Prostate MRI using a rigid two-channel phased-array endorectal coil: comparison with phased array coil acquisition at 3 T

Abstract

Background To compare image quality, lesion detection and patient comfort of 3T prostate MRI using a combined rigid two-channel phased-array endorectal coil and an external phased-array coil (ERC-PAC) compared to external PAC acquisition in the same patients. Methods Thirty three men (mean age 65.3y) with suspected (n = 15) or biopsy-proven prostate cancer (PCa, n = 18) were prospectively enrolled in this exploratory study. 3T prostate MRI including T2-weighted imaging (T2WI) and diffusion-weighted imaging (DWI) was performed using an ERC-PAC versus PAC alone, in random order. Image quality, lesion detection and characterization (biparametric PI-RADSv2.1) were evaluated by 2 independent observers. Estimated signal-to-noise ratio (eSNR) was measured in identified lesions and the peripheral zone (PZ). Patient comfort was assessed using a questionnaire. Data were compared between sequences and acquisitions. Inter/intra-observer agreement for PI-RADS scores was evaluated. Results Twenty four prostate lesions (22 PCa) were identified in 20/33 men. Superior image quality was found for ERC-PAC compared to PAC for T2WI for one observer (Obs.1, p < 0.03) and high b-value DWI for both observers (p < 0.05). The sensitivity of PI-RADS for lesion detection for ERC-PAC and PAC acquisitions was 79.2 and 75% for Obs.1, and 79.1 and 66.7%, for Obs.2, without significant difference for each observer (McNemar p-values ≥0.08). Inter−/intra-observer agreement for PI-RADS scores was moderate-to-substantial (kappa = 0.52–0.84). Higher eSNR was observed for lesions and PZ for T2WI and PZ for DWI using ERC-PAC (p < 0.013). Most patients (21/33) reported discomfort at ERC insertion. Conclusion Despite improved image quality and eSNR using the rigid ERC-PAC combination, no significant improvement in lesion detection was observed, therefore not supporting the routine use of ERC for prostate MRI.

Use of Imaging and Biopsy in Prostate Cancer Diagnosis: A Survey From the Asian Prostate Imaging Working Group

“…The preference for 3T without ERC is consistent with the results of a survey by the Society of Abdominal Radiology (SAR) prostate disease-focused panel (DFP) [ 6 ]. The rare use of ERC after its application for decades may be explained by patient discomfort, increased cost, possible susceptibility artifacts, and the need for software to correct the high signal intensity ratio bias [ 17 ].…”

Picture Perfect: The Status of Image Quality in Prostate MRI

“…At the same magnetic field strength, it is of no surprise that there are numerous studies that validate the proposed superiority in SNR that results from using an ERC 30–32 . Few studies have evaluated CNR, although Ullrich et al 31 did not observe improved discrimination between benign and cancerous prostate.…”

“…29 At the same magnetic field strength, it is of no surprise that there are numerous studies that validate the proposed superiority in SNR that results from using an ERC. [30][31][32] Few studies have evaluated CNR, although Ullrich et al 31 did not observe improved discrimination between benign and cancerous prostate. The expected improvement in image quality, as rated by radiologists on a 1 to 5 Likert scale, has been shown for DWI 32,33 and T2-WI 34 but certain studies have also reported no superiority with either set up.…”

“…Mirak et al 37 found a lower detection rate of anterior and transition zone tumors, but they used a cut-off point of GG 3 + 3 and ≥1 cm size. For any cancer, both Costa et al 38 and Turkbey et al 39 demonstrated greater sensitivity for cancer detection with ERC although a number of studies 32,34,36,40 have been unable to replicate this improvement. No differences in the ability of radiologists to accurately stage cancer, for example through detecting extraprostatic extension (EPE) and seminal vesicle (SV) invasion has been shown at 1.5 T 41 or 3 T. 42 Naturally, the majority of research reflects the reality of clinical practice where 1.5 T scanners using an ERC have been pitted against 3 T scanners without, in the aim of identifying which option may be more prone to artifacts and inferior image quality.…”

“…Few studies have evaluated CNR, although Ullrich et al 31 did not observe improved discrimination between benign and cancerous prostate. The expected improvement in image quality, as rated by radiologists on a 1 to 5 Likert scale, has been shown for DWI 32,33 and T2‐WI 34 but certain studies have also reported no superiority with either set up 30,35 . Where a 1 to 4 Likert scale was used to assess image quality, no significant difference was found 31 .…”