利用大数据建立乳腺浸润性微乳头状癌患者术前淋巴结转移预测模型

乳腺浸润性微乳头状癌是一种不同寻常的浸润性乳腺肿瘤亚型，具有高度的淋巴引流区域转移倾向，故有必要对乳腺浸润性微乳头状癌患者术前淋巴结转移进行预测，以决定术前新辅助治疗和腋窝淋巴结清扫的必要性。不过，目前仍然缺乏大样本研究结果。

2018年11月8日，施普林格·自然旗下《生物医学中心》癌症分册发表复旦大学附属肿瘤医院、中南大学湘雅医学院附属肿瘤医院、福建医科大学附属协和医院的研究报告，利用大数据确定乳腺浸润性微乳头状癌淋巴结转移的风险因素，并且建立术前预测淋巴结转移可能性的列线图。

列线图又称线规图、列线图、诺谟图，在其他相关因素变量已知时给出一个因素变量数值、用数字刻度反映因素变量之间函数关系的专用图，主要用于医学、测绘学、大气科学、古生物学。

该研究利用美国国家癌症研究所监测流行病学最终结果（SEER）数据库对2003～2014年所有诊断为浸润性微乳头状癌的1407例患者临床和病理记录进行回顾分析，将该组患者分为训练集（又称演算组，包括2003～2009年确诊的527例患者）和验证集（又称验算组，包括2009年后确诊的880例患者）。通过逻辑回归模型，对训练集构建列线图，随后对验证集进行验证。通过R软件3.4.1版对评定列线图性能的区分度和校准度进行量化。

结果，逻辑回归分析表明，区域淋巴结转移显著相关风险因素为病变较大、诊断年龄较小、黑人种族、激素受体表达缺乏。列线图的区分度曲线下面积为0.735（95%置信区间：0.692～0.777）表现出良好的预测性能。列线图的校准度曲线斜率接近1，证实了列线图的优异校准度。列线图的性能得到验证集的进一步验证，区分度曲线下面积为0.748（95%置信区间：0.701～0.767）。

因此，浸润性微乳头状癌与浸润性导管癌之间的显著差异仍为淋巴结转移增加，故有必要加强治疗。根据临床病理参数建立的列线图，术前可以准确预测区域淋巴结状态，将有助于术前淋巴结状态评估以及个体患者治疗决策。

BMC Cancer. 2018 Nov 8;18(1):1085.

Nomogram for predicting preoperative lymph node involvement in patients with invasive micropapillary carcinoma of breast: a SEER population-based study.

Ye FG, Xia C, Ma D, Lin PY, Hu X, Shao ZM.

Fudan University Shanghai Cancer Center, Shanghai, China; Shanghai Medical College, Fudan University, Shanghai, China; The Affiliated Cancer Hospital of Xiangya School of Medicine, Central South University, Changsha, China; Affiliated Union Hospital, Fujian Medical University, Fuzhou, China; Institutes of Biomedical Science, Fudan University, Shanghai, China.

BACKGROUND: Invasive micropapillary carcinoma (IMPC) is an unusual and distinct subtype of invasive breast tumor with high propensity for regional lymph node metastases. This study was to identify risk factors accounting for IMPC of the breast and to develop a nomogram to preoperatively predict the probability of lymph node involvement.

METHODS: A retrospective review of the clinical and pathology records was performed in patients diagnosed with IMPC between 2003 and 2014 from Surveillance, Epidemiology, and End Results (SEER) database. The cohort was divided into training and validation sets. Training set comprised patients diagnosed between 2003 and 2009, while validation set included patients diagnosed thereafter. A logistic regression model was used to construct the nomogram in the training set and then varified in the validation set. Nomogram performance was quantified with respect to discrimination and calibration using R 3.4.1 software.

RESULTS: Overall, 1407 patients diagnosed with IMPC were enrolled, of which 527 in training set and 880 in validation set. Logistic regression analysis indicated larger lesions, younger age at diagnosis, black ethnic and lack of hormone receptor expression were significantly related to regional nodes involvement. The AUC of the nomogram was 0.735 (95% confidence interval (CI) 0.692 to 0.777), demonstrating a good prediction performance. Calibration curve for the nomogram was plotted and the slope was close to 1, which demonstrated excellent calibration of the nomogram. The performance of the nomogram was further validated in the validation set, with AUC of 0.748 (95% CI 0.701 to 0.767).

CONCLUSIONS: The striking difference between IMPC and IDC remains the increased lymph node involvement in IMPC and therefore merits aggressive treatment. The nomogram based on the clinicalpathologic parameters was established, which could accurately preoperatively predict regional lymph node status. This nomogram would facilitate evaluating lymph node state preoperatively and thus treatment decision-making of individual patients.

KEYWORDS: Breast cancer; IMPC; Lymph nodes involvement; Predict; Preoperative

以下广告内容与本微信公众号无关