基于对比增强计算机断层扫描并结合机器学习与纹理特征的肾上腺皮质癌、嗜铬细胞瘤及肾上腺腺瘤分类：横断面研究

封面图片

如何引用文章

全文:

详细

论证：基于对比增强计算机断层扫描对肾上腺皮质癌、嗜铬细胞瘤及肾上腺腺瘤进行鉴别诊断仍然是一项复杂的任务，这主要与上述病变在影像学特征上的显著重叠有关。基于标准形态学特征的现有分类方法诊断准确性有限，可能导致诊断错误及不恰当的治疗策略选择。

目的：开发一种基于对比增强计算机断层扫描影像并结合纹理特征的机器学习模型，用于肾上腺肿瘤（腺瘤、肾上腺皮质癌及嗜铬细胞瘤）的多分类。

方法：本研究为单中心横断面研究，其中计算机断层扫描数据的收集为回顾性，影像结果的再分析为前瞻性。采用PyRadiomics软件对对比增强计算机断层扫描各期图像进行处理，计算纹理特征。为减少扫描参数差异的影响，对数据进行了标准化处理。使用LightGBM、XGBoost和CatBoost梯度提升模型，并采用分层五折交叉验证进行训练。模型诊断性能通过recall、precision、F1-score、宏平均F1-score、特异度、Balanced Accuracy以及各诊断类别的曲线下面积（AUC）进行评估。

结果：共纳入425例经组织学验证的肾上腺肿瘤患者，其中肾上腺皮质癌42例、嗜铬细胞瘤204例、腺瘤179例。所构建的机器学习模型在交叉验证中对腺瘤（XGBoost，F1-score最高达0.916）和嗜铬细胞瘤（XGBoost，F1-score最高达0.855）的分类表现出较高准确性，而对肾上腺皮质癌的分类性能明显较低（CatBoost，F1-score最高为0.521）。AUC的最佳值分别为：腺瘤0.971（LightGBM）、嗜铬细胞瘤0.924（LightGBM）和肾上腺皮质癌0.879（CatBoost）。平衡准确率最高达0.773，宏平均F1-score 达到0.747（CatBoost）。对最具信息量特征的分析表明，用于分类的重要参数包括表征不同对比增强期中纹理元素均一性及强度的指标。

结论：放射组学结合机器学习方法在基于对比增强计算机断层扫描对肾上腺腺瘤和嗜铬细胞瘤进行多分类方面表现出较高的诊断准确性。然而，肾上腺皮质癌的分类准确性仍然较低，这可能与肿瘤的高度异质性及病例数量有限有关。

作者简介

Almaz V. Manaev

Endocrinology Research Centre; National Research Nuclear University “MEPhI”

Email: a.manaew2016@yandex.ru
ORCID iD: 0009-0003-8035-676X
SPIN 代码: 2902-9767
俄罗斯联邦, Moscow; Moscow

Natalia V. Tarbaeva

Endocrinology Research Centre

Email: ntarbaeva@inbox.ru
ORCID iD: 0000-0001-7965-9454
SPIN 代码: 5808-8065

MD, Cand. Sci. (Medicine)

俄罗斯联邦, Moscow

Svetlana A. Buryakina

Endocrinology Research Centre

Email: sburyakina@yandex.ru
ORCID iD: 0000-0001-9065-7791
SPIN 代码: 5675-0651

MD, Cand. Sci. (Medicine)

俄罗斯联邦, Moscow

Liliya D. Kovalevich

Endocrinology Research Centre

Email: liliyakovalevich@gmail.com
ORCID iD: 0000-0001-8958-8223
SPIN 代码: 1642-5694
俄罗斯联邦, Moscow

Angelina V. Khairieva

Endocrinology Research Centre

Email: komarito@mail.ru
ORCID iD: 0000-0002-6758-5918
SPIN 代码: 4516-8297
俄罗斯联邦, Moscow

Liliya S. Urusova

Endocrinology Research Centre

Email: liselivanova89@yandex.ru
ORCID iD: 0000-0001-6891-0009
SPIN 代码: 5151-3675

MD, Dr. Sci. (Medicine)

俄罗斯联邦, Moscow

Nano V. Pachuashvili

Endocrinology Research Centre

Email: npachuashvili@bk.ru
ORCID iD: 0000-0002-8136-0117
SPIN 代码: 3477-8994

MD, Cand. Sci. (Medicine)

俄罗斯联邦, Moscow

Galina A. Mel'nichenko

Endocrinology Research Centre

编辑信件的主要联系方式.
Email: Melnichenko.Galina@endocrincentr.ru
ORCID iD: 0000-0002-5634-7877
SPIN 代码: 8615-0038

MD, Dr. Sci. (Medicine), Professor

俄罗斯联邦, Moscow

Natalia G. Mokrysheva

Endocrinology Research Centre

Email: mokrisheva.natalia@endocrincentr.ru
ORCID iD: 0000-0002-9717-9742
SPIN 代码: 5624-3875

MD, Dr. Sci. (Medicine), Professor

俄罗斯联邦, Moscow

Valentin E. Sinitsyn

Lomonosov Moscow State University; Research and Practical Clinical Center for Diagnostics and Telemedicine Technologies

Email: vsini@mail.ru
ORCID iD: 0000-0002-5649-2193
SPIN 代码: 8449-6590

MD, Dr. Sci. (Medicine), Professor

俄罗斯联邦, Moscow; Moscow

参考

  1. Albano D, Agnello F, Midiri F, et al. Imaging features of adrenal masses. Insights into Imaging. 2019;10(1):1–16. doi: 10.1186/s13244-019-0688-8 EDN: ETVIBX
  2. Fassnacht M, Tsagarakis S, Terzolo M, et al. European Society of Endocrinology clinical practice guidelines on the management of adrenal incidentalomas, in collaboration with the European Network for the Study of Adrenal Tumors. European Journal of Endocrinology. 2023;189(1):G1–G42. doi: 10.1093/ejendo/lvad066 EDN: EZKFAO
  3. Ebbehoj A, Li D, Kaur RJ, et al. Epidemiology of adrenal tumours in Olmsted County, Minnesota, USA: a population-based cohort study. The Lancet Diabetes & Endocrinology. 2020;8(11):894–902. doi: 10.1016/s2213-8587(20)30314-4 EDN: ICZBID
  4. Bancos I, Taylor AE, Chortis V, et al. Urine steroid metabolomics for the differential diagnosis of adrenal incidentalomas in the EURINE-ACT study: a prospective test validation study. The Lancet Diabetes & Endocrinology. 2020;8(9):773–781. doi: 10.1016/S2213-8587(20)30218-7 EDN: ZXDDAR
  5. Hong AR, Kim JH, Park KS, et al. Optimal follow-up strategies for adrenal incidentalomas: reappraisal of the 2016 ESE-ENSAT guidelines in real clinical practice. European Journal of Endocrinology. 2017;177(6):475–483. doi: 10.1530/EJE-17-0372
  6. Schloetelburg W, Ebert I, Petritsch B, et al. Adrenal wash-out CT: moderate diagnostic value in distinguishing benign from malignant adrenal masses. European Journal of Endocrinology. 2022;186(2):183–193. doi: 10.1530/EJE-21-0650
  7. Marty M, Gaye D, Perez P, et al. Diagnostic accuracy of computed tomography to identify adenomas among adrenal incidentalomas in an endocrinological population. European Journal of Endocrinology. 2018;178(5):439–446. doi: 10.1530/EJE-17-1056
  8. Vilar L, Freitas M, Canadas V, et al. Adrenal Incidentalomas: Diagnostic Evaluation and Long-Term Follow-up. Endocrine Practice. 2008;14(3):269–278. doi: 10.4158/EP.14.3.269
  9. Winoker JS, Ahlborn DT, Omidele OO, et al. Minimally invasive adrenal surgery: virtue or vice. Future Oncology. 2018;14(3):267–276. doi: 10.2217/fon-2017-0420
  10. Elhassan YS, Alahdab F, Prete A, et al. Natural history of adrenal incidentalomas with and without mild autonomous cortisol excess. Annals of Internal Medicine. 2019;171(2):107–116. doi: 10.7326/M18-3630
  11. Collienne M, Timmesfeld N, Bergmann S, et al. Adrenal incidentaloma and subclinical Cushing’s syndrome: a longitudinal follow-up study by endoscopic ultrasound. Ultraschall in der Medizin - European Journal of Ultrasound. 2015;38(04):411–419. doi: 10.1055/s-0041-107996 EDN: YCAZGU
  12. Corwin MT, Chalfant JS, Loehfelm TW, et al. Incidentally detected bilateral adrenal nodules in patients without cancer: is further workup necessary? American Journal of Roentgenology. 2018;210(4):780–784. doi: 10.2214/AJR.17.18543
  13. Goh Z, Phillips I, Hunt PJ, et al. Three-year follow up of adrenal incidentalomas in a New Zealand centre. Internal Medicine Journal. 2020;50(3):350–356. doi: 10.1111/imj.14332 EDN: ULXDXA
  14. Leung K, Stamm M, Raja A, Low G. Pheochromocytoma: the range of appearances on ultrasound, CT, MRI, and functional imaging. American Journal of Roentgenology. 2013;200(2):370–378. doi: 10.2214/AJR.12.9126
  15. Eloyan A, Yue MS, Khachatryan D. Tumor heterogeneity estimation for radiomics in cancer. Statistics in Medicine. 2020;39(30):4704–4723. doi: 10.1002/sim.8749 EDN: HJELUM
  16. Zhang H, Lei H, Pang J. Diagnostic performance of radiomics in adrenal masses: a systematic review and meta-analysis. Frontiers in Oncology. 2022;12:975183. doi: 10.3389/fonc.2022.975183 EDN: OJYRFN
  17. Ferro M, Tataru OS, Carrieri G, et al. Artificial intelligence and radiomics applications in adrenal lesions: a systematic review. Therapeutic Advances in Urology. 2025;17. doi: 10.1177/17562872251352553
  18. Cao L, Yang H, Wu H, et al. Adrenal indeterminate nodules: CT-based radiomics analysis of different machine learning models for predicting adrenal metastases in lung cancer patients. Frontiers in Oncology. 2024;14. doi: 10.3389/fonc.2024.1411214 EDN: YPYTOT
  19. Qi S, Zuo Y, Chang R, et al. Using CT radiomic features based on machine learning models to subtype adrenal adenoma. BMC Cancer. 2023;23(1):1411214. doi: 10.1186/s12885-023-10562-6 EDN: TAHRXY
  20. Tucci L, Vara G, Morelli V, et al. Prediction of adrenal masses nature through texture analysis and deep learning: Preliminary results from ENS@T RADIO-AI multicentric study. In: Endocrine Abstracts. Harrogate: Society for Endocrinology; 2024. doi: 10.1530/endoabs.99.oc11.3 EDN: SYRWPF
  21. Weiss LM. Comparative histologic study of 43 metastasizing and nonmetastasizing adrenocortical tumors. The American Journal of Surgical Pathology. 1984;8(3):163–170. doi: 10.1097/00000478-198403000-00001
  22. Bisceglia M, Ludovico O, Di Mattia A, et al. Adrenocortical oncocytic tumors: report of 10 cases and review of the literature. International Journal of Surgical Pathology. 2004;12(3):231–243. doi: 10.1177/106689690401200304
  23. Thompson LDR. Pheochromocytoma of the adrenal gland scaled score (PASS) to separate benign from malignant neoplasms. The American Journal of Surgical Pathology. 2002;26(5):551–566. doi: 10.1097/00000478-200205000-00002
  24. Urusova LS, Kletskaya IS, Porubayeva EE, Beltsevich DG. Adrenocortical carcinoma: modern concepts of morphological diagnosis and classification. Russian Journal of Archive of Pathology. 2023;85(4):32–38. doi: 10.17116/patol20238504132 EDN: XEDMRQ
  25. van Griethuysen JJM, Fedorov A, Parmar C, et al. Computational radiomics system to decode the radiographic phenotype. Cancer Research. 2017;77(21):e104–e107. doi: 10.1158/0008-5472.CAN-17-0339
  26. Varghese BA, Cen SY, Jensen K, et al. Investigating the role of imaging factors in the variability of CT-based texture analysis metrics. Journal of Applied Clinical Medical Physics. 2023;25(4):e14192. doi: 10.1002/acm2.14192
  27. Brodersen KH, Ong CS, Stephan KE, Buhmann JM. The balanced accuracy and its posterior distribution. In: Proceedings of the 20th International Conference on Pattern Recognition. Istanbul; 2010. P. 3121–3124. doi: 10.1109/ICPR.2010.764
  28. Saito T, Rehmsmeier M. The Precision-Recall Plot Is More Informative than the ROC plot when evaluating binary classifiers on imbalanced datasets. PLOS ONE. 2015;10(3):e0118432. doi: 10.1371/journal.pone.0118432 EDN: YBEDRE
  29. Fassnacht M, Dekkers OM, Else T, et al. European Society of Endocrinology Clinical Practice Guidelines on the management of adrenocortical carcinoma in adults, in collaboration with the European Network for the Study of Adrenal Tumors. European Journal of Endocrinology. 2018;179(4):G1–G46. doi: 10.1530/EJE-18-0608
  30. Sonoda S, Murata N. Neural network with unbounded activation functions is universal approximator. Applied and Computational Harmonic Analysis. 2017;43(2):233–268. doi: 10.1016/j.acha.2015.12.005
  31. Chen T, Guestrin C. XGBoost: a scalable tree boosting system. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. San Francisco; 2016. P. 785–794. doi: 10.1145/2939672.2939785
  32. Grinsztajn L, Oyallon E, Varoquaux G. Why do tree-based models still outperform deep learning on typical tabular data? In: Proceedings of 36th Conference on Neural Information Processing Systems (NeurIPS 2022) Track on Datasets and Benchmarks. 2022. Available from: https://papers.neurips.cc/paper_files/paper/2022/
  33. Walz MK, Metz KA, Theurer S, et al. Differentiating benign from malignant adrenocortical tumors by a single morphological parameter—a clinicopathological study on 837 adrenocortical neoplasias. Indian Journal of Surgical Oncology. 2020;11(4):705–710. doi: 10.1007/s13193-020-01205-4 EDN: OBEYXF
  34. Robertson-Tessi M, Gillies RJ, Gatenby RA, Anderson ARA. Impact of metabolic heterogeneity on tumor growth, invasion, and treatment outcomes. Cancer Research. 2015;75(8):1567–1579. doi: 10.1158/0008-5472.CAN-14-1428
  35. Kotelnikova LP, Zhizhilev YV. Computed tomography in the differential diagnosis of adrenal tumors. Endocrine Surgery. 2022;15(4):38–43. doi: 10.14341/serg12761 EDN: NPFHOR
  36. Phadte A, Krishnappa B, Memon SS, et al. High diagnostic accuracy of arterial phase CT in differentiating pheochromocytoma in good/poor washout adrenal masses. Journal of the Endocrine Society. 2024;9(1):bvae199. doi: 10.1210/jendso/bvae199
  37. Altay C, Basara Akin I, Ozgul AH, et al. Machine learning analysis of adrenal lesions: Preliminary study evaluating texture analysis in the differentiation of adrenal lesions. Diagnostic and Interventional Radiology. 2023;29(2):234–243. doi: 10.5152/dir.2022.21266 EDN: TEUYQK
  38. Florek P, Zagdański A. Benchmarking state-of-the-art gradient boosting algorithms for classification. arXiv. 2023. (in press). doi: 10.48550/arXiv.2305.17094
  39. Lau SK, Weiss LM. The Weiss system for evaluating adrenocortical neoplasms: 25 years later. Human Pathology. 2009;40(6):757–768. doi: 10.1016/j.humpath.2009.03.010
  40. Lam AK. Update on adrenal tumours in 2017 World Health Organization (WHO) of endocrine tumours. Endocrine Pathology. 2017;28(3):213–227. doi: 10.1007/s12022-017-9484-5 EDN: DZHAOK
  41. Manaev AV, Tarbaeva NV, Roslyakova AA, et al. Predicting high proliferative Ki-67 index in patients with adrenocortical carcinoma based on texture analysis of contrast-enhanced computed tomography images: a cross-sectional study. Digital Diagnostics. 2025;6(3):360–372. doi: 10.17816/DD643532 EDN: JRXXMQ

补充文件

附件文件
动作
1. JATS XML
下载
2. Fig. 1. Heat map of texture feature values for different adrenal gland formations. The standardised mean value (equal to the ratio of the difference between the mean value of a specific diagnosis and the mean value of the entire sample to the standard deviation of the entire sample) and the initial mean value (in brackets) of the corresponding texture feature indicator are given. ACC — adrenocortical carcinoma.

下载 (315KB)
索引源数据
3. Fig. 2. Classification accuracy of test sample observations (for all three models considered): 0 — adenomas; 1 — pheochromocytomas; 2 — adrenocortical carcinoma.

下载 (53KB)
索引源数据

版权所有 © Eco-Vector, 2025

Creative Commons License
此作品已接受知识共享署名-非商业性使用-禁止演绎 4.0国际许可协议的许可。

Согласие на обработку персональных данных

 

Используя сайт https://journals.rcsi.science, я (далее – «Пользователь» или «Субъект персональных данных») даю согласие на обработку персональных данных на этом сайте (текст Согласия) и на обработку персональных данных с помощью сервиса «Яндекс.Метрика» (текст Согласия).