Obviousness

Obviousness Analysis of US Patent 9814431 Under 35 U.S.C. § 103

This analysis identifies combinations of prior art references that would render the independent claims of US Patent 9814431 obvious to a person having ordinary skill in the art (PHOSITA) at the time of the invention (priority date 2007-05-04). The primary prior art references considered are US7359535B2 and US7574249B2, both of which predate the critical date of US9814431B2.

Claim 1: A method for retrospective internal gating

Claim 1 details a method comprising several steps for generating motion-corrected images using retrospective internal gating. A PHOSITA would find this claim obvious in light of the combination of US7359535B2 and US7574249B2.

Prior Art References:

1. US7359535B2 (GE Medical Systems Global Technology Company, Llc): "Systems and methods for retrospective internal gating"

   • Disclosure: This patent describes reconstructing images from 4D data by correlating a respiratory signal to image data. An exemplary method includes reconstructing a composite image, determining a breathing pattern based on activity in a region of interest (ROI), and reconstructing a gated image set by associating image data with respective phases of the breathing pattern. It also mentions analyzing data of a selected voxel in the composite image over time and filtering time data to remove high and low frequencies.
   • Relevance: This reference directly teaches the overarching concept of "retrospective internal gating" using image-derived motion signals, including acquiring image series, extracting time-activity information (from ROIs/voxels), filtering, mapping to motion phases, and generating motion-corrected images.

2. US7574249B2 (General Electric Company): "Device-less gating of physiological movement for improved image detection"

   • Disclosure: This patent discloses a method for device-less gating of physiological movement by acquiring images, processing them to identify an ROI related to physiological movement, generating a motion function based on the ROI, and gating the images using this motion function to reduce artifacts. It explicitly states that an ROI may correspond to a voxel, and a single voxel or a plurality of voxels can be selected to derive the motion function. Importantly, it teaches selecting these voxels based on criteria such as "signal amplitude, frequency, and signal to noise ratio," and filtering the motion function to remove unwanted noise.
   • Relevance: This reference complements US7359535B2 by further elaborating on the "device-less" aspect and specifically teaching how to derive a motion function from individual or multiple voxels within the image data, including criteria for selecting these voxels.

Obviousness Argument for Claim 1:

All limitations of Claim 1 are taught or rendered obvious by the combination of US7359535B2 and US7574249B2:

• "acquiring a series of images at times t 1 . . . tn;" is taught by US7359535B2, which discloses "acquiring time-varying image data" and "4D data," and by US7574249B2, which teaches "acquiring a set of images of an object".
• "extracting time-activity information for individual voxels;" is taught by US7359535B2, which mentions analyzing "the data of a selected voxel in the composite image over the entire time period". US7574249B2 explicitly teaches that an ROI can be a voxel and that "a single voxel... or a plurality of voxels may be selected and used to derive the motion function," directly covering the extraction of time-activity information for individual voxels.
• "prioritizing voxels for phase analysis, and assigning weighting factors;" US7574249B2 teaches "selecting voxels" based on "various criteria, such as, but not limited to, signal amplitude, frequency, and signal to noise ratio". This selection process effectively prioritizes voxels for their contribution to the motion function, which is subsequently used for phase analysis (gating). A PHOSITA, seeking to improve the accuracy of the motion function, would find it obvious to apply explicit "weighting factors" based on these selection criteria (e.g., higher weight for voxels with better SNR or amplitude, or zero weight for unimportant voxels as described in US9814431B2's dependent claims) to quantify and emphasize the importance of more reliable voxels. This is a common and predictable optimization technique in signal processing when combining data from multiple sources.
• "applying frequency filter to voxel time-activity curves;" is taught by US7359535B2 ("filtering the time data to remove high and low frequencies") and US7574249B2 ("filtering the motion function to remove unwanted noise and non-physiological frequencies").
• "using prioritization, combining voxel time-activity information into a time varying object motion function;" Both references teach combining voxel/ROI information to create a motion signal. US7359535B2 forms a "respiratory pattern signal" from filtered time-activity data. US7574249B2 generates a "motion function based on the region of interest" (which can be multiple voxels). The "prioritization" aspect is addressed by the motivation to explicitly weight voxels as discussed above.
• "using the time varying object motion function for the mapping of image data to corresponding motion phases;" is taught by US7359535B2 ("associating image data with respective phases of the breathing pattern") and US7574249B2 ("gating the set of images...utilizing the motion function," including identifying "gating points" corresponding to different phases).
• "using the mapping of image data to corresponding motion phases to generate at least one motion corrected image." is taught by US7359535B2 ("reconstructing a gated image set...corrected for patient motion") and US7574249B2 ("gating the set of images to reduce motion artifacts").

Motivation for a PHOSITA to Combine:

A PHOSITA in the field of medical imaging, seeking to improve the accuracy, robustness, and automation of retrospective internal gating techniques, would have been motivated to combine the teachings of US7359535B2 and US7574249B2. US7359535B2 provides the fundamental framework for image-based retrospective gating. US7574249B2 offers specific, detailed methods for enhancing the derivation of the internal motion signal by considering individual voxels and their characteristics.

The motivation to explicitly "prioritize voxels and assign weighting factors" (as described in Claim 1 and its dependent claims 3, 4, 6 of US9814431B2) when combining voxel time-activity information would stem directly from the desire to leverage US7574249B2's teaching on selecting voxels based on signal quality (e.g., amplitude, frequency, SNR). A PHOSITA would recognize that formalizing this selection into a weighting scheme would allow for a more nuanced and potentially more accurate contribution of each voxel to the final "time varying object motion function." Furthermore, standard signal processing techniques for combining multiple signals, such as considering phase differences (implicitly addressed by addition/subtraction scenarios in dependent Claim 14 of US9814431B2) and optimizing for signal strength or periodicity (e.g., by maximizing standard deviation as in dependent Claim 15), are well-known approaches that a PHOSITA would routinely apply to improve the quality of a derived physiological motion signal.

Claims 21 and 22: Non-transitory computer-readable medium and System

Claims 21 and 22 cover a non-transitory computer-readable medium encoded with a program and a system, respectively, that when executed, cause a machine to perform the method of Claim 1.

Obviousness Argument for Claims 21 and 22:

Given that the method described in Claim 1 would be obvious to a PHOSITA in light of the combination of US7359535B2 and US7574249B2, implementing this method on a computer system or storing the instructions on a computer-readable medium would also be obvious. At the time of the invention (priority date 2007-05-04), it was well-established practice in the field of medical imaging to implement complex image processing and reconstruction algorithms as software programs executable by general-purpose computers. The patent itself states the advantages of "software derived respiratory signals" as being "image based, and thus machine independent," and capable of being used with "existing scans, or scanners." This indicates that the implementation of such algorithms in software was a known and desired approach for PHOSITAs. Therefore, Claims 21 and 22, being directed to the computerized implementation of an obvious method, would themselves be obvious.