Abstract
Background/Objectives: There is a growing need for efficient prostate MRI protocols due to their increasing use in managing prostate cancer (PCa) and potential inclusion in screening. Deep learning reconstruction (DLR) may enhance MR acquisitions and improve image quality compared to conventional acceleration techniques. This systematic review examines DLR approaches to prostate MRI. Methods: A search of PubMed, Web of Science, and Google Scholar identified eligible studies comparing DLR to conventional reconstruction for prostate imaging. A narrative synthesis was performed to summarize the impact of DLR on acquisition time, image quality, and diagnostic performance. Results: Thirty-three studies showed that DLR can reduce acquisition times for T2w and DWI imaging while maintaining or improving image quality. It did not significantly affect clinical tasks, such as biopsy decisions, and performed comparably to human readers in PI-RADS scoring and the detection of extraprostatic extension. However, AI models trained on conventional data might be less accurate with DLR images. The heterogeneity in image quality metrics among the studies prevented quantitative synthesis. Discussion: DLR has the potential to achieve substantial time savings in prostate MRI while maintaining image quality, which is especially relevant because of increased MRI demands. Future research should address the effect of DLR on clinically relevant downstream tasks, including AI algorithms’ performances and biopsy decisions, and explore task-specific accelerated protocols for screening, image-guided biopsy, and treatment.
1. Introduction
The demand for magnetic resonance imaging (MRI) is increasing as its role in prostate cancer (PCa) early diagnosis, staging, active surveillance, follow-up, radiation treatment planning, and guidance expands [1]. This rise in demand is reflected in growing numbers of examinations, while overall waiting times for imaging services are increasing [2]. Including MRI in screening pathways would dramatically increase the demand for MRI capacity [3,4,5]. In the Goteborg-2 trial, second-stage MRI testing following PSA pre-selection included 6.7% and 17.9% of all invited men in the first round when PSA thresholds of 3.0 and 1.8 ng/mL were applied [3,6].
Multiple strategies can reduce the demand for MRI in diagnosis and screening, including using biomarkers [7]. Efficiency can be increased by acquiring only the necessary MRI sequences for the diagnostic task [8], such as omitting dynamic contrast-enhanced imaging [9,10]. Additional measures include accelerated MRI acquisition techniques, where time savings can lead to higher patient throughput [11]. Regardless of the strategies used, it is important to maintain or improve image quality and reduce the need for repeated acquisitions and patient recalls.
There are two main conventional accelerated MRI techniques: Parallel Imaging (PI) and Compressed Sensing (CS). PI uses the sensitivity profile of coil arrays to speed up image acquisition, while CS utilizes incoherent sampling of sparse information in MR images. Both techniques are commonly used in clinical applications, particularly T2w imaging [12]. Nevertheless, the signal-to-noise ratio and appearance of artifacts restrict the extent of acceleration achievable with these methods.
Deep learning reconstruction (DLR) is a technique that accelerates the process of MRI scanning by saving time, improving quality, and correcting motion [13]. DLR is offered by all major vendors of MR scanners [14] and can be used alongside conventional accelerated MRI methods. A notable aspect of DLR is its computational efficiency [15]. There are different approaches to implementing DLR in the reconstruction process. Image-based learning can improve the quality of reconstructed images after using conventional Inverse Fourier Transform (IFT) (Figure 1). K-space learning can enhance partially sampled raw data before reconstructions with IFT. Direct mapping achieves improvement and reconstruction in one step, using raw data as input to produce an improved image without IFT.
Figure 1.
Visual summary of accelerated MRI in prostate imaging using conventional and deep learning reconstruction (DLR). Furthermore, potential use cases, challenges, and future questions for the implementation of accelerated prostate imaging protocols.
This review examines deep learning (DL)-based reconstruction approaches aimed at either accelerating the acquisition or enhancing the quality of prostate MRI. First, sequence-specific acceleration techniques are evaluated. Second, accelerated screening protocols are reviewed. Third, image quality metrics are analyzed. Finally, we discuss the opportunities and challenges associated with future trials utilizing DL acceleration for protocol personalization in MRI.
2. Materials and Methods
This systematic review was registered in the PROSPERO International Prospective Registry of Systematic Reviews (CRD420251032136), where the protocol is available, and follows the PRISMA-DTA (Preferred Reporting Items for Systematic Reviews and Meta-analysis for Diagnostic Test Accuracy) guidelines (Appendix A) [16,17]. The electronic databases PubMed, Web of Science, and Google Scholar were searched for primary publications in English assessing DLR for MRI of the prostate published up to 17 March 2025. The first 400 entries from Google Scholar were included. The search terms were: prostate AND MRI AND (deep learning OR DL OR artificial intelligence OR AI) AND (reconstruct* OR sequence).
Two researchers (SU and GA) screened titles, abstracts, and full texts for eligibility. References of included reports were searched for dataset completion. Discrepancies were resolved by consensus. Original reports using artificial intelligence (AI) to improve prostate MRI image quality and/or accelerate acquisition in humans were eligible for inclusion. Conference abstracts, case reports, and non-English reports were excluded. Contact with the authors was sought if reports were not available.
Information from the individual reports was transferred into a data collection form by one reviewer and verified by the second reviewer (Appendix B). Data were included as presented in the included studies. First, sequence-specific acceleration techniques were assessed. Second, accelerated screening protocols were reviewed. Third, image quality metrics were analyzed. Narrative synthesis was used to summarize the effects of DLR on image quality. The reduction in acquisition time achieved with DLR was summarized with descriptive statistics. Narrative synthesis summarized the diagnostic performance of human readers and AI with DLR. One reviewer performed and one verified risk of bias assessment using QUADAS-2.
Quality metrics comparing reference and test sequences were recorded in the data extraction instrument. Due to the high methodological heterogeneity of the studies and the use of numerous quality metrics (Appendix B), quantitative synthesis was restricted to comparing the subjective overall image quality among the studies. The subjective image quality ratings for every reconstruction were recorded (e.g., median and inter-quartile range, mean and standard deviation, or frequency tables). Additionally, the relative acquisition time of the DLR sequence relative to the conventional comparator was recorded.
Consistent with most of the included studies, image quality was treated as a pseudo-continuous variable. The within-group standardized mean difference served as the effect size, considering the varying number of levels in the Likert scales for quality assessment and the paired nature of observations. The correlation between the paired observations, necessary for estimating the standard error, was determined empirically as r = 0.85 on in-house data.
The effect size could not be calculated directly for studies reporting median and interquartile range. To increase the number of studies in a secondary meta-analysis, the means and standard deviations of image quality were computed for these studies as described by Wan et al. This method assumes a normal distribution of the measurements [18]. The inclusion of these additional studies was assessed using a sensitivity analysis. An additional sensitivity analysis was conducted, excluding studies at a high risk of bias.
A three-level, random-effects meta-analysis computed the summary effect size, taking multiple, non-independent observations on a single cohort into account. The meta-analysis was performed separately for T2w imaging and diffusion-weighted imaging (DWI) and implemented with the “metafor” package (Version 4.8-0) in R language for statistical computing (Version 4.4.1) [19]. A funnel plot visualized the risk of publication bias. The I2 statistic estimated the heterogeneity among the studies that was not attributable to random sampling. Appendix E includes a more detailed description of the statistical methods.
3. Results
The search identified 510 unique records, 478 of which were excluded. In particular, 309 investigated unsuitable outcomes, i.e., they did not assess the impact of deep learning image reconstruction on image quality or acquisition time. Examples include studies focusing on using AI for prostate cancer detection with MRI, image segmentation, radiomics analyses, or genomic profiling. Fifty-five studies were classified as evaluating the wrong population because they investigated conditions other than prostate cancer. Examples include imaging studies related to orthopedic, cerebral, or non-prostate pelvic diseases. Eventually, 33 studies (Figure 2) fulfilled the inclusion criteria, investigating possible image quality improvements for individual sequences [20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52], the application of DLR in abbreviated protocols [30,37,40], and the utility of DLR for radiotherapy [52]. Twenty-five reports investigated DLR for T2w imaging, while 11 assessed DLR for DWI (Figure 3). Studies reported on men undergoing MRI for PCa diagnosis (18), patients with biopsy-proven PCa (3), patients undergoing MRI for multiple indications (11), or healthy volunteers (1). One study [42] included publicly available data from the fastMRI initiative, a dataset of T2w and DWI containing the k-space data of 312 participants [53,54], complemented by data obtained from the vendor. The remaining 32 studies included only in-house data, with nine stating that the data were available from the authors upon request and one indicating that the data were unavailable.
Figure 2.
PRISMA study selection flowchart.
Figure 3.
Characteristics of the included studies with (A) investigated sequence(s), (B) vendor of MRI scanners, (C) type of acceleration, (D) type of deep learning, (E) prospective or retrospective study design, and (F) MRI field strength.
Studies on T2w imaging and DWI investigated image-based learning (12), k-space learning (6), or direct mapping (14). Algorithms were developed in-house (9) and by vendors (23), 61% of which were product sequences. Studies most commonly employed 3T machines (85%) from a variety of vendors: Siemens (15), GE (7), Philips (5), Canon (3), United Imaging (1), and Elekta (1) (Figure 3).
3.1. Risk of Bias Assessment
Overall, the risk of bias was low, although seven studies were identified as having a high risk of bias or reduced applicability related to participant selection. This was due to the inclusion of participants with lesions visible on conventional MRI [28,39], those who had undergone biopsy or subsequently underwent prostatectomy [35,36,41,48], or healthy volunteers [44]. These cohorts were either enriched in patients with csPCa or consisted of younger participants with less csPCa. Therefore, they only partially represent clinical target populations. One study was at a high risk of bias in performing the reference test (conventional sequences) and index test (DLR sequences) [34]. The unblinded side-by-side comparison of the two sequences may result in expectation bias, where the readers favor one of the techniques, and anchoring bias, where previous observations influence subsequent evaluations. Appendix D includes the full risk of bias assessment.
3.2. Deep Learning Reconstruction
3.2.1. DLR in T2w Imaging
DLR was employed to accelerate T2w imaging or improve the image quality sequences without additional acceleration. Nine studies applied DLR without additional acceleration compared to the reference T2w sequence. DLR achieved a higher image quality in seven reports, while two observed a lower quality. Seven studies used objective image quality metrics. The most commonly calculated image quality metrics were the SNR and the contrast-to-noise ratios (CNR) in three and four studies, respectively. One study used other metrics. The objective image quality improved in five studies, was unaltered in one study, and deteriorated in another study when acquisition times were constant.
DLR can potentially accelerate image acquisition while maintaining or increasing image quality. This was investigated with accelerated acquisitions (15) or simulations (3; Table 1). The acquisition time was reduced to a minimum of 13–64% of the reference sequence, representing a 1.56–7.7-fold acceleration. Acceleration did not lead to a deterioration in objective image quality metrics in any of the studies. Subjective preference at low acceleration factors favored DLR. At 3-fold and higher acceleration, DLR was of comparable quality to conventional sequences.
Table 1.
Summary of acceleration approach (simulated or acquired) and domain of the input data (image space, k-space, or k-space with direct mapping).
Six studies evaluated the effect of DLR on clinically relevant tasks. These included cancer detection by AI algorithms and human diagnostic performance, and detecting extraprostatic extension (EPE). AI algorithms trained on conventional MRI data performed worse on DL-reconstructed T2w data [47,50]. DLR had a less significant impact on human diagnosis, showing comparable performance for transition zone lesions, despite a 3-fold acceleration [35]. Another study with 10 patients showed similar diagnostic performance [30]. Evidence on EPE detection varied: one report showed similar performance despite a 37–70% reduction in acquisition time [35,41], while another showed decreased performance with high-resolution DLR T2w imaging [36].
Evaluation cohorts included a median of 46 participants (IQR 30–96). The study by van Lohuizen et al., a multicenter trial of three centers in the Netherlands, employed the largest testing cohort, with 306 participants [50]. It used reconstructed images and Fast Fourier Transform (FFT) to obtain sub-sampled k-space-like data that were sub-sampled, avoiding the prospective collection of k-space data. Some ADC values derived from DLR images varied from the reference but showed a strong linear correlation. Yet, these differences were typically modest and within the test–retest limits defined by the Quantitative Imaging Biomarkers Alliance (QIBA) profile 2024 [55].
3.2.2. DLR in Diffusion-Weighted Imaging
Without additional acceleration, the subjective image quality of DLR DWI was comparable in three studies and improved in two studies. All five studies found a higher signal-to-noise ratio (SNR) in DLR images (1.24–2.17-fold). The studies had no other shared objective image quality metric.
DLR was employed to accelerate image acquisition in eight studies. Acquisition time reductions ranged between 13% and 69% of the reference sequences. The subjective image quality was independent of the degree of acceleration. However, retrospective sub-sampling of k-space simulated acceleration in five studies, and only three studies acquired accelerated sequences, potentially reducing gains in image quality from reduced motion artifacts. The effects of DLR in DWI on downstream tasks remain unexplored. Models were evaluated on relatively small cohorts with a median of 51 participants (IQR 36–60), lowering clinical generalizability.
3.2.3. DLR in AI-Accelerated Screening/Diagnostic Protocols
Three studies investigated complete accelerated bi-parametric protocols [30,37,40], comparable to proposed screening protocols.
In a retrospective study, Johnson et al. simulated undersampling of axial and coronal T2w and axial DWI sequences, validating a simulated acceleration from 11:48 min to 3:12 min (− 73%) at maintained image quality in 20 participants. No difference in diagnostic performance was detected, albeit only a few lesions were present in the cohort [30].
Lee et al. conducted a retrospective analysis of 40 prospectively recruited patients who underwent axial T2w and DWI. Standard of care imaging, T2w with a reduced number of averages and standard or higher resolution, and DWI with a reduced number of averages were reconstructed with different DLR strengths. With all of the sequences, DLR increased SNR and CNR ratios. Two radiologists rated the overall image quality as similar to the standard of care for most reconstructions. Only the accelerated T2w sequence reconstructed with medium-strength DLR showed a higher quality [37].
Oerther et al. studied 77 men suspected of prostate cancer using an accelerated and abbreviated bi-parametric MRI protocol with T2w imaging and DWI (3:28 min) compared to standard mpMRI with 3-planar T2w TSE and DCE (25:45 min). The abbreviated protocol had better image quality, higher inter-rater agreement in PI-RADS 2.1 scoring, and comparable diagnostic performance to the standard protocol [40].
3.2.4. DLR in PI-RADS Assessment
Thirteen studies assessed the effect of DLR on PI-RADS scoring, including studies investigating bi-parametric screening protocols (DLR T2w and DLR DWI; 3), DLR T2w with conventional DWI (9), or DLR DWI with conventional T2w (1). Most studies found good-to-excellent agreement or no significant difference between PI-RADS scoring performed on conventional and DLR images. Only one study observed significant upgrading with DLR DWI [29]. Inter-reader agreement increased with DLR in three studies [25,29,50]. In a prospective study of DLR for T2w imaging, Bischoff et al. observed a change in PI-RADS scores in 6% of patients that reduced the number of PI-RADS 3 cases by 78–86% through up- or down-scoring [21].
In a study comparing conventional multi-parametric MRI to DLR bi-parametric MRI, one of two readers had a lower specificity for lesions ≥ PI-RADS 3 and a lower sensitivity for lesions ≥ PI-RADS 4 on DLR bi-parametric MRI [43]. Only one study offered sufficient details to calculate an effect size [21]. Therefore, a meta-analysis of diagnostic performance was not possible.
3.3. Methods for Image Quality Assessment
3.3.1. Objective Image Quality Metrics
Twenty-six studies employed objective image quality metrics. However, the specific metrics used in the studies varied widely (Table 2). Image quality metrics most commonly included SNR, CNR, and edge rise distance/slope profile in 13, 12, and 6 studies, respectively. The most commonly used fully referenced quality metrics were structural similarity index, peak SNR, and root mean square error. Those metrics compare DLR images against a high-quality ground truth.
Table 2.
Summary of objective and subjective image quality metrics used in the included studies.
3.3.2. Subjective Image Quality Metrics
Twenty-nine studies assessed subjective image quality with one to five readers (median 2). Quality metrics, while differently defined, fell into one of six categories: overall image quality, noise, artifacts, sharpness, conspicuity of organs, and diagnostic confidence. Binary classifications, 4- and 5-tier Likert classifications were used.
Appendix B lists all subjective and objective image quality metrics for individual studies.
3.4. Quantitative Synthesis
Twenty-six studies reported subjective overall image quality and were potentially eligible for inclusion in the meta-analysis. Five studies did not report sufficient details to obtain an effect size. Two studies were excluded because they reported an interquartile range of 0, which did not allow for approximating a standard error. The final meta-analysis included eight studies in the primary analysis. An additional 11 studies with effect sizes computed indirectly were included in the extended cohort for the sensitivity analysis (Appendix F). The studies reported 26 comparisons in the primary and 52 in the extended analysis. One study reported a combined quality score for T2w imaging and DWI included in the meta-analysis of both sequences.
The meta-analysis revealed a non-significant trend towards a higher image quality with DLR on T2w imaging (summary effect size of 0.49, 95% confidence interval −0.39–1.39, p = 0.25, df = 13, Figure 4a). However, the overall heterogeneity of between-study comparisons was high, with an I2 of 77%. The heterogeneity of within-study comparisons was moderate at I2 = 22%, resulting in a total heterogeneity of I2 = 98%. The funnel plot reveals some asymmetry favoring DLR, indicative of possible publication bias (Figure 5). The sensitivity analysis, including studies where the effect size was approximated from median and interquartile range, showed a borderline significant trend towards higher image quality for DLR (summary effect size of 1.38, 95% confidence interval 0.00–2.04, p = 0.05, df = 32). However, this increase in quality was determined by two studies with extreme results (Appendix F). Image quality was not associated with the relative acquisition time (Figure 6a).
Figure 4.
Forest plot of studies investigating DLR for T2w imaging (a) and DWI (b) [29,30,37,39,40,47,48,51].
Figure 5.
Funnel plots of the studies included in the meta-analysis of subjective image quality in T2w (a) and DWI (b).
Figure 6.
Bubble plot of the meta-regression examining the association between subjective image quality and the relative acquisition time for T2w (a) and DWI (b). Each bubble represents a study, and its size is proportional to the precision of the effect size estimate. The 95% confidence interval of the regression line is shaded in gray.
The analysis of DWI revealed no significant association of image quality with the reconstruction method (summary effect size of 0.67, 95% confidence interval −0.41–1.74, p = 0.21, df = 13, Figure 4b). The overall heterogeneity of between-study comparisons was high, with an I2 of 78%. The heterogeneity of within-study comparisons was moderate at I2 = 20%, resulting in a total heterogeneity of I2 = 99%. The funnel plot showed a moderate asymmetry favoring DLR (Figure 5), indicative of possible publication bias. The sensitivity analysis, including studies where the effect size was approximated from median and interquartile range, did not change the results (summary effect size of 0.62, 95% confidence interval −0.26–1.50, p = 0.16, df = 19, Appendix F). Meta-regression showed no significant association between the relative acquisition time and image quality Figure 6.
4. Discussion
This systematic review evaluated the rapidly increasing literature on DLR for prostate MRI. Most of the research took place in the screening and early diagnostic setting. However, cohorts were frequently heterogeneous. The assessment of DLR primarily focused on image quality evaluation, and research on the effects of DLR on downstream tasks is only just beginning to accrue. Methodological heterogeneity and considerable variability in quality metrics limited the scope of quantitative comparisons. The potential of DLR for prostate imaging has been recognized and reaches beyond diagnostic imaging with initial research on the use of DLR for treatment guidance in radiotherapy.
4.1. Benefits of DLR
DLR can impact MRI protocol development and aid in creating MRI screening strategies for prostate cancer. This review demonstrates that DLR can speed up T2w and DWI acquisition of the prostate by up to 8-fold without sacrificing image quality [50]. A meta-analysis of the subjective image quality revealed no significant difference between conventional and DLR images among the included studies. However, there was considerable heterogeneity among the studies, with some aiming for image quality improvement and some for accelerated scanning time. Faster scans improve patient comfort and reduce motion artifacts. DLR can enhance image quality on lower-specification MRI systems without longer acquisition times. Cost-effectiveness studies indicate that DLR increases scanner capacity more effectively than buying another MRI system or extending weekend hours [11].
4.2. Risks of DLR
While image quality in conventional reconstruction was a trade-off between acquisition time, SNR, and resolution, additional factors play into DLR (Figure 7). One key risk is hallucinations, the addition or removal of realistic-looking image details, which may arise when there are multiple plausible solutions to the inverse image reconstruction problem. A deep learning model, influenced by the priors learned from the training data, may favor solutions introducing hallucinated features. The risk of hallucinations increases with stronger acceleration, i.e., decreasing sampling rates [56]. Reduced image noise may also reveal artifacts that are not ordinarily visible, leading to false-positive findings. A second phenomenon is instability, where small perturbations in the input (e.g., noise) lead to severe artifacts or degradation of image quality [57]. Importantly, hallucination and instability are inevitable in many cases of DLR. Achieving perfect accuracy, stability, and hallucination-free reconstructions simultaneously is not possible [58]. Nonetheless, DLR and conventional acceleration techniques require similar caution in their clinical implementation.
Figure 7.
Classically, image quality was seen as a trade-off between the signal-to-noise ratio, acquisition time, and spatial resolution (blue). Deep learning reconstruction (DLR) may increase image quality while reducing the acquisition time. However, as the sampled data become increasingly sparse, hallucinations, over-smoothing, and instability (yellow) become a greater concern.
Further artifacts specific to DLR include over-smoothing, which can mimic the charcoal sign in the transition zone, leading to false positives. Conversely, the loss of subtle findings adjacent to the capsule may cause false negatives. These artifacts depend on algorithm settings, vendors, scanners, and patient anatomy.
4.3. Effect of DLR on PI-RADS Assessment and Downstream Clinical Tasks
Image quality is only a surrogate marker for the aim of imaging: improving outcomes in a diagnostic or therapeutic pathway. Therefore, it is essential to evaluate the effect of DLR on downstream tasks such as the diagnosis of prostate cancer, AI-assisted lesion detection, biopsy guidance, or automated segmentation for radiotherapy (Figure 1). Initiatives to probe the effect of DLR on downstream tasks were ongoing.
There was no evidence that DLR would significantly impact PI-RADS scoring for human readers. The agreement between conventional and DLR images is generally high, and the significant difference in one study was modest in magnitude [40]. As only one comparison used DLR in T2w imaging and DWI simultaneously, data on fully DLR protocols are still scarce. Nonetheless, no evidence suggested that DLR significantly impacts PI-RADS scoring.
It is also relevant to go beyond comparing two imaging outcomes (e.g., PI-RADS). Instead, we should relate outcomes from imaging to a reference standard, e.g., follow-up or histological diagnosis of PCa. Among the studies that evaluated DLR on a downstream task not based on an imaging outcome (e.g., detection of histologically proven cancer or extraprostatic extension), the performance of conventional and DLR images was similar [29,30,34,35,41,48]. Where significant differences occurred, they were generally small [36]. This non-inferior evidence must still be considered preliminary, as the studies were retrospective, often included small patient cohorts, were at a high or uncertain risk of selection bias, and used biopsies guided by standard MRI as the reference standard. Biopsy guidance with DLR sequences remains unexplored.
ADC values derived from DLR may exhibit slight deviations from those obtained using conventional imaging techniques. These differences can be attributed to several factors: (1) denoising effects, where noise reduction—particularly at high b-values—can unmask underlying signal; (2) the characteristics of the DLR algorithm and its training data; (3) variations in sequence design and imaging protocols; and (4) tissue-specific effects. While the clinical significance of these differences is likely limited, given that relative ADC contrasts between tissues are generally preserved and the absolute changes are minor, sequence-specific validation is advisable before quantitative analysis.
4.4. Effect of DLR on Diagnostic AI
Diagnostic AI for PCa detection on MRI has been a focus of intense research, and models for clinical use are available from multiple vendors. Therefore, the combined diagnostic performance of DLR images as input for diagnostic AI algorithms is of great importance.
Van Lohuizen et al. and Tong et al. investigated the effect of DLR of T2w images on diagnostic prostate AI [47,50]. While on objective and subjective metrics, image quality was higher than or comparable to the reference standard in both studies, the performance of diagnostic AI on DLR images was inferior. This highlights the challenge posed by the out-of-domain nature of DLR images. One solution is to incorporate DLR images into the training data for AI. Nonetheless, use cases of AI where multiple AI algorithms interact may lead to unexpected outcomes (the double black box phenomenon), warranting careful evaluation in clinical practice.
4.5. Heterogeneity in the Evaluation of DLR Image Quality
Various metrics were employed for subjective and objective image quality assessment, limiting the comparability of the results across the studies included in this systematic review. An objective quantification of image quality is attractive as it may allow a comparison of outcomes across different DLR algorithms. However, the apparent precision of a numerical metric may correlate only poorly with the visual quality of an image. The SNR and CNR were the most common quality metrics. While both relate to image noise and signal, they neglect sharpness and image distortions. Similarly, the most common fully referenced metrics, peak SNR and structural similarity index measure (SSIM), agree poorly with radiologists’ subjective image quality assessments [59]. Furthermore, they are strongly dependent on image normalization [60].
No single image quality metric is sufficiently sensitive to all possible causes of image distortions. In response to this, Dohmen et al. suggested a combination of SSIM, learned perceptual image patch similarity (LPIPS), mean squared error (MSE), and normalized mutual information (NMI) to cover a large range of possible distortions [60]. Still, using fully referenced metrics will be unable to detect improvements in the quality of DLR images compared to the reference standard.
4.6. DLR Beyond Diagnosis
The increasing demand for prostate MRIs will likely require shorter and more personalized protocols that will have to be adapted more closely to downstream tasks. Consequently, novel sequences need to be evaluated in concert with the intended use case. For example, a diagnostic protocol requires high sensitivity to identify clinically significant cancer. Meanwhile, a simulation/planning scan for radiation oncology might weigh accurate anatomical representation higher, as this will impact treatment planning and dose delivery to known lesions. Finally, MRI for adaptive radiotherapy serves as the confirmation of existing treatment plans. Therefore, radiation oncologists may accept a slightly reduced image quality if the shorter acquisition time reduces patient motion and discomfort.
4.7. Limitations
This systematic review has some limitations. The heterogeneity among the included studies restricted the scope of the quantitative synthesis to the subjective overall image quality. This analysis confirmed the high methodological heterogeneity among the studies. Studies varied in terms of the included patient population, the operating point of the AI models (quality improvement vs. acceleration), and metrics for image quality assessment. Additional variation was introduced by different scanners from different vendors, with variable field strengths, and different clinical outcomes. Such variation represents the heterogeneity among medical centers, as the centers’ role in the healthcare system, long-term infrastructure investment, and software/hardware integration differ. It is also representative of a multi-faceted challenge where image quality is only the first layer. Only a few studies investigated the impact of DLR on the diagnostic performance of prostate MRI or biopsy guidance as clinically relevant endpoints. Finally, the review only included studies published in English. This decision was based on the limited proficiency of the research team in other relevant languages, which would have hindered devising a search strategy and assessing manuscripts consistently and reliably across languages.
4.8. Summary and Recommendations
Box 1 summarizes the principal findings of this systematic review. Box 2 presents recommendations and suggestions for the implementation of DLR for prostate MRI based on the findings of this systematic review.
Box 1. Principal findings of the systematic review.
Image quality in T2w and DWI sequences
- Non-accelerated DLR improves the subjective and objective image quality of T2w sequences, with maintained quality at 3-fold and stronger acceleration.
- DLR achieves a higher SNR in DWI sequences.
- Diagnostic AI performs invariably less well on DLR images, even if objective and subjective image quality metrics were higher.
DLR screening protocols and diagnostic performance
- DLR may reduce acquisition times without a loss in image quality.
- PI-RADS scoring is comparable between conventional and DLR sequences.
- Bi-parametric DLR protocols show similar diagnostic accuracy in small study cohorts.
Methods image quality assessment
- Common image quality metrics include no-reference and fully-referenced metrics (structural similarity index, peak SNR, and root mean square error).
- Subjective image quality metrics assess overall image quality, noise, artifacts, sharpness, conspicuity of structures, and diagnostic confidence.
Risks and limitations of DLR
- Stronger acceleration increases the risk of hallucinations and instability.
- Perfect accuracy, stability, and hallucination-free reconstructions are not achievable with DLR, requiring caution in clinical implementation.
Box 2. Recommendations and suggestions on the implementation of DLR.
Recommendations:
- Select the acceleration factor of DLR judiciously to balance the desired acquisition time savings with the risk of hallucinations and instability [56].
- Implement deep learning image reconstruction in coordination with the entire clinical team, including referring physicians if necessary, e.g., by means of side-by-side review sessions.
- Monitor the compatibility of DLR with downstream workflows such as diagnostic AI [47,50].
- Adapt DLR to the specific clinical context. The trade-off between image quality and acquisition time may vary between use cases [52].
- Monitor diagnostic performance and biopsy yield, especially when biopsy decision making transitions from conventional to DLR sequences [21].
Considerations:
- Implement DLR in a step-wise fashion rather than full DLR protocols.
- Implement DLR in addition to conventional sequences initially and while fine-tuning sequence parameters to achieve the desired diagnostic quality.
- Fully replace secondary sequences (coronal and sagittal T2w, axial T1w and DCE) before replacing axial T2w and DWI.
- Standardize parameters for image quality assessment or find a consensus on which one should be used.
- Investigate the effect DLR would have had on biopsy decisions in patients undergoing conventional and DLR imaging.
- Perform a comparative analysis of diagnostic AI using conventional and DLR-acquired images to assess consistency and potential performance shifts.
- Record the use of DLR in reports to facilitate audits and AI monitoring.
5. Conclusions
Image quality analyses indicate that DLR can accelerate high-quality prostate MRI in various use cases. Future research should prioritize prospective, reader/blinded, and paired study designs with histological validation as a reference standard to determine the effect of DLR on diagnostic accuracy. Developing task-specific protocols will be key to successfully implementing DLR, particularly when additional AI models are applied to images further downstream in the analysis pathway.
Author Contributions
Conceptualization, S.U. and I.G.S.; methodology, S.U., G.A. and I.G.S.; validation, G.A. and I.G.S.; formal analysis, S.U. and G.A.; investigation, S.U. and G.A.; resources, I.G.S.; data curation, S.U.; writing—original draft preparation, S.U. and I.G.S.; writing—review and editing, all co-authors; visualization, S.U.; supervision, R.G.H.B.-T. and I.G.S.; project administration, I.G.S. All authors have read and agreed to the published version of the manuscript.
Funding
This research received no external funding.
Institutional Review Board Statement
Not applicable.
Informed Consent Statement
Not applicable.
Data Availability Statement
No new data were created or analyzed in this study. Data sharing is not applicable to this article.
Conflicts of Interest
T.N.B. received free access to software and servers from Siemens Healthineers for their research. The other authors declare no conflicts of interest.
Abbreviations
The following abbreviations are used in this manuscript:
| ADC | Apparent diffusion coefficient |
| AI | Artificial intelligence |
| CNR | Contrast-to-noise ratio |
| CS | Compressed sensing |
| DLR | Deep learning reconstruction |
| DWI | Diffusion-weighted imaging |
| EPE | Extra-prostatic extension |
| IFT | Inverse Fourier transform |
| IQR | Interquartile range |
| (mp)MRI | (multi-parametric) Magnetic resonance imaging |
| PCa | Prostate cancer |
| PI | Parallel imaging |
| PI-RADS | Prostate imaging reporting and data system |
| PSA | Prostate-specific antigen |
| SNR | Signal-to-noise ratio |
| T2w | T2-weighted |
Appendix A. PRISMA Checklist
Figure A1.
PRISMA Checklist.
Appendix B. Data Extraction Instrument
Table A1.
Data extraction instrument.
Table A1.
Data extraction instrument.
| # | First Author | Journal | Year | Vendor | Field Strength |
|---|---|---|---|---|---|
| 1 | Belue | European Journal of Radiology | 2024 | Philips | 3 |
| 2 | Bischoff | Radiology | 2023 | Philips | 3 |
| 3 | Boschheidgen | Magnetic Resonance Imaging | 2025 | Siemens | 3 |
| 4 | Cochran | Clinical Imaging | 2025 | GE | 3 |
| 5 | Gassenmaier | Cancers | 2023 | Siemens | 3 |
| 6 | Gassenmaier | European Journal of Radiology | 2021 | Siemens | 3 |
| 7 | Gassenmaier | Cancers | 2021 | Siemens | 3 |
| 8 | Harder | Cancers | 2022 | Philips | 3 |
| 9 | Hu | Frontiers in Oncology | 2021 | Siemens | 3 |
| 10 | Jeong | Quantitative imagin in medicine and surgery | 2024 | Siemens | 3 |
| 11 | Johnson | Journal of Magnetic Resonance Imaging | 2022 | Siemens | 3 |
| 12 | Jung | British Journal of Radiology | 2022 | Siemens | 3 |
| 13 | Jurka | Quantitative imagin in medicine and surgery | 2024 | Philips | 3 |
| 14 | Kaye | Radiology: Artificial Intelligence | 2020 | GE | 3 |
| 15 | Kim | Cancers | 2024 | GE | 3 |
| 16 | Kim | Nature Scientifc Reports | 2024 | Siemens | 3 |
| 17 | Kim | European Journal of Radiology | 2021 | Siemens | 3 |
| 18 | Lee | European Journal of Radiology | 2023 | GE | 3 |
| 19 | Liu | Journal of Digital Imaging | 2021 | Siemens | 3 |
| 20 | Nishioka | European Journal of Radiology Open | 2024 | Philips | 3 |
| 21 | Oerther | European Radiology | 2024 | Siemens | 3 |
| 22 | Park | Journal of Magnetic Resonance Imaging | 2022 | GE | 3 |
| 23 | Pfaff | Nature Scientifc Reports | 2024 | Siemens | 0.55, 1.5 and 3 |
| 24 | Riederer | Abdominal Radiology | 2024 | GE | 3 |
| 25 | Sato | Radiological Physics and Technology | 2024 | Canon | 1.5 |
| 26 | Shen | Academic Radiology | 2024 | United Imaging | 3 |
| 27 | Shirashi | Magnetic Resonance Imaging | 2024 | Canon | 3 |
| 28 | Tong | Journal of Magnetic Resonance Imaging | 2023 | Siemens | 3 |
| 29 | Ueda | Radiology | 2022 | Canon | 3 |
| 30 | Ursprung | European Journal of Radiology | 2023 | Siemens | 3 |
| 31 | van Lohuizen | European Radiology | 2024 | NA | NA |
| 32 | Wang | Abdominal Radiology | 2021 | GE | 1.5 and 3 |
| 33 | Zhu | Radiation Oncology | 2023 | Elekta | 1.5 |
Table A1.
Cont.
Table A1.
Cont.
| # | Sequence | DL Type | Space | Direct Mapping | Participants |
|---|---|---|---|---|---|
| 1 Belue | T2w | in-house | Image | NA | 96 |
| 2 Bischoff | T2w | product | k-space | yes | 109 |
| 3 Boschheidgen | T2w | product | k-space | yes | 120 |
| 4 Cochran | DWI | product | k-space | no | 52 |
| 5 Gassenmaier | T2w | vendor research | k-space | yes | 30 |
| 6 Gassenmaier | T2w | vendor research | k-space | yes | 30 |
| 7 Gassenmaier | T2w | vendor research | k-space | yes | 60 |
| 8 Harder | T2w | product | k-space | yes | 33 |
| 9 Hu | DWI | in-house | Image | NA | Total 210, validation 60 |
| 10 Jeong | DWI | vendor research | k-space | yes | 70 |
| 11 Johnson | T2w and DWI | in-house | k-space | yes | Total 113, validation 20 |
| 12 Jung | T2w | in-house | Image | NA | 40 |
| 13 Jurka | T2w | vendor research | k-space | yes | 47 |
| 14 Kaye | DWI | in-house | Image | NA | Total 140, validation 37 |
| 15 Kim | T2w | product | k-space | no | 88 |
| 16 Kim | T2w | NA | NA | NA | 162 |
| 17 Kim | T2w | vendor research | k-space | yes | 46 |
| 18 Lee | T2w and DWI | product | k-space | no | 40 |
| 19 Liu | T2w | in-house | Image | NA | 346 |
| 20 Nishioka | DWI | product | k-space | yes | 32 |
| 21 Oerther | T2w and DWI | product | k-space | yes | 77 |
| 22 Park | T2w | product | k-space | no | 109 |
| 23 Pfaff | DWI | in-house | Image | NA | 409 training, 112 validation, 51 testing |
| 24 Riederer | T2w | product | k-space | no | 17 |
| 25 Sato | T2w | product | Image | NA | 13 |
| 26 Shen | T2w | product | Image | NA | 40 |
| 27 Shirashi | T2w | vendor research | Image | NA | 28 |
| 28 Tong | T2w | vendor research | k-space | yes | 80 |
| 29 Ueda | DWI | product | Image | NA | 60 |
| 30 Ursprung | DWI | vendor research | k-space | yes | 35 |
| 31 van Lohuizen | T2w | in-house | Image | NA | Total 1536, validation 306 |
| 32 Wang | T2w | product | k-space | no | 31 |
| 33 Zhu | T2w | in-house | Image | NA | 19 |
Table A1.
Cont.
Table A1.
Cont.
| # | Acceleration | Study Design | Clinical Setting |
|---|---|---|---|
| 1 Belue | NA | retrospective | Diagnostic prostate MRI |
| 2 Bischoff | acquired | prospective | Diagnostic prostate MRI |
| 3 Boschheidgen | acquired | prospective | Diagnostic prostate MRI |
| 4 Cochran | simulated k-space | retrospective | Diagnostic prostate MRI |
| 5 Gassenmaier | acquired | prospective | Diagnostic or biochemical recurrence |
| 6 Gassenmaier | acquired | retrospective | Diagnostic prostate MRI |
| 7 Gassenmaier | acquired | prospective | Diagnostic prostate MRI |
| 8 Harder | acquired | prospective | Biopsy proven prostate cancer |
| 9 Hu | NA | prospective | Diagnostic prostate MRI |
| 10 Jeong | acquired | retrospective | Biopsy proven prostate cancer |
| 11 Johnson | simulated k-space | retrospective | Any prostate MRI |
| 12 Jung | acquired and simulated | both | Diagnostic prostate MRI |
| 13 Jurka | acquired | prospective | Diagnostic prostate MRI |
| 14 Kaye | simulated k-space | retrospective | Any prostate MRI |
| 15 Kim | NA | prospective | Diagnostic prostate MRI |
| 16 Kim | acquired | retrospective | Diagnostic prostate MRI |
| 17 Kim | acquired | retrospective | Any prostate MRI |
| 18 Lee | acquired | prospective | Any prostate MRI |
| 19 Liu | NA | unclear | Diagnostic prostate MRI |
| 20 Nishioka | NA | retrospective | Diagnostic prostate MRI |
| 21 Oerther | acquired | prospective | Diagnostic prostate MRI |
| 22 Park | acquired | retrospective | Diagnostic prostate MRI |
| 23 Pfaff | simulated k-space | retrospective | Healthy volunteers + any prostate MRI |
| 24 Riederer | NA | prospective | Any prostate MRI |
| 25 Sato | NA | prospective | Healthy volunteers |
| 26 Shen | acquired | prospective | Any prostate MRI |
| 27 Shirashi | NA | retrospective | Any prostate MRI |
| 28 Tong | acquired | retrospective | Diagnostic or AS prostate MRI |
| 29 Ueda | NA | retrospective | Diagnostic prostate MRI |
| 30 Ursprung | simulated k-space | retrospective | Any prostate MRI |
| 31 van Lohuizen | simulated k-like space after FFT | retrospective | Diagnostic prostate MRI |
| 32 Wang | NA | prospective | Diagnostic prostate MRI |
| 33 Zhu | acquired | prospective | Radiotherapy MRI for adaptive treatment |
Table A1.
Cont.
Table A1.
Cont.
| # | # Readers | Acquisition Time Reduction |
|---|---|---|
| 1 Belue | 5 | 100% |
| 2 Bischoff | 2 | 64% |
| 3 Boschheidgen | 2 | 56–72% |
| 4 Cochran | 2 | 32–100% |
| 5 Gassenmaier | 2 | 110% |
| 6 Gassenmaier | 2 | 35% |
| 7 Gassenmaier | 2 | 37% |
| 8 Harder | 4 | 41–100% |
| 9 Hu | 1 | not specified |
| 10 Jeong | 2 | 42–100% |
| 11 Johnson | 4 | 18–27% |
| 12 Jung | 2 | 29–58% |
| 13 Jurka | 3 | 50% |
| 14 Kaye | 2 | 13% |
| 15 Kim | 2 | 100% |
| 16 Kim | 2 | 31% |
| 17 Kim | 2 | 24–31% |
| 18 Lee | 2 | For DWI: 51–100%, For T2w 67–100% |
| 19 Liu | 1 | 100% |
| 20 Nishioka | 2 | 100% |
| 21 Oerther | 2 | 37% |
| 22 Park | 3 | 63% |
| 23 Pfaff | NA | 58% |
| 24 Riederer | 3 | 100% |
| 25 Sato | 2 | 100% |
| 26 Shen | 2 | 47–53% |
| 27 Shirashi | 2 | 100% |
| 28 Tong | 3 | 30–50% |
| 29 Ueda | 2 | 100% |
| 30 Ursprung | 3 | 69% |
| 31 van Lohuizen | NA | 13–25% |
| 32 Wang | 3 | 100% |
| 33 Zhu | NA | 28% |
Table A1.
Cont.
Table A1.
Cont.
| # | Subjective Image Quality Metrics | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Motion | Artifacts | Sharpness | Noise | Lesion Conspicuity | Structural Conspicuity | Diagnostic Confidence | Overall Quality | Others | |
| 1 Belue | X | X | X | ||||||
| 2 Bischoff | X | X | X | X | X | ||||
| 3 Boschheidgen | X | X | X | X | |||||
| 4 Cochran | X | X | X | X | X | ||||
| 5 Gassenmaier | X | X | X | X | X | X | |||
| 6 Gassenmaier | X | X | X | X | X | X | |||
| 7 Gassenmaier | X | X | X | X | X | Natural impression | |||
| 8 Harder | X | X | X | X | X | X | |||
| 9 Hu | X | ||||||||
| 10 Jeong | X | X | X | ||||||
| 11 Johnson | X | X | X | ||||||
| 12 Jung | X | X | X | Perceived SNR | |||||
| 13 Jurka | X | X | X | X | X | Contrast | |||
| 14 Kaye | X | X | X | X | X | ||||
| 15 Kim | X | X | X | ||||||
| 16 Kim | |||||||||
| 17 Kim | X | X | X | X | |||||
| 18 Lee | X | X | |||||||
| 19 Liu | X | ||||||||
| 20 Nishioka | X | X | X | ||||||
| 21 Oerther | X | X | X | ||||||
| 22 Park | X | X | |||||||
| 23 Pfaff | |||||||||
| 24 Riederer | X | X | X | X | X | Subjective SNR | |||
| 25 Sato | X | X | X | X | |||||
| 26 Shen | X | X | X | X | X | ||||
| 27 Shirashi | X | X | X | X | Conrast | ||||
| 28 Tong | X | X | |||||||
| 29 Ueda | X | ||||||||
| 30 Ursprung | X | X | X | X | X | ||||
| 31 van Lohuizen | |||||||||
| 32 Wang | X | X | X | ||||||
| 33 Zhu | |||||||||
Studies reporting a subjective image quality metrics are marked with ’X’.
Table A1.
Cont.
Table A1.
Cont.
| # | Objective Image Quality Metrics | |||||||
|---|---|---|---|---|---|---|---|---|
| SNR | CNR | Intensity Gradient Metric | SSIM | pSNR | RMSE | AI | Others | |
| 1 Belue | X | |||||||
| 2 Bischoff | X | X | X | |||||
| 3 Boschheidgen | X | X | X | |||||
| 4 Cochran | X | |||||||
| 5 Gassenmaier | ||||||||
| 6 Gassenmaier | ||||||||
| 7 Gassenmaier | ||||||||
| 8 Harder | X | X | X | |||||
| 9 Hu | X | X | X | Feature similarity index | ||||
| 10 Jeong | X | X | ||||||
| 11 Johnson | ||||||||
| 12 Jung | X | X | X | |||||
| 13 Jurka | X | X | ||||||
| 14 Kaye | X | X | ||||||
| 15 Kim | X | X | ||||||
| 16 Kim | ||||||||
| 17 Kim | X | X | ||||||
| 18 Lee | X | X | ||||||
| 19 Liu | X | X | X | Perceptual index | ||||
| 20 Nishioka | X | X | ||||||
| 21 Oerther | ||||||||
| 22 Park | X | X | ||||||
| 23 Pfaff | Variance, Gaussian log-likelihood | |||||||
| 24 Riederer | ||||||||
| 25 Sato | X | X | ||||||
| 26 Shen | X | X | X | |||||
| 27 Shirashi | X | X | Contrast | |||||
| 28 Tong | X | |||||||
| 29 Ueda | X | X | ||||||
| 30 Ursprung | X | |||||||
| 31 van Lohuizen | X | X | X | |||||
| 32 Wang | Signal ratio | |||||||
| 33 Zhu | X | X | MAE, edge keeping distance | |||||
Abbreviations: CNR: contrast-to-noise ratio, DWI: diffusion-weighted imaging, MAE: mean absolute error, (p)SNR: (peak) signal-to-noise ratio, RMSE: root mean square error, SSIM: structural similarity index, T2w: T2-weighted. Studies reporting an objective image quality metrics are marked with ’X’.
Appendix C. Study Categorization
Table A2.
Categorization of the included studies.
Table A2.
Categorization of the included studies.
| Data Space | ||
|---|---|---|
| Image | k-Space | k-Space w/ Direct Mapping |
| 1 Belue | 4 Cochran | 2 Bischoff |
| 9 Hu | 15 Kim | 3 Boschheidgen |
| 12 Jung | 18 Lee | 5 Gassenmaier |
| 14 Kaye | 22 Park | 6 Gassenmaier |
| 19 Liu | 24 Riederer | 7 Gassenmaier |
| 23 Pfaff | 32 Wang | 8 Harder |
| 25 Sato | 10 Jeong | |
| 26 Shen | 11 Johnson | |
| 27 Shirashi | 13 Jurka | |
| 29 Ueda | 17 Kim | |
| 31 van Lohuizen | 20 Nishioka | |
| 33 Zhu | 21 Oerther | |
| 28 Tong | ||
| 30 Ursprung | ||
Table A2.
Cont.
Table A2.
Cont.
| Sequence | ||
|---|---|---|
| T2w | DWI | DWI & T2w |
| 1 Belue | 4 Cochran | 11 Johnson |
| 2 Bischoff | 9 Hu | 18 Lee |
| 3 Boschheidgen | 10 Jeong | 21 Oerther |
| 5 Gassenmaier | 14 Kaye | |
| 6 Gassenmaier | 20 Nishioka | |
| 7 Gassenmaier | 23 Pfaff | |
| 8 Harder | 29 Ueda | |
| 12 Jung | 30 Ursprung | |
| 13 Jurka | ||
| 15 Kim | ||
| 16 Kim | ||
| 17 Kim | ||
| 19 Liu | ||
| 22 Park | ||
| 24 Riederer | ||
| 25 Sato | ||
| 26 Shen | ||
| 27 Shirashi | ||
| 28 Tong | ||
| 31 van Lohuizen | ||
| 32 Wang | ||
| 33 Zhu | ||
Table A2.
Cont.
Table A2.
Cont.
| Study Design | ||
|---|---|---|
| Prospective | Retrospective | |
| 2 Bischoff | 1 Belue | |
| 3 Boschheidgen | 4 Cochran | |
| 5 Gassenmaier | 6 Gassenmaier | |
| 7 Gassenmaier | 10 Jeong | |
| 8 Harder | 11 Johnson | |
| 9 Hu | 12 Jung | |
| 12 Jung | 14 Kaye | |
| 13 Jurka | 16 Kim | |
| 15 Kim | 17 Kim | |
| 18 Lee | 20 Nishioka | |
| 21 Oerther | 22 Park | |
| 24 Riederer | 23 Pfaff | |
| 25 Sato | 27 Shirashi | |
| 26 Shen | 28 Tong | |
| 32 Wang | 29 Ueda | |
| 33 Zhu | 30 Ursprung | |
| 31 van Lohuizen | ||
Table A2.
Cont.
Table A2.
Cont.
| Deep Learning Type | ||
|---|---|---|
| In-House Research | Vendor Research | Product |
| 1 Belue | 5 Gassenmaier | 2 Bischoff |
| 9 Hu | 6 Gassenmaier | 3 Boschheidgen |
| 11 Johnson | 7 Gassenmaier | 4 Cochran |
| 12 Jung | 10 Jeong | 8 Harder |
| 14 Kaye | 13 Jurka | 15 Kim |
| 19 Liu | 17 Kim | 18 Lee |
| 23 Pfaff | 27 Shirashi | 20 Nishioka |
| 31 van Lohuizen | 28 Tong | 21 Oerther |
| 33 Zhu | 30 Ursprung | 22 Park |
| 24 Riederer | ||
| 25 Sato | ||
| 26 Shen | ||
| 29 Ueda | ||
| 32 Wang | ||
Table A2.
Cont.
Table A2.
Cont.
| Acceleration | ||
|---|---|---|
| Simulated | Acquired | |
| 4 Cochran | 2 Bischoff | |
| 11 Johnson | 3 Boschheidgen | |
| 12 Jung | 5 Gassenmaier | |
| 14 Kaye | 6 Gassenmaier | |
| 23 Pfaff | 7 Gassenmaier | |
| 30 Ursprung | 8 Harder | |
| 31 van Lohuizen | 10 Jeong | |
| 12 Jung | ||
| 13 Jurka | ||
| 16 Kim | ||
| 17 Kim | ||
| 18 Lee | ||
| 21 Oerther | ||
| 22 Park | ||
| 26 Shen | ||
| 28 Tong | ||
| 33 Zhu | ||
Table A2.
Cont.
Table A2.
Cont.
| Clinical Validation | |||
|---|---|---|---|
| PI-RADS | EPE Detection | Diagnostic Performance | Diagnostic AI Performance |
| 2 Bischoff | 15 Kim | 10 Jeong | 28 Tong |
| 5 Gassenmaier | 16 Kim | 11 Johnson | 31 van Lohuizen |
| 6 Gassenmaier | 22 Park | 29 Ueda | |
| 7 Gassenmaier | |||
| 8 Harder | |||
| 10 Jeong | |||
| 11 Johnson | |||
| 12 Jung | |||
| 16 Kim | |||
| 18 Lee | |||
| 21 Oerther | |||
| 26 Shen | |||
| 27 Shirashi | |||
Full references to the manuscript numbers and first authors can be found in Appendix B. Abbreviations: DWI: diffusion-weighted imaging, EPE: extraprostatic extension, T2w: T2-weighted.
Appendix D. QUADAS-2 Assessment
High risk of bias was assessed in patient selection seven articles as described in the Section 3, with an unclear risk in five articles. Unclear risks of bias or applicability were identified for the index and reference test for 13 and 10 studies, respectively. The most common area of potential bias was the simulation of an accelerated acquisition as it might not produce identical results to truly accelerated acquisitions. There were no concerns about the flow and timing of the studies.
Table A3.
QUADAS-2 analysis. of the risk of bias and applicability of the included studies.
Table A3.
QUADAS-2 analysis. of the risk of bias and applicability of the included studies.
| # | Objective Image Quality Metrics | Index Test | Reference Test | Flow and Timing | ||||
|---|---|---|---|---|---|---|---|---|
| Risk of Bias | Applicability | Risk of Bias | Applicability | Risk of Bias | Applicability | Risk of Bias | Applicability | |
| 1 Belue | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 2 Bischoff | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 3 Boschheidgen | 😐︎ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 4 Cochran | ☺ | ☺ | 😐︎ | 😐︎ | 😐︎ | ☺ | ☺ | ☺ |
| 5 Gassenmaier | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 6 Gassenmaier | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 7 Gassenmaier | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 8 Harder | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 9 Hu | ☹ | ☹ | 😐︎ | ☺ | 😐︎ | ☺ | ☺ | ☺ |
| 10 Jeong | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 11 Johnson | 😐︎ | 😐︎ | 😐︎ | 😐︎ | ☺ | ☺ | ☺ | ☺ |
| 12 Jung | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 13 Jurka | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 14 Kaye | 😐︎ | 😐︎ | 😐︎ | 😐︎ | ☺ | ☺ | ☺ | ☺ |
| 15 Kim | 😐︎ | ☹ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 16 Kim | 😐︎ | ☹ | 😐︎ | 😐︎ | 😐︎ | ☺ | ☺ | ☺ |
| 17 Kim | ☺ | ☺ | ☹ | ☺ | ☹ | ☺ | ☺ | ☺ |
| 18 Lee | 😐︎ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 19 Liu | ☺ | ☺ | 😐︎ | 😐︎ | 😐︎ | 😐︎ | ☺ | ☺ |
| 20 Nishioka | ☹ | ☹ | 😐︎ | 😐︎ | 😐︎ | 😐︎ | ☺ | ☺ |
| 21 Oerther | ☺ | ☺ | ☺ | ☺ | 😐︎ | ☺ | ☺ | ☺ |
| 22 Park | 😐︎ | ☹ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 23 Pfaff | 😐︎ | 😐︎ | 😐︎ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 24 Riederer | ☺ | ☺ | 😐︎ | ☺ | 😐︎ | ☺ | ☺ | ☺ |
| 25 Sato | ☺ | ☹ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 26 Shen | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 27 Shirashi | ☺ | ☺ | 😐︎ | ☺ | 😐︎ | ☺ | ☺ | ☺ |
| 28 Tong | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
| 29 Ueda | ☹ | ☹ | 😐︎ | ☺ | 😐︎ | ☺ | ☺ | ☺ |
| 30 Ursprung | ☺ | ☺ | 😐︎ | 😐︎ | ☺ | ☺ | ☺ | ☺ |
| 31 van Lohuizen | ☺ | ☺ | 😐︎ | 😐︎ | ☺ | ☺ | ☺ | ☺ |
| 32 Wang | ☺ | ☺ | ☺ | ☺ | ☺ | 😐︎ | ☺ | ☺ |
| 33 Zhu | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ | ☺ |
☺ on green background: low risk of bias or applicability. 😐︎ on orange background: unclear risk of bias or applicability. ☹ on red background: high risk of bias or applicability.
Appendix E. Statistical Methods
Supplementary statistical methods complementing the main methods: To quantify the heterogeneity between the studies in the meta-analysis, the I2 statistic was calculated. I2 is an estimate of the percentage of the variability in effect estimators that is attributable to heterogeneity between studies and cannot be explained by chance arising from sampling error. I2 values of 25% and less are usually considered to be low or unimportant, 25–50% moderate and values above 70% are considered high (Higgins JPT, Thompson SG (2002) Quantifying heterogeneity in a meta-analysis.). The computation of inter-study dispersion assumes that, effect sizes would follow a chi-squared distribution if variation was only due to random selection of participants and studies were methodologically identical (Higgins J, Thomas J, Chandler J, et al Cochrane Handbook for Systematic Reviews of Interventions version 6.0.).
A funnel plot was employed as a tool to explore the presence of publication bias among the studies included in the meta-analysis. The standardized mean difference of each study is plotted against its standard error, which is inversely related to sample size. An asymmetric distribution of studies on the funnel plot can indicate publication bias as a result of preferential reporting of significant studies. The dotted diagonal lines border the sector where 95% of all studies were assumed to lie (Sterne JAC, Egger M (2001) Funnel plots for detecting bias in meta-analysis: Guidelines on choice of axis).
Appendix F. Supplementary Data Meta-Analysis
Figure A2.
Forest plot of all studies investigating subjective image quality of DLR for T2w sequences, including studies where the effect size had to be approximated from the median and interquartile range.
Figure A3.
Forest plot of all studies investigating subjective image quality of DLR for DWI sequences, including studies where the effect size had to be approximated from the median and interquartile range.
References
- Hofmann, B.; Haug, E.S.; Andersen, E.R.; Kjelle, E. Increased magnetic resonance imaging in prostate cancer management-What are the outcomes? J. Eval. Clin. Pract. 2023, 29, 893–902. [Google Scholar] [CrossRef] [PubMed]
- Hofmann, B.; Øfsti Brandsaeter, I.; Kjelle, E. Variations in wait times for imaging services: A register-based study of self-reported wait times for specific examinations in Norway. BMC Health Serv. Res. 2023, 23, 1287. [Google Scholar] [CrossRef] [PubMed]
- Hugosson, J.; Månsson, M.; Wallström, J.; Axcrona, U.; Carlsson, S.V.; Egevad, L.; Geterud, K.; Khatami, A.; Kohestani, K.; Pihl, C.G.; et al. Prostate Cancer Screening with PSA and MRI Followed by Targeted Biopsy Only. N. Engl. J. Med. 2022, 387, 2126–2137. [Google Scholar] [CrossRef] [PubMed]
- Hugosson, J.; Godtman, R.A.; Wallstrom, J.; Axcrona, U.; Bergh, A.; Egevad, L.; Geterud, K.; Khatami, A.; Socratous, A.; Spyratou, V.; et al. Results after Four Years of Screening for Prostate Cancer with PSA and MRI. N. Engl. J. Med. 2024, 391, 1083–1095. [Google Scholar] [CrossRef]
- Nordström, T.; Annerstedt, M.; Glaessgen, A.; Carlsson, S.; Clements, M.; Abbadi, A.; Grönberg, H.; Jäderling, F.; Eklund, M.; Discacciati, A. Repeated Prostate Cancer Screening Using Prostate-Specific Antigen Testing and Magnetic Resonance Imaging: A Secondary Analysis of the STHLM3-MRI Randomized Clinical Trial. JAMA Netw. Open 2024, 7, E2354577. [Google Scholar] [CrossRef]
- Möller, F.; Månsson, M.; Wallström, J.; Hellström, M.; Hugosson, J.; Godtman, R.A. Prostate Cancers in the Prostate-specific Antigen Interval of 1.8–3 ng/ml: Results from the Göteborg-2 Prostate Cancer Screening Trial. Eur. Urol. 2024, 86, 95–100. [Google Scholar] [CrossRef]
- Farha, M.W.; Salami, S.S. Biomarkers for prostate cancer detection and risk stratification. Ther. Adv. Urol. 2022, 14, 1–14. [Google Scholar] [CrossRef]
- Padhani, A.R.; Schoots, I.G.; Barentsz, J.O. Fast Magnetic Resonance Imaging as a Viable Method for Directing the Prostate Cancer Diagnostic Pathway. Eur. Urol. Oncol. 2021, 4, 863–865. [Google Scholar] [CrossRef]
- Schoots, I.G.; Haider, M.A.; Punwani, S.; Padhani, A.R. MRI in Prostate Cancer Screening: A Review and Recommendations, From the AJR Special Series on Screening. AJR Am. J. Roentgenol. 2025. [Google Scholar] [CrossRef]
- Agrotis, G.; Ursprung, S.; Schoots, I. Personalized Prostate MRI: Tailoring Contrast Medium Use in an Era of Increasing Demand. Eur. Radiol. 2025. [Google Scholar] [CrossRef]
- Brix, M.A.; Järvinen, J.; Bode, M.K.; Nevalainen, M.; Nikki, M.; Niinimäki, J.; Lammentausta, E. Financial impact of incorporating deep learning reconstruction into magnetic resonance imaging routine. Eur. J. Radiol. 2024, 175, 111434. [Google Scholar] [CrossRef] [PubMed]
- Hamilton, J.; Franson, D.; Seiberlich, N. Recent advances in parallel imaging for MRI. Prog. Nucl. Magn. Reson. Spectrosc. 2017, 101, 71–95. [Google Scholar] [CrossRef] [PubMed]
- Cui, L.; Song, Y.; Wang, Y.; Wang, R.; Wu, D.; Xie, H.; Li, J.; Yang, G. Motion artifact reduction for magnetic resonance imaging with deep learning and k-space analysis. PLoS ONE 2023, 18, e0278668. [Google Scholar] [CrossRef]
- Kiryu, S.; Akai, H.; Yasaka, K.; Tajima, T.; Kunimatsu, A.; Yoshioka, N.; Akahane, M.; Abe, O.; Ohtomo, K. Clinical Impact of Deep Learning Reconstruction in MRI. Radiographics 2023, 43, e220133. [Google Scholar] [CrossRef]
- Ravishankar, S.; Ye, J.C.; Fessler, J.A. Image Reconstruction: From Sparsity to Data-adaptive Methods and Machine Learning. Proc. IEEE. Inst. Electr. Electron. Eng. 2020, 108, 86–109. [Google Scholar] [CrossRef]
- Page, M.J.; McKenzie, J.E.; Bossuyt, P.M.; Boutron, I.; Hoffmann, T.C.; Mulrow, C.D.; Shamseer, L.; Tetzlaff, J.M.; Akl, E.A.; Brennan, S.E.; et al. The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. Syst. Rev. 2021, 4, 89. [Google Scholar] [CrossRef]
- McInnes, M.; Moher, D.; Thombs, B.; McGrath, T.; Bossuyt, P.; Clifford, T.; Cohen, J.; Deeks, J.; Gatsonis, C.; Hooft, L.; et al. Preferred Reporting Items for a Systematic Review and Meta-analysis of Diagnostic Test Accuracy Studies. JAMA 2018, 10, 388–396. [Google Scholar] [CrossRef]
- Wan, X.; Wang, W.; Liu, J.; Tong, T. Estimating the sample mean and standard deviation from the sample size, median, range and/or interquartile range. BMC Med. Res. Methodol. 2014, 14, 135. [Google Scholar] [CrossRef]
- Viechtbauer, W. Conducting Meta-Analyses in R with the metafor Package. J. Stat. Softw. 2010, 91, 121108. [Google Scholar] [CrossRef]
- Belue, M.J.; Harmon, S.A.; Masoudi, S.; Barrett, T.; Law, Y.M.; Purysko, A.S.; Panebianco, V.; Yilmaz, E.C.; Lin, Y.; Jadda, P.K.; et al. Quality of T2-weighted MRI re-acquisition versus deep learning GAN image reconstruction: A multi-reader study. Eur. J. Radiol. 2024, 170, 111259. [Google Scholar] [CrossRef]
- Bischoff, L.M.; Peeters, J.M.; Weinhold, L.; Krausewitz, P.; Ellinger, J.; Katemann, C.; Isaak, A.; Weber, O.M.; Kuetting, D.; Attenberger, U.; et al. Deep Learning Super-Resolution Reconstruction for Fast and Motion-Robust T2-weighted Prostate MRI. Radiology 2023, 308, e230427. [Google Scholar] [CrossRef] [PubMed]
- Boschheidgen, M.; Drewes, L.; Valentin, B.; Ullrich, T.; Trappe, S.; Al-Monajjed, R.; Radtke, J.P.; Albers, P.; Wittsack, H.J.; Antoch, G.; et al. Use of deep learning-accelerated T2 TSE for prostate MRI: Comparison with and without hyoscine butylbromide admission. Magn. Reson. Imaging 2025, 118, 110358. [Google Scholar] [CrossRef] [PubMed]
- Cochran, R.L.; Milshteyn, E.; Ghosh, S.; Nakrour, N.; Mercaldo, N.D.; Guidon, A.; Harisinghani, M.G. Minimizing prostate diffusion weighted MRI examination time through deep learning reconstruction. Clin. Imaging 2025, 117, 110341. [Google Scholar] [CrossRef] [PubMed]
- Gassenmaier, S.; Afat, S.; Nickel, M.D.; Mostapha, M.; Herrmann, J.; Almansour, H.; Nikolaou, K.; Othman, A.E. Accelerated t2-weighted tse imaging of the prostate using deep learning image reconstruction: A prospective comparison with standard t2-weighted tse imaging. Cancers 2021, 13, 3593. [Google Scholar] [CrossRef]
- Gassenmaier, S.; Afat, S.; Nickel, D.; Mostapha, M.; Herrmann, J.; Othman, A.E. Deep learning–accelerated T2-weighted imaging of the prostate: Reduction of acquisition time and improvement of image quality. Eur. J. Radiol. 2021, 137, 109600. [Google Scholar] [CrossRef]
- Gassenmaier, S.; Warm, V.; Nickel, D.; Weiland, E.; Herrmann, J.; Almansour, H.; Wessling, D.; Afat, S. Thin-Slice Prostate MRI Enabled by Deep Learning Image Reconstruction. Cancers 2023, 15, 578. [Google Scholar] [CrossRef]
- Harder, F.N.; Weiss, K.; Amiel, T.; Peeters, J.M.; Tauber, R.; Ziegelmayer, S.; Burian, E.; Makowski, M.R.; Sauter, A.P.; Gschwend, J.E.; et al. Prospectively Accelerated T2-Weighted Imaging of the Prostate by Combining Compressed SENSE and Deep Learning in Patients with Histologically Proven Prostate Cancer. Cancers 2022, 14, 5741. [Google Scholar] [CrossRef]
- Hu, L.; Zhou, D.W.; Fu, C.X.; Benkert, T.; Xiao, Y.F.; Wei, L.M.; Zhao, J.G. Calculation of Apparent Diffusion Coefficients in Prostate Cancer Using Deep Learning Algorithms: A Pilot Study. Front. Oncol. 2021, 11, 697721. [Google Scholar] [CrossRef]
- Jeong, J.; Yeom, S.K.; Choi, I.Y.; Cha, S.H.; Han, J.S.; Lee, C.H.; Sung, D.J.; Choi, J.W.; Bae, J.H.; Choi, H.; et al. Deep learning image reconstruction of diffusion-weighted imaging in evaluation of prostate cancer focusing on its clinical implications. Quant. Imaging Med. Surg. 2024, 14, 3432–3446. [Google Scholar] [CrossRef]
- Johnson, P.M.; Tong, A.; Donthireddy, A.; Melamud, K.; Petrocelli, R.; Smereka, P.; Qian, K.; Keerthivasan, M.B.; Chandarana, H.; Knoll, F. Deep Learning Reconstruction Enables Highly Accelerated Biparametric MR Imaging of the Prostate. J. Magn. Reson. Imaging 2022, 56, 184–195. [Google Scholar] [CrossRef]
- Jung, W.; Kim, E.H.; Ko, J.; Jeong, G.; Choi, M.H. Convolutional neural network-based reconstruction for acceleration of prostate T2 weighted MR imaging: A retro-and prospective study. Br. J. Radiol. 2022, 95, 20211378. [Google Scholar] [CrossRef] [PubMed]
- Jurka, M.; Macova, I.; Wagnerova, M.; Capoun, O.; Jakubicek, R.; Ourednicek, P.; Lambert, L.; Burgetova, A. Deep-learning-based reconstruction of T2-weighted magnetic resonance imaging of the prostate accelerated by compressed sensing provides improved image quality at half the acquisition time. Quant. Imaging Med. Surg. 2024, 14, 3534–3543. [Google Scholar] [CrossRef] [PubMed]
- Kaye, E.A.; Aherne, E.A.; Duzgol, C.; Häggström, I.; Kobler, E.; Mazaheri, Y.; Fung, M.M.; Zhang, Z.; Otazo, R.; Vargas, H.A.; et al. Accelerating prostate diffusion-weighted MRI using a guided denoising convolutional neural network: Retrospective feasibility study. Radiol. Artif. Intell. 2020, 2, 1–9. [Google Scholar] [CrossRef] [PubMed]
- Kim, E.H.; Choi, M.H.; Lee, Y.J.; Han, D.; Mostapha, M.; Nickel, D. Deep learning-accelerated T2-weighted imaging of the prostate: Impact of further acceleration with lower spatial resolution on image quality. Eur. J. Radiol. 2021, 145, 110012. [Google Scholar] [CrossRef]
- Kim, D.H.; Choi, M.H.; Lee, Y.J.; Rha, S.E.; Nickel, M.D.; Lee, H.S.; Han, D. Deep learning-accelerated T2WI of the prostate for transition zone lesion evaluation and extraprostatic extension assessment. Sci. Rep. 2024, 14, 29249. [Google Scholar] [CrossRef]
- Kim, M.; Kim, S.H.; Hong, S.; Kim, Y.J.; Kim, H.R.; Kim, J.Y. Evaluation of Extra-Prostatic Extension on Deep Learning-Reconstructed High-Resolution Thin-Slice T2-Weighted Images in Patients with Prostate Cancer. Cancers 2024, 16, 413. [Google Scholar] [CrossRef]
- Lee, K.L.; Kessler, D.A.; Dezonie, S.; Chishaya, W.; Shepherd, C.; Carmo, B.; Graves, M.J.; Barrett, T. Assessment of deep learning-based reconstruction on T2-weighted and diffusion-weighted prostate MRI image quality. Eur. J. Radiol. 2023, 166, 111017. [Google Scholar] [CrossRef]
- Liu, Y.; Liu, Y.; Vanguri, R.; Litwiller, D.; Liu, M.; Hsu, H.Y.; Ha, R.; Shaish, H.; Jambawalikar, S. 3D Isotropic Super-resolution Prostate MRI Using Generative Adversarial Networks and Unpaired Multiplane Slices. J. Digit. Imaging 2021, 34, 1199–1208. [Google Scholar] [CrossRef]
- Nishioka, N.; Fujima, N.; Tsuneta, S.; Yoshikawa, M.; Kimura, R.; Sakamoto, K.; Kato, F.; Miyata, H.; Kikuchi, H.; Matsumoto, R.; et al. Enhancing the image quality of prostate diffusion-weighted imaging in patients with prostate cancer through model-based deep learning reconstruction. Eur. J. Radiol. Open 2024, 13, 100588. [Google Scholar] [CrossRef]
- Oerther, B.; Engel, H.; Nedelcu, A.; Strecker, R.; Benkert, T.; Nickel, D.; Weiland, E.; Mayrhofer, T.; Bamberg, F.; Benndorf, M.; et al. Performance of an ultra-fast deep-learning accelerated MRI screening protocol for prostate cancer compared to a standard multiparametric protocol. Eur. Radiol. 2024, 34, 7053–7062. [Google Scholar] [CrossRef]
- Park, J.C.; Park, K.J.; Park, M.Y.; hyun Kim, M.; Kim, J.K. Fast T2-Weighted Imaging With Deep Learning-Based Reconstruction: Evaluation of Image Quality and Diagnostic Performance in Patients Undergoing Radical Prostatectomy. J. Magn. Reson. Imaging 2022, 55, 1735–1744. [Google Scholar] [CrossRef] [PubMed]
- Pfaff, L.; Darwish, O.; Wagner, F.; Thies, M.; Vysotskaya, N.; Hossbach, J.; Weiland, E.; Benkert, T.; Eichner, C.; Nickel, D.; et al. Enhancing diffusion-weighted prostate MRI through self-supervised denoising and evaluation. Sci. Rep. 2024, 14, 24292. [Google Scholar] [CrossRef] [PubMed]
- Riederer, S.J.; Borisch, E.A.; Froemming, A.T.; Kawashima, A.; Takahashi, N. Comparison of model-based versus deep learning-based image reconstruction for thin-slice T2-weighted spin-echo prostate MRI. Abdom. Radiol. 2024, 49, 2921–2931. [Google Scholar] [CrossRef] [PubMed]
- Sato, Y.; Ohkuma, K. Verification of image quality improvement by deep learning reconstruction to 1.5 T MRI in T2-weighted images of the prostate gland. Radiol. Phys. Technol. 2024, 17, 756–764. [Google Scholar] [CrossRef]
- Shen, L.; Xu, H.; Liao, Q.; Yuan, Y.; Yu, D.; Wei, J.; Yang, Z.; Wang, L. A Feasibility Study of AI-Assisted Compressed Sensing in Prostate T2-Weighted Imaging. Acad. Radiol. 2024, 31, 5022–5033. [Google Scholar] [CrossRef]
- Shiraishi, K.; Nakaura, T.; Kobayashi, N.; Uetani, H.; Nagayama, Y.; Kidoh, M.; Yatsuda, J.; Kurahashi, R.; Kamba, T.; Yamahita, Y.; et al. Enhancing thin slice 3D T2-weighted prostate MRI with super-resolution deep learning reconstruction: Impact on image quality and PI-RADS assessment. Magn. Reson. Imaging 2025, 117, 110308. [Google Scholar] [CrossRef]
- Tong, A.; Bagga, B.; Petrocelli, R.; Smereka, P.; Vij, A.; Qian, K.; Grimm, R.; Kamen, A.; Keerthivasan, M.B.; Nickel, M.D.; et al. Comparison of a Deep Learning-Accelerated vs. Conventional T2-Weighted Sequence in Biparametric MRI of the Prostate. J. Magn. Reson. Imaging 2023, 58, 1055–1064. [Google Scholar] [CrossRef]
- Ueda, T.; Ohno, Y.; Yamamoto, K.; Murayama, K.; Ikedo, M.; Yui, M.; Hanamatsu, S.; Tanaka, Y.; Obama, Y.; Ikeda, H.; et al. Deep Learning Reconstruction of Diffusion-weighted MRI Improves Image Quality for Prostatic Imaging. Radiology 2022, 303, 373–381. [Google Scholar] [CrossRef]
- Ursprung, S.; Herrmann, J.; Joos, N.; Weiland, E.; Benkert, T.; Almansour, H.; Lingg, A.; Afat, S.; Gassenmaier, S. Accelerated diffusion-weighted imaging of the prostate using deep learning image reconstruction: A retrospective comparison with standard diffusion-weighted imaging. Eur. J. Radiol. 2023, 165, 110953. [Google Scholar] [CrossRef]
- van Lohuizen, Q.; Roest, C.; Simonis, F.F.; Fransen, S.J.; Kwee, T.C.; Yakar, D.; Huisman, H. Assessing deep learning reconstruction for faster prostate MRI: Visual vs. diagnostic performance metrics. Eur. Radiol. 2024, 34, 7364–7372. [Google Scholar] [CrossRef]
- Wang, X.; Ma, J.; Bhosale, P.; Rovira, J.J.I.; Qayyum, A.; Sun, J.; Bayram, E.; Szklaruk, J. Novel deep learning-based noise reduction technique for prostate magnetic resonance imaging. Abdom. Radiol. 2021, 46, 3378–3386. [Google Scholar] [CrossRef] [PubMed]
- Zhu, J.; Chen, X.; Liu, Y.; Yang, B.; Wei, R.; Qin, S.; Yang, Z.; Hu, Z.; Dai, J.; Men, K. Improving accelerated 3D imaging in MRI-guided radiotherapy for prostate cancer using a deep learning method. Radiat. Oncol. 2023, 18, 108. [Google Scholar] [CrossRef] [PubMed]
- Tibrewala, R.; Dutt, T.; Tong, A.; Ginocchio, L.; Lattanzi, R.; Keerthivasan, M.B.; Baete, S.H.; Chopra, S.; Lui, Y.W.; Sodickson, D.K.; et al. FastMRI Prostate: A public, biparametric MRI dataset to advance machine learning for prostate cancer imaging. Sci. Data 2024, 11, 404. [Google Scholar] [CrossRef]
- Health, N.L. fastMRI Dataset. Available online: https://fastmri.med.nyu.edu/ (accessed on 26 May 2025).
- Boss, M.A.; Malyarenko, D.; Partridge, S.; Obuchowski, N.; Shukla-Dave, A.; Winfield, J.M.; Fuller, C.D.; Miller, K.; Mishra, V.; Ohliger, M.; et al. The QIBA Profile for Diffusion-Weighted MRI: Apparent Diffusion Coefficient as a Quantitative Imaging Biomarker. Radiology 2024, 313, e233055. [Google Scholar] [CrossRef]
- Edupuganti, V.; Mardani, M.; Cheng, J.; Vasanawala, S.; Pauly, J. Exploring the Hallucination Risk of Deep Generative Models in MR Image Recovery. In Proceedings of the 27th ISMRM Proceedings, ISMRM, Montreal, QC, Canada, 11–16 May 2019. [Google Scholar]
- Antun, V.; Renna, F.; Poon, C.; Adcock, B.; Hansen, A.C. On instabilities of deep learning in image reconstruction and the potential costs of AI. Proc. Natl. Acad. Sci. USA 2020, 117, 30088–30095. [Google Scholar] [CrossRef]
- Gottschling, N.M.; Antun, V.; Hansen, A.C.; Adcock, B. The Troublesome Kernel: On Hallucinations, No Free Lunches, and the Accuracy-Stability Tradeoff in Inverse Problems. SIAM Rev. 2025, 67, 73–104. [Google Scholar] [CrossRef]
- Kastryulin, S.; Zakirov, J.; Pezzotti, N.; Dylov, D.V. Image Quality Assessment for Magnetic Resonance Imaging. IEEE Access 2023, 11, 14154–14168. [Google Scholar] [CrossRef]
- Dohmen, M.; Klemens, M.A.; Baltruschat, I.M.; Truong, T.; Lenga, M. Similarity and quality metrics for MR image-to-image translation. Sci. Rep. 2025, 15, 3853. [Google Scholar] [CrossRef]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).