Tuesday, December 1 | 10:10 a.m.-10:20 a.m. | T3-SSGU03-7 | Room S502
Deep-learning reconstruction (DLR) of super-resolution T2-weighted imaging (T2WI) MR images at 5-tesla produces superior image quality and quantitative metrics compared to both conventional and non-DLR reconstruction super-resolution sequences, a research team from Nanjing, China, has found.
"The marked improvement in edge definition and lesion conspicuity offered by DLR-enhanced T2WI has direct implications for improving diagnostic confidence in assessing critical features such as capsular integrity and small lesions, potentially elevating the technical standard for prostate biparametric MRI," presenter Xiangpan Meng, MD, of Southeast University and colleagues noted.
Meng will share findings from a study that included 103 men (mean age, 68 years) with suspected prostate cancer who underwent 5-tesla prostate MRI on a system equipped with a 24-channel super-flex body coil and a 48-channel spinal coil; the exam included a conventional high-resolution (T2C) phase, a super-resolution without DLR (T2NDL) phase, and a super-resolution with DLR (T2DL) phase. Two readers assessed image quality for the three phases using a five-point Likert scale, and the researchers assessed quantitative factors such as signal-to-noise ratio (SNR) and contrast-to-noise ratio CNR).
The super-resolution with DLR (T2DL) phase produced the highest qualitative scores compared to T2C and T2NDL (p<0.001). It also produced significantly higher mean SNR and CNR than both T2C and T2NDL (all p<0.001).
Check out this session to hear more about how the study results could affect clinical practice.
