Dynamic contrast-enhanced (DCE) MRI is an attractive technique in quantitative in vivo microvascular characterization by pharmacokinetic modeling for diagnosing disease and monitoring its progress^1-2. Additionally, a quantitative approach with arterial input function (AIF) improves the reproducibility and reliability of DCE-MRI³. In general, gradient-echo based T1-weighted sequence for DCE-MRI is conventionally used with the careful control of administration dose of contrast agent to keep the concentration low enough to neglect T2*-relaxation artifact. On the other hand, low dose assumption is not satisfied simultaneously both in arterial and tissue regions whose different T2*-relaxation would hamper accurate pharmacokinetic analysis. Furthermore, AIF measurement in a rodent model is challenging due to their high heart rates (up to 600beats/min) and requirement for high spatial resolution4. In this study, we propose the new strategy to achieve sub-second temporal resolution without compromising slice coverage of DCE-MRI applications by replacing turbo factor with compressed sensing (CS) accelerations for calibration-free turbo spin echo (TSE) acquisitions.The sparse sampling scheme was retrospectively optimized from fully sampled dynamic TSE; the resulting scheme was implemented for DCE-MRI applications (0.96 s, 128×128, 4 slices). The feasibility of calibration-free quantification of Gd-concentration and the degree of contrast enhancement were compared among fast low-angle shot (FLASH), TSE, and CS-TSE acquisitions with multiple phantoms (0.1–6 mM). The kidney-feeding AIF was measured at multiple administration doses (0.1–0.3 mmol/kg) to evaluate the benefit of CS-TSE for quantifying rapidly changing high Gd-concentrations in a rodent model.In phantom studies, both TSE and CS-TSE showed significantly improved contrast enhancement over FLASH from reduced T2* relaxation as shown in Fig 1(A). In case of FLASH in Fig 1(B), Gd-concentrations above 2mM were underestimated due to T2*-relaxation effect. In contrast, both calibration-free and calibrated approaches estimated equivalent Gd-concentrations for CS-TSE (scatter plot slope = 0.9801, r2 = 0.9998) plotted in Fig 1(C). In in vivo studies, four-fold higher temporal resolution (0.96 s) of CS-TSE over the corresponding TSE enabled robust measurement of AIF first-pass peak and resulting peak enhancement with CS-TSE were observed in Fig 2(A), with 1.1439 and 2.1258-fold times higher than those with TSE and FLASH acquisitions, respectively (0.1 mmol/kg dose) as plotted in Fig 2(B). Calibration-free estimates of AIF peak concentration with CS-TSE were in good agreement with the calibrated approach at multiple administration doses (scatter plot slope = 0.7800, r2 = 0.8014) as shown in Fig 2(C).The results of this study show the benefit of CS-TSE acquisition for DCE-MRI applications, including (1) enhanced temporal resolution without compromising slice coverage with respect to conventional TSE acquisition, (2) calibration-free estimation of Gd-concentration, and (3) improved contrast enhancements over FLASH acquisitions. The dose-independency in AIF with CS-TSE would be beneficial as the concentration gradient between the arterial and tissue signal would not further hamper accurate pharmacokinetic analysis, potentially leading to more accurate quantification of MR perfusion. In conclusion, the present work validates the feasibility of practical CS-TSE for its application to calibration-free DCE-MRI for sub-second acquisition, without compromising slice coverage, by complementing turbo factor with CS accelerations.