基于机器学习构建放射性碘治疗疗效的预测模型

doi:10.6040/j.issn.1671-7554.0.2022.1111

摘要/Abstract

摘要： 目的基于分化型甲状腺癌(DTC)患者的临床资料及放射学参数,通过机器学习算法构建放射性碘治疗(RAI)疗效的预测模型。方法选取2015年12月至2020年12月于山东大学第二医院核医学科接受RAI治疗的1 642例DTC患者为研究对象,在RAI治疗结束的6个月后评估其治疗疗效,筛选与疗效相关的核心特征进行机器学习建模。将研究对象按就诊时间(2019年7月)划分为训练集(n=973)和验证集(n=669),于训练集中利用Logistic、随机森林、支持向量机、Adaboost 4种方法进行模型构建,利用受试者工作特征曲线下面积(AUC)、灵敏度和特异度评估模型的性能,绘制校准曲线及决策曲线评估模型的准确度和临床受益性,并在验证集中评估模型外部稳定性。结果 4种模型的预测性能较高,稳定性较好,预测精度和净收益高于目前临床常规应用的肿瘤-淋巴结-转移(TNM)分期和复发风险分层。Logistic模型表现最佳,其AUC在训练集中为0.827、验证集中为0.869。结论基于机器学习构建的术后RAI治疗疗效预测模型有较高的预测性能,由此构建的列线图可实现个体化精准预测。

关键词: 放射性碘治疗, 机器学习, 分化型甲状腺癌, 预测模型, 治疗疗效

Abstract: Objective To construct the predictive models for radioactive iodine therapy(RAI)efficacy by machine learning algorithms based on clinical data and radiological parameters of patients with differentiated thyroid cancer(DTC). Methods A total of 1,642 DTC patients treated with RAI during Dec. 2015 and Dec. 2020 were collected. The efficacy was evaluated 6 months after RAI and the core features associated with efficacy were screened out for machine-learning modeling. The patients were divided into the training set(n=973)and validation set(n=669)according to the consultation time(July 2019). In the training set, the Logistic model, random forest model, support vector machine model and Adaboost model were constructed, whose performances were evaluated with the area under the receiver operating characteristic curve(AUC), accuracy and specificity. The calibration and decision curves were drawn to evaluate the accuracy and clinical benefits of the models. The external stability was assessed in the validation set. Results The four predictive models had great predictive performance, having higher predictive accuracy and net benefits than the tumor-node-metastasis(TNM)staging and recurrence risk stratification. The Logistic model was the most effective, with an AUC of 0.827 in the training set and 0.869 in the validation set. Conclusion Predictive models for RAI efficacy built on machine learning have great predictive performance. The developed nomogram can realize individualized and accurate prediction for the therapeutic efficacy of RAI for DTC.

Key words: Radioactive iodine therapy, Machine learning, Differentiated thyroid cancer, Predictive model, Therapeutic efficacy

中图分类号:

R736.1

巨艳丽,王丽华,成芳,黄凤艳,陈学禹,贾红英. 基于机器学习构建放射性碘治疗疗效的预测模型[J]. 山东大学学报 (医学版), 2023, 61(1): 94-99.

JU Yanli, WANG Lihua, CHENG Fang, HUANG Fengyan, CHEN Xueyu, JIA Hongying. Construction of predictive models of radioiodine therapy based on machine learning[J]. Journal of Shandong University (Health Sciences), 2023, 61(1): 94-99.

参考文献

[1] Cabanillas ME, McFadden DG, Durante C. Thyroid cancer [J]. Lancet, 2016, 388(10061): 2783-2795.
[2] Jonklaas J, Sarlis NJ, Litofsky D, et al. Outcomes of patients with differentiated thyroid carcinoma following initial therapy [J]. Thyroid, 2006, 16(12): 1229-1242.
[3] Ito Y, Higashiyama T, Takamura Y, et al. Prognosis of patients with papillary thyroid carcinoma showing postoperative recurrence to the central neck [J]. World J Surg, 2011, 35(4): 767-772.
[4] Yang X, Liang J, Li T, et al. Preablative stimulated thyroglobulin correlates to new therapy response system in differentiated thyroid cancer [J]. J Clin Endocrinol Metab, 2016, 101(3): 1307-1313.
[5] Van CB, Wynants L. Machine learning in medicine [J]. N Engl J Med, 2019, 380(26): 2588. doi: 10.1161/CIRCULATIONAHA.115.001593.
[6] May M. Eight ways machine learning is assisting medicine [J]. Nat Med, 2021, 27(1): 2-3.
[7] Stephen JS, Mireia CO, Suet FC, et al. Multi-omic machine learning predictor of breast cancer therapy response [J]. Nature, 2022, 601(7894): 623-629.
[8] Chaudhary K, Poirion OB, Lu L, et al. Deep learning-based multi-omics integration robustly predicts survival in liver cancer [J]. Clin Cancer Res, 2018, 24(6): 1248-1259.
[9] Kather JN, Pearson AT, Halama N, et al. Deep learning can predict microsatellite instability directly from histology in gastrointestinal cancer [J]. Nat Med, 2019, 25(7): 1054-1056.
[10] Gould MK, Huang BZ, Tammemagi MC, et al. Machine learning for early lung cancer identification using routine clinical and laboratory data [J]. Am J Respir Crit Care Med, 2021, 204(4): 445-453.
[11] Haugen BR, Alexander EK, Bible KC, et al. 2015 American thyroid association management guidelines for adult patients with thyroid nodules and differentiated thyroid cancer: the American thyroid association guidelines task force on thyroid nodules and differentiated thyroid cancer [J]. Thyroid, 2016, 26(1): 1-133.
[12] 中华医学会核医学分会. ¹³¹I治疗分化型甲状腺癌指南(2021版)[J]. 中华核医学与分子影像杂志, 2021, 41(4): 218-241. Chinese Society of Nuclear Medicine. Guidelines for radioiodine therapy of differentiated thyroid cancer(2021 edition)[J]. Chinese Journal of Nuclear Medicine and Molecular Imaging, 2021, 41(4): 218-241.
[13] Tuttle RM, Haugen B, Perrier ND. Updated American joint committee on cancer/tumor-node-metastasis staging system for differentiated and anaplastic thyroid cancer(eighth edition): what changed and why [J]. Thyroid, 2017, 27(6): 751-756.
[14] Cooper DS, Doherty GM, Haugen BR, et al. Revised American thyroid association management guidelines for patients with thyroid nodules and differentiated thyroid cancer [J]. Thyroid, 2009, 19(11): 1167-1214.
[15] Grani G, Zatelli MC, Alfò M, et al. Real-world performance of the American Thyroid Association risk estimates in predicting 1-year differentiated thyroid cancer outcomes: a prospective multicenter study of 2000 patients [J]. Thyroid, 2021, 31(2): 264-271.
[16] Tuttle RM, Alzahrani AS. Risk stratification in differentiated thyroid cancer: from detection to final follow-up [J]. J Clin Endocrinol Metab, 2019, 104(9): 4087-4100.
[17] 林岩松, 李娇. 2015年美国甲状腺学会《成人甲状腺结节与分化型甲状腺癌诊治指南》解读:分化型甲状腺癌~¹³¹I治疗新进展[J]. 中国癌症杂志, 2016, 26(1): 1-12. LIN Yansong, LI Jiao. The interpretation of 2015 American thyroid association management guidelines for adult patients with thyroid nodules and differentiated thyroid carcinoma: new progress in radioactive iodine therapy of differentiated thyroid carcinoma [J]. China Oncology, 2016, 26(1): 1-12.
[18] Shen FC, Hsieh CJ, Huang IC, et al. Dynamic risk estimates of outcome in Chinese patients with well-differentiated thyroid cancer after total thyroidectomy and radioactive iodine remnant ablation [J]. Thyroid, 2017, 27(4): 531-536.
[19] 周倩, 王瑞华, 刘保平, 等. 高危分化型甲状腺癌手术及¹³¹I治疗后疗效分类及影响因素分析[J]. 中华核医学与分子影像杂志, 2021, 41(11): 664-669. ZHOU Qian, WANG Ruihua, LIU Baoping, et al. Classification of therapeutic effect and influencing factors in patients with high-risk differentiated thyroid carcinoma after surgery and ¹³¹I treat [J]. Chinese Journal of Nuclear Medicine and Molecular Imaging, 2021, 41(11): 664-669.
[20] Prpi c M, Franceschi M, Romi c M, et al. Thyroglobulin as a tumor marker in differentiated thyroid cancer-clinical considerations [J]. Acta Clin Croat, 2018,57(3): 518-527.
[21] Wong KCW, Ng TY, Yu KS, et al. The use of post-ablation stimulated thyroglobulin in predicting clinical outcomes in differentiated thyroid carcinoma-what cut-off values should we use [J]. Clin Oncol(R Coll Radiol), 2019, 31(2): e11-e20.
[22] Jayasekara J, Jonker P, Lin JF, et al. Early postoperative stimulated serum thyroglobulin quantifies risk of recurrence in papillary thyroid cancer [J]. Surgery, 2020, 167(1): 40-45.
[23] Nóbrega G, Cavalcanti M, Leite V, et al. Value of stimulated pre-ablation thyroglobulin as a prognostic marker in patients with differentiated thyroid carcinoma treated with radioiodine [J]. Endocrine, 2022, 76(3): 642-647.
[24] Pan M, Li Z, Jia M, et al. Combination of stimulated thyroglobulin and antithyroglobulin antibody predicts the efficacy and prognosis of(131)I therapy in patients with differentiated thyroid cancer following total thyroidectomy: a retrospective study [J]. Front Endocrinol(Lausanne), 2022, 13: 857057. doi: 10.3389/fendo.2022.857057.
[25] Prpic M, Kust D, Kruljac I, et al. Prediction of radioactive iodine remnant ablation failure in patients with differentiated thyroid cancer: a cohort study of 740 patients [J]. Head Neck, 2017, 39(1): 109-115.

相关文章 15

[1]	贺士卿,李皖皖,董书晴,牟婧怡,刘宇莹,魏思雨,刘钊,张家新. 基于数据库构建乳腺癌焦亡相关基因的预后风险模型[J]. 山东大学学报 (医学版), 2022, 60(8): 34-43.
[2]	况利,徐小明,曾琪. 机器学习用于自杀研究的综述[J]. 山东大学学报 (医学版), 2022, 60(4): 10-16.
[3]	姜震,孙静,邹雯,王唱唱,高琦. 基于两种机器学习算法的双相情感障碍患者自杀行为影响因素模型比较研究[J]. 山东大学学报 (医学版), 2022, 60(1): 101-108.
[4]	李皖皖,周文凯,董书晴,贺士卿,刘钊,张家新,刘斌. 利用数据库信息构建乳腺癌免疫关联lncRNAs风险评估模型[J]. 山东大学学报 (医学版), 2021, 59(7): 74-84.
[5]	田瑶天,王宝,李叶琴,王滕,田力文,韩波,王翠艳. 基于可解释性心脏磁共振参数的机器学习模型预测儿童心肌炎的预后[J]. 山东大学学报 (医学版), 2021, 59(7): 43-49.
[6]	赵洁,李岩,李明,于德新. 螺旋CT对黏液性软组织肿瘤良恶性鉴别的价值[J]. 山东大学学报 (医学版), 2021, 59(4): 100-107.
[7]	吴强,何泽鲲,刘琚,崔晓萌,孙双,石伟. 基于机器学习的脑胶质瘤多模态影像分析[J]. 山东大学学报 (医学版), 2020, 1(8): 81-87.
[8]	张伟,谭文浩,李贻斌. 基于深度强化学习的四足机器人运动控制发展现状与展望[J]. 山东大学学报 (医学版), 2020, 1(8): 61-66.
[9]	李潘,李月月,李延青. 个体化肠道准备对肠道准备质量的影响[J]. 山东大学学报 (医学版), 2020, 58(3): 113-117.
[10]	林浩添,李龙辉,陈睛晶. 儿童眼病的人工智能研究进展[J]. 山东大学学报 (医学版), 2020, 58(11): 11-16.
[11]	徐源佑,杨亚超,王春霞,马晓天,薛付忠,刘言训,王萍. 基于体检队列的胃炎风险预测模型[J]. 山东大学学报 (医学版), 2019, 57(6): 112-116.
[12]	周苗,夏同耀,孙爱玲,李明,申振伟,卞伟玮,蒋正,康凤玲,柳晓涓,薛付忠,刘静. 健康管理人群慢性肾脏病风险预测模型[J]. 山东大学学报（医学版）, 2017, 55(6): 98-103.
[13]	孙苑潆,杨亚超,曲明苓,陈雁敏,李敏,王淑康,薛付忠,刘云霞. 健康管理人群代谢综合征发病风险预测模型[J]. 山东大学学报（医学版）, 2017, 55(6): 87-92.
[14]	李吉庆,赵焕宗,宋炳红,张理纯,李向一,陈亚飞,王萍,薛付忠. 基于健康管理队列的心血管事件风险预测模型[J]. 山东大学学报（医学版）, 2017, 55(6): 56-60.
[15]	苏萍,杨亚超,杨洋,季加东,阿力木·达依木,李敏,薛付忠,刘言训. 健康管理人群2型糖尿病发病风险预测模型[J]. 山东大学学报（医学版）, 2017, 55(6): 82-86.

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed