DETAILED ACTION
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1 and 12 are rejected under 35 U.S.C. 103 as being unpatentable over VAZQUEZ ROMAGUERA et al. (Pub. No. US 2023/0129194 hereinafter “VAZ”) in view of Liao et al. (Pub. No. US 2017/0337682).
Regarding claims 1 and 12, VAZ teaches a system for fast deformable image registration using 2D cine MRI image frames (image acquisition device) acquired during radiation therapy comprising: a source (image acquisition device) of cine MRI images (2D cine-MRI images) received in approximate real-time (acquired in real-time) from a patient undergoing MRI-guided radiation therapy (MRgRT) [Para. 42 “The methods and systems described herein provide a computational framework for reliable and rapid organ movement and deformation tracking.”; Para. 41 “Indeed, the methods and systems described herein allow to track targets that displace in 3D using only 2D cine-MRI images acquired in real-time.”; Para. 42 “In some embodiments, the methods and system described herein are integrated with Magnetic Resonance Imaging Guided Linear Accelerator (MRI-Linac) applications, a hybrid technology that combines radiation and high-resolution MRI in a single machine”; Para. 43 “An image-guided radiation therapy system 602 comprises an image acquisition device 604 for obtaining 2D images of the patient 600 in real-time”].
VAZ teaches a processor (processing unit 502) that receives the images and operates a trained deep learning (DL) [Para. 20, 22, 37, and 43] and predicts frame by frame (each time step) motion from images [Para. 20, 22, 37, and 43].
However, VAZ doesn’t explicitly teach employs a reference image (R) to a moving image as inputs, and outputs a dense motion vector field (MVF) that aligns the images.
Liao teaches employs a reference image (R) to a moving image as inputs, and outputs a dense motion vector field (MVF) (dense deformation field) that aligns the images (registered using the dense correspondence) [Para. 50 “One of the medical images is designated as the reference image I.sub.ref and the other is designated as the moving image I.sub.mov.” and Par. 144 “In particular, the first and second medical images are input to a DNN trained using the method of FIG. 19, and the trained DNN outputs a 3D vector field representing the dense correspondence (i.e., dense deformation field) between the first and second images. At step 2206, the first and second medical images are registered using the dense correspondence”].
It would have been obvious to one of ordinary skill in the art before the effective filing date to modify VAZ’s radiotherapy motion-prediction system by incorporating Liao’s teaching of receiving a reference image and moving image as inputs and outputting a dense motion vector field that aligns the images to register each incoming 2D cine MRI images to the reference image before motion prediction. This medication improves VAZ by reducing spatial misalignment in the images used to predict the tumor target position, thereby improving the accuracy of therapy beam positioning.
VAZ also teaches a guidance process (target tracking and beam positioning device) that controls therapy beam operation (determines where to position the radiation beam) based upon the prediction (predicted position of a tumor target) [Para. 43 “The reconstructed 3D volume is provided to a target tracking and beam positioning device 608 which determines where to position the radiation beam 610 based on a predicted position of a tumor target 616 in the anatomical structure 614.”].
Claims 2, 10, 13 and 21 are rejected under 35 U.S.C. 103 as being unpatentable over VAZQUEZ ROMAGUERA et al. (Pub. No. US 2023/0129194 hereinafter “VAZ”) in view of Liao et al. (Pub. No. US 2017/0337682) further in view of Alex et al. (Pub. No. US 2015/0309135).
Regarding claims 2 and 13, VAZ in view of Liao doesn’t explicitly teach the claim limitations.
However, Alex teaches wherein both cardiac and respiratory motion (cardiac and respiratory motion) are visible in the images (real time cine images) [Para. 3 “However, the superposition of respiratory and cardiac motion limits temporal sparsity”; Para. 23 “The exemplary system, method and computer-accessible medium can include a free-breathing cine MRI framework that can sort and synchronize cardiac and respiratory motion into two separate dimensions, followed by an exemplary joint multi-coil compressed sensing reconstruction (see, e.g., Reference 6) (e.g., image reconstruction) on the higher dimensional data set, using different sparsity constraints on respiratory and cardiac motion dimensions.].
It would have been obvious to one of ordinary still in the art before the effective filing date to modify VAZ’s MRI-Linac cine-image tracking system, as supplemented by Liao’s trained dense-registration process, by using Alex’s cine MRI image acquisition in which cardiac and respiratory motion are both present and synchronized. This modification improves VAZ by allowing the registration and beam-guidance process to account for both cardiac and respiratory anatomical motion visible in the treatment images.
Regarding claims 10 and 21, VAZ in view of Liao doesn’t explicitly teach the claim limitations.
However, Alex teaches comprising a combination process that coordinates an EKG signal derived from the patient with the cine MRI images so as to provide data related to cardiac and respiratory motion to the DL process [Para. 4 “the electrocardiogram (“ECG”) signal can usually be monitored such that the “ectopic” cardiac cycles can be discarded before image synchronization and reconstruction”; Para. 23 “sort and synchronize cardiac and respiratory motion into two separate dimensions”].
It would have been obvious to one of ordinary still in the art before the effective filing date to modify VAZ’s MRI-Linac cine-image tracking system, as supplemented by Liao’s trained dense-registration process, by using Alex’s cine MRI image acquisition in which cardiac and respiratory motion are both present and synchronized. This modification improves VAZ by allowing the registration and beam-guidance process to account for both cardiac and respiratory anatomical motion visible in the treatment images.
Claims 3, 4, 14, and 15 are rejected under 35 U.S.C. 103 as being unpatentable over VAZQUEZ ROMAGUERA et al. (Pub. No. US 2023/0129194 hereinafter “VAZ”) in view of Liao et al. (Pub. No. US 2017/0337682) further in view of Alex et al. (Pub. No. US 2015/0309135) further in view of Schaefferkoetter (Pub. No. US 2025/0037327 hereinafter “Scha”).
Regarding claims 3 and 14, VAZ in view of Liao doesn’t explicitly teach the claim limitations.
However, Scha teaches wherein the DL process (first CNN and second CNN) extracts, from the images (anatomical image and functional image), a predetermined number of hierarchical feature maps (first and second feature map) at multiple spatial resolutions which are reconstructed into the MVF with spatial resolution (each resolution block) equal to the inputs (deformation vector filed) [Para. 43, 44, and 46 “ Referring to FIG. 4, second CNN 406b receives as input the sets of feature maps (410, 412) generated based on both original input images (402, 404), and produces a three-dimensional (3D) deformation vector field (DVF) 416 characterizing the relative deformation of the pair of images”].
It would have been obvious to one of ordinary still in the art before the effective filing date to modify VAZ’s real-time MRI-Linc motion tracking system as modified by Liao’s dense pairwise image registration process and Alex’s cardio MRI teaching by using Scha’s DL process to extract feature maps at multiple spatial resolutions and reconstruction an MVF. This medication improves VAZ’s deformable registration accuracy for moving treatment anatomy.
Regarding claims 4 and 15, VAZ in view of Liao doesn’t explicitly teach the claim limitations.
However, Scha teaches wherein the MVF (DVF) is applied by shifting the pixels in the images by associated motion vectors (interpolation grid) to generate a transformed image (registered anatomical image) [Para. 45-47].
It would have been obvious to one of ordinary still in the art before the effective filing date to modify VAZ’s real-time MRI-Linc motion tracking system as modified by Liao’s dense pairwise image registration process and Alex’s cardio MRI teaching by using Scha’s DL process to extract feature maps at multiple spatial resolutions and reconstruction an MVF. This medication improves VAZ’s deformable registration accuracy for moving treatment anatomy.
Claims 5-9 and 16-20 are rejected under 35 U.S.C. 103 as being unpatentable over VAZQUEZ ROMAGUERA et al. (Pub. No. US 2023/0129194 hereinafter “VAZ”) in view of Liao et al. (Pub. No. US 2017/0337682) further in view of Alex et al. (Pub. No. US 2015/0309135) further in view of Schaefferkoetter (Pub. No. US 2025/0037327 hereinafter “Scha”) further in view of Balakrishnan (“VoxelMorph: A learning framework for deformable medical image registration”).
Regarding claims 5 and 16, VAZ teaches wherein the transformed image is generated via a spatial transform (spatial transformation) [Para. 22 “A spatial transformation is applied to the 3D reference volume based on the set of predicted deformations, and the reference volume post-spatial transformation is displayed as the motion-compensated anatomical structure for each time step i ϵ T.sub.out”].
Regarding claims 6 and 17, VAZ in view of Liao further in view of Alex and further in view of Scha doesn’t explicitly teach the claim limitations.
Bala (“VoxelMorph: A learning framework for deformable medical image registration”) teaches wherein the DL process defines a plurality of network layers (convolutional layers) and at least one of the network layers is based on a U-Net architecture (UNet) of a convolutional neural network (CNN) [section IV-A “The parametrization of gθ(·,·) is based on a convolutional neural network architecture similar to UNet [63], [64], which consists of encoder and decoder sections with skip connections.”; and “We apply 3D convolutions in both the encoder and decoder stages using a kernel size of 3, and a stride of 2. Each convolution is followed by a LeakyReLU layer with parameter 0.2. The convolutional layers capture hierarchical features of the input image pair, used to estimate φ”].
I would have been obvious to one of ordinary skill in the art before the effective filing date to modify VAZ in view of Liao further in view of Alex and further in view of Scha’s trained radiotherapy image registration process by incorporating Bala’s U-net architecture (UNet) implemented with network layers in a CNN for estimating the deformation field from image pairs. This modification improves VAZ by using encoder-decoder skip-connected convolutional processing to preserve spatial detail while extracting registration features.
Regarding claims 7 and 18, VAZ in view of Liao further in view of Alex and further in view of Scha doesn’t explicitly teach the claim limitations.
Bala teaches comprising a loss function with a hyperparameter (λ) (regularization parameter) defined as: LOSStotal=LOSSdissimilarity + λ LOSSgradients where, LOSSdissimiarity computes the post-registration mean square error (mean squared voxelwise difference) between T and R and LOSSgradients computes the magnitude of spatial gradients (spatial gradients) in the MVF [Section IV-C. “Unsupervised Loss Function: The unsupervised loss Lus(·,·,·) consists of twocomponents:Lsim that penalizes differences inappearance, andLsmooth that penalizes local spatialvariations inφ:”; “We encourage a smooth displacement field φ using a diffusion regularizer on the spatial gradients of displacement u:”].
I would have been obvious to one of ordinary skill in the art before the effective filing date to modify VAZ in view of Liao further in view of Alex and further in view of Scha’s trained radiotherapy image registration process by incorporating Bala’s U-net architecture (UNet) implemented with network layers in a CNN for estimating the deformation field from image pairs. This modification improves VAZ by using encoder-decoder skip-connected convolutional processing to preserve spatial detail while extracting registration features.
Regarding claims, 8 and 19, VAZ in view of Liao further in view of Alex and further in view of Scha doesn’t explicitly teach the claim limitations.
Balakrishnan teaches wherein the network provides network layers in the following order, for each successive layer, of spatial dimensions (spatial rsolution): 1, ½ ,14, 1/8, 1/16, 1/8,/2, 1 [Section IV-A, fig. 3 and related description].
I would have been obvious to one of ordinary skill in the art before the effective filing date to modify VAZ in view of Liao further in view of Alex and further in view of Scha’s trained radiotherapy image registration process by incorporating Bala’s U-net architecture (UNet) implemented with network layers in a CNN for estimating the deformation field from image pairs. This modification improves VAZ by using encoder-decoder skip-connected convolutional processing to preserve spatial detail while extracting registration features
Regarding claims 9 and 20, VAZ in view of Liao further in view of Alex and further in view of Scha doesn’t explicitly teach the claim limitations.
Balakrishnan teaches wherein the loss function is defined by:
PNG
media_image1.png
221
506
media_image1.png
Greyscale
Wherein m and n represent the height and width of images [Section IV-C].
I would have been obvious to one of ordinary skill in the art before the effective filing date to modify VAZ in view of Liao further in view of Alex and further in view of Scha’s trained radiotherapy image-registration model by incorporating Bala’s LOSS function (unsupervised loss) including LOSS-DISSIMILARITY (mean squared voxelwise difference) and LOSS_GRADIENTS (diffusion regularizer) applied over image locations, with m and n corresponding to image height and width for 2D cine MRI frames. This medication improves VAZ by providing an explicitly optimization objective for aligning transformed image T to reference image R while maintaining smooth motion-vector fields.
Claims 11 and 22 are rejected under 35 U.S.C. 103 as being unpatentable over VAZQUEZ ROMAGUERA et al. (Pub. No. US 2023/0129194 hereinafter “VAZ”) in view of Liao et al. (Pub. No. US 2017/0337682) further in view of Alex et al. (Pub. No. US 2015/0309135) further in view of Ward et al (Pub. No. US 2017/0067976).
Regarding claims 11 and 22, VAZ in view of Liao further in view of Alex doesn’t explicitly teach the claim limitations.
However, Ward teaches wherein the EKG signal is derived from one or two legs of the patient [Para. 34, “The ECG cable bundle 100 is configured to transfer ECG signals from the electrodes 16 placed on the patient during performance of an MRI scan, such that the ECG signals are substantially similar to ECG signals taken outside of the MRI scan”].
It would have been obvious to one of ordinary still in the art before the effective filing date to modify VAZ in view of Liao further in view of Alex’s cine-MRI radiotherapy tracking system, as modified by Liao and Axel by driving the ECG signal from one or two leg electrode locations as taught by Ward. This medication improves VAZ by enabling ECG acquisition during MRI scanning while reducing chest-electrode interference risk in the MR imaging region.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to SOLOMON G BEZUAYEHU whose telephone number is (571)270-7452. The examiner can normally be reached on Monday-Friday 10 AM-7 PM.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, O’Neal Mistry can be reached on 313-446-4912. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-0101 (IN USA OR CANADA) or 571-272-1000.
/SOLOMON G BEZUAYEHU/ Primary Examiner, Art Unit 2666
Prosecution Insights