Evaluation of MRI-only based online adaptive radiotherapy of abdominal region on MR-linac.

PURPOSE: A hybrid magnetic resonance linear accelerator (MRL) can perform magnetic resonance imaging (MRI) with high soft-tissue contrast to be used for online adaptive radiotherapy (oART). To obtain electron densities needed for the oART dose calculation, a computed tomography (CT) is often deformably registered to MRI. Our aim was to evaluate an MRI-only based synthetic CT (sCT) generation as an alternative to the deformed CT (dCT)-based oART in the abdominal region. METHODS: The study data consisted of 57 patients who were treated on a 0.35 T MRL system mainly for abdominal tumors. Simulation MRI-CT pairs of 43 patients were used for training and validation of a prototype convolutional neural network sCT-generation algorithm, based on HighRes3DNet, for the abdominal region. For remaining test patients, sCT images were produced from simulation MRIs and daily MRIs. The dCT-based plans were re-calculated on sCT with identical calculation parameters. The sCT and dCT were compared in terms of geometric agreement and calculated dose. RESULTS: The mean and one standard deviation of the geometric agreement metrics over dCT-sCT-pairs were: mean error of 8 ± 10 HU, mean absolute error of 49 ± 10 HU, and Dice similarity coefficient of 55 ± 12%, 60 ± 5%, and 82 ± 15% for bone, fat, and lung tissues, respectively. The dose differences between the sCT and dCT-based dose for planning target volumes were 0.5 ± 0.9%, 0.6 ± 0.8%, and 0.5 ± 0.8% at D2% , D50% , and D98% in physical dose and 0.8 ± 1.4%, 0.8 ± 1.2%, and 0.6 ± 1.1% in biologically effective dose (BED). For organs-at-risk, the dose differences of all evaluated dose-volume histogram points were within [-4.5%, 7.8%] and [-1.1 Gy, 3.5 Gy] in both physical dose and BED. CONCLUSIONS: The geometric agreement metrics were within typically reported values and most average relative dose differences were within 1%. Thus, an MRI-only sCT-based approach is a promising alternative to the current clinical practice of the abdominal oART on MRL.

Planar dGEMRIC Maps May Aid Imaging Assessment of Cartilage Damage in Femoroacetabular Impingement.

BACKGROUND: Three-dimensional (3-D) delayed gadolinium-enhanced MRI of cartilage (dGEMRIC) helps quantify biochemical changes in articular cartilage that correlate with early-stage osteoarthritis. However, dGEMRIC analysis is performed slice by slice, limiting the potential of 3-D data to give an overall impression of cartilage biochemistry. We previously developed a computational algorithm to produce unfolded, or "planar," dGEMRIC maps of acetabular cartilage, but have neither assessed their application nor determined whether MRI-based grading of cartilage damage or dGEMRIC measurements predict intraoperative findings in hips with symptomatic femoroacetabular impingement (FAI). QUESTIONS/PURPOSES: (1) Does imaging-based assessment of acetabular cartilage damage correlate with intraoperative findings in hips with symptomatic FAI? (2) Does the planar dGEMRIC map improve this correlation? (3) Does the planar map improve the correlation between the dGEMRIC index and MRI-based grading of cartilage damage in hips with symptomatic FAI? (4) Does the planar map improve imaging-based evaluation time for hips with symptomatic FAI? METHODS: We retrospectively studied 47 hips of 45 patients with symptomatic FAI who underwent hip surgery between 2009 and 2013 and had a 1.5-T 3-D dGEMRIC scan within 6 months preoperatively. Our cohort included 25 males and 20 females with a mean ± SD age at surgery of 29 ± 11 years. Planar dGEMRIC maps were generated from isotropic, sagittal oblique TrueFISP and T1 sequences. A pediatric musculoskeletal radiologist with experience in hip MRI evaluated studies using radially reformatted sequences. For six acetabular subregions (anterior-peripheral [AP]; anterior-central [AC]; superior-peripheral [SP]; superior-central [SC]; posterior-peripheral [PP]; posterior-central [PC]), modified Outerbridge cartilage damage grades were recorded and region-of-interest T1 averages (the dGEMRIC index) were measured. Beck's intraoperative cartilage damage grades were compared with the Outerbridge grades and dGEMRIC indices. For a subset of 26 hips, 13 were reevaluated with the map and 13 without the map, and total evaluation times were recorded. RESULTS: There were no meaningful differences in the correlations obtained with versus without referencing the planar maps. Planar map-independent Outerbridge grades had a notable (p < 0.05) Spearman's rank correlation (ρ) with Beck's grades that was moderate in AP, SC, and PC (0.3 < ρ < 0.5) and strong in SP (ρ > 0.5). For map-dependent Outerbridge grades, ρ was moderate in AP, AC, and SC and strong in SP. Map-independent dGEMRIC indices had a ρ with Beck's grades that was moderate in AP and SC (-0.3 > ρ > -0.5) and strong in SP (ρ < -0.5). For map-dependent dGEMRIC indices, ρ was moderate in SC and strong in SP. Similarly, there were no meaningful, map-dependent differences in the correlations. When comparing Outerbridge grades and dGEMRIC indices, there were notable correlations across all subregions. Without the planar map, ρ was moderate in AC and PC and strong in AP, SP, SC, and PP. With the map, ρ was strong in all six subregions. In AC, there was a notable map-dependent improvement in this correlation (p < 0.001). Finally, referencing the planar dGEMRIC map during evaluation was associated with a decrease in mean evaluation time, from 207 ± 32 seconds to 152 ± 33 seconds (p = 0.001). CONCLUSIONS: Our work challenges the weak correlation between dGEMRIC and intraoperative findings of cartilage damage that was previously reported in hips with symptomatic FAI, suggesting that dGEMRIC has potential diagnostic use for this patient population. The planar dGEMRIC maps did not meaningfully alter the correlation of imaging-based evaluation of cartilage damage with intraoperative findings; however, they notably improved the correlation of dGEMRIC and MRI-based grading in AC, and their use incurred no additional time cost to imaging-based evaluation. Therefore, the planar maps may improve dGEMRIC's use as a continuous proxy for an otherwise discrete and simplified MRI-based grade of cartilage damage in hips with symptomatic FAI. LEVEL OF EVIDENCE: Level III, diagnostic study.

Delayed gadolinium-enhanced MRI of meniscus (dGEMRIM) and cartilage (dGEMRIC) in healthy knees and in knees with different stages of meniscus pathology.

BACKGROUND: Lesions in the meniscus are risk factors for developing knee osteoarthritis (OA), not least because of the role of the meniscus in the pathological progression of OA. Delayed gadolinium enhanced MRI of cartilage (dGEMRIC) has extensively been used to identify pre-radiographic cartilage changes in OA. In contrast, its counterpart with regard to examination of the meniscus, gadolinium enhanced MRI of meniscus (dGEMRIM), has been less utilized. In this study we use 3D dGEMRIM in patients with meniscus lesions and compare them with previous results of healthy individuals. METHODS: Eighteen subjects with MRI-verified posteromedial meniscus lesions and 12 healthy subjects with non-injured and non-symptomatic knee joints, together 30 volunteers, were examined using 3D Look-Locker sequence after intravenous injection of Gd-DTPA2- (0.2 mmol/kg body weight). Relaxation time (T1) was measured in the posterior meniscus and femoral cartilage before and 60, 90, 120 and 180 min after injection. Relaxation rate (R1 = 1/T1) and change in relaxation rate (ΔR1) were calculated. For statistical analyses, Student's t-test and Analysis of Variance (ANOVA) were used. RESULTS: The pre-contrast diagnostic MRI identified two sub-cohorts in the 18 patients with regard to meniscus injury: 1) 11 subjects with MRI verified pathological intrameniscal changes (grade 2) in the posteromedial meniscus only and no obvious cartilage changes. The lateral meniscus showed no pathology. 2) 7 subjects with MRI verified pathological rupture (grade 3) of the posteromedial meniscus and pathological changes in the lateral meniscus and/or medial and lateral joint cartilage. Comparisons of pathological and healthy posteromedial meniscus revealed opposite patterns in both T1Gd and ΔR1 values between pathological meniscus grade 2 and grade 3. The concentration of the contrast agent was lower than in healthy meniscus in grade 2 lesions (p = 0.046) but tended to increase in grade 3 lesions (p = 0.110). Maximum concentration of contrast agent was reached after 180 min in both cartilage and menisci (except for grade 3 menisci where the maximum concentration was reached after 90 min). CONCLUSION: dGEMRIM and dGEMRIC may be feasible to combine in vivo, preferably with one examination before and one 2 h after contrast injection. Possible different dGEMRIM patterns at different stages of meniscus lesions must be taken into account when evaluating meniscus pathology.

Unthreaded Fixation of Slipped Capital Femoral Epiphysis Leads to Continued Growth of the Femoral Neck.

BACKGROUND: The optimal treatment for slipped capital femoral epiphysis (SCFE) remains controversial. In Sweden, the standard treatment is unthreaded fixation over the physis, with the purpose to permit continued growth of the femoral neck. The aim of the present study was to verify and quantify longitudinal growth of the femoral neck after in situ pinning with the Hansson hook-pin. METHODS: We performed a retrospective study of 54 patients treated with the Hansson hook-pin for SCFE between 2001 and 2009. The immediate postoperative radiograph and the radiograph after physeal closure (mean interval, 34 mo) were analyzed. Because the smooth Hansson hook-pin only has a grip fixation in the epiphysis, the femoral neck growth was determined as the difference in nail protrusion from the lateral cortex between the 2 radiographs. The femoral neck offset was also measured in all radiographs. RESULTS: Significant longitudinal growth occurred both in the slipped and the prophylactically treated contralateral hip by mean 7.1 mm (P<0.001) and 10.0 mm (P<0.001), respectively. There was no difference in growth between genders and no correlation between the amount of longitudinal growth and slip severity (range, 4.0 to 71,6 degrees; mean 27.3 degrees). Young patients (less than 11 y) grew more than older patients (more than 14 y), 12.1 vs. 4.2 mm, P=0.002. The femoral offset increased by mean 16% from mean 30.0 to 35.2 mm (P<0.001). The longitudinal growth of the femoral neck correlated with the increase in femoral offset (R=0.51, P<0.001). CONCLUSIONS: Unthreaded fixation of SCFE with the Hansson hook-pin allows continued growth of the femoral neck. The remaining growth enables the patient to achieve an almost anatomic offset of the hip. This is essential to optimize the abduction forces that stabilize the pelvis during gait. Future studies need to establish whether the longitudinal growth also results in improved remodelling of the proximal femur. LEVEL OF EVIDENCE: Level III-retrospective comparative study.

In vivo transport of Gd-DTPA2- into human meniscus and cartilage assessed with delayed gadolinium-enhanced MRI of cartilage (dGEMRIC).

BACKGROUND: Impaired stability is a risk factor in knee osteoarthritis (OA), where the whole joint and not only the joint cartilage is affected. The meniscus provides joint stability and is involved in the early pathological progress of OA. Delayed gadolinium-enhanced MRI of cartilage (dGEMRIC) has been used to identify pre-radiographic changes in the cartilage in OA, but has been used less commonly to examine the meniscus, and then using only a double dose of the contrast agent. The purpose of this study was to enable improved early OA diagnosis by investigate the temporal contrast agent distribution in the meniscus and femoral cartilage simultaneously, in healthy volunteers, using 3D dGEMRIC at two different doses of the contrast agent Gd-DTPA2-. METHODS: The right knee in 12 asymptomatic volunteers was examined using a 3D Look-Locker sequence on two occasions after an intravenous injection of a double or triple dose of Gd-DTPA2- (0.2 or 0.3 mmol/kg body weight). The relaxation time (T1) and relaxation rate (R1 = 1/T1) were measured in the meniscus and femoral cartilage before, and 60, 90, 120 and 180 minutes after injection, and the change in relaxation rate (ΔR1) was calculated. Paired t-test and Analysis of Variance (ANOVA) were used for statistical evaluation. RESULTS: The triple dose yielded higher concentrations of Gd-DTPA2- in the meniscus and cartilage than the double dose, but provided no additional information. The observed patterns of ΔR1 were similar for double and triple doses of the contrast agent. ΔR1 was higher in the meniscus than in femoral cartilage in the corresponding compartments at all time points after injection. ΔR1 increased until 90-180 minutes in both the cartilage and the meniscus (p < 0.05), and was lower in the medial than in the lateral meniscus at all time points (p < 0.05). A faster increase in ΔR1 was observed in the vascularized peripheral region of the posterior medial meniscus, than in the avascular central part of the posterior medial meniscus during the first 60 minutes (p < 0.05). CONCLUSION: It is feasible to examine undamaged meniscus and cartilage simultaneously using dGEMRIC, preferably 90 minutes after the injection of a double dose of Gd-DTPA2- (0.2 mmol/kg body weight).

Effects of B1 inhomogeneity correction for three-dimensional variable flip angle T1 measurements in hip dGEMRIC at 3 T and 1.5 T.

Delayed gadolinium-enhanced MRI of cartilage is a technique for studying the development of osteoarthritis using quantitative T(1) measurements. Three-dimensional variable flip angle is a promising method for performing such measurements rapidly, by using two successive spoiled gradient echo sequences with different excitation pulse flip angles. However, the three-dimensional variable flip angle method is very sensitive to inhomogeneities in the transmitted B(1) field in vivo. In this study, a method for correcting for such inhomogeneities, using an additional B(1) mapping spin-echo sequence, was evaluated. Phantom studies concluded that three-dimensional variable flip angle with B(1) correction calculates accurate T(1) values also in areas with high B(1) deviation. Retrospective analysis of in vivo hip delayed gadolinium-enhanced MRI of cartilage data from 40 subjects showed the difference between three-dimensional variable flip angle with and without B(1) correction to be generally two to three times higher at 3 T than at 1.5 T. In conclusion, the B(1) variations should always be taken into account, both at 1.5 T and at 3 T.

Synthetic computed tomography data allows for accurate absorbed dose calculations in a magnetic resonance imaging only workflow for head and neck radiotherapy.

BACKGROUND AND PURPOSE: Few studies on magnetic resonance imaging (MRI) only head and neck radiation treatment planning exist, and none using a generally available software. The aim of this study was to evaluate the accuracy of absorbed dose for head and neck synthetic computed tomography data (sCT) generated by a commercial convolutional neural network-based algorithm. MATERIALS AND METHODS: For 44 head and neck cancer patients, sCT were generated and the geometry was validated against computed tomography data (CT). The clinical CT based treatment plan was transferred to the sCT and recalculated without re-optimization, and differences in relative absorbed dose were determined for dose-volume-histogram (DVH) parameters and the 3D volume. RESULTS: For overall body, the results of the geometric validation were (Mean ± 1sd): Mean error -5 ± 10 HU, mean absolute error 67 ± 14 HU, Dice similarity coefficient 0.98 ± 0.05, and Hausdorff distance difference 4.2 ± 1.7 mm. Water equivalent depth difference for region Th1-C7, mid mandible and mid nose were -0.3 ± 3.4, 1.1 ± 2.0 and 0.7 ± 3.8 mm respectively. The maximum mean deviation in absorbed dose for all DVH parameters was 0.30% (0.12 Gy). The absorbed doses were considered equivalent (p-value < 0.001) and the mean 3D gamma passing rate was 99.4 (range: 95.7-99.9%). CONCLUSIONS: The convolutional neural network-based algorithm generates sCT which allows for accurate absorbed dose calculations for MRI-only head and neck radiation treatment planning. The sCT allows for statistically equivalent absorbed dose calculations compared to CT based radiotherapy.

Clinical validation of a commercially available deep learning software for synthetic CT generation for brain.

BACKGROUND: Most studies on synthetic computed tomography (sCT) generation for brain rely on in-house developed methods. They often focus on performance rather than clinical feasibility. Therefore, the aim of this work was to validate sCT images generated using a commercially available software, based on a convolutional neural network (CNN) algorithm, to enable MRI-only treatment planning for the brain in a clinical setting. METHODS: This prospective study included 20 patients with brain malignancies of which 14 had areas of resected skull bone due to surgery. A Dixon magnetic resonance (MR) acquisition sequence for sCT generation was added to the clinical brain MR-protocol. The corresponding sCT images were provided by the software MRI Planner (Spectronic Medical AB, Sweden). sCT images were rigidly registered and resampled to CT for each patient. Treatment plans were optimized on CT and recalculated on sCT images for evaluation of dosimetric and geometric endpoints. Further analysis was also performed for the post-surgical cases. Clinical robustness in patient setup verification was assessed by rigidly registering cone beam CT (CBCT) to sCT and CT images, respectively. RESULTS: All sCT images were successfully generated. Areas of bone resection due to surgery were accurately depicted. Mean absolute error of the sCT images within the body contour for all patients was 62.2 ± 4.1 HU. Average absorbed dose differences were below 0.2% for parameters evaluated for both targets and organs at risk. Mean pass rate of global gamma (1%/1 mm) for all patients was 100.0 ± 0.0% within PTV and 99.1 ± 0.6% for the full dose distribution. No clinically relevant deviations were found in the CBCT-sCT vs CBCT-CT image registrations. In addition, mean values of voxel-wise patient specific geometric distortion in the Dixon images for sCT generation were below 0.1 mm for soft tissue, and below 0.2 mm for air and bone. CONCLUSIONS: This work successfully validated a commercially available CNN-based software for sCT generation. Results were comparable for sCT and CT images in both dosimetric and geometric evaluation, for both patients with and without anatomical anomalies. Thus, MRI Planner is feasible to use for radiotherapy treatment planning of brain tumours.

Repeatability of T1-quantification in dGEMRIC for three different acquisition techniques: two-dimensional inversion recovery, three-dimensional look locker, and three-dimensional variable flip angle.

PURPOSE: To evaluate the repeatability of the dGEMRIC (delayed gadolinium enhanced MRI of cartilage) method in osteoarthritis-prone knee joints for three different T1 quantification techniques: two-dimensional inversion recovery (2D-IR), three-dimensional Look-Locker (3D-LL), and three-dimensional variable flip angle (3D-VFA). MATERIALS AND METHODS: Nine subjects were examined twice, with a 2-week interval, using all three measurement techniques. Four regions of interest were defined in the central medial and lateral femoral cartilage. The repeatability was evaluated for each measurement technique. For the 3D techniques, the variation between different slices was also evaluated. RESULTS: Repeatability expressed by root-mean-square coefficient of variation (CV(RMS)) showed similar results for 2D-IR and 3D-LL (5.4-8.4%). For 3D-VFA CV(RMS) was higher (9.3-15.2%). Intraclass correlation coefficient showed both 2D-IR and 3D-LL reliability to be moderate, while 3D-VFA reliability was low. Inter-slice CV(RMS) and ICC was of the same magnitude as the repeatability. No clear differences could be interpreted between the condyles. CONCLUSION: Both 2D-IR and 3D-LL perform well in generating repeatable dGEMRIC results, while 3D-VFA results are somewhat inferior. Furthermore, repeatability results in this study are similar to previously published results for healthy subjects. Finally, the positioning of the analyzed images is crucial to generate reliable repeatability results.

MR-OPERA: A Multicenter/Multivendor Validation of Magnetic Resonance Imaging-Only Prostate Treatment Planning Using Synthetic Computed Tomography Images.

PURPOSE: To validate the dosimetric accuracy and clinical robustness of a commercially available software for magnetic resonance (MR) to synthetic computed tomography (sCT) conversion, in an MR imaging-only workflow for 170 prostate cancer patients. METHODS AND MATERIALS: The 4 participating centers had MriPlanner (Spectronic Medical), an atlas-based sCT generation software, installed as a cloud-based service. A T2-weighted MR sequence, covering the body contour, was added to the clinical protocol. The MR images were sent from the MR scanner workstation to the MriPlanner platform. The sCT was automatically returned to the treatment planning system. Four MR scanners and 2 magnetic field strengths were included in the study. For each patient, a CT-treatment plan was created and approved according to clinical practice. The sCT was rigidly registered to the CT, and the clinical treatment plan was recalculated on the sCT. The dose distributions from the CT plan and the sCT plan were compared according to a set of dose-volume histogram parameters and gamma evaluation. Treatment techniques included volumetric modulated arc therapy, intensity modulated radiation therapy, and conventional treatment using 2 treatment planning systems and different dose calculation algorithms. RESULTS: The overall (multicenter/multivendor) mean dose differences between sCT and CT dose distributions were below 0.3% for all evaluated organs and targets. Gamma evaluation showed a mean pass rate of 99.12% (0.63%, 1 SD) in the complete body volume and 99.97% (0.13%, 1 SD) in the planning target volume using a 2%/2-mm global gamma criteria. CONCLUSIONS: Results of the study show that the sCT conversion method can be used clinically, with minimal differences between sCT and CT dose distributions for target and relevant organs at risk. The small differences seen are consistent between centers, indicating that an MR imaging-only workflow using MriPlanner is robust for a variety of field strengths, vendors, and treatment techniques.

Technical Note: MRI only prostate radiotherapy planning using the statistical decomposition algorithm.

PURPOSE: In order to enable a magnetic resonance imaging (MRI) only workflow in radiotherapy treatment planning, methods are required for generating Hounsfield unit (HU) maps (i.e., synthetic computed tomography, sCT) for dose calculations, directly from MRI. The Statistical Decomposition Algorithm (SDA) is a method for automatically generating sCT images from a single MR image volume, based on automatic tissue classification in combination with a model trained using a multimodal template material. This study compares dose calculations between sCT generated by the SDA and conventional CT in the male pelvic region. METHODS: The study comprised ten prostate cancer patients, for whom a 3D T2 weighted MRI and a conventional planning CT were acquired. For each patient, sCT images were generated from the acquired MRI using the SDA. In order to decouple the effect of variations in patient geometry between imaging modalities from the effect of uncertainties in the SDA, the conventional CT was nonrigidly registered to the MRI to assure that their geometries were well aligned. For each patient, a volumetric modulated arc therapy plan was created for the registered CT (rCT) and recalculated for both the sCT and the conventional CT. The results were evaluated using several methods, including mean average error (MAE), a set of dose-volume histogram parameters, and a restrictive gamma criterion (2% local dose/1 mm). RESULTS: The MAE within the body contour was 36.5 ± 4.1 (1 s.d.) HU between sCT and rCT. Average mean absorbed dose difference to target was 0.0% ± 0.2% (1 s.d.) between sCT and rCT, whereas it was -0.3% ± 0.3% (1 s.d.) between CT and rCT. The average gamma pass rate was 99.9% for sCT vs rCT, whereas it was 90.3% for CT vs rCT. CONCLUSIONS: The SDA enables a highly accurate MRI only workflow in prostate radiotherapy planning. The dosimetric uncertainties originating from the SDA appear negligible and are notably lower than the uncertainties introduced by variations in patient geometry between imaging sessions.

Local flip angle correction for improved volume T1-quantification in three-dimensional dGEMRIC using the Look-Locker technique.

PURPOSE: To present an evaluation method for three-dimensional Look-Locker (3D-LL) based T1 quantification, calculating correct T1 values independent of local flip angle (FA) variations. The method was evaluated both in phantoms and in vivo in a delayed Gadolinium Enhanced MRI of Cartilage (dGEMRIC) study with 33 subjects. MATERIALS AND METHODS: T1 was measured with 3D-LL, using both local FA correction and a precalculated FA slice profile, and compared with standard constant FA correction, for all slices in phantoms and in both femur condyles in vivo. T1 measured using two-dimensional Inversion Recovery (2D-IR) was used as gold standard. RESULTS: Due to the FA being slice dependent, the standard constant FA correction results in erroneous T1 (systematic error = 109.1 ms in vivo), especially in the outer slices. With local FA correction, the calculated T1 is excellent for all slices in phantoms (<5% deviation from 2D-IR). In vivo the performance is lower (systematic error = -57.5 ms), probably due to imperfect inversion. With precalculated FA correction the performance is very good also in vivo (systematic error = 13.3 ms). CONCLUSION: With the precalculated FA correction method, the 3D-LL sequence is robust enough for in vivo dGEMRIC, even outside the centermost slices.