The introduction of the birdcage body coil in 1985[1] has enabled unprecedented B1-field homogeneity. However, at 3T the use of the birdcage body coil has some considerable disadvantages. These include sometimes suboptimal B1+ homogeneity for abdominal imaging and high local SAR levels[2] but most of all: very large power requirements. A phantom study has shown that the use of a local transmit array of loop coils severely reduces these power requirements [3]. In this study, we intend to explore the use of dipole antennas as transceive array elements at 3T. Dipole antennas are being used frequently at ultrahigh field strengths. However, at 3T, the use of dipole antennas requires the addition of large inductances to the antenna legs to keep them at a reasonable size. We compared three designs: a segmented dipole antenna where lumped elements inductors are added between the segments, an antenna where the inductance was added in the shape of 9 meanders and a similar antenna with 12 (more narrow) meanders. Elements are evaluated by a single channel B1+ measurement on a pelvis shaped phantom with tissue-resembling properties (εr=34, σ=0.4). AFI B1+ maps have been acquired to compare the performance of each design. The single element measurements resulted in a clear advantage for the antenna with 12 meanders . A transceive array of eight of these dipole antennas has been realized for a 8-channel multi-transmit 3T system (Philips Healthcare, Best, The Netherlands). In addition, the system was equipped with an 8-channel Tx/Rx switch box with integrated preamps (MR Coils BV, Drunen, The Netherlands) and the software was adapted to constantly detune the present 8-channel TEM body coil. The elements are driven in the center through a lattice balun that simultaneously performs impedance matching. Numerical simulations of the array on Virtual Family model ‘Duke’ [4]. have been performed in Sim4Life (ZMT, Zurich, Switzerland) to assess the 10g averaged SAR distribution (SAR10g). Combined with careful calibration and real-time monitoring of the power emitted by the amplifiers, safe local and global SAR exposure was ensured. One volunteer (31y, 1.82m, 73kg, informed consent) was scanned using 8x200W peak power and T2w TSE images (TR/TE=2000/90 ms, 0.75x1x3 mm3, 3 slices, TSE-factor: 13) have been obtained. The in-depth B1+ profiles for the investigated elements are presented in figure 1. The lumped element design is weaker than the others and was extremely sensitive to loading variations. The element with 12 meanders performed the best and also showed the least sensitivity towards loading variations. This design has been used to realize an eight-element transceive array (figure 2). Note that although dipole antennas have been presented as particularly suitable for ultrahigh field strengths[5,6], we now explore their applicability at 3T mainly because of the relatively homogeneous B1+ field patterns (less steep fall-off) and reduced inter-element coupling. From the array simulations, the sum-of-magnitude electric field for all antenna elements has been used to calculate the worst-case SAR distribution as presented in figure 3. Clearly, the maximum SAR is located at the anterior side of the thighs where the dipoles end. Using an average power of 1W per channel, the maximum SAR10g value is 2.2 W/kg. The simulated B1 efficiency for the array was 10.6 µT using 8x200 W input power. Figure 5 shows prostate imaging results for the dipole antenna array at 3T. Using 8x200W, a B1+ level of 12 µT was achieved inside the prostate. This corresponds reasonably well to the simulated value and is indeed much more efficient than the birdcage body coil or the 8-channel TEM body coil that is integrated in this MRI platform. The current implementation used a simple transmit-receive operation, limiting reception to only 8 dipoles, which is less than the state of the art for modern prostate imaging This will be amended once the array is extended by receive or transceive loop coil elements underneath the dipoles, similar to Voogt et al [6]. Dipole antennas at 3T can be realized with the addition of extra inductance in the antenna legs. If the extra inductance is realized as meanders instead of lumped element inductors, the performance is better and the antennas are not very sensitive to loading variations. A design consisting of eight meandering dipole antennas has been realized. Simulations show a worst-case maximum SAR10g of 2.2 W/kg for 8x1 W accepted input power. In an actual MRI experiment, the array realizes 12 µT with 8x200W on an average-sized human subject. T2w prostate images have been successfully acquired.