibility. In recent years，the evolution of artificial intelligence（AI），particularly large language models（LLMs），has offered novel pathways to address these limitations. This paper systematically reviews the structural characteristics and practical constraints of major RADS frameworks and synthesizes empirical progress regarding AI in lesion identification，risk re-stratification，and workflow standardization. Current evidence suggests that AI is best positioned as an“augmentative layer”for RADS to bolster objectivity and reproducibility. Furthermore，this article explores the potential of LLMs in semantic understanding，automated reasoning，and quality control，while projecting the evolution of human-computer collaboration from a regulatory perspective. This trend holds significant implications for promoting the homogenization of diagnostic imaging，empowering primary healthcare services，and establishing intelligent regulatory systems in China.

关键词

影像报告与数据系统 / 人工智能 / 结构化影像报告 / 风险分层

Key words

reporting and data system / artificial intelligence / structured imaging reporting / risk stratification

参考文献

[1] YADAV A, SINGH A, MADAAN S, et al. International academy of cytology Yokohama system for reporting breast cytology and the ACR breast imaging reporting and data system（BIRADS）：are they concordant?[J]. Iran J Pathol, 2024, 19(4):400-407

[2] SPAK D A, PLAXCO J S, SANTIAGO L, et al. BI—RADS® fifth edition：a summary of changes[J]. Diagn Interv Imaging, 2017, 98(3):179-190

[3] MERCADO C L. BI—RADS update[J]. Radiol Clin North Am, 2014, 52(3):481-487

[4] COZZI A, PINKER K, HIDBER A, et al. BI—RADS category assignments by GPT—3.5，GPT—4，and google bard：a multilanguage study[J]. Radiology, 2024, 311(1):e232133

[5] YOON S H, KIM Y J, DOH K, et al. Interobserver variability in Lung CT Screening Reporting and Data System categorisation in subsolid nodule-enriched lung cancer screening CTs[J]. Eur Radiol, 2021, 31(9):7184-7191

[6] KIRI L, ABDOLELL M, COSTA A F, et al. US LI—RADS visualization score：interobserver variability and association with cause of liver disease，sex，and body mass index[J]. J L’association Can Des Radiol, 2022, 73(1):68-74

[7] LEE A Y, WISNER D J, AMINOLOLAMA—SHAKERI S, et al. Inter—reader variability in the use of BI—RADS descriptors for suspicious findings on diagnostic mammography：a multi-institution study of 10 academic radiologists[J]. Acad Radiol, 2017, 24(1):60-66

[8] 姜艳, 项霞青, 周婉. 美国放射学会甲状腺影像报告和数据系统分类在甲状腺结节中的观察者一致性研究[J]. 医疗装备, 2022, 35(21):114-117

[9] JIANG Y, XIANG L Q, ZHOU W. Observer consistency of American college of radiology thyroid imaging Reporting and data system classification in thyroid nodules[J]. Medical Equipment, 2022, 35(21):114-117

[10] SEVIM S, DICLE O, GEZER N S, et al. How high is the inter-observer reproducibility in the LIRADS reporting system?[J]. Pol J Radiol, 2019, 84:e464-e469

[11] ABDEL RAZEK A A K, EL—SEROUGY L G, SALEH G A, et al. Reproducibility of LI—RADS treatment response algorithm for hepatocellular carcinoma after locoregional therapy[J]. Diagn Interv Imaging, 2020, 101(9):547-553

[12] MULÉ S, RONOT M, GHOSN M, et al. Automated CT LI—RADS v2018 scoring of liver observations using machine learning：a multivendor，multicentre retrospective study[J]. JHEP Rep, 2023, 5(10):100857

[13] GALGANO S J, SMITH E N. LI—RADS version 2018 for MRI and CT：interreader agreement in real—world practice[J]. Radiology, 2023, 307(5):e231212

[14] SHUKLA—DAVE A, OBUCHOWSKI N A, CHENEVERT T L, et al. Quantitative imaging biomarkers alliance（QIBA）recommendations for improved precision of DWI and DCE—MRI derived biomarkers in multicenter oncology trials[J]. J Magn Reson Imaging, 2019, 49(7):e101-e121

[15] WANG C D, SHAO J, HE Y C, et al. Data-driven risk stratification and precision management of pulmonary nodules detected on chest computed tomography[J]. Nat Med, 2024, 30(11):3184-3195

[16] YANG D D, LEE L K, TSUI J M G, et al. AI—derived tumor volume from multiparametric MRI and outcomes in localized prostate cancer[J]. Radiology, 2024, 313(1):e240041

[17] 陈敏. 智能赋影，融合创新：人工智能时代下医学影像学科的发展与展望[J]. 中国医学影像学杂志, 2026, 34（ 1):1-2

[18] CHEN M. Intelligent imaging empowerment and integrative innovation：development and prospects of the medical imaging discipline in the era of artificial intelligence[J]. Chinese Journal of Medical Imaging, 2026, 34（ 1):1-2

[19] YAN B, PEI M T, ZHAO M, et al. Prior guided transformer for accurate radiology reports generation[J]. IEEE J Biomed Health Inform, 2022, 26(11):5631-5640

[20] BHAYANA R. Chatbots and large language models in radiology：a practical primer for clinical and research applications[J]. Radiology, 2024, 310(1):e232756

[21] 刘再毅, 赵致禾, 石镇维. 医学影像大模型：肿瘤精准诊疗的范式革新[J]. 协和医学杂志, 2025, 16(4):805-811

[22] LIU Z Y, ZHAO Z H, SHI Z W. Large medical imaging models：a paradigm innovation for precision diagnosis and treatment of tumors[J]. Medical Journal of Peking Union Medical College Hospital, 2025, 16(4):805-811

[23] AIKEN A H, RATH T J, ANZAI Y, et al. ACR neck imaging reporting and data systems（NI—RADS）：a white paper of the ACR NI—RADS committee[J]. J Am Coll Radiol, 2018, 15(8):1097-1108

[24] REINHOLD C, ROCKALL A, SADOWSKI E A, et al. Ovarian—adnexal reporting lexicon for MRI：a white paper of the ACR ovarian-adnexal reporting and data systems MRI committee[J]. J Am Coll Radiol, 2021, 18(5):713-729

[25] HOANG J K, MIDDLETON W D, TESSLER F N. Update on ACR TI—RADS：successes，challenges，and future directions，from the AJR special series on radiology reporting and data systems[J]. AJR Am J Roentgenol, 2021, 216(3):570-578

[26] TESSLER F N, MIDDLETON W D, GRANT E G, et al. ACR thyroid imaging，reporting and data system（TI-RADS）：white paper of the ACR TI—RADS committee[J]. J Am Coll Radiol, 2017, 14(5):587-595

[27] WEINREB J C, BARENTSZ J O, CHOYKE P L, et al. PI—RADS prostate imaging-reporting and data system：2015，version 2[J]. Eur Urol, 2016, 69(1):16-40

[28] AGRAWAL R, AHUJA J, STRANGE C D, et al. Lung-RADS v2022 update[J]. Radiol Clin North Am, 2025, 63(4):507-516

[29] SCHIMA W, KOPF H, EISENHUBER E. LI—RADS made easy[J]. Rofo, 2023, 195(6):486-494

[30] TESSLER F N, MIDDLETON W D, GRANT E G. Thyroid imaging reporting and data system（TI—RADS）：a user’s guide[J]. Radiology, 2018, 287(1):29-36

[31] GHASEMI A, AHLAWAT S. Bone reporting and data system（bone—RADS）and other proposed practice guidelines for reporting bone tumors[J]. Rofo, 2024, 196(11):1134-1142

[32] CURY R C, LEIPSIC J, ABBARA S, et al. CAD—RADS™ 2.0—2022 coronary artery disease—reporting and data system an expert consensus document of the society of cardiovascular computed tomography（SCCT），the American college of cardiology（ACC），the American college of radiology（ACR）and the North America society of cardiovascular imaging（NASCI）[J]. Radiol Cardiothorac Imag, 2022, 4(5):e220183

[33] YEE J, DACHMAN A, KIM D H, et al. CT colonography reporting and data system（C—RADS）：version 2023 update[J]. Radiology, 2024, 310(1):e232007

[34] 乔敏, 冯尚勇, 沈德娟, 等. 中国甲状腺影像报告和数据系统对甲状腺结节良恶性的鉴别诊断[J]. 中国医学影像学杂志, 2021, 29(11):1070-1075

[35] QIAO M, FENG S Y, SHEN D J, et al. Differential diagnosis of benign and malignant thyroid nodules using the Chinese Thyroid Imaging Reporting and Data System[J]. Chinese Journal of Medical Imaging, 2021, 29(11):1070-1075

[36] KAHN C E Jr, LANGLOTZ C P, BURNSIDE E S, et al. Toward best practices in radiology reporting[J]. Radiology, 2009, 252(3):852-856

[37] KNIGHTLY D V, PLACE J N. Best practices in computerized tomography[J]. Radiol Manage, 2010, 32(1):16-23

[38] GRANATA V, DE MUZIO F, CUTOLO C, et al. Structured reporting in radiological settings：pitfalls and perspectives[J]. J Pers Med, 2022, 12(8):1344

[39] 张紫欣, 吕志彬, 王宇新, 等. LI—RADS 2018版MRI辅助征象在肝癌分类及预后评估中的应用[J]. 临床放射学杂志, 2024, 43(5):858-861

[40] ZHANG Z X, LV Z B, WANG Y X, et al. Application of LI—RADS 2018 MRI ancillary features in hepatocellular carcinoma classification and prognostic evaluation[J]. Journal of Clinical Radiology, 2024, 43(5):858-861

[41] AN J Y, UNSDORFER K M L, WEINREB J C. BI—RADS，C—RADS，CAD—RADS，LI—RADS，lung—RADS，NI—RADS，O—RADS，PI—RADS，TI—RADS：reporting and data systems[J]. RadioGraphics, 2019, 39(5):1435-1436

[42] 唐雅伦, 李瑞, 高磊, 等. 人工智能辅助诊断系统与Lung—RADS对不同临床特征肺结节的良恶性预测效能[J]. 分子影像学杂志, 2025, 48(6):668-677

[43] TANG Y L, LI R, GAO L, et al. Predictive performance of an artificial intelligence-assisted diagnostic system and Lung-RADS for malignancy of pulmonary nodules with different clinical characteristics[J]. Journal of Molecular Imaging, 2025, 48(6):668-677

[44] 刘丹丹, 韦平, 高彦, 等. 联合Kaiser评分、DWI及MRS对乳腺BI—RADS3—5类病变的诊断效能[J]. 临床放射学杂志, 2025, 44(12):2290-2294

[45] LIU D, WEI P, GAO Y, et al. Diagnostic performance of combined Kaiser Score，diffusion—weighted imaging，and magnetic resonance spectroscopy for breast BI—RADS category 3—5 lisions[J]. Journal of Clinical Radiology, 2025, 44(12):2290-2294

[46] VAN RIEL S J, JACOBS C, SCHOLTEN E T, et al. Observer variability for Lung-RADS categorisation of lung cancer screening CTs：impact on patient management[J]. Eur Radiol, 2019, 29(2):924-931

[47] BERG W A, CAMPASSI C, LANGENBERG P, et al. Breast Imaging Reporting and Data System：inter— and intraobserver variability in feature analysis and final assessment[J]. AJR Am J Roentgenol, 2000, 174(6):1769-1777

[48] ALIKHASSI A, ESMAILI GOURABI H, BAIKPOUR M. Comparison of inter-and intra-observer variability of breast density assessments using the fourth and fifth editions of Breast Imaging Reporting and Data System[J]. Eur J Radiol Open, 2018, 5:67-72

[49] BECKER A S, BARTH B K, MARQUEZ P H, et al. Increased interreader agreement in diagnosis of hepatocellular carcinoma using an adapted LI—RADS algorithm[J]. Eur J Radiol, 2017, 86:33-40

[50] CHOI H H, KIM S, SHUM D J, et al. Assessing adherence to US LI—RADS follow—up recommendations in vulnerable patients undergoing hepatocellular carcinoma surveillance[J]. Radiol Imaging Cancer, 2024, 6(1):e230118

[51] BOSS M A, MALYARENKO D, PARTRIDGE S, et al. The QIBA profile for diffusion—weighted MRI：apparent diffusion coefficient as a quantitative imaging biomarker[J]. Radiology, 2024, 313:e233055

[52] RAUNIG D L, MCSHANE L M, PENNELLO G, et al. Quantitative imaging biomarkers：a review of statistical methods for technical performance assessment[J]. Stat Methods Med Res, 2015, 24(1):27-67

[53] HAGIWARA A, FUJITA S, OHNO Y, et al. Variability and standardization of quantitative imaging：monoparametric to multiparametric quantification，radiomics，and artificial intelligence[J]. Invest Radiol, 2020, 55(9):601-616

[54] 程枫, 艾慧俊, 冯娅琴, 等. 描述性标准化超声报告对甲状腺乳头状癌诊断的影响[J]. 现代实用医学, 2018, 30(11):1507-1509

[55] CHENG F, AI H J, FENG Y Q, et al. Impact of descriptive standardized ultrasound reporting on the diagnosis of papillary thyroid carcinoma[J]. Modern Practical Medicine, 2018, 30(11):1507-1509

[56] LIN Y N, FU M Z, DING R W, et al. Patient adherence to lung CT screening reporting & data system—recommended screening intervals in the United States：a systematic review and meta—analysis[J]. J Thorac Oncol, 2022, 17(1):38-55

[57] LACSON R, WANG A J, COCHON L, et al. Factors associated with optimal follow—up in women with BI—RADS 3 breast findings[J]. J Am Coll Radiol, 2020, 17(4):469-474

[58] LIN X H, LIAO T T, YANG Y T, et al. Value of deep learning model for predicting breast imaging reporting and data system 3 and 4A lesions on mammography[J]. Quant Imaging Med Surg, 2025, 15(5):4047-4058

[59] WU H C, LIU F, YANG Q S, et al. Automated MRI system for clinically significant prostate cancer detection development validation and real—world implementation[J]. Nat Commun, 2025, 16(1):11583

[60] XING Z Y, CHEN J, PAN L, et al. Enhanced detection of prostate cancer lesions on biparametric MRI using artificial intelligence：a multicenter，fully—crossed，multi—reader multi—case trial[J]. Acad Radiol, 2025, 32(10):5954-5963

[61] CHEN H, YANG B W, QIAN L, et al. Deep learning prediction of ovarian malignancy at US compared with O—RADS and expert assessment[J]. Radiology, 2022, 304(1):106-113

[62] WU Y N, WHITE G M, CORNELIUS T, et al. Deep learning LI—RADS grading system based on contrast enhanced multiphase MRI for differentiation between LR—3 and LR—4/LR—5 liver tumors[J]. Ann Transl Med, 2020, 8(11):701

[63] HENDRIX W, HENDRIX N, SCHOLTEN E T, et al. Deep learning for the detection of benign and malignant pulmonary nodules in non—screening chest CT scans[J]. Commun Med, 2023, 3(1):156

[64] DONG X Y, CHEN G L, ZHU Y P, et al. Artificial intelligence in skeletal metastasis imaging[J]. Comput Struct Biotechnol J, 2024, 23:157-164

[65] HUANG R Q, MENG X L, ZHANG X X, et al. Artificial intelligence-driven change redefining radiology through interdisciplinary innovation[J]. Interdiscip Med, 2025, 3(1):e20240063

[66] SU M J, LIANG X, ZENG X T, et al. Using a large language model for breast imaging reporting and data system classification and malignancy prediction to enhance breast ultrasound diagnosis：retrospective study[J]. JMIR Med Inform, 2025, 13:e70924

[67] GERTZ R J, DRATSCH T, BUNCK A C, et al. Potential of GPT—4 for detecting errors in radiology reports：implications for reporting accuracy[J]. Radiology, 2024, 311(1):e232714

[68] WELLS B J, NGUYEN H M, MCWILLIAMS A, et al. A practical framework for appropriate implementation and review of artificial intelligence（FAIR—AI）in healthcare[J]. NPJ Digit Med, 2025, 8(1):514

[69] YOU J G, HERNANDEZ—BOUSSARD T, PFEFFER M A, et al. Clinical trials informed framework for real world clinical implementation and deployment of artificial intelligence applications[J]. NPJ Digit Med, 2025, 8(1):107

[70] 唐蛰雨, 李绍钦, 贾中芝. 基于CT图像的深度学习在主动脉夹层中的应用进展[J]. 南京医科大学学报（自然科学版）, 2024, 44(8):1174-1178

[71] TANG H Y, LI S M, JIA Z Z. Application progress of deep learning based on CT images in aortic dissection[J]. Journal of Nanjing Medical University（Natural Sciences）, 2024, 44(8):1174-1178

[72] AL KUWAITI A, NAZER K, AL—REEDY A, et al. A review of the role of artificial intelligence in healthcare[J]. J Pers Med, 2023, 13(6):951

[73] YANG Y, CUI Y U, WANG Y T, et al. Interpretation of the WHO’s“ethics and governance of artificial intelligence for health：guidance on large multi—modal models”and its implications for China[J]. Chin J Prev Med, 2025, 59(6):960-969

[74] 汪洋, 马玉, 刘颖. 人工智能背景下医疗器械创新监管研究与启示[J]. 中国设备工程, 2025,(23):272-274

[75] WANG Y, MA Y, LIU Y. Research and implications of innovative regulation of medical devices in the context of artificial intelligence[J]. China Plant Engineering, 2025,(23):272-274

[76] 许琳琳, 黄德群. 人工智能应用于医疗领域的法律法规及技术挑战[J]. 临床医学工程, 2025, 32(11):1177-1182

[77] XU L L, HUANG D Q. Legal regulations and technical challenges of artificial intelligence applications in the medical field[J]. Clinical Medical Engineering, 2025, 32(11):1177-1182

基金资助

广东省自然科学基金项目(2023A1515012499)

广东省自然科学基金项目(2025A1515010318)

广州市科技计划重点项目(2024B03J0781)

广东省医学装备学会科研基金项目(YZXH2025KT10)

广东省基础与应用基础研究基金(2024A1515220081)