Intrauterine growth restriction (IUGR) is associated with an increased perinatal mortality and morbidity and hence is of significant clinical impact. (1). One reason for fetal growth abnormalities originates from placental abnormalities or maternal disease which is associated with an decreased oxygen delivery to the fetus (2). With the relationship between T2 relaxation time and oxygen saturation (sO2), magnetic resonance oximetry represents a valuable method for a direct noninvasive determination of fetal oxygen saturation (3). The purpose of this work was to establish a relationship between fetal sO2 and T2 relaxation time based on the Luz and Meiboom model (4) using in-vitro fetal blood samples. Parameters describing the T2 relaxation of fetal blood were consecutively validated in-vivo.In order to measure T2 signal intensities insensitive to inflow effects or turbulences, a three dimensional balanced steady-state free precession (SSFP) sequence in combination with T2 preparation pulses (5) was applied at 1.5 Tesla (Achieva, Philips Healthcare, Best, The Netherlands). The sequence was used in a multishot manner whereas the interval between T2 preparation and SSFP shot was used for dephasing. The TE-times of the T2-preparation ranged from 30 to 150 ms, 7 TE-times were measured. The SSFP shot had the following parameters: (TE/TR: 2.4 ms / 4.8 ms, field of view: 150 mm, acq. Matrix: 96 x 96, voxel size: 156 x 156 x 8.00, slices: 8). Fetal blood for in-vitro measurements was derived from the umbilical cord during abdominal delivery from three different fetuses. The blood from each fetus was heparinized and divided into 5 samples ( each 7.5 ml) with different oxygen saturation levels ranging from 30 % to 100%. For MRI, the samples were placed in a customized holder into a waterbath tempered to 36°, whereas the position of the samples was turned after each scan to prevent sedimentation. As the used sequence is critically dependent on locally correct pulse angles, a B1 map based on the two angle method was used to ensure a correct transmit field. If necessary, the radiofrequency (RF) field was adjusted using the scaling factor of the transmit RF field. In order to ensure a correct T2 weighting, calculated T2 values using the SSFP sequence were compared to a conventional multi echo spin echo sequence. The relationship between T2 and blood can be expressed as a function of fully oxygenated blood (T20), hematocrit (hct), susceptibility of hemoglobin (a) and exchange time (tex) ( S=A*exp(-TE/T20+hct (1-hct)*a*tex)x(1-O2/100)^2*f(tex,TE/4). Measured signal intensities were fitted to the equation using hct and sO2 values measured using an blood gas analyzer (Radiometer, ABL system). The determined relationship between T2 and sO2 based on fitted values of tex, a, and T20 was consecutively applied to one subject (34 gestation week) in the left ventricle using the same SSFP sequence. Furthermore, as the hct value is unknown for in-vivo measurements, the impact of the hct value (ranging from 0.3 to 0.6) was calculated and compared to measured sO2 values.The validation of the SFFP sequence with the conventional T2 multi echo sequence resulted in a high correlation of r = 0.99 and an offset of 3 ms (Fig.1). The blood samples from each fetus were fitted simultaneously to the equation (Fig.2), yielding mean parameters for T20: 160±10 ms, tex: 4.5 ms±1.2 and a: 0.032±0.002 103sec-1. The mean parameters were retrospectively used to calculate the sO2 for each sample for verification (Fig.3), yielding a high agreement between measured and fitted sO2 for each sample ( r = 0.9 for all blood samples). Results of measured signal intensities in the left fetal ventricle using the mean calculated parameters resulted in a sO2 value of 98 %. However, only three TE values (30 ms,60 ms,90 ms) were considered due to motion during the remaining measurements (Fig.4). The influence of hct on the sO2 calculation is shown in Fig. 5. The maximum difference between measured and calculated sO2 over the range of all hct values was 5 %.The T2-prepared SSFP sequence was successfully validated using a conventional multi-echo T2 sequence. Calculated parameters to determine the relationship between T2 and sO2 were similar for each fetal blood sample and could be used to calculate the sO2 value for each sample with a high accuracy compared to measured sO2 values and were successfully evaluated in-vivo. In conclusion, MR oximetry is a promising method for a noninvasively determination of oxygen saturation. In future, the calculated parameters have to be validated using a higher number of blood samples and the application has to be validated in a larger patient population.