LAUSR

The last decade has seen the rapid growth of multiple pelvic magnetic resonance (MR) examinations performed for either cancer detection or staging, including the evaluation of prostate, rectal, cervical and endometrial cancer. All of the protocols used for these examinations hinge on the performance of high-resolution T2-weighted anatomical imaging. In developing MR protocols, we often reference the colloquial phrase ‘‘there ain’t no such thing as a free lunch,’’ and while the exact origin of this statement is unknown, its usage seems quite appropriate in this context even though it most certainly was never intended to be applied to magnetic resonance. Resolution, signal-to-noise ratio (SNR) and scan time are 3 parameters which are inexorably linked, and for the most part we are unable to improve one without compromising another. Doubling the number of phase encodes to increase the image matrix results in a doubling of scan time and a reduction in SNR, for example. Periodically in the history of magnetic resonance, major new technologies have been developed which tantalizingly offer up the prospect of parameter improvement without a significant downside, but there is usually a catch. There is little question that a modern multichannel system in 2021 provides superior SNR to a clinical scanner in use in the 1990s, although those improvements have been slow and incremental. Parallel imaging brought with it major improvements to scanning time but at the expense of reduced SNR and new artifacts. The paper by Dr. Czyzewska and colleagues in this issue of the CARJ looks at whether a pulse sequence initially designed to reduce motion artifact in neurologic and cardiac imaging (PROPELLER) can provide the same benefit in the pelvis, and if so at what (or any) cost. Prostate MRI is one of the most technically demanding studies that radiologists perform, due to the need for very high resolution images and diffusion imaging with high b-values in a region prone to susceptibility artifact and, to some extent, motion (bowel peristaltic/spasm and bulk patient motion). Many techniques have been described to reduce artifacts, including the use of antispasmodic pharmaceutical agents and interventions to reduce the amount of fecal material and gas in the rectum (ranging from low residue diets to enemas and manual aspiration of gas prior to the exam). Currently, PIRADS V2.1 does not strongly endorse any of these techniques, but rather comments that they ‘‘may be beneficial.’’ Returning to the free lunch, although these methods can show modest benefit, this comes at a cost. Antispasmodics add expense and logistical complexity to an exam as well as the possibility of patient side effects, and efforts to reduce rectal stool and gas can be poorly tolerated by patients. A pulse sequence that improves image quality by reducing these artifacts then would be most welcome, particularly if there was no downside. Rather than filling k-space with data in sequential lines, PROPELLER samples k-space in strips of phase-encoded lines which rotate around the center, and each block of typically 8-32 lines form one of the ‘‘propellers’’ or ‘‘blades’’ of the acquisition. The center of k-space is oversampled, which theoretically should improve signal-to-noise. The redundant data allows for comparison of the blocks and correction, or rejection, of blocks affected by patient motion. The major MR vendors all have their own version of PROPELLER (General Electric), including BLADE (Siemens), MultiVane (Philips), and JET (Canon). The PROPELLER technique has been used for many years as a backup sequence in neuroradiology, held in reserve for patients who cannot hold still for MRI of the brain. In more recent years, this T2-weighted sequence has been applied to imaging of the female pelvis, rectum and prostate in a particular effort to reduce motion artifacts. Despite a long history of use, PROPELLER sequences have not routinely supplanted conventional T2-weighted fast/turbo spin echo imaging. A paper by Dr. Rosenkrantz and colleagues helps explain why. In a study of 49 prostate cancer patients, it was noted that while a T2-weighted BLADE sequence reduced motion artifacts, those images also demonstrated lower contrast, particularly a reduction in tumor-to-peripheral-zone contrast. Anecdotally, this corresponds to our real-world experience with this sequence in pelvic MRI, where there can be a reduction in