MR-only treatment planning requires the knowledge of electron density to account for medium heterogeneities during dose calculation. This work introduces a novel perspective for electron density estimation by utilizing quantitative water/fat imaging. Water/fat phantoms with different percentages were scanned on MR and CT. Water/fat separation was performed while correcting (or minimizing) major sources of signal bias. A linear regression model between CT and corrected MR signals was calculated and used to derive MR-based electron density curve. This approach targets radiotherapy applications that require sensitive soft-tissue heterogeneity correction such as prostate and breast low-dose-rate (LDR) brachytherapy.