Background: Neoadjuvant chemotherapy (NACT) has arisen as a treatment option for breast cancer (BC). However, the response to NACT is still unpredictable and dependent on cancer subtype. Metabolomics is a tool for predicting biomarkers and chemotherapy response. We used plasma to verify metabolomic alterations in BC before NACT, relating to clinical data. Methods: Liquid chromatography coupled to mass spectrometry (LC-MS) was performed on pre-NACT plasma from patients with BC (n= 75). After data filtering, an SVM model for classification was built and validated with 75%/25% of the data, respectively. Results: The model composed of 19 identified metabolites effectively predicted NACT response for training/validation sets with high sensitivity (95.4%/93.3%), specificity (91.6%/100.0%), and accuracy (94.6%/94.7%). In both sets, the panel correctly classified 95% of resistant and 94% of sensitive females. Most compounds identified by the model were lipids and amino acids and revealed pathway alterations related to chemoresistance. Conclusion: We developed a model for predicting patient response to NACT. These metabolite panels allow clinical gain by building precision medicine strategies based on tumor stratification.
背景：新辅助化疗（NACT）已成为乳腺癌（BC）的一种治疗选择。然而，对NACT的反应仍难以预测，且取决于癌症亚型。代谢组学是预测生物标志物和化疗反应的工具。我们利用血浆验证了乳腺癌患者在接受NACT前的代谢组学变化，并将其与临床数据相关联。方法：对乳腺癌患者（n=75）的NACT前血浆样本进行液相色谱-质谱联用（LC-MS）分析。数据过滤后，构建了支持向量机（SVM）分类模型，并分别使用75%和25%的数据进行训练和验证。结果：由19种已鉴定代谢物组成的模型能有效预测训练集/验证集的NACT反应，具有高灵敏度（95.4%/93.3%）、特异性（91.6%/100.0%）和准确率（94.6%/94.7%）。在两个数据集中，该代谢物组合正确分类了95%的耐药患者和94%的敏感患者。模型鉴定的大多数化合物为脂质和氨基酸，揭示了与化疗耐药相关的通路改变。结论：我们开发了一个预测患者对NACT反应的模型。这些代谢物组合通过建立基于肿瘤分层的精准医疗策略，具有临床获益潜力。
Plasma Metabolome Signatures to Predict Responsiveness to Neoadjuvant Chemotherapy in Breast Cancer
肿瘤(癌症)患者之家