Deep learning technologies and applications demonstrate one of the most important upcoming developments in radiology. The impact and influence of these technologies on image acquisition and reporting might change daily clinical practice. The aim of this review was to present current deep learning technologies, with a focus on magnetic resonance image reconstruction. The first part of this manuscript concentrates on the basic technical principles that are necessary for deep learning image reconstruction. The second part highlights the translation of these techniques into clinical practice. The third part outlines the different aspects of image reconstruction techniques, and presents a review of the current literature regarding image reconstruction and image post-processing in MRI. The promising results of the most recent studies indicate that deep learning will be a major player in radiology in the upcoming years. Apart from decision and diagnosis support, the major advantages of deep learning magnetic resonance imaging reconstruction techniques are related to acquisition time reduction and the improvement of image quality. The implementation of these techniques may be the solution for the alleviation of limited scanner availability via workflow acceleration. It can be assumed that this disruptive technology will change daily routines and workflows permanently.

Deep learning–reconstructed turbo spin-echo (TSE) acquisition was interchangeable with standard TSE for detecting various spinal abnormalities at MRI, providing excellent quality and a 70% reductio... Background Deep learning (DL)–based MRI reconstructions can reduce examination times for turbo spin-echo (TSE) acquisitions. Studies that prospectively employ DL-based reconstructions of rapidly ac...

Multiparametric MRI (mpMRI) of the prostate has become the standard of care in prostate cancer evaluation. Recently, deep learning image reconstruction (DLR) methods have been introduced with promising results regarding scan acceleration. Therefore, the aim of this study was to investigate the impact of deep learning image reconstruction (DLR) in a shortened acquisition process of T2-weighted TSE imaging, regarding the image quality and diagnostic confidence, as well as PI-RADS and T2 scoring, as compared to standard T2 TSE imaging. Sixty patients undergoing 3T mpMRI for the evaluation of prostate cancer were prospectively enrolled in this institutional review board-approved study between October 2020 and March 2021. After the acquisition of standard T2 TSE imaging (T2_S), the novel T2 TSE sequence with DLR (T2_DLR) was applied in three planes. Overall, the acquisition time for T2_S resulted in 10:21 min versus 3:50 min for T2_DLR. The image evaluation was performed by two radiologists independently using a Likert scale ranging from 1–4 (4 best) applying the following criteria: noise levels, artifacts, overall image quality, diagnostic confidence, and lesion conspicuity. Additionally, T2 and PI-RADS scoring were performed. The mean patient age was 69 ± 9 years (range, 49–85 years). The noise levels and the extent of the artifacts were evaluated to be significantly improved in T2_DLR versus T2_S by both readers (p < 0.05). Overall image quality was also evaluated to be superior in T2_DLR versus T2_S in all three acquisition planes (p = 0.005–<0.001). Both readers evaluated the item lesion conspicuity to be superior in T2_DLR with a median of 4 versus a median of 3 in T2_S (p = 0.001 and <0.001, respectively). T2-weighted TSE imaging of the prostate in three planes with an acquisition time reduction of more than 60% including DLR is feasible with a significant improvement of image quality.

The objective of this study is to conduct a qualitative and a quantitative image quality and lesion evaluation in patients undergoing MR-guided radiation therapy (MRgRT) for prostate cancer on a hybrid magnetic resonance imaging and linear accelerator system (MR-Linac or MRL) at 1.5 Tesla. This prospective study was approved by the institutional review board. A total of 13 consecutive patients with biopsy-confirmed prostate cancer and an indication for MRgRT were included. Prior to radiation therapy, each patient underwent an MR-examination on an MRL and on a standard MRI scanner at 3 Tesla (MRI_3T). Three readers (two radiologists and a radiation oncologist) conducted an independent qualitative and quantitative analysis of T2-weighted (T2w) and diffusion-weighted images (DWI). Qualitative outcome measures were as follows: zonal anatomy, capsule demarcation, resolution, visibility of the seminal vesicles, geometric distortion, artifacts, overall image quality, lesion conspicuity, and diagnostic confidence. All ratings were performed on an ordinal 4-point Likert scale. Lesion conspicuity and diagnostic confidence were firstly analyzed only on MRL. Afterwards, these outcome parameters were analyzed in consensus with the MRI_3T. Quantitative outcome measures were as follows: anteroposterior and right left diameter of the prostate, lesion size, PI-RADS score (Prostate Imaging—Reporting and Data System) and apparent diffusion coefficient (ADC) of the lesions. Intergroup comparisons were computed using the Wilcoxon-sign rank test and t tests. A post-hoc regression analysis was computed for lesion evaluation. Finally, inter-/intra-reader agreement was analyzed using the Fleiss kappa and intraclass correlation coefficient. For T2w images, the MRL showed good results across all quality criteria (median 3 and 4). Furthermore, there were no significant differences between MRL and MRI_3T regarding capsule demarcation or geometric distortion. For the DWI, the MRL performed significantly less than MRI_3T across most image quality criteria with a median ranging between 2 and 3. However, there were no significant differences between MRL and MRI_3T regarding geometric distortion. In terms of lesion conspicuity and diagnostic confidence, inter-reader agreement was fair for MRL alone (Kappa = 0.42) and good for MRL in consensus with MRI_3T (Kappa = 0.708). Thus, lesion conspicuity and diagnostic confidence could be significantly improved when reading MRL images in consensus with MRI_3T (Odds ratio: 9- to 11-fold for the T2w images and 5- to 8–fold for the DWI) (p < 0.001). For measures of lesion size, anterior-posterior and right-left prostate diameter, inter-reader and intersequence agreement were excellent (ICC > 0.90) and there were no significant differences between MRL and MRI_3T among all three readers. In terms of Prostate Imaging Reporting and Data System (PIRADS) scoring, no significant differences were observed between MRL and MRI_3T. Finally, there was a significant positive linear relationship between lesion ADC measurements (r = 0.76, p < 0.01) between the ADC values measured on both systems. In conclusion, image quality for T2w was comparable and diagnostic even without administration of spasmolytic- or contrast agents, while DWI images did not reach diagnostic level and need to be optimized for further exploitation in the setting of MRgRT. Diagnostic confidence and lesion conspicuity were significantly improved by reading MRL in consensus with MRI_3T which would be advisable for a safe planning and treatment workflow. Finally, ADC measurements of lesions on both systems were comparable indicating that, lesion ADC as measured on the MRL could be used as a biomarker for evaluation of treatment response, similar to examinations using MRI_3T.

Purpose: The purpose of this study was to test the technical feasibility and the impact on the image quality of a deep learning-based super-resolution reconstruction algorithm in 1.5 T abdominopelvic MR imaging. Methods: 44 patients who underwent abdominopelvic MRI were retrospectively included, of which 4 had to be subsequently excluded. After the acquisition of the conventional volume interpolated breath-hold examination (VIBE_Std), images underwent postprocessing, using a deep learning-based iterative denoising super-resolution reconstruction algorithm for partial Fourier acquisitions (VIBE_SR). Image analysis of 40 patients with a mean age of 56 years (range 18–84 years) was performed qualitatively by two radiologists independently using a Likert scale ranging from 1 to 5, where 5 was considered the best rating. Results: Image analysis showed an improvement of image quality, noise, sharpness of the organs and lymph nodes, and sharpness of the intestine for pre- and postcontrast images in VIBE_SR compared to VIBE_Std (each p < 0.001). Lesion detectability was better for VIBE_SR (p < 0.001), while there were no differences concerning the number of lesions. Average acquisition time was 16 s (±1) for the upper abdomen and 15 s (±1) for the pelvis for VIBE_Std, and 15 s (±1) for the upper abdomen and 14 s (±1) for the pelvis for VIBE_SR. Conclusion: This study demonstrated the technical feasibility of a deep learning-based super-resolution algorithm including partial Fourier technique in abdominopelvic MR images and illustrated a significant improvement of image quality, noise, and sharpness while reducing TA.

Objectives: Thin-slice prostate MRI might be beneficial for prostate cancer diagnostics. However, prolongation of acquisition time is a major drawback of thin-slice imaging. Therefore, the purpose of this study was to investigate the impact of a thin-slice deep learning accelerated T2-weighted (w) TSE imaging sequence (T2_DLR) of the prostate as compared to conventional T2w TSE imaging (T2_S). Materials and Methods: Thirty patients were included in this prospective study at one university center after obtaining written informed consent. T2_S (3 mm slice thickness) was acquired first in three orthogonal planes followed by thin-slice T2_DLR (2 mm slice thickness) in axial plane. Acquisition time of axial conventional T2_S was 4:12 min compared to 4:37 min for T2_DLR. Imaging datasets were evaluated by two radiologists using a Likert-scale ranging from 1–4, with 4 being the best regarding the following parameters: sharpness, lesion detectability, artifacts, overall image quality, and diagnostic confidence. Furthermore, preference of T2_S versus T2_DLR was evaluated. Results: The mean patient age was 68 ± 8 years. Sharpness of images and lesion detectability were rated better in T2_DLR with a median of 4 versus a median of 3 in T2_S (p < 0.001 for both readers). Image noise was evaluated to be significantly worse in T2_DLR as compared to T2_S (p < 0.001 and p = 0.021, respectively). Overall image quality was also evaluated to be superior in T2_DLR versus T2_S with a median of 4 versus 3 (p < 0.001 for both readers). Both readers chose T2_DLR in 29 cases as their preference. Conclusions: Thin-slice T2_DLR of the prostate provides a significant improvement of image quality without significant prolongation of acquisition time.