Fluctuations in serum lipid levels during neoadjuvant treatment as novel predictive and prognostic biomarkers for locally advanced breast cancer: a retrospective analysis based on a prospective cohort.

Background: With increasing attention given to host-specific lipid metabolism status, it is of urgent need to identify lipid metabolism indices with predictive or prognostic value in locally advanced breast cancer patients treated with neoadjuvant chemotherapy (NAC), and to evaluate the performance improvement by incorporating them into the existing Neo-Bioscore staging system.

Methods: Patients from a prospectively maintained database of locally advanced breast cancer patients who received radical surgery after NAC between January 2014 to December 2020 were enrolled in this study. The enrolled patients were randomly divided into a training set and a test set at a ratio of 6:4. The random forest algorithm was applied to rank the importance of prognostic factors, top-ranked lipid metabolism indices of which were then incorporated into Neo-Bioscore to construct an updated prognostic model. The performances of these two models were compared in both training set and test set from multiple perspectives. Study outcomes included disease-free survival (DFS), relapse-free survival (RFS), distance-recurrence-free survival (DRFS), locoregional-recurrence-free survival (LRFS) and overall survival (OS).

Results: A total of 200 eligible patients were included in this study. After a median follow-up of 4.73 years, it was demonstrated that the relative increase in total cholesterol (TC; DFS: HR = 4.782, 95%CI 1.410 ~ 16.217, P = 0.012) and low-density lipoprotein (LDL; DFS: HR = 4.622, 95%CI 1.517 ~ 14.088, P = 0.007) during NAC led to poorer survival outcomes. Patients with either a higher body mass index (BMI) or elevated LDL during NAC had a worse prognosis (DFS: HR = 6.351, 95%CI 1.938 ~ 20.809, P = 0.002; OS, HR = 6.919, 95%CI 1.296 ~ 36.932, P = 0.024). Incorporating BMI and LDL fluctuations during NAC into Neo-Bioscore improved the prognostic stratification, especially in terms of LRFS (P = 0.046 vs. P = 0.65) and OS (P = 0.013 vs. P = 0.61). Multidimensional evaluation confirmed the improvement in model fit and clinical use for the updated model in both training set and test set.

Conclusions: This is the first study to illustrate the relative elevation of LDL and TC levels during NAC as independent prognosticators for locally advanced breast cancer. This is also the first attempt to incorporate lipid metabolism indices into the original Neo-Bioscore staging system, which further improves the prognostic stratification of patients receiving NAC.