To investigate the clinical feasibility of a quantitative sodium 23 (23Na) magnetic resonance (MR) imaging protocol developed for breast tumor assessment and to compare it with 7-T diffusion-weighted imaging (DWI). Materials and Methods: Written informed consent in this institutional review board-approved study was obtained from eight healthy volunteers and 17 patients with 20 breast tumors (five benign, 15 malignant). To achieve the best image quality and reproducibility, the 23Na sequence was optimized and tested on phantoms and healthy volunteers. For in vivo quantification of absolute tissue sodium concentration (TSC), an external phantom was used. Static magnetic field, or B0, and combined transmit and receive radiofrequency field, or B1, maps were acquired, and image quality, measurement reproducibility, and accuracy testing were performed. Bilateral 23Na and DWI sequences were performed before contrast material-enhanced MR imaging in patients with breast tumors. TSC and apparent diffusion coefficient (ADC) were calculated and correlated for healthy glandular tissue and benign and malignant lesions. Results: The 23Na MR imaging protocol is feasible, with 1.5-mm inplane resolution and 16-minute imaging time. Good image quality was achieved, with high reproducibility (mean TSC values 6 standard deviation for the test, 36 mmol per kilogram of wet weight ± 2 [range, 34-37 mmol/kg ± for the retest, 37 mmol/kg ±1 [range, 35-39 mmol/kg]; P = .610) and accuracy (r = 0.998, P <.001). TSC values in normal glandular and adipose breast tissue were 35 mmol/kg 6 3 and 18 mmol/kg ± 3, respectively. In malignant lesions (mean size, 31 mm ± 24; range, 6-92 mm), the TSC of 69 mmol/kg 6 10 was, on average, 49% higher than that in benign lesions (mean size, 14 mm 6 12; range, 6-35 mm), with a TSC of 47 mmol/kg ± 8 (P = .002). There were similar ADC differences between benign ([1.78 6 0.23] 3 10-3 mm2/sec) and malignant ([1.03 ± 0.23] 3 10-3mm2/sec) tumors (P = .002). ADC and TSC were inversely correlated (r = 20.881, P < .001). Conclusion: Quantitative 23Na MR imaging is clinically feasible, may provide good differentiation between malignant and benign breast lesions, and demonstrates an inverse correlation with ADC.