School of Medicine, University of Electronic Science and Technology of China, Chengdu, China; Department of Breast Surgery, Sichuan Clinical Research Center for Cancer, Sichuan Cancer Hospital & Institute, Sichuan Cancer Center, Affiliated Cancer Hospital of University of Electronic Science and Technology of China, Chengdu, China.
Department of Radiation Oncology, Radiation Oncology Key Laboratory of Sichuan Province, Sichuan Clinical Research Center for Cancer, Sichuan Cancer Hospital & Institute, Sichuan Cancer Center, Affiliated Cancer Hospital of University of Electronic Science and Technology of China, Chengdu, China.
Clin Breast Cancer. 2024 Jun;24(4):376-383. doi: 10.1016/j.clbc.2024.02.008. Epub 2024 Feb 16.
The incidence of breast cancer ranks highest among cancers and is exceedingly heterogeneous. Immunohistochemical staining is commonly used clinically to identify the molecular subtype for subsequent treatment and prognosis.
Raman spectroscopy and support vector machine (SVM) learning algorithm were utilized to identify blood samples from breast cancer patients in order to investigate a novel molecular typing approach.
Tumor tissue coarse needle aspiration biopsy samples, and peripheral venous blood samples were gathered from 459 invasive breast cancer patients admitted to the breast department of Sichuan Cancer Hospital between June 2021 and September 2022. Immunohistochemical staining and in situ hybridization were performed on the coarse needle aspiration biopsy tissues to obtain their molecular typing pathological labels, including: 70 cases of Luminal A, 167 cases of Luminal B (HER2-positive), 57 cases of Luminal B (HER2-negative), 84 cases of HER2-positive, and 81 cases of triple-negative. Blood samples were processed to obtained Raman spectra taken for SVM classification models establishment with machine algorithms (using 80% of the sample data as the training set), and then the performance of the SVM classification models was evaluated by the independent validation set (20% of the sample data).
The AUC values of SVM classification models remained above 0.85, demonstrating outstanding model performance and excellent subtype discrimination of breast cancer molecular subtypes.
Raman spectroscopy of serum samples can promptly and precisely detect the molecular subtype of invasive breast cancer, which has the potential for clinical value.
乳腺癌的发病率在所有癌症中位居首位，且具有极高的异质性。临床上常通过免疫组织化学染色来鉴定分子亚型，以便为后续治疗和预后提供依据。
本研究旨在探讨一种新的分子分型方法，利用拉曼光谱和支持向量机（SVM）学习算法来识别乳腺癌患者的血液样本。
本研究收集了 2021 年 6 月至 2022 年 9 月期间在四川省肿瘤医院乳腺科就诊的 459 例浸润性乳腺癌患者的肿瘤组织粗针穿刺活检样本和外周静脉血样本。对粗针穿刺活检组织进行免疫组织化学染色和原位杂交，获得其分子分型病理标签，包括：Luminal A 型 70 例、Luminal B（HER2 阳性）型 167 例、Luminal B（HER2 阴性）型 57 例、HER2 阳性型 84 例和三阴性型 81 例。对血液样本进行处理，获取拉曼光谱，并用机器算法建立 SVM 分类模型（使用 80%的样本数据作为训练集），然后通过独立验证集（20%的样本数据）评估 SVM 分类模型的性能。
SVM 分类模型的 AUC 值均保持在 0.85 以上，表明模型性能优异，对乳腺癌分子亚型具有良好的区分能力。
血清样本的拉曼光谱能够快速准确地检测浸润性乳腺癌的分子亚型，具有潜在的临床应用价值。
