Abstract P2-08-05: Plasma lipidomics analysis to identify potential non-invasive biomarkers for breast white adipose inflammation and aromatase expression levels

Abstract

Abstract Background: The presence of white adipose tissue inflammation (WATi) in the breast has been associated with increased breast cancer risk and a worse clinical course. Elevated body mass index (BMI) and the post-menopausal state are both associated with breast WATi. Breast WATi has also been associated with increased expression of aromatase, the rate-limiting enzyme for estrogen biosynthesis. Currently, WATi is diagnosed in surgical samples of breast tissue by the identification of crown-like structures (CLS), which are comprised of a dead or dying adipocyte enveloped by macrophages. In contrast to surgical specimens, core biopsies have been shown to be inadequate for assessing WATi. Hence, there is an unmet need for strategies to non-invasively diagnose WATi. Here we used a novel lipidomics platform to identify potential non-invasive blood signatures of breast WATi. Methods: We conducted a cross-sectional study which included 100 patients who underwent mastectomy for breast cancer treatment or risk reduction. WATi was detected by CD68 immunohistochemistry to identify CLS. Breast aromatase expression levels were measured by qPCR. Clinicopathologic data were abstracted from electronic medical records. Lipidomic data were measured from blood plasma in collaboration with Waters Technologies Corporation. Lipid levels in association with breast WATi (presence/absence) and levels of aromatase expression (high/low) in non-tumorous breast tissue were examined using Welch’s t-test. P-values were adjusted for multiple comparisons by controlling the false discovery rate (FDR) using the Benjamini-Hochberg method. A logistic regression model was used to develop predictive models that evaluated potential lipid biomarkers of the presence of breast WATi and high breast aromatase expression levels. Stepwise regression was used for variable selection. AUC of the ROC curves was used to evaluate the performance of the predictive models. Results: Among 140 lipids analyzed, 13 were identified to be associated with breast WATi (P<0.05, |log2FC|>0.3). Specifically, 8 lipids had lower levels, and 6 lipids had higher levels in patients with breast WATi compared to those without. Levels of 7 lipids were significantly higher in patients with an increased level of aromatase (P<0.05, |log2FC|>0.3). After variable selection, LPE(22:6) (P=0.018), LPE(20:3) ES-(P=0.006), along with menopausal status and BMI provided an 86.0% (95% CI, 77.6%-94.5%) accuracy in predicting higher breast aromatase levels. Combining the effect of two lipids improved the accuracy by 10.4%(P=0.030) compared to the model only using menopausal status and BMI. A model with 5 lipids and menopausal status provided an 88.8%(95% CI, 81.9%-95.8%) accuracy for predicting breast WATi. The model performance improved by 9.2% (P=0.026) compared to the model only using menopausal status and BMI. Conclusions: Our study identified several lipid species that showed significant changes in association with breast WATi and levels of aromatase expression. Further validation of these blood signatures could provide non-invasive assessment of WATi and aromatase levels. The availability of such a diagnostic algorithm could help, in turn, to both identify women at elevated risk for breast cancer and for monitoring the efficacy of interventions aimed at reducing inflammation and aromatase levels. Citation Format: Xiao Cai, Siqi Wei, Zizhuo Xu, Xi K Zhou, Andrew Peck, Steven Lai, Giorgis Isaac, Hernando Olivos, Nayasha Munjoma, Suraj Dhungana, Rob Plumb, Andrew J Dannenberg, Neil M Iyengar. Plasma lipidomics analysis to identify potential non-invasive biomarkers for breast white adipose inflammation and aromatase expression levels [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P2-08-05.