MR Deep Resolve - Numaris X Online Training

Continue Continue Continue Continue Continue Continue Continue Deep Resolve Master Template HILS2218 | Effective Date: 25 Mar 2022 MR Deep Resolve – Numaris X Online Training This online training focuses on how MR Deep Resolve accelerates MRI examinations, boosts efficiency, while improving the signal-to-noise ratio. The benefits and differences between Deep Resolve Gain, Deep Resolve Sharp, Deep Resolve Boost, and Deep Resolve Swift Brain will be reviewed. Deep Resolve Gain 1 Deep Resolve Boost 3 Deep Resolve Sharp 2 Deep Resolve Swift Brain 4 ? Welcome Audio: This online training focuses on how M-R Deep Resolve accelerates M-R-I examinations, boosts efficiency, while improving the signal-to-noise ratio. The benefits and differences between Deep Resolve Gain, Sharp, Boost, and Deep Resolve Swift Brain will be reviewed. ? MR Deep Resolve Deep Resolve Gain Deep Resolve Sharp Deep Resolve Boost MR Deep Resolve (DR) Audio: Deep Resolve is an acceleration and AI-powered advanced image reconstruction technology with the following variants: Deep Resolve Gain Deep Resolve Sharp Deep Resolve Boost Deep Resolve1 Improving Challenges in Acquisition and Reconstruction ? Deep learning reconstruction simultaneously improves: 1 Optional license Image Resolution Signal-to-Noise Ratio Acquisition Time DR Improving Challenges Audio: Deep Resolve aids in improving the everyday challenges Technologist face when acquiring and reconstructing image data. With Deep Resolve, the Deep Learning Reconstruction simultaneously improves Resolution, and Signal-to-Noise, enabling a time savings during data acquisition time. Deep Resolve Gain (TSE) Intelligent Denoising Deep Resolve Sharp1 (TSE) Deep Learning Reconstruction Deep Resolve 1 Optional license Two Applications – Deep Resolve Gain and Deep Resolve Sharp ? DR Gain & DR Sharp Audio: There are two applications that can be used with Deep Resolve, Deep Resolve Gain and Deep Resolve Sharp. Deep Resolve Gain, addresses locally varying noise with intelligent denoising, resulting in higher signal-to-noise ratio. While Deep Revolve Sharp, uses a deep neural network to increase the image sharpness and resolution. Let's take a closer look at these two innovative techniques. ? MR Deep Resolve Deep Resolve Gain Deep Resolve Sharp Deep Resolve Boost MR Deep Resolve - DR Gain What is it? ? Deep Resolve Gain (DR Gain) Deep Resolve Gain is an intelligent reconstruction method that uses acquired MR data to generate a noise map that reflects spatial noise variations. MR image and corresponding noise map are used as an input for an iterative process. Results in an intelligent denoised image with a higher signal-to-noise ratio (SNR). Low SNR input data High SNR output data DR Gain - What is it? Audio: In the Deep Resolve family, different levels of the model-based approach were applied for different tasks. Deep Resolve Gain is an intelligent reconstruction method that uses the acquired M-R data to generate a noise map which reflects spatial noise variations and indicates which regions have high and which regions have low noise levels. Then, the M-R image and the corresponding noise map are used as input for an intelligent denoising process. This means the noise will be reduced in an iterative process. This denoising effect is stronger in regions with low SNR compared to regions with a high SNR. The result is an intelligent denoised image with a higher signal-to-noise-ratio. Deep Resolve Gain (DR Gain) What problem does it solve? ? This impairs SNR and decreases image quality. The noise has local variations – and conventional standard noise filters do not address that. Coil array geometry and acceleration techniques cause noise in MRI images. With this, we bring the iterative principle of Compressed Sensing to a broader use. Address local noise variations This is the first time we target the local variation in noise. Broad application It is applicable for 2D and 3D cartesian sequences. DR Gain applies intelligent, localized denoising. DR Gain - What problem does it solve? Audio: Coil array geometry and acceleration techniques create noise in the M-R images. Since the noise in M-R images is not evenly distributed, coil geometries and acceleration techniques can amplify the noise in central locations in the M-R images. This usually results in local signal-to-noise variations. It is important that these are considered during denoising since the additional noise reduces signal-to-noise ratio, and effects image quality. This is where Deep Resolve Gain is a benefit, since it uses a targeted denoising method to increase Signal-to-Noise ratio in the M-R images. The acquired M-R-I data is used to generate a noise map which reflects spatial noise variations. The noise map is used as input for an iterative process, reducing noise in a targeted manner, thus addressing the local variations in image noise, due to the coil array geometry. Deep Resolve Gain (DR Gain) How does it work? ? Noisy Image Data-Specific Noise Map Final Image DR Gain - How does it work? Audio: Let’s look at how Deep Resolve Gain works? Deep Resolve Gain uses the acquired M-R data to generate noise maps which reflect spatial noise variations. These noise maps are used as prior information in iterative reconstruction processes, like compressed sensing. They are generated without spending additional scan time and are determined using scan-specific adjustment data, acquired as part of the normal imaging process.  The reconstruction algorithm takes local noise variations into account and enables stronger denoising where noise would be most dominant when reconstructing with conventional methods. This results in improved image quality and higher signal-to-noise ratio. No neural networks are used in Deep Resolve Gain. . ? MR Deep Resolve Deep Resolve Gain Deep Resolve Sharp Deep Resolve Boost MR Deep Resolve - DR Sharp ? Deep Resolve Sharp (DR Sharp) What is it? AI Deep Neural Network Cross-check with acquired data Low Resolution Input Data High Resolution Reconstruction k-space Data Input DR Sharp - What is it? Base Layer Audio: Another Deep Resolve Reconstruction technique is Deep Resolve Sharp.   It uses a deep convolutional neural network to increase image sharpness and improve image quality. With Deep Resolve Sharp you can acquire high resolution images with shorter scan times.   The Neural network at the core of Deep Resolve Sharp was trained by comparing a vast amount of high and low-resolution M-R data. Within the reconstruction pipeline Deep Resolve Sharp generates a high-resolution image from a low-resolution input. By cross-checking the high-resolution image with the acquired data, clinically robust results are achieved.   Deep Resolve Sharp is always performed in combination with Deep Resolve Gain, the result is an image with a high signal-to-noise ratio, sharp edges, and high resolution. Deep Resolve Sharp What problem does it solve? ? Well-trained It was trained on large amounts of pairs of low-res and high-res images. AI & deep learning The neural network is able to anticipate where to expect a sharp edge in an image. Hi-res from low-res The neural network generates high-res output from low-res input. Consistency ensured Consistency with raw data is ensured by a cross check. We can either have a longer scan that acquires a lot of data and gives us high resolution. Or: A quicker scan, less data - but lower resolution. DR Sharp - What problem does it solve? Audio: In M-R we can either have a longer scan time that acquires a lot of data and give us high resolution, or, a faster scan that acquires less data, but has lower resolution. Deep Resolve Sharp enables high-resolution images with shorter scan times consistent with low-resolution datasets. Deep Resolve Sharp uses Artificial Intelligence and a deep learning neural network to be able to anticipate where to expect sharp edges in the image. Enabling high-resolution images to be reconstructed from low-resolution images all the while ensuring consistency through cross-checking the acquired k-space data. How does it work? ? Deep Resolve Sharp Original Image Deep Neural Network Network Image Final Image Reinforce IQ with Original Data Click here for additional information on network training DR Sharp - How does it work? Audio: Now, Let’s look at how Deep Resolve Sharp works? During image acquisition, the number of phase and frequency encoding scans can be reduced, thus reducing the amount of data needed to acquire and speed up scan time. The data with reduced phase and frequency encoding scans are used to generate a low-resolution image. Then Artificial Intelligence comes into play, and the low-resolution image is passed as input to a deep neural network that has been trained on vast numbers of pairs of low- and high-resolution M-R data. The network generates a high-resolution image and collates it with the original low-resolution data. The result is an image with sharp edges and high resolution, which is still fully consistent with the acquired data. Deep Resolve Sharp offers up to a factor of two in-plane-resolution. Click the button on the left to view additional information on how deep resolve sharp was trained. DRS_Training Layer: The network in Deep Resolve Sharp was trained and validated on more than 10,000 images covering a broad range of contrasts, body regions, sampling patterns, and field strengths. The performance was characterized by numerous quality metrics including peak signal-to-noise ratio, structural similarity index, and perceptual loss. Table with 2 columns and 4 rows Dataset Size > 10K images Dataset Composition Broad range of different body regions, contrasts, fat suppression techniques, orientations, and field strengths (1.5T and 3T). Evaluation Criteria Quality metrics included PSNR, SSIM, and perceptual loss. Testing also included visual rating and evaluation of image sharpness by intensity profile comparisons of reconstructions with and without DR Sharp Reference Standard Input data retrospectively generated from ground-truth via data manipulation and augmentation, including downsampling (i.e., removal of high-resolution k-space data). How was it trained? The network in Deep Resolve Sharp was trained and validated on more than 10,000 images covering a broad range of contrasts, body regions, sampling patterns, and field strengths. The performance was characterized by numerous quality metrics including peak signal-to-noise ratio, structural similarity index, and perceptual loss. ? Deep Resolve Gain & Sharp Applications Practical Considerations Image Examples DR Gain & Sharp - Applications ? Deep Resolve – Protocol Transformation How to transform a protocol into a Deep Resolve protocol Under the Resolution parameter card, click on the Acceleration card. Step 2 Step 2 Step 3 Step 3 Step 4 Step 4 Tab 1 Tab 1 Step 1 Click the numbered tabs to view each step. DR Protocol Transformation Audio: To modify a protocol and add Deep Resolve Gain and Sharp, select the Resolution, Acceleration parameter card. Click on the Deep Resolve drop-down menu and select the option On. Deep Resolve Gain and Sharp can only be combined with specific image filters. Once Deep Resolve is selected other filters may be deactivated. If additional changes are required, the Scan Assistant window opens and displays the required filter changes.  If the Deep Resolve Sharp license is available, it is automatically selected when Deep Resolve is activated. The user can use Deep Resolve Sharp along with Deep Resolve Gain, or they can use Deep Resolve Gain without Deep Resolve Sharp. Step 4 If the Deep Resolve Sharp license is available, it is automatically selected when Deep Resolve is activated. NOTE: The user has the option use DR Sharp along with DR Gain or use DR Gain only. Step 1 Step 1 Step 3 Deep Resolve Gain and Sharp can only be combined with specific image filters. Other filters might get deactivated. If additional changes are required, the Scan Assistant window opens and shows the required filter changes. Step 1 Step 1 Step 2 In the Deep Resolve drop-down menu, select . Step 1 Step 1 ? Deep Resolve Gain – Application Strategies 1 1 1 2 2 2 3 3 3 4 4 4 5 5 5 6 6 6 7 7 7 How to apply Deep Resolve Gain without Deep Resolve Sharp How to apply Deep Resolve Gain without Deep Resolve Sharp How to improve image quality with less noise and more signal-to-noise ratio How to improve image quality with less noise and more signal-to-noise ratio Improve image quality and resolution without changing scan time Improve image quality and resolution without changing scan time DR Gain - Application Strategies Audio: 0 App Strategies Select the numbered icons below to view application strategies such as how to apply Deep Resolve Gain without deep resolve sharp, how to improve image quality with less noise and more signal to noise ratio, as well as improve image quality and resolution without changing scan time. 1 App of DR G w/o DR S 1 of 2 Let’s take a look at some application strategies that can be used with Deep Resolve. First, we will explore how to use Deep Resolve Gain without Deep Resolve Sharp. Once Deep Resolve has been activated, you can modify the denoising strength and edge enhancement to optimize the effects of Deep Resolve Gain. Select the Resolution then Acceleration Parameter cards. Click the ellipsis next to the Deep Resolve drop down menu. 2 App of DR G w/o DR S 2 of 2 For the Denoising Strength, the lowest setting is one and the highest is eight. For the Edge Enhancement, the lowest setting is one and the highest is five. Remember, a higher Denoising Strength results in an increase in Signal-to-Noise. In some cases, using Deep Resolve Gain may lead to a decrease in the perceived image sharpness, however, increasing the edge enhancement and image resolution may help. Another suggestion would be too lower the denoising strength to reduce image smoothing and a fake image appearance. If the Deep Resolve Sharp license is available and you want to use Deep Resolve Gain without Deep Resolve Sharp, remove the checkmark from the Deep Resolve Sharp checkbox. 3 Increase image quality w/ less noise Next let’s look at how to increase image quality with less noise and more signal-to-noise when using Deep Resolve Gain. To increase image quality with less noise and more signal-to-noise. First, select the Resolution, Acceleration parameter card, select Deep Resolve, then in the drop-down menu select On. Next, click on the ellipsis next to the drop-down menu to open the Deep Resolve window and modify the default denoising value. A higher Denoising Strength results in higher Signal-to-noise. Please note: In some cases, using Deep Resolve Gain may lead to a decrease in the perceived image sharpness. Applying increased edge enhancement and increasing the image resolution can be helpful here. You can use a lower denoising strength to reduce image smoothing. 4 Improve image quality & resolution w/o changing scan time 1 of 2 Now let’s look at how to improved image quality and resolution without changing the scan time with Deep Resolve Gain and Deep Resolve Sharp. To improve image quality and resolution without changing scan time. First, select Deep Resolve in the drop-down menu, and select On. Next, click on the ellipsis next to the drop-down menu, and choose the Denoising Strength and Enhancement Levels in the Deep Resolve window. 5 Improve image quality & resolution w/o changing scan time 2 of 2 These image examples in the lumbar spine, show the benefit of using Deep Resolve Gain and Sharp on this T-two, Turbo Spin Echo sequence. Note the increase in resolution with the higher matrix size without an increase in scan time when Deep Resolve Gain and Sharp were implemented, especially, when compared to the traditional T-two, Turbo Spin Echo sequence. 6 Improve IQ & reduce scan time 1 To improve the image quality and reduce scan time. First, select Deep Resolve in the drop-down menu. Next, click on the ellipsis next to the drop-down menu and choose the denoising strength and enhancement levels in the Deep Resolve window. Then open the Resolution, Common parameter card and change the value in the Base Resolution field to reduce both the matrix size and scan time. Additional adjustments could include decreasing the number of Averages, or increasing the Acceleration Factor, however, these adjustments may cause some artifacts to become more severe. Remember is it always important to avoid any significant signal-to-noise loss. 7 Improve IQ & reduce scan time 2 These image examples show the benefit of using Deep Resolve Gain and Sharp for a Proton Density, Turbo Spin Echo sequence with Fat Saturation. Note the reduction in scan time, reduced averages, and higher Matrix size when Deep Resolve Gain and Sharp were implemented, especially, when compared to the traditional Proton Density Turbo Spin Echo sequence with Fat Saturation. Improve image quality and reduce scan time (2 of 2) 7 1 1 1 2 2 2 3 3 3 4 4 4 5 5 5 6 6 6 7 7 7 Original: MAGNETOM Vida, 3T PD TSE fs, 2 Av, TA 3:33. Matrix size: 307 x 512 Deep Resolve Gain & Sharp: MAGNETOM Vida PD TSE fs, 1 Av, TA 1:50 Matrix size: 608 x 608 Improve image quality and reduce scan time (1 of 2) 6 From the Resolution / Acceleration parameter cards, select On in the Deep Resolve field. Click the ellipsis button to open the Deep Resolve window. Select the denoising strength and enhancement levels. Select the Resolution then Common parameter cards. Change the value in the Base Resolution field to reduce both the matrix size and scan time. Decreasing the number of Averages or increasing the Acceleration Factor may cause some artifacts to become more severe. Improve image quality and resolution without changing scan time (2 of 2) 5 1 1 1 2 2 2 3 3 3 4 4 4 5 5 5 6 6 6 7 7 7 Original: MAGNETOM Sola T2 TSE, TA 3:45 min. Matrix size: 256 x 320 Deep Resolve Gain & Sharp: MAGNETOM Sola T2 TSE, TA 3:45 min. Matrix size: 512 x 640 Improve image quality and resolution without changing scan time (1 of 2) 4 Select from the Deep Resolve drop list menu. Click on the ellipsis next to the drop-down menu. From the Deep Resolve window, choose the Denoising Strength and Enhancement Levels. Increase image quality with less noise and more SNR 3 Select the Resolution and Acceleration parameter cards. Select from the Deep Resolve drop list. Click the ellipsis to open the Deep Resolve window. Modify the default denoising value. Keep in mind, a higher Denoising Strength results in higher SNR. Use lower denoising strengths to reduce image smoothing. NOTE: In some cases, using DR Gain may lead to a decrease in the perceived image sharpness. Applying increased edge enhancement and increasing the image resolution can be helpful. Application of DR Gain without DR Sharp (2 of 2) 2 Adjust the following fields: Denoising Strength: low (=1) to high (=8) Edge Enhancement: low (=1) to high (=5) If the Deep Resolve Sharp license is available, de-select the Deep Resolve Sharp checkbox to use DR Gain without DR Sharp. Application of DR Gain without DR Sharp (1 of 2) 1 After Deep Resolve has been activated, you can adjust the denoising strength and the edge enhancement to fine-tune the effects of Deep Resolve Gain. ? Deep Resolve Gain & Sharp Applications Practical Considerations Image Examples DR Gain & Sharp - Practical Considerations Deep Resolve Gain Retrospective Reconstruction (1 of 3) ? Deep Resolve Gain settings can be modified after the scan through Retrospective Reconstruction to trial and fine-tune Deep Resolve settings. Once the images have been acquired, select the Job View icon in the top right corner. Select the MR Reconstructions tab. The protocol list opens. Select the required protocol, then select Edit to edit a protocol for Retro Recon. DR Gain - Retro Recon 1 Audio: Deep Resolve Gain settings can be modified after the scan through Retrospective Reconstruction to trial and fine-tune Deep Resolve settings. Once the images are acquired, go to the upper right corner and select Job View, then select the MR Reconstructions tab. The Protocol list opens, next select the required protocol and click on the Edit button to edit a protocol for Retro-Recon. Deep Resolve Gain Retrospective Reconstruction (2 of 3) ? NOTE: You can perform a Retro Recon with each data set that is generated with or without DRG. Under certain circumstances, data sets generated without DRG cannot be reconstructed with Retro Recon. Under the Resolution parameter card, select the Acceleration card. Click on the ellipsis button next to Deep Resolve. Adjust the Deep Resolve values associated with the protocol. After adapting the parameters, add the adapted parameters by selecting the arrow on the right side of the window. The updated parameter list is displayed below. Adding an image comment may also be helpful. Select the Start button to close the protocol. The adapted protocol is listed in the protocol list. A green indicator shows the changes that have been saved. DR Gain - Retro Recon 2 Audio: Next, under the Resolution, Acceleration parameter card, click on the ellipsis next to Deep Resolve and adjust the Deep Resolve values used for that protocol. Please note: You can perform a Retro-Recon with each data set that is generated with or without Deep Resolve Gain. Under certain circumstances data sets that are generated without Deep Resolve Gain cannot be reconstructed using Retro-Recon. After adapting the parameters, add the parameters by selecting the arrow on the right side of the window. The updated parameter list is displayed below, adding an image comment may also be helpful. Next, click on the Start button to close the protocol. The adapted protocol is listed in the protocol list and a green indicator shows that the changes have been saved. ? Deep Resolve Gain Retrospective Reconstruction (3 of 3) Interpolation without DR Sharp NOTE: Retro Recon cannot be used when actively running a scan. DR Sharp Denoising: 4 Edge Enhancement: 2 DR Sharp Denoising: 8 Edge Enhancement: 5 DR Gain - Retro Recon 3 Audio: These images compare: Deep Resolve Sharp without Interpolation, Deep Resolve Sharp with a Denoising Value of four and Edge Enhancement of two, And, Deep Resolve Sharp with a Denoising value of eight and an Edge Enhancement of five. Please note, Retro Recon cannot be used when actively running a scan. ? Deep Resolve Sharp Gibbs Ringing Artifacts In TSE protocols, when using Deep Resolve Sharp with high TE values (~>80ms), ringing artifacts can occur. This depends on the TE, echo-spacing, and Turbo Factor selected. To remove the ringing for these protocols, calculate a new Turbo Factor using this formula: (TE * 2 / echo-spacing) -1 DR Sharp - Gibbs Ringing Artifacts Audio: When using Deep Resolve Sharp in a Turbo Spin Echo sequence, if the T-E values are equal-to-or-greater-than eighty milliseconds, ringing artifacts may occur. This not only depends on the T-E, but also the echo-spacing and Turbo factor selected. The root cause of the ringing artifacts is the reordering pattern of the Turbo Spin Echo sequence. Especially, when using some combinations of T-E, echo-spacing and Turbo Factor with the Deep Resolve Sharp technique. To remove the ringing artifacts for these protocols, calculate a new Turbo-Factor using this formula: Take the T-E, times-two, then divide by the Echo-spacing, and subtract one. Then enter the new Turbo Factor in the protocol. ? Deep Resolve Practical Considerations for DR Gain & DR Sharp Use lower denoising strengths to reduce image smoothing and a change in contrast. Decreasing the number of averages or increasing the acceleration factor can cause a decrease in Signal-to-Noise Ratio (SNR). Deep Resolve Sharp may enhance existing, and sometimes hidden artifacts and bring back a little bit of noise. Good starting parameters are key to good results. Image results may be poor if using a sequence that has a lower SNR. If time savings is the primary goal, allow the Deep Learning Neural Network of Deep Resolve Sharp to provide the required resolution by using a lower starting resolution. Practical Considerations for DR Gain & DR Sharp Audio: The image shown here is an example of using inadequate starting parameters for the Deep Resolve Gain and Sharp sequence. The suggestions listed can help avoid poor image quality results. Use a lower denoising strength to reduce image smoothing and a change in contrast. Decreasing the number of averages or increasing the acceleration factor can cause a decrease in the signal to noise ratio. Deep Resolve Sharp may enhance existing, and sometimes hidden artifacts and bring back a little bit of noise.  Good starting parameters are key to good results. It is important to note, that the image results may be poor if you use a sequence that has a lower SNR. If time savings are the primary goal, allow the Deep Learning Neural Network of Deep Resolve Sharp to provide the required resolution by using a lower starting resolution. ? Deep Resolve Gain & Sharp Applications Practical Considerations Image Examples DR Gain & Sharp - Image Examples Deep Resolve DR Gain + DR Sharp - Brain ? DR Gain & Sharp - Brain Audio: This Brain examination was performed on a MAGNETOM Altea. The scan times compare a traditional customer whole brain protocol with both Deep Resolve Gain and Deep Resolve Sharp implemented. The overall scan time is forty-nine percent faster for the Deep Resolve protocol compared to the traditional customer protocol. Deep Resolve DR Gain + DR Sharp - Lumbar ? DR Gain & Sharp - Lumbar Audio: This Lumbar examination was performed on a MAGNETOM Altea. The scan times compare a traditional customer whole spine protocol with both Deep Resolve Gain and Deep Resolve Sharp implemented. The overall scan time is forty-one percent faster for the Deep Resolve protocol compared to the traditional customer protocol. Deep Resolve DR Gain + DR Sharp - Knee ? DR Gain & Sharp - Knee Audio: This Knee examination was performed on a MAGNETOM Altea. The scan times compare a traditional customer whole knee protocol with both Deep Resolve Gain and Deep Resolve Sharp implemented. The overall scan time is fifty percent faster for the Deep Resolve protocol compared to the traditional customer protocol. ? MR Deep Resolve Deep Resolve Gain Deep Resolve Sharp Deep Resolve Boost MR Deep Resolve - DR Boost ? Deep Resolve Boost1,2 What is it? 1 Optional License 2 Prerequisite: High-end MaRS Deep Resolve Boost uses a raw-data-to-image deep learning reconstruction technology. The neural network behind Deep Resolve Boost enables effective denoising that improves the visual signal-to-noise ratio (SNR) of the images. Deep Resolve Boost significantly reduces scan times for Turbo Spin Echo (TSE) sequences and is applicable for head-to-toe applications. Deep Resolve Boost can be combined with Simultaneous Multi-Slice (SMS TSE). Original: PAT 1, TA 2:12 28 Slices, 0/4 x 0.4 x0.4 mm3 Deep Resolve Boost & Sharp Original: PAT 4, TA 36s 28 Slices, 0.2 x 0.2 x 0.4 mm3 DR Boost - What is it? Audio: What is Deep Resolve Boost? Deep Resolve Boost uses a raw-data-to-image deep learning reconstruction technology. The neural network behind Deep Resolve Boost enables effective denoising that improves the visual signal-to-noise ratio of the images. Deep Resolve Boost significantly reduces scan times for Turbo Spin Echo sequences and is applicable for head-to-toe applications. This k-space-to-image-space Deep Learning Reconstruction can also be combined with Simultaneous Multi-Slice Turbo Spin Echo . ? Deep Resolve Boost1,2 What problem does it solve? 1 Optional License 2 Prerequisite: High-end MARS Deep Resolve Boost Reconstruction enables the generation of images with higher signal-to-noise ratios (SNRs) and faster image acquisition simultaneously. Higher SNR Ì Faster Acquisition Times Conventional Reconstruction Methods a highly accelerated image acquisition will lead to strong noise contamination. DR Boost - What problem does it solve? Audio: In M-R with conventional reconstruction methods, a highly accelerated image acquisition leads to strong noise contamination. Deep Resolve Boost reconstruction, generates images with higher signal-to-noise ratios and faster image acquisition simultaneously ? Deep Resolve Boost1,2 How does it work? 1 Optional License 2 Prerequisite: High-end MaRS Raw data input from accelerated acquisition Initial Recon Deep Neural Network Reinforce IQ with raw data Repeat n-times Final Image Click here for additional information on network training DR Boost - How does it work? Audio: A more advanced M-R-I acceleration using deep learning approach is applied by Deep Resolve Boost. An initial image is generated from highly under-sampled data using a standard (SENSE) parallel image scheme. This initial image is then passed as the input to an Unrolled Image Reconstruction Network. The unrolled recon repeatedly alternates between deep neural-network denoising and data consistency updates, (a pattern very similar to the reconstruction process used in Compressed Sensing). The neural-network updates drive high Image Quality by letting the reconstruction take advantage of information learned from datasets made of the raw data from a vast number of images, and the data-consistency updates incorporate the M-R physics of the acquisition process, reintegrating the measured data into the images to ensure high-consistency. The result, images with substantially reduced noise. The integration of the raw data along the entire reconstruction process preserves the consistency with the measured data. Please note, a pre-requisite for Deep Resolve Boost license is the High-end MaRS. Click the button on the right to view additional information on how deep resolve boost was trained. Layer 1: The network in Deep Resolve Boost was trained and validated on more than 25,000 images covering a broad range of contrasts, body regions, and field strengths. The performance was characterized by numerous quality metrics including peak signal to noise ratio and structural similarity index. The neural network in DR Boost was trained and validated on more than 25,000 images covering a broad range of contrasts, body regions, and field strengths. The network in Deep Resolve Boost was trained and validated on more than 25,000 images covering a broad range of contrasts, body regions, and field strengths. The performance was characterized by numerous quality metrics including peak signal-to-noise ratio and structural similarity index. Table with 2 columns and 4 rows Dataset Size > 25K Dataset Composition Broad range of different body regions, contrasts, fat suppression techniques, orientations, and field strengths (1.5T and 3T). Evaluation Criteria Quality metrics included PSNR and SSIM. Radiological quality verification was also performed in a clinical setting on >400 patients. Reference Standard Input data retrospectively generated from ground-truth via data manipulation and augmentation, including: undersampling, noise addition, and data mirroring. How was it trained? ? Deep Resolve Boost1,2 Parameter Card Hint: GRAPPA must be used with Boost. If GRAPPA is not already in use, the Scan Assistant shows the mandatory change. To activate Deep Resolve Boost Select the Resolution, Acceleration parameter card. Select Deep Resolve, then in the drop-down menu select On. Click on the ellipsis next to the drop-down menu to open the Deep Resolve window. Select Boost from the Denoising Mode drop-down menu. 1 Optional License 2 Prerequisite: High-end MARS DR Boost - Parameter Card Audio: To activate Deep Resolve Boost in your sequence, open the sequence and select the Resolution, Acceleration parameter card. Next, select Deep Resolve, then in the drop-down menu select On. Click on the ellipsis next to the drop-down menu to open the Deep Resolve window. Then select the Denoising Mode, Boost from the drop-down menu. Please note: GRAPPA must be used with Deep Resolve Boost. If GRAPPA is not already in use, the Scan Assistant will show the mandatory changes. ? Deep Resolve Boost1,2 1 Optional License 2 Prerequisite: High-end MARS Benefits Enables effective denoising that improves the visual signal-to-noise ratio (SNR). Reduces scan times for Turbo Spin Echo (TSE) sequences. Combines Deep Resolve Boost and Simultaneous Multi-Slice (SMS TSE). Adapts the Denoising Strength by choosing from three levels. Low, Medium, and High Head-to-toe applications in clinical routines without compromising image quality. (e.g., neuro, spine, and MSK) DR Boost - Benefits Audio: Some of the benefits of Deep Resolve Boost include: Effective denoising that improves the visual signal-to-noise ratio. Reduced scan times for Turbo Spin Echo sequences. Combining Deep Resolve Boost with Simultaneous Multi-Slice Adapting the Denoising Strength by choosing from three levels, low, medium, and high. And the ability to perform head-to-toe applications for routine clinical examinations without compromising image quality. For example, neuro, spine, and musculoskeletal. ? Deep Resolve Boost Practical Considerations Image Examples DR Boost - Practical Considerations ? Deep Resolve Boost Practical Considerations (1 of 2) Data on file Deep Resolve Boost uses a SENSE-like reconstruction, therefore, the FoV must be larger than the object to prevent aliasing of tissue from outside the FoV. Sufficient phase over-sampling can also be applied to prevent aliasing. Consider using additional phase over-sampling with higher acceleration factors. GRAPPA Deep Resolve Boost DR Boost - Practical Considerations 1 Audio: Deep Resolve Boost uses a SENSE-like reconstruction; therefore, the Field-of-View must be larger than the object to prevent aliasing from tissue that is outside the Field-of-View. To prevent aliasing, apply sufficient phase over-sampling to the sequence, especially when using higher acceleration factors. ? Deep Resolve Boost Practical Considerations (2 of 2) It is important to consider the acceleration reference scan type when using Deep Resolve Boost.  Streaking, banding or shading artifacts are all possible with poor selection of the reference scan type.  Integrated reference scan is preferred in most Body examinations. TSE/Separate reference scan may perform better in Neuro examinations. TSE/Separate reference scan is the default selection when SMS acceleration is selected. TSE/Separate Reference Scan Integrated Reference Scan DR Boost - Practical Considerations 2 Audio: It is also important to consider what acceleration reference scan type is being used with Deep Resolve Boost.  Streaking, banding, or shading artifacts are all possible when the incorrect reference scan type is selected.  For most body examinations, Integrated Reference Scan is preferred. However, for Neuro examinations, T-S-E, Separate Reference Scan may perform better. If Simultaneous Multi-Slice acceleration is applied, then T-S-E, Separate Reference Scan is automatically selected. ? Deep Resolve Boost Practical Considerations Image Examples DR Boost - Image Examples Deep Resolve Boost Brain ? DR Boost - Brain Audio: This whole brain examination without contrast was performed on a MAGNETOM Vida in under ten-minutes. Deep Resolve Boost Cervical Spine ? DR Boost - Cervical Spine Audio: This Cervical Spine examination was performed on a MAGNETOM Vida in just over five-minutes. Deep Resolve Boost Thoracic Spine ? DR Boost - Thoracic Spine Audio: This Thoracic Spine examination was performed on a MAGNETOM Vida in just under six-minutes. Deep Resolve Boost Lumbar Spine ? DR Boost - Lumbar Spine Audio: This Lumbar Spine examination was performed on a MAGNETOM Vida in just over six-minutes. Deep Resolve Boost Shoulder ? DR Boost - Shoulder Audio: This Shoulder examination was performed on a MAGNETOM Vida in just over six-minutes. ? Deep Resolve Boost and SMS TSE Faster Knee Exams PD TSE fs p3 s2 0.2 x 0.2 x 3 mm3 TA: 39 sec PD TSE fs p3 s2 0.2 x 0.2 x 3 mm3 TA: 36 sec PD TSE fs p3 s2 0.2 x 0.2 x 3 mm3 TA: 28 sec PD TSE fs p3 s2 0.2 x 0.2 x 3 mm3 TA: 15 sec Total Scan Time: 1:58 min DR Boost - SMS TSE Audio: Acquire a knee examination in under two-minutes, simply by incorporating Simultaneous Multi-Slice into your Deep Resolve Boost Turbo Spin Echo protocol. ? 1Requires optional license 2High-end computing required Deep Resolve Boost for EPI Diffusion1,2 What is it? What is it? Highlights & Benefits Highlights & Benefits Example Example PEPconnect Resources PEPconnect Resources Select each box for more information. DRB for EPI Diffusion Audio: Deep Resolve Boost and Deep Resolve Sharp are now available for the E-P-I Diffusion sequence. Deep Resolve Boost takes Deep Learning Reconstruction to the next level. Highlights and Benefits include: The ability to improve image quality and lesion conspicuity and reduce acquisition time for body Diffusion sequences. Using Deep Resolve on an E-P-I Diffusion sequence and combining it with simultaneous multi-slice. The capability to perform bi-parametric PI-RADS two-point-one, Prostate examinations in six-minutes or less with Deep Resolve. Example Layer These image examples compare a T-two Turbo Spin Echo with Deep Resolve to a Diffusion with ZOOM-it-Pro and Deep Resolve. ? 1Requires optional license 2High-end computing required TA - 0:58 min PAT 3 TA - 0:58 min PAT 4 TA - 1:02 min PAT 4 b50 b800 Calc b1400 ADC T2 TSE Deep Resolve 0.3 x 0.3 x 3.0 mm3 TA = 2:58 min Diffusion with ZOOMitPRO Deep Resolve 0.8 x 0.8 x 3.0 mm3 TA = 2:02 min Deep Resolve Boost for EPI Diffusion1,2 Online Training: MR Deep Resolve - Numaris X Job Aid: Deep Resolve Boost - USA PEPconnect Resources Improve image quality and lesion conspicuity Reduce acquisition time for body Diffusion sequences Deep Resolve for EPI Diffusion is combinable with Simultaneous Multi-Slice (SMS) Perform bi-parametric PI-RADS 2.1 Prostate examinations in 6 minutes or less with Deep Resolve Highlights & Benefits Deep Resolve Boost and Deep Resolve Sharp are now available for the EPI Diffusion sequence. Deep Resolve Boost takes Deep Learning Reconstruction to the next level. What is it? ? 1Requires optional license 2High-end computing required Deep Resolve Boost for HASTE1,2 What is it? What is it? Highlights & Benefits Highlights & Benefits PEPconnect Resources PEPconnect Resources Select each box for more information. Example Example DRB for HASTE Audio: Deep Resolve Boost and Deep Resolve Sharp are now available for the HASTE sequence taking Deep Learning Reconstruction to the next level. Achieve excellent image quality, diagnostic confidence, and lesion detection for HASTE body imaging with Deep Resolve HASTE. Highlights and Benefits include: the ability to enable higher i-PAT factors, resulting in decreased scan times. The use of shorter Echo Train durations and reduced T-two blurring Reducing acquisition times for body T-two-weighted sequences, in a single breath-hold Example Layer This t-2 BLADE fat sat sequence was acquired with 4 breath holds for a total of one minute and 56 seconds. When compared to the Deep Resolve HASTE sequence acquired using a single breath hold in 23 seconds, the benefit of this technique is apparent. ? 1Requires optional license 2High-end computing required Deep Resolve Boost for HASTE1,2 Conventional PAT 3, TA 1:56 min; .9 x .9 x 6.0 mm3 PAT 3, TA 0:23 sec; 1.1 x 1.1 x 6.0 mm3 T2 BLADE FS Tra 4 Breath-holds Deep Resolve BOOST for HASTE Single Breath-hold Online Training: MR Deep Resolve - Numaris X Job Aid: Deep Resolve Boost - USA PEPconnect Resources Enables higher iPAT factors, resulting in decreased scan times Enable shorter Echo Train [ET] durations and reduced T2 blurring Reduced acquisition times for body T2-weighted sequences, in a single breath-hold Highlights & Benefits Deep Resolve Boost and Deep Resolve Sharp are now available for the HASTE sequence taking Deep Learning Reconstruction to the next level. Achieve excellent image quality, diagnostic confidence, and lesion detection for HASTE body imaging with Deep Resolve HASTE. What is it? ? 1Requires optional license 2High-end computing required Deep Resolve Boost for EPI Diffusion & HASTE1,2 DTI, 64 directions, b1000 PAT 2, SMS 3, Deep Resolve 0.9 x 0.9 x 3.0 mm3 TA = 3:24 min T2 HASTE PAT 4, Deep Resolve1 0.3 x 0.3 x 4.0 mm3 TA = 0:32 seconds DRB EPI Diffusion & HASTE Audio The Deep Resolve Boost EPI Diffusion Tensor image was acquired with 64 directions, a b-value of 1000, in 3 minutes and 24 seconds. The Deep Resolve Boost, T2, HASTE sequence was acquired with a PAT factor of 4, in 32 seconds. ? Deep Resolve Swift Brain DR Swift Brain Audio: Deep Resolve Swift Brain is an Ultra-fast Brain examination with a Net scan time approximately two-minutes. Select the tab arrows to learn more about these Deep Resolve Swift Brain. ? Deep Resolve Swift Brain1,2 What is it? *Requires SMS Package instead of Simultaneous Multi-Slice; 1 Pre-requisite: High-end MaRS 2 Exclusive on 3T Systems for Brain Imaging Deep Resolve Swift Brain is the combination of smart acquisition and a deep learning reconstruction method that delivers all relevant neuro contrasts and orientations with a total acquisition time of only two minutes. Contrasts: T1, T2 +T2*, T2 Dark-Fluid, and Diffusion Techniques: Multi-shot EPI Deep Learning Based Reconstruction Fast GRE Reference Scan Optional: Simultaneous Multi-Slice* Please Note: 3T only, a dedicated license: N_ms_epi, and high-end MaRS is recommended. DR Swift Brain - What is it? Audio: What is Deep Resolve Swift Brain? Deep Resolve Swift Brain is an ultrafast protocol leveraging the fastest available imaging sequence, E-P-I. It is the combination of smart acquisition and a deep learning reconstruction method that delivers T-one, T-two-plus-T-two-star, T-two-dark-fluid, and diffusion contrasts that can be acquired in any orientation, with a total acquisition time of only two minutes. The techniques incorporated into Deep Resolve Swift Brain includes: Multi-shot Echo Planar Imaging, Deep Learning Based Reconstruction, and a Fast G-R-E Reference Scan In addition, Simultaneous Multi-Slice can also be used in the Deep Resolve Swift Brain protocol. ? Deep Resolve Swift Brain1,2 How does it work? Deep Resolve Swift Brain utilizes a multi-shot Echo Planar Imaging (msEPI) approach that enables a faster acquisition. In a very SWIFT acquisition, all typical contrasts for routine brain imaging are generated. T2 and T2* are measured simultaneously, which can save additional time. A static field correction is applied to reduce distortions. 1 Pre-requisite: High-end MaRS; 2 Exclusive on 3T Systems for Brain Imaging *Requires SMS Package instead of Simultaneous Multi-Slice T1 21 s T2* 25 s T2 0 s Diffusion 21 s T2 Dark-Fluid 51 s Ì Click here for additional information on network training DR Swift Brain - How does it work? Base Layer Audio: So how does Deep Resolve Swift Brain work? Deep Resolve Swift Brain utilizes a multi-shot Echo Planar Imaging based approach that enables a faster acquisition. In a very SWIFT acquisition, all typical contrasts for routine brain imaging are generated. The T-one sequence is based on a fast G-R-E imaging sequence, and the diffusion sequence is single-shot E-P-I based. The remaining contrasts are multi-shot E-P-I based.  Multi-shot-shot E-P-I-together with a new static field correction minimizes geometric distortions that are intrinsic to E-P-I imaging.  Within this multi-shot E-P-I approach, T-two and T-two-star are measured simultaneously by a Siemens Healthineers exclusive E-P-I implementation, further accelerating this protocol. Multi-shot E-P-I-based-T-two dark-fluid reduces imaging times substantially compared to standard Turbo Spin Echo imaging and offers true Dark-Fluid contrast due to magnetization transfer preparation. Highest image quality is ensured by a Deep Learning Reconstruction to push the limits of parallel imaging acceleration. The unique combination of multi-shot E-P-I imaging and Deep Learning Reconstruction can drive exam times of approximately two minutes. The total table time including pre-scans and adjustments can be completed in approximately five minutes with this protocol. Click the button on the right to view additional information on how deep resolve swift brain was trained DRSB Training Layer Audio: The networks in Deep Resolve Swift Brain were trained and validated on more than 25,000 images covering a broad range of contrasts and orientations. The performance was characterized by numerous quality metrics including peak signal-to-noise ratio and structural similarity index. Table with 2 columns and 4 rows Dataset Size > 25K images Dataset Composition Broad range of 3T brain images with different contrasts and orientations. Evaluation Criteria Quality metrics included PSNR and SSIM. Radiological quality verification was performed in a clinical setting on >100 patients. Reference Standard Input data retrospectively generated from ground truth via data manipulation and augmentation including undersampling and noise addition. How was it trained? The networks in Deep Resolve Swift Brain were trained and validated on more than 25,000 images covering a broad range of contrasts and orientations. The performance was characterized by numerous quality metrics including peak signal-to-noise ratio and structural similarity index. Deep Resolve Swift Brain1,2 ? Deep Resolve Swift Brain1,2 Benefits 1 Pre-requisite: High-end MaRS 2 Exclusive on 3T Systems for Brain Imaging Ultra-fast Brain examination with a Net scan time < 2 minutes and a Total Table time including prescans, adjustments, and imaging < 5 minutes. Multi-shot EPI implementation supported by Deep Learning Reconstruction. Fast GRE Reference Scan to speed up the acceleration reference data acquisition. New static field correction to further minimize EPI-intrinsic geometric distortions. New flow attenuation leads to reduced flow artifacts through optimized RF pulses and inverted interleaved slice ordering. DR Swift Brain - Benefits Audio: The benefits of Deep Resolve Swift Brain include: The ability to acquire an ultra-fast Brain examination in approximately two minutes, with a total table time including pre-scans, adjustments, and imaging in approximately five-minutes. Multi-shot E-P-I implementation supported by Deep Learning Reconstruction. Fast G-R-E Reference Scan to speed up the acceleration reference data acquisition. New static field correction to further minimize E-P-I-intrinsic geometric distortions. And New flow attenuation that leads to reduced flow artifacts through optimized Radio-Frequency pulses and inverted interleaved slice ordering. ? Deep Resolve Swift Brain1,2 Protocols 1 Pre-requisite: High-end MaRS 2 Exclusive on 3T Systems for Brain Imaging Deep Resolve Swift Brain protocols are in the following areas: Head > Advanced Applications > Deep Resolve Swift Brain Library > T1, T2, Diffusion Perform a quick search to locate the protocol. In the search field, enter Swift. All protocols related to Deep Resolve Swift will be highlighted yellow. DR Swift Brain - Protocols Audio: To locate the Deep Resolve Swift Brain protocols, open the myExam Cockpit. Select Head, Advanced Applications Libraries, and Deep Resolve Swift Brain. Select the Library, to locate the T-one, T-two, and Diffusion sequences. A Quick Search can also be performed to locate the protocol. In the Search field, enter Swift. All the protocols related to Deep Resolve Swift Brain will be highlighted in yellow. ? Deep Resolve Swift Brain1,2 Impressive Speed with No Compromise on Image Quality 1 Pre-requisite: High-end MaRS 2 Exclusive on 3T Systems for Brain Imaging *Requires SMS Package instead of Simultaneous Multi-Slice; Deep Resolve Swift Brain – when time is critical T1 0:21 min 0:21 min Net imaging time Diffusion 0:21 min SMS DWI 0:42 min Net imaging time T2 Dark-fluid 0:51 min 1:33 min Net imaging time T2* T2 Simultaneous Acquisition Ì 0:25 min 1:58 min Net imaging time DR Swift Brain - Speed Audio: When time is critical the Deep Resolve Swift Brain protocol can be implemented to reduce overall scan time without compromising image quality. ? Deep Learning and AI Techniques Acceleration from Head-to-Toe Learn more about our head-to-toe coverage by exploring the interactive body map. Click here to visit the interactive site: Deep Learning and AI Audio: In summary, Siemens Healthineers offers a variety of M-R-I acceleration techniques. The addition of Deep Resolve introduces Deep Learning and Artificial Intelligence to our already expansive portfolio taking clinical imaging to next level. To learn more about Siemens Healthineers acceleration techniques, click on the link provided: Acceleration Techniques Interactive Body Map. ? Course Review Congratulations. You have completed the MR Deep Resolve – Numaris X course. Select the numbered buttons below to review the material before proceeding to the final assessment. MR Deep Resolve Swift Brain MR Deep Resolve Boost MR Deep Resolve Sharp MR Deep Resolve Gain 1 1 2 2 2 3 3 3 4 4 4 Course Review Audio: Congratulations. You have completed the M-R, Deep Resolve – Numaris X online training. Select the objectives listed below to review the material before proceeding to the final assessment. The T-one sequence is based on a fast G-R-E imaging sequence, and the diffusion sequence is single-shot E-P-I based. The remaining contrasts are multi-shot E-P-I based.  Deep Resolve Swift Brain What is it? Deep Resolve Swift Brain is an ultra fast Brain protocol leveraging the fastest available imaging sequence, EPI. It is the combination of smart acquisition and a deep learning reconstruction method that delivers T1, T2 + T2*, T2 Dark-Fluid, and Diffusion contrasts that can be acquire in any orientation, with a total acquisition time of only two minutes. How does it work? Deep Resolve Swift Brain utilizes a multi-shot Echo Planar Imaging approach that enables a faster acquisition. All contrasts for a routine brain exam are generated. The T1 sequence is based on a fast GRE imaging sequence, and the diffusion sequence is single-shot EPI-based. The remaining contrasts are multi-shot EPI-based. Multi-shot EPI together with a new static field correction minimizes geometric distortions. Within this multi-shot EPI approach, T2 and T2* are measured simultaneously. What are the benefits? Ultra-fast Brain examination Fast GRE Reference Scan to speed up the acceleration reference data acquisition New static field correction to minimize EPI-intrinsic geometric distortions New flow attenuation leads to reduced flow artifacts through optimized RF pulses and inverted interleaved slice ordering Deep Resolve Boost What is it? Deep Resolve Boost delivers especially strong denoising for fast acquisitions from head-to-toe. Enabled by Siemens Healthineers first raw-data-to-image deep learning reconstruction. How does it work? Deep Resolve Boost uses raw data from a reduced, and thus faster scan as input. In an iterative process, a deep neural network is applied multiple times to generate an image with significantly reduced noise. The integration of the raw data along the entire reconstruction process leads to an unmatched performance. The result is images with higher signal-to-noise ratios which are consistent with the measured data. What are the benefits? Increased Signal-to-Noise (SNR) Faster scan times for Turbo Spin Echo (TSE) sequences Reduced noise Combine with Simultaneous Multi-Slice (SMS) Turbo Spin Echo (TSE) Deep Resolve Sharp What is it? A deep neural network that improves MR Image quality by increasing image sharpness. How does it work? During image acquisition, phase and frequency encoding lines can be reduced, thus reducing the amount of data and speeding up scan time. From the raw data, a low-resolution image is reconstructed in the first step. Next, Artificial Intelligence comes into play. The deep neural network in Deep Resolve Sharp has been trained on a vast number of pairs of low and high-resolution MR data. It reconstructs a high-resolution image from low resolution data and cross-checks this image with the measured data to ensure consistency. The result is an image with sharp edges and high resolution. What are the benefits? Increased image sharpness Reduces scan time Artificial Intelligence and a deep learning neural network Deep Resolve Gain What is it? Deep Resolve Gain is a targeted denoising method to increase the signal-to-noise ratio (SNR) of images. How does it work? Deep Resolve Gain uses the acquired MRI data to generate a noise map which reflects spatial noise variations. These noise maps are used as prior information in iterative reconstruction processes, like compressed sensing. They are generated without needing to spend additional scan time and can be extracted from the raw data. The reconstruction algorithm takes local noise variations into account and enables stronger denoising where noise would be most dominant when reconstructing with conventional methods. This results in improved image quality and a higher signal-to-noise ratio. No neural networks are used by Deep Resolve Gain. What are the benefits? Increased signal-to-noise ratio (SNR) Shortened scan time Improved image quality Deep Resolve Additional References PEPconnect – Videos Deep Resolve - XA31 PEPconnect – Job Aids Deep Resolve XA31 Deep Resolve Boost – USA Deep Resolve Swift Brain - USA Siemens Healthineers - Videos Deep Resolve Gain Deep Resolve Sharp Deep Resolve Boost Deep Resolve Swift Brain ? DR Additional Resources Disclaimer Please note that the learning material is for training purposes only. For the proper use of the software or hardware, please always use the Operator Manual or Instructions for Use (hereinafter collectively “Operator Manual”) issued by Siemens Healthineers. This material is to be used as training material only and shall by no means substitute the Operator Manual. Any material used in this training will not be updated on a regular basis and does not necessarily reflect the latest version of the software and hardware available at the time of the training. The Operator Manual shall be used as your main reference, in particular for relevant safety information like warnings and cautions. Please note: Some functions shown in this material are optional and might not be part of your system. Certain products, product related claims or functionalities (hereinafter collectively “Functionality”) may not (yet) be commercially available in your country. Due to regulatory requirements, the future availability of said Functionalities in any specific country is not guaranteed. Please contact your local Siemens Healthineers sales representative for the most current information. The reproduction, transmission or distribution of this training or its contents is not permitted without express written authority. Offenders will be liable for damages. All names and data of patients, parameters and configuration dependent designations are fictional and examples only. All rights, including rights created by patent grant or registration of a utility model or design, are reserved. © Siemens Healthcare GmbH 2023 Siemens Healthineers Headquarters\Siemens Healthcare GmbH\Henkestr. 127\ 91052 Erlangen, Germany\Telephone: +49 9131 84-0\siemens-healthineers.com Disclaimer Assessment Welcome to the assessment. For each question, select your answer and then select Submit. You will have 3 attempts to take this assessment. You must receive a score of 80% or higher to successfully pass this course. You will receive your score once you complete the assessment. Start Assessment Select the best answer. Deep Resolve Gain Deep Resolve Swift Brain Question 1 of 10 _____ is a targeted denoising method to increase signal-to-noise-ratio. Deep Resolve Sharp Deep Resolve Boost Multiple Choice Slides 6 & 49 Select the best answer. Deep Resolve Gain Deep Resolve Swift Brain Question 2 of 10 _____ uses a deep neural network that improves image quality by increasing image sharpness. Deep Resolve Sharp Deep Resolve Boost Multiple Choice Slides 9 & 50 Select the best answer. Deep Resolve Gain Deep Resolve Swift Brain Question 3 of 10 _____ delivers strong denoising for a fast acquisition from head-to-toe. Deep Resolve Sharp Deep Resolve Boost Multiple Choice Slides 28 & 51 Select the best answer. Deep Resolve Gain Deep Resolve Swift Brain Question 4 of 10 _____ is an ultra fast Brain protocol leveraging the fastest available imaging sequence, EPI. Deep Resolve Sharp Deep Resolve Boost Multiple Choice Slides 43 & 52 Select the best answer. Proton Density BLADE Question 5 of 10 The protocol for Deep Resolve Swift Brain includes a T1, T2+T2*, _____ and Diffusion. T2 Dark-Fluid T1Dark-Fluid Multiple Choice Slides 43, 46 & 52 Select the best answer. True Question 6 of 10 Retro Recon can be used retrospectively to postprocess Deep Resolve Gain. False Multiple Choice slides 19 & 20 Select the best answer. Deep Resolve Gain Deep Resolve Swift Brain Question 7 of 10 A pre-requisite for the _____ license is the High-end MaRS. Deep Resolve Sharp Deep Resolve Boost Multiple Choice Slides 30 Select the best answer. .55T 7T Question 8 of 10 Deep Resolve Swift Brain is only available on the MAGNETOM _____ Tesla systems. 1.5T 3T Multiple Choice Slides 42 Select the best answer. Resolution > Acceleration Geometry > Saturation Question 9 of 10 To activate Deep Resolve, select the _____ parameter cards. Contrast > Common Sequence > Part 2 Multiple Choice Slide 12 Select the best answer. Deep Resolve Gain Deep Resolve Swift Brain Question 10 of 10 What Deep Resolve technique uses low, medium, and high denoising strengths? Deep Resolve Sharp Deep Resolve Boost Multiple Choice Slide 32 Retry Assessment Results %Results.ScorePercent%% %Results.PassPercent%% Continue YOUR SCORE: PASSING SCORE: Assessment Results You have exceeded your number of assessment attempts. Exit You did not pass the course. Select Retry to continue. Congratulations. You passed the course. Exit To access your Certificate, select the Launch button drop down on the course overview page. You can also access the certificate from your PEPconnect transcript. You have completed the MR Deep Resolve - Numaris X online training. Completion Navigation Help Select the icon above to open the table of contents. Click Next to continue. Next Welcome Slide The timeline displays the slide progression. Slide the orange bar backwards to rewind the timeline. Click Next to continue. Next Tmeline Select the CC icon to display closed captioning (subtitles). Click Next to continue. Next Caption Icon Select the buttons to learn more about a topic. Be sure to review all topics before navigating to the next slide. Click Next to continue. Next Tab Arrow Slide Select the X to close the pop-up. Click Next to continue. Next Layer Slide Some images may have a magnifier icon. Select the image to see an enlarged view. Select it again to return to the normal view. Click Next to continue. Next Zoom Slide Some images have a magnifier icon in the bottom-left corner. Select these image to see an enlarged view of the image. Select the image again to return to the normal view. Videos should default to the optimal resolution. However, to change the video resolution select the gear icon. Click Next to continue. Next Video Slide Some images have a magnifier icon in the bottom-left corner. Select these image to see an enlarged view of the image. Select the image again to return to the normal view. Select Submit to record your response. Click the X in the upper right corner to exit the navigation help. Assessment Slide Question Bank 1 QR700003973_EffDate_27Dec2023 1.1 Welcome 1.2 MR Deep Resolve (DR) 1.3 DR Improving Challenges 1.4 DR Gain & DR Sharp 1.5 MR Deep Resolve - DR Gain 1.6 DR Gain - What is it? 1.7 DR Gain - What problem does it solve? 1.8 DR Gain - How does it work? 1.9 MR Deep Resolve - DR Sharp 1.10 DR Sharp - What is it? 1.11 DR Sharp - What problem does it solve? 1.12 DR Sharp - How does it work? 1.13 DR Gain & Sharp - Applications 1.14 DR Protocol Transformation 1.15 DR Gain - Application Strategies 1.16 DR Gain & Sharp - Practical Considerations 1.17 DR Gain - Retro Recon 1 1.18 DR Gain - Retro Recon 2 1.19 DR Gain - Retro Recon 3 1.20 DR Sharp - Gibbs Ringing Artifacts 1.21 Practical Considerations for DR Gain & DR Sharp 1.22 DR Gain & Sharp - Image Examples 1.23 DR Gain & Sharp - Brain 1.24 DR Gain & Sharp - Lumbar 1.25 DR Gain & Sharp - Knee 1.26 MR Deep Resolve - DR Boost 1.27 DR Boost - What is it? 1.28 DR Boost - What problem does it solve? 1.29 DR Boost - How does it work? 1.30 DR Boost - Parameter Card 1.31 DR Boost - Benefits 1.32 DR Boost - Practical Considerations 1.33 DR Boost - Practical Considerations 1 1.34 DR Boost - Practical Considerations 2 1.35 DR Boost - Image Examples 1.36 DR Boost - Brain 1.37 DR Boost - Cervical Spine 1.38 DR Boost - Thoracic Spine 1.39 DR Boost - Lumbar Spine 1.40 DR Boost - Shoulder 1.41 DR Boost - SMS TSE 1.42 DRB for EPI Diffusion 1.43 DRB for HASTE 1.44 DRB EPI Diffusion & HASTE 1.45 DR Swift Brain 1.46 DR Swift Brain - What is it? 1.47 DR Swift Brain - How does it work? 1.48 DR Swift Brain - Benefits 1.49 DR Swift Brain - Protocols 1.50 DR Swift Brain - Speed 1.51 Deep Learning and AI 1.52 Course Review 1.53 DR Additional Resources 1.54 Disclaimer 1.55 Assessment