本文選題：乳腺癌 + 危險因素��； 參考：《中華腫瘤防治雜志》2017年12期

【摘要】：目的乳腺癌風(fēng)險預(yù)測模型可將人群分為不同的風(fēng)險等級,有助于降低篩查成本,使乳腺癌篩查效益最大化。本研究分析了上海市女性乳腺癌的危險因素,初步建立了符合該人群流行病學(xué)特征的風(fēng)險預(yù)測模型,為乳腺癌高危人群的篩選提供依據(jù)。方法 2008-05-23-2012-09-30,采用調(diào)查表對上海市閔行區(qū)149 577名35~74歲女性開展乳腺癌初篩,內(nèi)容包括人口學(xué)、月經(jīng)生育史、乳腺疾病史和家族史等信息,具備任一明確定義危險因素者為初篩陽性。將所有對象的個人信息與上海市腫瘤登記系統(tǒng)和生命統(tǒng)計系統(tǒng)進(jìn)行記錄聯(lián)動,收集2015-06-30前乳腺癌確診和全死因死亡信息。采用Cox比例風(fēng)險模型,建立乳腺癌風(fēng)險預(yù)測模型,計算乳腺癌5年發(fā)病風(fēng)險,并采用5折交叉驗證法,分別計算期望病例數(shù)與觀察病例數(shù)比值(ratio of the expected to the observed number,E/O)和受試者工作特征曲線下面積(areas under the receiver operating characteristic curve,AUC),評價模型的校準(zhǔn)度和區(qū)分力。結(jié)果經(jīng)過774 333人年(中位隨訪人年5.05年)隨訪,共發(fā)現(xiàn)新發(fā)乳腺癌病例973例,粗發(fā)病率(crude incidence rate,CIR)和年齡標(biāo)化率(age-standardized incidence rate,ASR)分別為125.66/10萬和112.55/10萬,初篩陽性者的粗率和標(biāo)化率分別為133.91/10萬和121.83/10萬,顯著高于初篩陰性者的119.76/10萬和106.91/10萬。年齡、教育程度、乳腺癌家族史、患重度乳腺小葉增生、有乳房腫塊、患乳腺導(dǎo)管內(nèi)乳頭狀瘤與乳腺癌呈正向關(guān)聯(lián),哺乳和月經(jīng)周期規(guī)律與乳腺癌呈負(fù)向關(guān)聯(lián)�；谶@些因素建立的風(fēng)險預(yù)測模型估計該人群乳腺癌5年絕對發(fā)病風(fēng)險高峰出現(xiàn)在55歲,在0.19%~1.10%之間變化。模型的E/O值為0.98(95%CI為0.92,1.04),AUC為0.596(95%CI為0.538,0.654)。進(jìn)一步按年齡分層,發(fā)現(xiàn)55歲以下組和55歲及以上組的E/O值分別為0.96(0.88,1.03)和1.01(0.91,1.16),AUC分別為0.627(0.514,0.701)和0.567(0.518,0.630)。結(jié)論本研究建立的風(fēng)險評估模型主要基于自我報告的乳腺癥狀及體征,總體校準(zhǔn)度較好,而總體區(qū)分力不理想,但在55歲以下女性中有所提高,可用于社區(qū)人群尤其是55歲以下人群的乳腺癌風(fēng)險分級。
[Abstract]:Objective Breast cancer risk prediction model can divide the population into different risk levels, which is helpful to reduce the screening cost and maximize the efficiency of breast cancer screening. This study analyzed the risk factors of female breast cancer in Shanghai, and established a risk prediction model according to the epidemiological characteristics of the population, which provided the basis for screening high risk population of breast cancer. Methods 149,577 women aged 35 to 74 years old in Minhang District of Shanghai were screened for breast cancer by questionnaire from May to September 2009.The information included demography, menstrual history, breast disease history and family history. All subjects' personal information was recorded with Shanghai Cancer Registration system and vital Statistics system to collect the diagnosis and death information of breast cancer before 2015-06-30. Cox proportional risk model was used to predict breast cancer risk. The 5 year risk of breast cancer was calculated, and the 5% cross validation method was used. The ratio of expected cases to observed cases and the area under the operating characteristic curve were calculated to evaluate the calibration degree and distinguishing power of the model. Results after a follow-up of 774,333 person-years (median 5.05 years), 973 new breast cancer cases were found. The crude incidence rate and age-standardized incidence rate were 125.66 / 100 and 112.55 / 100 million, respectively. The crude rate and standardized rate of positive primary screening were 133.91 / 100 and 121.83 / 100, respectively, which were significantly higher than those of negative screening of 1.19 76 / 100 million and 106.91% / 100 000 respectively. Age, education, family history of breast cancer, severe breast lobule hyperplasia, breast mass, breast intraductal papilloma were positively correlated with breast cancer, lactation and menstrual cycle were negatively correlated with breast cancer. The risk prediction model based on these factors estimated that the absolute risk of breast cancer in this population peaked at 55 years old and varied between 0.19% and 1.10%. The E / O value of the model was 0.92 ~ 1.04A ~ (UC) = 0.538A ~ (0.654) ~ (-1). Furthermore, according to the age, it was found that the E / O values of the group under 55 and the group aged 55 and over were 0.96 ~ 0.881.03) and 1.01U _ (0.91) ~ (1.16) A ~ (UC) were 0.627 ~ 0.514 ~ (0.701) and 0.567n ~ (0.5) ~ (18) ~ (0.630), respectively. Conclusion the risk assessment model established in this study is mainly based on self-reported breast symptoms and signs. The overall calibration degree is good, but the overall discrimination is not ideal, but it is improved in women under 55 years of age. Can be used for the community population, especially under 55 years of age for breast cancer risk classification.
【作者單位】： 復(fù)旦大學(xué)公共衛(wèi)生學(xué)院流行病學(xué)教研室公共安全教育部重點實驗室;閔行區(qū)疾病預(yù)防控制中心衛(wèi)生科;閔行區(qū)疾病預(yù)防控制中心慢性病防治科;
【基金】：美國中華醫(yī)學(xué)基金會(China Medical Board,HPSS 09-991) 上海市第四輪公共衛(wèi)生計劃重點學(xué)科建設(shè)課題(15GWZK0801) 上海市自然科學(xué)基金青年項目(12ZR1448700)
【分類號】：R737.9

【相似文獻(xiàn)】

相關(guān)期刊論文 前2條

1 李杰清;牛桂芬;;乳腺癌危險度預(yù)測模型用于婦女乳腺癌的預(yù)測[J];中國處方藥;2014年05期

2 ;[J];;年期

相關(guān)博士學(xué)位論文 前1條

1 白云鵬;華法林穩(wěn)態(tài)劑量預(yù)測模型在瓣膜置換術(shù)后抗凝治療中的應(yīng)用[D];天津醫(yī)科大學(xué);2016年

，

本文編號：1971573