乳腺密度自动测量与乳腺癌术后他莫昔芬治疗预后方法

乳腺癌严重威胁人类生命健康，乳腺癌患者手术后辅以他莫昔芬治疗是常用的治疗方法。然而患者治疗后仍然面临复发或转移的风险，所以需要有效的预后方法来预测疗效。为了探索一种基于钼靶X线影像的乳腺癌术后他莫昔芬治疗预后分析方法，本文通过基于通道注意力机制的SE-CNN(Squeeze-and-Excitation Convolutional Neural Network)方法研究了钼靶X线影像中的乳腺密度自动提取模型，提出预后影像标志物——乳腺密度变化率（MDCR），并进行生存分析，研究其对乳腺癌术后他莫昔芬治疗的预后能力。研究表明：SE-CNN的阈值绝对误差为9.92±4.78，决对系数为0.77，结果表明该方法能够准确提取阈值。生存分析中得到无进展生存期为HR：2.654(95%CI,1.102-6.395)，p=0.030。MDCR值高的患者预后较好，反之则较差。可见乳腺密度变化率可以作为乳腺癌术后他莫昔芬治疗预后影像标志物。

Automatic breast density measurement and postoperative tamoxifen therapy for breast cancer prognosis

Breast cancer is a serious threat to human life and health. Tamoxifen is a common treatment for breast cancer patients after surgery. However, patients still face the risk of recurrence or metastasis after treatment, so effective prognostic methods are needed to predict efficacy. In order to explore a method for prognostic analysis of postoperative tamoxifen treatment for breast cancer based on molybdenum target X-ray image. In this paper, the model for mammographic density extraction from molybdenum target X-ray image is studied by using the Squeeze-and- Excitation Convolutional Neural Network (SE-CNN) method. A prognostic imaging marker, mammographic density change rate (MDCR), was proposed and survival analysis was performed to study its prognostic ability for postoperative tamoxifen treatment of breast cancer. The results show that the threshold absolute error of SE-CNN is 9.92±4.78, and the determination coefficient is 0.77. The results show that this method can accurately extract the threshold. The progression-free survival was HR 2.654(95%CI,1.102-6.395), P =0.030.Patients with high MDCR had a better prognosis, while those with low MDCR had a worse prognosis. It is concluded that the rate of mammographic density change can be used as a prognostic imaging marker of postoperative tamoxifen treatment for breast cancer.