Scholarlink

Background <p>Doxorubicin (DOX)-based chemotherapy has improved survival outcomes in breast cancer patients but is often limited by doxorubicin-induced cardiotoxicity (DIC). Currently, no validated biomarkers can predict early DIC. Identifying novel biomarkers is essential for detecting patients at higher risk and enable timely interventions before irreversible cardiac injury occurs.</p> Methods <p>Twenty-seven breast cancer patients treated with DOX-containing chemotherapy were stratified by change in left ventricular ejection fraction (LVEF): 19 patients who maintained normal cardiac function (normal, decline &lt; 10%) and 8 who developed cardiotoxicity (abnormal, decline &gt; 10%). Plasma samples were collected at baseline and after chemotherapy for untargeted metabolomic profiling. Both baseline and pre–post designs were employed to capture static and dynamic metabolic alterations associated with DIC. Stepwise logistic regression was used to filter non-informative metabolites, and predictive performance was further validated using Random Forest modeling.</p> Results <p>A well-marked separation of plasma metabolomic profiles was observed between normal and abnormal cardiotoxicity groups at baseline (T0). Statistical analysis identified 100 significant metabolites at baseline (T0) and 78 metabolites after the first cycle of chemotherapy (T0-T1), with 10 metabolites common to both time-points: 3-phosphoglycerate, 2-hydroxyphenylacetate, inosine, taurine, suberate (C8-DC), sebacate (C10-DC), sphingadienine, oxindolylalanine. Machine learning models identified key metabolites (e.g., sebacate [C10-DC], 2-hydroxyhippurate, orotate, picolinate, and suberate [C8-DC]) as candidate predictors of cardiotoxicity, achieving moderate discriminatory performance in cross-validation, with higher specificity than sensitivity, indicating limited detection of abnormal cases.</p> Conclusions <p>Metabolomic profiling shows potential for early detection of DIC in breast cancer patients, supporting personalized interventions to prevent irreversible cardiac damage.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Metabolic phenotypes of doxorubicin-induced cardiotoxicity among patients with breast cancer

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月4日

查看全文
  • Amarnath Singh,
  • Se-Ran Jun,
  • Katherine Wallis,
  • Renny S. Lan,
  • Valentina Todorova,
  • L. Joseph Su,
  • Sam Makhoul,
  • Ping-Ching Hsu

摘要

Background

Doxorubicin (DOX)-based chemotherapy has improved survival outcomes in breast cancer patients but is often limited by doxorubicin-induced cardiotoxicity (DIC). Currently, no validated biomarkers can predict early DIC. Identifying novel biomarkers is essential for detecting patients at higher risk and enable timely interventions before irreversible cardiac injury occurs.

Methods

Twenty-seven breast cancer patients treated with DOX-containing chemotherapy were stratified by change in left ventricular ejection fraction (LVEF): 19 patients who maintained normal cardiac function (normal, decline < 10%) and 8 who developed cardiotoxicity (abnormal, decline > 10%). Plasma samples were collected at baseline and after chemotherapy for untargeted metabolomic profiling. Both baseline and pre–post designs were employed to capture static and dynamic metabolic alterations associated with DIC. Stepwise logistic regression was used to filter non-informative metabolites, and predictive performance was further validated using Random Forest modeling.

Results

A well-marked separation of plasma metabolomic profiles was observed between normal and abnormal cardiotoxicity groups at baseline (T0). Statistical analysis identified 100 significant metabolites at baseline (T0) and 78 metabolites after the first cycle of chemotherapy (T0-T1), with 10 metabolites common to both time-points: 3-phosphoglycerate, 2-hydroxyphenylacetate, inosine, taurine, suberate (C8-DC), sebacate (C10-DC), sphingadienine, oxindolylalanine. Machine learning models identified key metabolites (e.g., sebacate [C10-DC], 2-hydroxyhippurate, orotate, picolinate, and suberate [C8-DC]) as candidate predictors of cardiotoxicity, achieving moderate discriminatory performance in cross-validation, with higher specificity than sensitivity, indicating limited detection of abnormal cases.

Conclusions

Metabolomic profiling shows potential for early detection of DIC in breast cancer patients, supporting personalized interventions to prevent irreversible cardiac damage.