Assessing deep learning reconstruction for faster prostate MRI: visual vs. diagnostic performance metrics.

OBJECTIVE: Deep learning (DL) MRI reconstruction enables fast scan acquisition with good visual quality, but the diagnostic impact is often not assessed because of large reader study requirements. This study used existing diagnostic DL to assess the diagnostic quality of reconstructed images. MATERIALS AND METHODS: A retrospective multisite study of 1535 patients assessed biparametric prostate MRI between 2016 and 2020. Likely clinically significant prostate cancer (csPCa) lesions (PI-RADS ≥ 4) were delineated by expert radiologists. T2-weighted scans were retrospectively undersampled, simulating accelerated protocols. DL reconstruction (DLRecon) and diagnostic DL detection (DLDetect) were developed. The effect on the partial area under (pAUC), the Free-Response Operating Characteristic (FROC) curve, and the structural similarity (SSIM) were compared as metrics for diagnostic and visual quality, respectively. DLDetect was validated with a reader concordance analysis. Statistical analysis included Wilcoxon, permutation, and Cohen's kappa tests for visual quality, diagnostic performance, and reader concordance. RESULTS: DLRecon improved visual quality at 4- and 8-fold (R4, R8) subsampling rates, with SSIM (range: -1 to 1) improved to 0.78 ± 0.02 (p < 0.001) and 0.67 ± 0.03 (p < 0.001) from 0.68 ± 0.03 and 0.51 ± 0.03, respectively. However, diagnostic performance at R4 showed a pAUC FROC of 1.33 (CI 1.28-1.39) for DL and 1.29 (CI 1.23-1.35) for naive reconstructions, both significantly lower than fully sampled pAUC of 1.58 (DL: p = 0.024, naïve: p = 0.02). Similar trends were noted for R8. CONCLUSION: DL reconstruction produces visually appealing images but may reduce diagnostic accuracy. Incorporating diagnostic AI into the assessment framework offers a clinically relevant metric essential for adopting reconstruction models into clinical practice. CLINICAL RELEVANCE STATEMENT: In clinical settings, caution is warranted when using DL reconstruction for MRI scans. While it recovered visual quality, it failed to match the prostate cancer detection rates observed in scans not subjected to acceleration and DL reconstruction.

Using deep learning to optimize the prostate MRI protocol by assessing the diagnostic efficacy of MRI sequences.

PURPOSE: To explore diagnostic deep learning for optimizing the prostate MRI protocol by assessing the diagnostic efficacy of MRI sequences. METHOD: This retrospective study included 840 patients with a biparametric prostate MRI scan. The MRI protocol included a T2-weighted image, three DWI sequences (b50, b400, and b800 s/mm2), a calculated ADC map, and a calculated b1400 sequence. Two accelerated MRI protocols were simulated, using only two acquired b-values to calculate the ADC and b1400. Deep learning models were trained to detect prostate cancer lesions on accelerated and full protocols. The diagnostic performances of the protocols were compared on the patient-level with the area under the receiver operating characteristic (AUROC), using DeLong's test, and on the lesion-level with the partial area under the free response operating characteristic (pAUFROC), using a permutation test. Validation of the results was performed among expert radiologists. RESULTS: No significant differences in diagnostic performance were found between the accelerated protocols and the full bpMRI baseline. Omitting b800 reduced 53% DWI scan time, with a performance difference of + 0.01 AUROC (p = 0.20) and -0.03 pAUFROC (p = 0.45). Omitting b400 reduced 32% DWI scan time, with a performance difference of -0.01 AUROC (p = 0.65) and + 0.01 pAUFROC (p = 0.73). Multiple expert radiologists underlined the findings. CONCLUSIONS: This study shows that deep learning can assess the diagnostic efficacy of MRI sequences by comparing prostate MRI protocols on diagnostic accuracy. Omitting either the b400 or the b800 DWI sequence can optimize the prostate MRI protocol by reducing scan time without compromising diagnostic quality.

Recent Developments in Speeding up Prostate MRI.

The increasing incidence of prostate cancer cases worldwide has led to a tremendous demand for multiparametric MRI (mpMRI). In order to relieve the pressure on healthcare, reducing mpMRI scan time is necessary. This review focuses on recent techniques proposed for faster mpMRI acquisition, specifically shortening T2W and DWI sequences while adhering to the PI-RADS (Prostate Imaging Reporting and Data System) guidelines. Speeding up techniques in the reviewed studies rely on more efficient sampling of data, ranging from the acquisition of fewer averages or b-values to adjustment of the pulse sequence. Novel acquisition methods based on undersampling techniques are often followed by suitable reconstruction methods typically incorporating synthetic priori information. These reconstruction methods often use artificial intelligence for various tasks such as denoising, artifact correction, improvement of image quality, and in the case of DWI, for the generation of synthetic high b-value images or apparent diffusion coefficient maps. Reduction of mpMRI scan time is possible, but it is crucial to maintain diagnostic quality, confirmed through radiological evaluation, to integrate the proposed methods into the standard mpMRI protocol. Additionally, before clinical integration, prospective studies are recommended to validate undersampling techniques to avoid potentially inaccurate results demonstrated by retrospective analysis. This review provides an overview of recently proposed techniques, discussing their implementation, advantages, disadvantages, and diagnostic performance according to PI-RADS guidelines compared to conventional methods. LEVEL OF EVIDENCE: 3 TECHNICAL EFFICACY: Stage 3.

Virtual Resection: A New Tool for Preparing for Nephron-Sparing Surgery in Wilms Tumor Patients.

Nephron-sparing surgery (NSS) in Wilms tumor (WT) patients is a surgically challenging procedure used in highly selective cases only. Virtual resections can be used for preoperative planning of NSS to estimate the remnant renal volume (RRV) and to virtually mimic radical tumor resection. In this single-center evaluation study, virtual resection for NSS planning and the user experience were evaluated. Virtual resection was performed in nine WT patient cases by two pediatric surgeons and one pediatric urologist. Pre- and postoperative MRI scans were used for 3D visualization. The virtual RRV was acquired after performing virtual resection and a questionnaire was used to assess the ease of use. The actual RRV was derived from the postoperative 3D visualization and compared with the derived virtual RRV. Virtual resection resulted in virtual RRVs that matched nearly perfectly with the actual RRVs. According to the questionnaire, virtual resection appeared to be straightforward and was not considered to be difficult. This study demonstrated the potential of virtual resection as a new planning tool to estimate the RRV after NSS in WT patients. Future research should further evaluate the clinical relevance of virtual resection by relating it to surgical outcome.

Breast MRI in patients after breast conserving surgery with sentinel node procedure using a superparamagnetic tracer.

BACKGROUND: A procedure for sentinel lymph node biopsy (SLNB) using superparamagnetic iron-oxide (SPIO) nanoparticles and intraoperative sentinel lymph node (SLN) detection was developed to overcome drawbacks associated with the current standard-of-care SLNB. However, residual SPIO nanoparticles can result in void artefacts at follow-up magnetic resonance imaging (MRI) scans. We present a grading protocol to quantitatively assess the severity of these artefacts and offer an option to minimise the impact of SPIO nanoparticles on diagnostic imaging. METHODS: Follow-up mammography and MRI of two patient groups after a magnetic SLNB were included in the study. They received a 2-mL subareolar dose of SPIO (high-dose, HD) or a 0.1-mL intratumoural dose of SPIO (low-dose, LD). Follow-up mammography and MRI after magnetic SLNB were acquired within 4 years after breast conserving surgery (BCS). Two radiologists with over 10-year experience in breast imaging assessed the images and analysed the void artefacts and their impact on diagnostic follow-up. RESULTS: A total of 19 patients were included (HD, n = 13; LD, n = 6). In the HD group, 9/13 patients displayed an artefact on T1-weighted images up to 3.6 years after the procedure, while no impact of the SPIO remnants was observed in the LD group. CONCLUSIONS: SLNB using a 2-mL subareolar dose of magnetic tracer in patients undergoing BCS resulted in residual artefacts in the breast in the majority of patients, which may hamper follow-up MRI. This can be avoided by using a 0.1-mL intratumoural dose.

Comparing the signal enhancement of a gadolinium based and an iron-oxide based contrast agent in low-field MRI.

Recently, there has been a renewed interest in low-field MRI. Contrast agents (CA) in MRI have magnetic behavior dependent on magnetic field strength. Therefore, the optimal contrast agent for low-field MRI might be different from what is used at higher fields. Ultra-small superparamagnetic iron-oxides (USPIOs), commonly used as negative CA, might also be used for generating positive contrast in low-field MRI. The purpose of this study was to determine whether an USPIO or a gadolinium based contrast agent is more appropriate at low field strengths. Relaxivity values of ferumoxytol (USPIO) and gadoterate (gadolinium based) were used in this research to simulate normalized signal intensity (SI) curves within a concentration range of 0-15 mM. Simulations were experimentally validated on a 0.25T MRI scanner. Simulations and experiments were performed using spin echo (SE), spoiled gradient echo (SGE), and balanced steady-state free precession (bSSFP) sequences. Maximum achievable SIs were assessed for both CAs in a range of concentrations on all sequences. Simulations at 0.25T showed a peak in SIs at low concentrations ferumoxytol versus a wide top at higher concentrations for gadoterate in SE and SGE. Experiments agreed well with the simulations in SE and SGE, but less in the bSSFP sequence due to overestimated relaxivities in simulations. At low magnetic field strengths, ferumoxytol generates similar signal enhancement at lower concentrations than gadoterate.

The diagnostic potential of low-field MRI in problematic total knee arthroplasties - a feasibility study.

PURPOSE: Low-field MRI, allowing imaging in supine and weight-bearing position, may be utilized as a non-invasive and affordable tool to differentiate between causes of dissatisfaction after TKA ('problematic TKA'). However, it remains unclear whether low-field MRI results in sufficient image quality with limited metal artefacts. Therefore, this feasibility study explored the diagnostic value of low-field MRI concerning pathologies associated with problematic TKA's' by comparing low-field MRI findings with CT and surgical findings. Secondly, differences in patellofemoral parameters between supine and weight-bearing low-field MRI were evaluated. METHODS: Eight patients with a problematic TKA were scanned using low-field MRI in weight-bearing and supine conditions. Six of these patients underwent revision surgery. Scans were analysed by a radiologist for pathologies associated with a problematic TKA. Additional patellofemoral and alignment parameters were measured by an imaging expert. MRI observations were compared to those obtained with CT, the diagnosis based on the clinical work-up, and findings during revision surgery. RESULTS: MRI observations of rotational malalignment, component loosening and patellofemoral arthrosis were comparable with the clinical diagnosis (six out of eight) and were confirmed during surgery (four out of six). All MRI observations were in line with CT findings (seven out of seven). Clinical diagnosis and surgical findings of collateral excessive laxity could not be confirmed with MRI (two out of eight). CONCLUSION: Low-field MRI shows comparable diagnostic value as CT and might be a future low cost and ionizing radiation free alternative. Differences between supine and weight-bearing MRI did not yield clinically relevant information. The study was approved by the Medical Research Ethics Committees of Twente (Netherlands Trial Register: Trial NL7009 (NTR7207). Registered 5 March 2018, https://www.trialregister.nl/trial/7009 ).

Improved repeatability of dynamic contrast-enhanced MRI using the complex MRI signal to derive arterial input functions: a test-retest study in prostate cancer patients.

PURPOSE: The arterial input function (AIF) is a major source of uncertainty in tracer kinetic (TK) analysis of dynamic contrast-enhanced (DCE)-MRI data. The aim of this study was to investigate the repeatability of AIFs extracted from the complex signal and of the resulting TK parameters in prostate cancer patients. METHODS: Twenty-two patients with biopsy-proven prostate cancer underwent a 3T MRI exam twice. DCE-MRI data were acquired with a 3D spoiled gradient echo sequence. AIFs were extracted from the magnitude of the signal (AIFMAGN ), phase (AIFPHASE ), and complex signal (AIFCOMPLEX ). The Tofts model was applied to extract Ktrans , kep and ve . Repeatability of AIF curve characteristics and TK parameters was assessed with the within-subject coefficient of variation (wCV). RESULTS: The wCV for peak height and full width at half maximum for AIFCOMPLEX (7% and 8%) indicated an improved repeatability compared to AIFMAGN (12% and 12%) and AIFPHASE (12% and 7%). This translated in lower wCV values for Ktrans (11%) with AIFCOMPLEX in comparison to AIFMAGN (24%) and AIFPHASE (15%). For kep , the wCV was 16% with AIFMAGN , 13% with AIFPHASE , and 13% with AIFCOMPLEX . CONCLUSION: Repeatability of AIFPHASE and AIFCOMPLEX is higher than for AIFMAGN , resulting in a better repeatability of TK parameters. Thus, use of either AIFPHASE or AIFCOMPLEX improves the robustness of quantitative analysis of DCE-MRI in prostate cancer.

Respiratory motion estimation of the liver with abdominal motion as a surrogate.

BACKGROUND: Respiratory-induced motion (RIM) causes uncertainties in localizing hepatic lesions, which could lead to inaccurate targeting during interventions. One approach to mitigate the problem is respiratory motion estimation (RME), in which the liver motion is estimated by measuring external signals called surrogates. METHODS: A learning-based approach has been developed and validated to estimate the RIM of hepatic lesions. External markers placed on the human's abdomen were chosen as surrogates. Accordingly, appropriate motion models (multivariate, Ridge and Lasso regression models) were designed to correlate the liver motion with the abdominal motion, and trained to estimate the superior-inferior (SI) motion of the liver. Three subjects volunteered for 6 sessions of such that liver images acquired by magnetic resonance imaging (MRI) were recorded alongside camera-tracked external markers. RESULTS AND CONCLUSIONS: The proposed machine learning approach was validated in MRI on human subjects and the results show that the approach could estimate the respiratory-induced SI motion of the liver with a mean absolute error (MAE) accuracy below 2 mm.

Improving the arterial input function in dynamic contrast enhanced MRI by fitting the signal in the complex plane.

PURPOSE: Dynamic contrast enhanced (DCE) imaging is a widely used technique in oncologic imaging. An essential prerequisite for obtaining quantitative values from DCE-MRI is the determination of the arterial input function (AIF). However, it is very challenging to accurately estimate the AIF using MR. A comprehensive model, which uses complex data instead of either magnitude or phase, was developed to improve AIF estimation. THEORY AND METHODS: The model was first applied to simulated data. Subsequently, the accuracy of the estimated contrast agent concentration was validated in a phantom. Finally the method was applied to existing DCE scans of 13 prostate cancer patients. RESULTS: The complex signal method combines the complementary strengths of the magnitude and phase method, increasing the precision and accuracy of concentration estimation in simulated and phantom data. The in vivo AIFs show a good agreement between arterial voxels (standard deviation in the peak and tail equal 0.4 mM and 0.12 mM, respectively). Furthermore, the dynamic behavior closely followed the AIF obtained with DCE-CT in the same patients (mean correlation coefficient: 0.92). CONCLUSION: By using the complex signal, the AIF estimation becomes more accurate and precise. This might enable patient specific AIFs, thereby improving the quantitative values obtained from DCE-MRI. Magn Reson Med 76:1236-1245, 2016. © 2015 Wiley Periodicals, Inc.