Prostate diffusion-weighted imaging (DWI) in MR-guided radiotherapy: Reproducibility assessment on 1.5 T MR-Linac and 1.5 T MR-simulator.

PURPOSE: Diffusion-weighted imaging (DWI) holds promise for image-guided radiotherapy (MRgRT) in prostate cancer. However, challenges persist due to image distortion, artifacts, and apparent diffusion coefficient (ADC) reproducibility issues. This study aimed to assess DWI image quality and ADC reproducibility on both a 1.5 T MR-simulator and a 1.5 T MR-Linac, employing measurements from both an ACR MRI phantom and prostate cancer patients undergoing MRgRT. METHODS: DW-MRI scans were conducted on 19 patients (mean age = 69 ± 8 years, with 23 MR-visible intra-prostatic lesions) and an ACR MRI phantom using a 1.5 T MR-simulator (b-values = 0, 800, 1400s/mm2) and a 1.5 T MR-Linac (b-values = 50, 500, 800 s/mm2). ADC homogeneity in the phantom was evaluated via 1D profile flatness (FL) in three directions. Image quality was assessed through qualitative 5-point Likert scale ratings and quantitative ADC and signal-to-noise ratio (SNR) measurements. Intra-observer reproducibility of image quality scores was evaluated using ICC(1, 2). Geometric distortion was measured by comparing landmark sizes on the ACR phantom against the ground truth. Mean ADC and reproducibility were assessed using Bland-Altman plots. RESULTS: Both MR-simulator and MR-Linac demonstrated high ADC homogeneity (FL > 87.5% - MR-simulator: 97.23 ± 0.62%, MR-Linac: 94.75 ± 0.62%, p < 0.05) in the phantom. Image quality scores revealed acceptable ratings (≥3) for capsule demarcation, image artifacts, and geometric distortion in patients. However, intra-prostatic lesions were barely discernible in b800 images for both MR-simulator (average score = 2.37 ± 1.33) and MR-Linac (average score = 2.16 ± 1.28). While MR-Linac DWI scans exhibited significantly more severe geometric distortion than MR-simulator scans (p < 0.01), most phantom measurements fell within the image in-plane resolution of 3 mm. Significant differences were noted in MR-simulator ADC (CTV: 1.20 ± 0.14 × 10-3 mm2/s (MR-simulator) vs 1.06 ± 0.10 × 10-3 mm2/s (MR-Linac); GTV: 1.05 ± 0.21 × 10-3 mm2/s vs 0.91 ± 0.16 × 10 mm2/s, all p < 0.05), with a small non-zero bias observed in the Bland-Altman analysis (CTV: 12.3%; GTV: 14.5%). CONCLUSION: The significantly larger MR-simulator ADC and the small non-zero bias hint at potential systematic differences in ADC values acquired from an MR-simulator and an MR-Linac, both at 1.5 T. Although acceptable ADC homogeneity was noted, caution is warranted in interpreting MR-Linac DWI images due to occasional severe artifacts. Further studies are essential to validate DWI and ADC as reliable imaging markers in prostate cancer MRgRT.

One-year clinical outcomes of MR-guided stereotactic body radiation therapy with rectal spacer for patients with localized prostate cancer.

BACKGROUND AND PURPOSE: This prospective study aimed to investigate adaptive magnetic resonance (MR)-guided stereotactic body radiation therapy (MRgSBRT) with rectal spacer for localized prostate cancer (PC) and report 1-year clinical outcomes. MATERIALS AND METHODS: Thirty-four consecutive patients with low- to high-risk localized PC that underwent 5-fraction adaptive MRgSBRT with rectal spacer were enrolled. The dosimetric comparison was performed on a risk- and age-matched cohort treated with MRgSBRT but without a spacer at a similar timepoint. Clinician-reported outcomes were based on Common Terminology Criteria for Adverse Events. Patient-reported outcomes were based on the Expanded Prostate Cancer Index Composite (EPIC) questionnaire at baseline, acute (1-3 months), subacute (4-12 months), and late (> 12 months) phases. RESULTS: The median follow-up was 390 days (range 28-823) and the median age was 70 years (range 58-82). One patient experienced rectal bleeding soon after spacer insertion that subsided before MRgSBRT. The median distance between the midline of the prostate midgland and the rectum after spacer insertion measured 7.8 mm (range 2.6-15.3), and the median length of the spacer was 45.9 mm (range 16.8-62.9) based on T2-weighted MR imaging. The use of spacer resulted in significant improvements in target coverage (V100% > 95% = 98.6% [range 93.4-99.8] for spacer vs. 97.8% [range 69.6-99.7] for non-spacer) and rectal sparing (V95% < 3 cc = 0.7 cc [range 0-4.6] for spacer vs. 4.9 cc [range 0-12.5] for non-spacer). Nine patients (26.5%) experienced grade 1 gastrointestinal toxicities, and no grade ≥ 2 toxicities were observed. During the 1-year follow-up period, EPIC scores for the bowel domain remained stable and were the highest among all other domains. CONCLUSIONS: MRgSBRT with rectal spacer for localized PC showed exceptional tolerability with minimal gastrointestinal toxicities and satisfactory patient-reported outcomes. Improvements in dosimetry, rectal sparing, and target coverage were achieved with a rectal spacer. Randomized trials are warranted for further validation.

A study of Bayesian deep network uncertainty and its application to synthetic CT generation for MR-only radiotherapy treatment planning.

BACKGROUND: The use of synthetic computed tomography (CT) for radiotherapy treatment planning has received considerable attention because of the absence of ionizing radiation and close spatial correspondence to source magnetic resonance (MR) images, which have excellent tissue contrast. However, in an MR-only environment, little effort has been made to examine the quality of synthetic CT images without using the original CT images. PURPOSE: To estimate synthetic CT quality without referring to original CT images, this study established the relationship between synthetic CT uncertainty and Bayesian uncertainty, and proposed a new Bayesian deep network for generating synthetic CT images and estimating synthetic CT uncertainty for MR-only radiotherapy treatment planning. METHODS AND MATERIALS: A novel deep Bayesian network was formulated using probabilistic network weights. Two mathematical expressions were proposed to quantify the Bayesian uncertainty of the network and synthetic CT uncertainty, which was closely related to the mean absolute error (MAE) in Hounsfield Unit (HU) of synthetic CT. These uncertainties were examined to demonstrate the accuracy of representing the synthetic CT uncertainty using a Bayesian counterpart. We developed a hybrid Bayesian architecture and a new data normalization scheme, enabling the Bayesian network to generate both accurate synthetic CT and reliable uncertainty information when probabilistic weights were applied. The proposed method was evaluated in 59 patients (13/12/32/2 for training/validation/testing/uncertainty visualization) diagnosed with prostate cancer, who underwent same-day pelvic CT- and MR-acquisitions. To assess the relationship between Bayesian and synthetic CT uncertainties, linear and non-linear correlation coefficients were calculated on per-voxel, per-tissue, and per-patient bases. For accessing the accuracy of the CT number and dosimetric accuracy, the proposed method was compared with a commercially available atlas-based method (MRCAT) and a U-Net conditional-generative adversarial network (UcGAN). RESULTS: The proposed model exhibited 44.33 MAE, outperforming UcGAN 52.51 and MRCAT 54.87. The gamma rate (2%/2 mm dose difference/distance to agreement) of the proposed model was 98.68%, comparable to that of UcGAN (98.60%) and MRCAT (98.56%). The per-patient and per-tissue linear correlation coefficients between the Bayesian and synthetic CT uncertainties ranged from 0.53 to 0.83, implying a moderate to strong linear correlation. Per-voxel correlation coefficients varied from -0.13 to 0.67 depending on the regions-of-interest evaluated, indicating tissue-dependent correlation. The R2 value for estimating MAE solely using Bayesian uncertainty was 0.98, suggesting that the uncertainty of the proposed model was an ideal candidate for predicting synthetic CT error, without referring to the original CT. CONCLUSION: This study established a relationship between the Bayesian model uncertainty and synthetic CT uncertainty. A novel Bayesian deep network was proposed to generate a synthetic CT and estimate its uncertainty. Various metrics were used to thoroughly examine the relationship between the uncertainties of the proposed Bayesian model and the generated synthetic CT. Compared with existing approaches, the proposed model showed comparable CT number and dosimetric accuracies. The experiments showed that the proposed Bayesian model was capable of producing accurate synthetic CT, and was an effective indicator of the uncertainty and error associated with synthetic CT in MR-only workflows.

Analysis of online plan adaptation for 1.5T magnetic resonance-guided stereotactic body radiotherapy (MRgSBRT) of prostate cancer.

PURPOSE: To analyze and characterize the online plan adaptation of 1.5T magnetic resonance-guided stereotactic body radiotherapy (MRgSBRT) of prostate cancer (PC). METHODS: PC patients (n = 107) who received adaptive 1.5 Tesla MRgSBRT were included. Online plan adaptation was implemented by either the adapt-to-position (ATP) or adapt-to-shape (ATS) methods. Patients were assigned to the ATS group if they underwent ≥ 1 ATS fraction (n = 51); the remainder were assigned to the ATP group (n = 56). The online plan adaptation records of 535 (107 × 5) fractions were retrospectively reviewed. Rationales for ATS decision-making were determined and analyzed using predefined criteria. Statistics of ATS fractions were summarized. Associations of patient characteristics and clinical factors with ATS utilization were investigated. RESULTS: There were 87 (16.3%) ATS fractions and 448 ATP fractions (83.7%). The numbers of ATS adoptions in fractions 1-5 were 29 (29/107, 27.1%), 18 (16.8%), 15 (14.0%), 16 (15.0%), and 9 (8.4%), respectively, with significant differences in adoption frequency between fractions (p = 0.007). Other baseline patient characteristics and clinical factors were not significantly associated with ATS classification (all p > 0.05). Underlying criteria for the determination of ATS implementation comprised anatomical changes (77 fractions in 50 patients) and discrete multiple targets (15 fractions in 3 patients). No ATS utilization was determined using dosimetric or online quality assurance criteria. CONCLUSIONS: This study contributes to facilitating the establishment of a standardized protocol for online MR-guided adaptive radiotherapy in PC.

A Prospective Study of Stereotactic Body Radiotherapy (SBRT) with Concomitant Whole-Pelvic Radiotherapy (WPRT) for High-Risk Localized Prostate Cancer Patients Using 1.5 Tesla Magnetic Resonance Guidance: The Preliminary Clinical Outcome.

Background: Conventionally fractionated whole-pelvic nodal radiotherapy (WPRT) improves clinical outcome compared to prostate-only RT in high-risk prostate cancer (HR-PC). MR-guided stereotactic body radiotherapy (MRgSBRT) with concomitant WPRT represents a novel radiotherapy (RT) paradigm for HR-PC, potentially improving online image guidance and clinical outcomes. This study aims to report the preliminary clinical experiences and treatment outcome of 1.5 Tesla adaptive MRgSBRT with concomitant WPRT in HR-PC patients. Materials and methods: Forty-two consecutive HR-PC patients (72.5 ± 6.8 years) were prospectively enrolled, treated by online adaptive MRgSBRT (8 Gy(prostate)/5 Gy(WPRT) × 5 fractions) combined with androgen deprivation therapy (ADT) and followed up (median: 251 days, range: 20−609 days). Clinical outcomes were measured by gastrointestinal (GI) and genitourinary (GU) toxicities according to the Common Terminology Criteria for Adverse Events (CTCAE) Scale v. 5.0, patient-reported quality of life (QoL) with EPIC (Expanded Prostate Cancer Index Composite) questionnaire, and prostate-specific antigen (PSA) responses. Results: All MRgSBRT fractions achieved planning objectives and dose specifications of the targets and organs at risk, and they were successfully delivered. The maximum cumulative acute GI/GU grade 1 and 2 toxicity rates were 19.0%/81.0% and 2.4%/7.1%, respectively. The subacute (>30 days) GI/GU grade 1 and 2 toxicity rates were 21.4%/64.3% and 2.4%/2.4%, respectively. No grade 3 toxicities were reported. QoL showed insignificant changes in urinary, bowel, sexual, and hormonal domain scores during the follow-up period. All patients had early post-MRgSBRT biochemical responses, while biochemical recurrence (PSA nadir + 2 ng/mL) occurred in one patient at month 18. Conclusions: To our knowledge, this is the first prospective study that showed the clinical outcomes of MRgSBRT with concomitant WPRT in HR-PC patients. The early results suggested favorable treatment-related toxicities and encouraging patient-reported QoLs, but long-term follow-up is needed to confirm our early results.

Acquisition repeatability of MRI radiomics features in the head and neck: a dual-3D-sequence multi-scan study.

Radiomics has increasingly been investigated as a potential biomarker in quantitative imaging to facilitate personalized diagnosis and treatment of head and neck cancer (HNC), a group of malignancies associated with high heterogeneity. However, the feature reliability of radiomics is a major obstacle to its broad validity and generality in application to the highly heterogeneous head and neck (HN) tissues. In particular, feature repeatability of radiomics in magnetic resonance imaging (MRI) acquisition, which is considered a crucial confounding factor of radiomics feature reliability, is still sparsely investigated. This study prospectively investigated the acquisition repeatability of 93 MRI radiomics features in ten HN tissues of 15 healthy volunteers, aiming for potential magnetic resonance-guided radiotherapy (MRgRT) treatment of HNC. Each subject underwent four MRI acquisitions with MRgRT treatment position and immobilization using two pulse sequences of 3D T1-weighed turbo spin-echo and 3D T2-weighed turbo spin-echo on a 1.5 T MRI simulator. The repeatability of radiomics feature acquisition was evaluated in terms of the intraclass correlation coefficient (ICC), whereas within-subject acquisition variability was evaluated in terms of the coefficient of variation (CV). The results showed that MRI radiomics features exhibited heterogeneous acquisition variability and uncertainty dependent on feature types, tissues, and pulse sequences. Only a small fraction of features showed excellent acquisition repeatability (ICC > 0.9) and low within-subject variability. Multiple MRI scans improved the accuracy and confidence of the identification of reliable features concerning MRI acquisition compared to simple test-retest repeated scans. This study contributes to the literature on the reliability of radiomics features with respect to MRI acquisition and the selection of reliable radiomics features for use in modeling in future HNC MRgRT applications.

A narrative review of MRI acquisition for MR-guided-radiotherapy in prostate cancer.

Magnetic resonance guided radiotherapy (MRgRT), enabled by the clinical introduction of the integrated MRI and linear accelerator (MR-LINAC), is a novel technique for prostate cancer (PCa) treatment, promising to further improve clinical outcome and reduce toxicity. The role of prostate MRI has been greatly expanded from the traditional PCa diagnosis to also PCa screening, treatment and surveillance. Diagnostic prostate MRI has been relatively familiar in the community, particularly with the development of Prostate Imaging - Reporting and Data System (PI-RADS). But, on the other hand, the use of MRI in the emerging clinical practice of PCa MRgRT, which is substantially different from that in PCa diagnosis, has been so far sparsely presented in the medical literature. This review attempts to give a comprehensive overview of MRI acquisition techniques currently used in the clinical workflows of PCa MRgRT, from treatment planning to online treatment guidance, in order to promote MRI practice and research for PCa MRgRT. In particular, the major differences in the MRI acquisition of PCa MRgRT from that of diagnostic prostate MRI are demonstrated and explained. Limitations in the current MRI acquisition for PCa MRgRT are analyzed. The future developments of MRI in the PCa MRgRT are also discussed.

3D-T2W-TSE radiotherapy treatment planning MRI using compressed sensing acceleration for prostate cancer: Image quality and delineation value.

PURPOSE: To assess the image quality and delineation value of compressed sensing (CS)-accelerated 3D T2W turbo-spin-echo (TSE) sequence for radiotherapy treatment planning (RTP) of prostate cancer. METHODS: An optimized CS-accelerated 3D-T2W-TSE was determined by volunteer imaging and applied for clinical RTP-MRI. This optimized CS-accelerated planning MRI and the standardized adaptive MRI acquired at 1.5T were retrospectively analyzed in 26 prostate cancer patients who were to receive MR-guided radiotherapy. Signal-to-noise ratio (SNR) and relative contrast ratio (CR) were quantitatively assessed. Image quality and artifacts were qualitatively assessed using a five-point scale rating. Delineation value in the prostate and organs-at-risk (OARs) was also rated and compared. Wilcoxon signed-rank test was used for SNR, relative CR, and rating comparisons. The interobserver rating agreement was evaluated by percent agreement. RESULTS: Significantly better SNR and relative CR in the prostate, rectum, bowel, penis, and penile bulb, while significantly worse in the cauda equina, were observed on the planning MRI. Significantly better ratings of image quality and artifacts were given to the planning MRI, with much less Gibbs ringing and reconstruction artifacts. Significantly better delineation value rating was achieved on the planning MRI in the prostate, seminal vesicle, rectum, penis, penile bulb, and testes, while significantly worse in the cauda equina. A strong to almost perfect interobserver rating agreement was obtained. CONCLUSION: This study suggested that CS acceleration is applicable and valuable in prostate RTP-MRI. CS-accelerated 3D-T2W-TSE images should benefit the delineation of prostate and many OARs.

3D T1-weighted turbo spin echo contrast-enhanced MRI at 1.5 T for frameless brain metastases radiotherapy.

PURPOSE: Performance of 3D-T1W-TSE has been proven superior to 3D-MP-GRE at 3 T on brain metastases (BM) contrast-enhanced (CE) MRI. However, its performance at 1.5 T is largely unknown and sparsely reported. This study aims to assess image quality, lesion detectability and conspicuity of 1.5 T 3D-T1W-TSE on planning MRI of frameless BM radiotherapy. METHODS: 94 BM patients to be treated by frameless brain radiotherapy were scanned using 3D-T1W-TSE with immobilization on multi-vendor 1.5 T MRI-simulators. BMs were jointly diagnosed by 4 reviewers. Enhanced lesion conspicuity was quantitatively assessed by calculating contrast ratio (CR) and contrast-to-noise ratio (CNR). Signal-to-noise ratio (SNR) reduction of white matter due to the use of flexible coil was assessed. Lesion detectability and conspicuity were compared between 1.5 T planning MRI and 3 T diagnostic MRI by an oncologist and a radiologist in 10 patients. RESULTS: 497 BMs were jointly diagnosed. The CR and CNR were 75.2 ± 39.9% and 14.2 ± 8.1, respectively. SNR reduced considerably from 31.7 ± 8.3 to 21.9 ± 5.4 with the longer distance to coils. 3 T diagnostic MRI and 1.5 T planning MRI yielded exactly the same detection of 84 BMs. Qualitatively, lesion conspicuity at 1.5 T was not inferior to that at 3 T. Quantitatively, lower brain SNR and lesion CNR were found at 1.5 T, while lesion CR at 1.5 T was highly comparable to that at 3 T. CONCLUSION: 1.5 T 3D-T1W-TSE planning MRI of frameless BM radiotherapy was comprehensively assessed. Highly comparable BM detectability and conspicuity were achieved by 1.5 T planning MRI compared to 3 T diagnostic MRI. 1.5 T 3D-T1W-TSE should be valuable for frameless brain radiotherapy planning.

1.5T Magnetic Resonance-Guided Stereotactic Body Radiotherapy for Localized Prostate Cancer: Preliminary Clinical Results of Clinician- and Patient-Reported Outcomes.

BACKGROUND: Magnetic resonance-guided stereotactic body radiotherapy (MRgSBRT) offers the potential for achieving better prostate cancer (PC) treatment outcomes. This study reports the preliminary clinical results of 1.5T MRgSBRT in localized PC, based on both clinician-reported outcome measurement (CROM) and patient-reported outcome measurement (PROM). METHODS: Fifty-one consecutive localized PC patients were prospectively enrolled with a median follow-up of 199 days. MRgSBRT was delivered in five fractions of 7.25-8 Gy with daily online adaptation. Clinician-reported gastrointestinal (GI) and genitourinary (GU) adverse events based on the Common Terminology Criteria for Adverse Events (CTCAE) Scale v. 5.0 were assessed. The Expanded Prostate Cancer Index Composite Questionnaire was collected at baseline, 1 month, and every 3 months thereafter. Serial prostate-specific antigen measurements were longitudinally recorded. RESULTS: The maximum cumulative clinician-reported grade ≥ 2 acute GU and GI toxicities were 11.8% (6/51) and 2.0% (1/51), respectively, while grade ≥ 2 subacute GU and GI toxicities were 2.3% (1/43) each. Patient-reported urinary, bowel, and hormonal domain summary scores were reduced at 1 month, then gradually returned to baseline levels, with the exception of the sexual domain. Domain-specific subscale scores showed similar longitudinal changes. All patients had early post-MRgSBRT biochemical responses. CONCLUSIONS: The finding of low toxicity supports the accumulation of clinical evidence for 1.5T MRgSBRT in localized PC.

Radiomics feature reliability assessed by intraclass correlation coefficient: a systematic review.

Radiomics research is rapidly growing in recent years, but more concerns on radiomics reliability are also raised. This review attempts to update and overview the current status of radiomics reliability research in the ever expanding medical literature from the perspective of a single reliability metric of intraclass correlation coefficient (ICC). To conduct this systematic review, Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed. After literature search and selection, a total of 481 radiomics studies using CT, PET, or MRI, covering a wide range of subject and disease types, were included for review. In these highly heterogeneous studies, feature reliability to image segmentation was much more investigated than reliability to other factors, such as image acquisition, reconstruction, post-processing, and feature quantification. The reported ICCs also suggested high radiomics feature reliability to image segmentation. Image acquisition was found to introduce much more feature variability than image segmentation, in particular for MRI, based on the reported ICC values. Image post-processing and feature quantification yielded different levels of radiomics reliability and might be used to mitigate image acquisition-induced variability. Some common flaws and pitfalls in ICC use were identified, and suggestions on better ICC use were given. Due to the extremely high study heterogeneities and possible risks of bias, the degree of radiomics feature reliability that has been achieved could not yet be safely synthesized or derived in this review. More future researches on radiomics reliability are warranted.

Quantitative assessment of acquisition imaging parameters on MRI radiomics features: a prospective anthropomorphic phantom study using a 3D-T2W-TSE sequence for MR-guided-radiotherapy.

BACKGROUND: MRI pulse sequences and imaging parameters substantially influence the variation of MRI radiomics features, thus impose a critical challenge on MRI radiomics reproducibility and reliability. This study aims to prospectively investigate the impact of various imaging parameters on MRI radiomics features in a 3D T2-weighted (T2W) turbo-spin-echo (TSE) pulse sequence for MR-guided-radiotherapy (MRgRT). METHODS: An anthropomorphic phantom was scanned using a 3D-T2W-TSE MRgRT sequence at 1.5T under a variety of acquisition imaging parameter changes. T1 and T2 relaxation times of the phantom were also measured. 93 first-order and texture radiomics features in the original and 14 transformed images, yielding 1,395 features in total, were extracted from 10 volumes-of-interest (VOIs). The percentage deviation (d%) of radiomics feature values from the baseline values and intra-class correlation coefficient (ICC) with the baseline were calculated. Robust radiomics features were identified based on the excellent agreement of radiomics feature values with the baseline, i.e., the averaged d% <5% and ICC >0.90 in all VOIs for all imaging parameter variations. RESULTS: The radiomics feature values changed considerably but to different degrees with different imaging parameter adjustments, in the ten VOIs. The deviation d% ranged from 0.02% to 321.3%, with a mean of 12.5% averaged for all original features in all ten VOIs. First-order and GLCM features were generally more robust to imaging parameters than other features in the original images. There were also significantly different radiomics feature values (ANOVA, P<0.001) between the original and the transformed images, exhibiting quite different robustness to imaging parameters. 330 out of 1395 features (23.7%) robust to imaging parameters were identified. GLCM and GLSZM features had the most (42.5%, 153/360) and least (3.8%, 9/240) robust features in the original and transformed images, respectively. CONCLUSIONS: This study helps better understand the quantitative dependence of radiomics feature values on imaging parameters in a 3D-T2W-TSE sequence for MRgRT. Imaging parameter heterogeneity should be considered as a significant source of radiomics variability and uncertainty, which must be well harmonized for reliable clinical use. The identified robust features to imaging parameters are helpful for the pre-selection of radiomics features for reliable radiomics modeling.

Reliability of radiomics features due to image reconstruction using a standardized T2 -weighted pulse sequence for MR-guided radiotherapy: An anthropomorphic phantom study.

PURPOSE: To prospectively investigate the impact of image reconstruction on MRI radiomics features. METHODS: An anthropomorphic phantom was scanned at 1.5 T using a standardized sequence for MR-guided radiotherapy under SENSE and compressed-SENSE reconstruction settings. A total of 93 first-order and texture radiomics features in 10 volumes of interest were assessed based on (1) accuracy measured by the percentage deviation from the reference, (2) robustness on reconstruction in all volumes of interest measured by the intraclass correlation coefficient, and (3) repeatability measured by the coefficient of variance over the repetitive acquisitions. Finally, reliable and unreliable radiomics features were comprehensively determined based on their accuracy, robustness, and repeatability. RESULTS: Better accuracy and robustness of the radiomics features were achieved under SENSE than compressed-SENSE reconstruction. The feature accuracy under SENSE reconstruction was more affected by acceleration factor than direction, whereas under compressed-SENSE reconstruction, accuracy was substantially impacted by the increasing denoising levels. Feature repeatability was dependent more on feature types than on reconstruction. A total of 45 reliable features and 13 unreliable features were finally determined for SENSE, compared with 22 reliable and 26 unreliable features for compressed SENSE. First-order and gray-level co-occurrence matrix features were generally more reliable than other features. CONCLUSION: Radiomics features could be substantially affected by MRI reconstruction, so precautions need to be taken regarding their reliability for clinical use. This study helps the guidance of the preselection of reliable radiomics features and the preclusion of unreliable features in MR-guided radiotherapy.

Longitudinal acquisition repeatability of MRI radiomics features: An ACR MRI phantom study on two MRI scanners using a 3D T1W TSE sequence.

PURPOSE: The purpose of this study was to quantitatively assess the longitudinal acquisition repeatability of MRI radiomics features in a three-dimensional (3D) T1-weighted (T1W) TSE sequence via a well-controlled prospective phantom study. METHODS: Thirty consecutive daily datasets of an ACR-MRI phantom were acquired on two 1.5T MRI simulators using a 3D T1W TSE sequence. Images were blindly segmented by two observers. Post-acquisition processing was minimized but an intensity discretization (fixed bin size of 25). One hundred and one radiomics features (shape n = 12; first order n = 16; texture n = 73) were extracted. Longitudinal repeatability of each feature was evaluated by Pearson correlation and coefficient of variance (CV68% ). Interobserver feature value agreement was also quantified using intraclass correlation coefficient (ICC) and Bland-Altman analysis. A most repeatable radiomics feature set on both scanners was determined by feature coefficient of variance (CV68% <5%), ICC (>0.75), and the ratio of the interobserver difference to the interobserver mean δ<5%. RESULTS: No trend of radiomics feature value changed with time. Longitudinal feature repeatability CV68% ranged 0.01-38.60% (mean/median: 12.5%/9.9%), and 0.01-40.47%, (8.49%/7.34%) on the scanners A and B. Shape features exhibited significantly better repeatability than first-order and texture features (all P < 0.01). Significant longitudinal repeatability difference was observed in texture features (P < 0.001) between the two scanners, but not in shape and first-order features (P > 0.30). First-order and texture features had smaller interobserver-dependent variation than acquisition-dependent variation. They also showed good interobserver agreement on both scanners (A:ICC = 0.80 ± 0.23; B:ICC = 0.80 ± 0.22), independent of acquisition repeatability. The repeatable radiomics features in common on both scanners, including 12 shape features, 0 first-order features, and 3 texture features, were determined as the most repeatable MRI radiomics feature set. CONCLUSIONS: Radiomics features exhibited heterogeneous longitudinal repeatability, while the shape features were the most repeatable, in this phantom study with a 3D T1W TSE acquisition. The most repeatable radiomics feature set derived in this study should be helpful for the selection of reliable radiomics features in the future clinical use.

A pilot study of highly accelerated 3D MRI in the head and neck position verification for MR-guided radiotherapy.

BACKGROUND: To evaluate the performance of a highly accelerated 3D MRI on inter-fractional positional measurement for MR-guided radiotherapy (MRgRT) in the head and neck (HN). METHODS: Fourteen healthy volunteers received 159 scans on a 1.5 T MR-sim to simulate MRgRT fractions. MRI acquisition included a high-resolution (HQI-MRI, voxel-size =1.05×1.05×1.05 mm3, duration =5 min) and a highly-accelerated low-resolution (true-LQI-MRI, acceleration-factor =9, voxel-size =1.4×1.4×1.4 mm3, duration =86 s) T1w spin-echo sequence (TR/TE =420/7.2 ms). The first session HQI-MRI was used as the reference to mimic planning MRI. Other HQI-MRI was also retrospectively down-sampled in K-space and GRAPPA reconstructed to generate pseudo-LQI-MRI. Inter-sessional positional shift calculated from HQI-MRI, true-LQI-MRI and pseudo-LQI-MRI rigidly registering to the reference were analyzed and compared in the overall HN and the sub-regions of brain, nasopharynx, oropharynx and hypopharynx. RESULTS: The calculated SD of systematic errors (Σ) from HQI-MRI/pseudo-LQI-MRI/true-LQI-MRI images for overall HN were 1.11/1.14/1.08, 0.28/0.26/0.29, 0.43/0.44/0.60, and 0.77/0.79/0.74 mm for translation in LR, AP, SI and 3D, respectively; The corresponding RMS of random errors (σ) were 0.97/0.98/0.96, 0.28/0.27/0.26, 0.77/0.77/0.72, and 0.85/0.87/0.85 mm. For all sub-regions, brain showed the smallest Σ and σ in 3D. Other sub-regions showed direction-dependent error patterns, but the positioning results were consistent, independent of the datasets used for registration. CONCLUSIONS: A highly-accelerated 3D-MRI could be used for MR-guided HN radiotherapy without compromising position verification accuracy.

A fast volumetric 4D-MRI with sub-second frame rate for abdominal motion monitoring and characterization in MRI-guided radiotherapy.

BACKGROUND: To propose a fast volumetric 4D-MRI based on 3D pulse sequence acquisition for abdominal motion monitoring and characterization in MRI-guided radiotherapy (MRgRT). METHODS: A 3D spoiled gradient echo sequence volumetric interpolated breath-hold examination (VIBE) [repetition time/echo time (TR/TE) =0.53/1.57 ms, flip-angle =5°, receiver bandwidth (RBW) =1,400 Hz/voxel] based 4D-MRI acquisition, accelerated by 4-fold controlled aliasing in parallel imaging results in higher acceleration (CAIPIRINHA), named CAIPIRINHA-VIBE 4D-MRI, was implemented on a 1.5T MRI simulator (MR-sim) and applied for abdominal imaging of nine healthy volunteers under free breathing. One hundred and forty-four dynamics of the entire abdomen volume (56 slices), in total 8,064 (144×56) images with a voxel size of 2.7×2.7×4.0 mm3, were acquired in 89 s for 4D-MRI. This CAIPIRINHA-VIBE 4D-MRI was qualitatively compared with a 2D half-Fourier acquisition single-shot turbo spin-echo (2D-HASTE) based 4D-MRI. The motions of liver dome, kidney and spleen were analyzed using the CAIPIRINHA-VIBE 4D-MRI data. The kidney motion was quantitatively characterized in terms of motion range and the correlations between left and right kidneys. RESULTS: CAIPIRINHA-VIBE 4D-MRI was successfully conducted in all subjects. CAIPIRINHA-VIBE 4D-MRI exhibited much higher effective volumetric temporal resolution (0.615 vs. ~5 s/volume) and better reconstructed volume consistency than 2D-HASTE 4D-MRI. CAIPIRINHA-VIBE 4D-MRI was able to characterize the respiratory motion of abdominal organs simultaneously in three orthogonal directions, and could potentially be used for whole abdomen deformable motion tracking. Renal motion range was most pronounced in superior-inferior (SI) direction (L: 10.03±2.65 mm; R: 10.38±2.80 mm), significantly larger (P<0.001) than that in anterior-posterior (AP) and the least in left-right (LR) directions. Right kidney had significantly larger mobility (4.18±2.19 vs. 2.32±1.34 mm, P=0.045) than left kidney in AP, but not in LR and SI directions. The Pearson correlation coefficients r between left and right kidney motion were 0.5063 (P=0.164), 0.6624 (P=0.052) and 0.5752 (P=0.105) in LR, AP and SI correspondingly. The correlation of renal motion in SI and AP was found significant in right kidney (r=0.843, P=0.004) but not in left kidney (r=0.467, P=0.205). CONCLUSIONS: A fast volumetric 4D-MRI was implemented for abdominal motion monitoring in MRgRT. A sub-second volumetric temporal resolution of 0.615 s, covering the entire abdomen, was demonstrated for respiratory motion monitoring and characterization. This technique holds potentials for MRgRT applications.

Assessment of positional reproducibility in the head and neck on a 1.5-T MR simulator for an offline MR-guided radiotherapy solution.

BACKGROUND: Recently, a shuttle-based offline magnetic resonance-guided radiotherapy (MRgRT) approach was proposed. This study aims to evaluate the positional reproducibility in the immobilized head and neck using a 1.5-T MR-simulator (MR-sim) on healthy volunteers. METHODS: A total of 159 scans of 14 healthy volunteers were conducted on a 1.5-T MR-sim with thermoplastic mask immobilization. MR images with isotropic 1.053 mm3 voxel size were rigidly registered to the first scan based on fiducial, anatomical and gross positions. Mean and standard deviation of positional displacements in translation and rotation were assessed. Systematic error and random errors of positioning in the head and neck on the MR-sim were determined in the translation of, and in the rotation of roll, pitch and yaw. RESULTS: The systematic error (Σ) of translation in left-right (LR), anterior-posterior (AP) and superior-inferior (SI) direction was 0.57, 0.22 and 0.26 mm for fiducial displacement, 0.28, 0.10 and 0.52 mm for anatomical displacement, and 0.53, 0.22 and 0.49 mm for gross displacement, respectively. The random error (σ) in corresponding translation direction was 2.07, 0.54 and 1.32 mm for fiducial displacement, 1.34, 0.73 and 2.04 mm for anatomical displacement, and 2.24, 0.86 and 2.61 mm for gross displacement. The systematic error and random error of rotation were generally smaller than 1°. CONCLUSIONS: Our results suggested that high gross positional reproducibility (<1 mm translational and <1° rotational systematic error) could be achieved on an MR-sim for the proposed offline MRgRT.

Evaluation and Minimization of the Pseudohepatic Anisotropy Artifact in Liver Intravoxel Incoherent Motion.

PURPOSE: The aim of this study was to evaluate the effect of the pseudohepatic anisotropy artifact on liver intravoxel incoherent motion (IVIM) metrics and whether the use of multiple gradient directions in the IVIM acquisition minimizes the artifact. MATERIALS AND METHODS: Multiple breath-holding and forced shallow free-breathing IVIM scans were performed on 8 healthy volunteers using 1 and 6 gradient directions. Cluster analysis was carried out to separate motion-contaminated parenchyma from liver parenchyma and vessels. Nonlinear motion analysis was also performed to look for a possible link between IVIM metrics and nonlinear liver motion. RESULTS: On the basis of the resulted clusters, motion-contaminated parenchyma is often noted in the left liver lobe, where the prominent pseudohepatic artifact has previously been identified. A significant reduction in outliers was obtained with the acquisition of 6 noncoplanar gradient directions and when using forced shallow free-breathing. CONCLUSION: The pseudohepatic anisotropy artifact can be minimized when using multiple diffusion-encoding gradient directions and forced free-breathing during IVIM acquisition.

Image quality assessment of a 1.5T dedicated magnetic resonance-simulator for radiotherapy with a flexible radio frequency coil setting using the standard American College of Radiology magnetic resonance imaging phantom test.

BACKGROUND: A flexible RF coil setting has to be used on an MR-simulator (MR-sim) in the head and neck simulation scan for radiotherapy (RT) purpose, while the image quality might be compromised due to the sub-optimized flexible coil compared to the normal diagnostic radiological (DR) head coil. In this study, we assessed the image quality of an MR-sim by conducting the standard American College of Radiology (ACR) MRI phantom test on a 1.5T MR-sim under RT-setting and comparing it to DR-setting. METHODS: A large ACR MRI phantom was carefully positioned, aligned and scanned 9 times for each under RT- and DR-setting on a 1.5T MR-sim, following the ACR scanning instruction. Images were analyzed following the ACR guidance. Measurement results under two coil settings were quantitatively compared. Inter-observer disagreements under RT-setting between two physicists were compared using Bland-Altman (BA) analysis and intra-class correlation coefficient (ICC). RESULTS: The MR-sim with RT-setting obtained sufficiently good image quality to pass all ACR recommended criteria. No significant difference was found in phantom length accuracy, high-contrast spatial resolution, slice thickness accuracy, slice position accuracy, and percent-signal ghosting. RT-setting significantly under-performed in low-contrast object detectability, while better performed in image intensity uniformity. BA analysis showed that 95% limit of agreement and biases of phantom test measurement under RT-setting between two observers were very small. Excellent inter-observer agreement (ICC >0.75) was achieved in all measurements except for slice thickness accuracy (ICC =0.42, moderate agreement) under RT-setting. CONCLUSIONS: Very good and highly reproducible image quality could be achieved on a 1.5T MR-sim with a flexible coil setting as revealed by the standard ACR MRI phantom test. The flexible RT-setting compromised in image signal-to-noise ratio (SNR) compared to the normal DR-setting, and resulted in reduced low-contrast object detectability.

Intrasession and Intersession Repeatability of Diffusion Tensor Imaging in Healthy Human Liver.

OBJECTIVE: The aim of this study was to evaluate the effect of signal to noise ratio (SNR) and number of gradient directions (NGD) on intra- and intersession repeatability of liver diffusion tensor imaging (DTI) metrics. METHODS: At each of 3 liver DTI scan sessions, liver diffusion was assessed in 5 healthy volunteers using a 6-direction DTI scan performed 9 separate times (ie, number of signal averages [NSA]). In addition, 4 combinations of NSA and NGD were acquired (NSA/NGD = 1/30, 3/10, 3/12, and 5/6) to determine the combined effect to DTI metrics, which was based on intersubject variability and intrasession (Vintra) and intersession (Vinter) repeatability. RESULTS: Intersubject variability was less than 20%, whereas Vintra and Vinter repeatability were less than 5% and less than 10%, respectfully. Vinter was not affected by the NGD used. Decreases in Vinter(FA), Vinter(λ1), Vinter(RD), and Vinter(MD) were observed with increasing NSA, and hence SNR. CONCLUSION: Increased SNR may improve intrasession and intersession repeatability of liver DTI metrics. Scan repeatability was not influenced by NGD.