Siemens Healthcare GmbH

Patent Trials and Appeals Board

Mar 9, 2022

2021001353 (P.T.A.B. Mar. 9, 2022)

UNITED STATES PATENT AND TRADEMARK OFFICE UNITED STATES DEPARTMENT OF COMMERCE United States Patent and Trademark Office Address: COMMISSIONER FOR PATENTS P.O. Box 1450 Alexandria, Virginia 22313-1450 www.uspto.gov APPLICATION NO. FILING DATE FIRST NAMED INVENTOR ATTORNEY DOCKET NO. CONFIRMATION NO. 15/968,836 05/02/2018 Lucian Mihai Itu 2017P27302US 7466 28524 7590 03/09/2022 SIEMENS CORPORATION IP Dept - Mail Code INT-244 3850 Quadrangle Blvd Orlando, FL 32817 EXAMINER RAHMJOO, MANUCHEHR ART UNIT PAPER NUMBER 2667 NOTIFICATION DATE DELIVERY MODE 03/09/2022 ELECTRONIC Please find below and/or attached an Office communication concerning this application or proceeding. The time period for reply, if any, is set in the attached communication. Notice of the Office communication was sent electronically on above-indicated "Notification Date" to the following e-mail address(es): ipdadmin.us@siemens.com PTOL-90A (Rev. 04/07) UNITED STATES PATENT AND TRADEMARK OFFICE ________________ BEFORE THE PATENT TRIAL AND APPEAL BOARD ________________ Ex parte LUCIAN MIHAI ITU, SAIKIRAN RAPAKA, TIZIANO PASSERINI, and PUNEET SHARMA ________________ Appeal 2021-001353 Application 15/968,836 Technology Center 2600 ________________ Before CAROLYN D. THOMAS, JASON V. MORGAN, and ADAM J. PYONIN, Administrative Patent Judges. MORGAN, Administrative Patent Judge. DECISION ON APPEAL STATEMENT OF THE CASE Pursuant to 35 U.S.C. § 134(a), Appellant1 appeals from the Examiner’s decision to reject claims 1-23. We have jurisdiction under 35 U.S.C. § 6(b). We reverse. 1 “Appellant” refers to “applicant” as defined in 37 C.F.R. § 1.42. Appellant identifies Siemens Healthcare GmbH as the real party in interest. Appeal Br. 1. Appeal 2021-001353 Application 15/968,836 2 SUMMARY OF THE DISCLOSURE Appellant’s claimed subject matter relates to “fast non-invasive computer-based computation of a hemodynamic index, such as fractional flow reserve (FFR)[,] from medical image data of a patient.” Abstract. As part of the claimed invention, “[r]egions in the automatically generated patient-specific anatomical model for which user feedback is required for accurate computation of a hemodynamic index are predicted using one or more trained machine learning models.” Id. ILLUSTRATIVE CLAIM Claim 1, which is illustrative with respect to claims 1-23, is reproduced below (disputed limitations emphasized and bracketing added). 1. A method for providing fast non-invasive computer- based computation of a hemodynamic index from medical image data of a patient, comprising: automatically generating a patient-specific anatomical model of one or more arteries of a patient based on medical image data of the patient; and [1] predicting regions in the automatically generated patient- specific anatomical model for which user feedback is required for accurate computation of a hemodynamic index using one or more trained machine learning models. REFERENCES The Examiner relies on the following references (only the first named inventor of each reference is listed): Name Reference Date Itu ’209 US 2016/0166209 A1 June 16, 2016 Gulsun US 2017/0258433 A1 Sept. 14, 2017 Itu ’336 US 10,463,336 B2 Nov. 5, 2019 Appeal 2021-001353 Application 15/968,836 3 REJECTIONS The Examiner rejects claims 1-23 as follows: Claims Rejected 35 U.S.C. § Reference(s)/Basis Citation 1-23 Obviousness-type double-patenting (Itu ’336) Final Act. 3-4 1, 14, 19 102(a)(1) Itu ’336 Final Act. 5 1-10, 12-23 102(a)(1) Itu ’209 Final Act. 5-11 11 103 Itu ’209, Gulsun Final Act. 11-14 ANALYSIS Obviousness-Type Double-Patenting The Examiner rejects claims 1-23 on grounds of obviousness-type double-patenting as being unpatentable over claims 1-36 of Itu ’336. Final Act. 3-4. In making this rejection, the Examiner cites to claims 1 and 13 of Itu ’336 as evidence. Id. Claims 1 and 13 of Itu ’336 are reproduced below. 1. A method for determining a hemodynamic index for one or more locations of interest in coronary arteries of a patient, comprising: receiving medical image data of the patient; extracting patient-specific coronary arterial tree geometry of the patient from the medical image data; extracting geometric features from the patient-specific coronary arterial tree geometry of the patient; and computing a hemodynamic index for one or more locations of interest in the patient-specific coronary arterial tree using a trained machine-learning based surrogate model and based purely on the extracted geometric features without considering features from patient-specific physiological measurements, the trained machine-learning based surrogate model trained based Appeal 2021-001353 Application 15/968,836 4 on geometric features extracted from synthetically generated coronary arterial tree geometries. Itu ’336 18:44-60. 13. The method of claim 1, wherein computing a hemodynamic index for one or more locations of interest in the patient- specific coronary arterial tree using a trained machine-learning based surrogate model and based purely on the extracted geometric features without considering features from patient- specific physiological measurements comprises: computing the hemodynamic index for the one or more locations of interest in the patient-specific coronary arterial tree in response to a user input identifying the one or more locations of interest. Itu ’336 21:12-22. Appellant contends the Examiner errs because the only recitation in the cited claims of Itu ’336 potentially pertinent to the claim 1 recitation of [1] “predicting regions in the automatically generated patient-specific anatomical model for which user feedback is required” is the Itu ’336 claim 13 recitation of “computing the hemodynamic index for the one or more locations of interest in the patient-specific coronary arterial tree in response to a user input identifying the one or more locations of interest.” Appeal Br. 6 (quoting Itu ’336 claim 13). In response, the Examiner submits that “‘predicting regions . . . for which user feedback is required for accurate computation of a hemodynamic index’ [is] a recitation of the intended use of the claimed invention.” Ans. 5. The Examiner finds there are “no structural differences” between disputed recitation [1] and claims 1 and 13 of Itu ’336. Id. at 6. Appellant, however, submits that claim 1 of the current application “does not provide for an intended use of the predicting, but instead explicitly defines what is predicted.” Reply Br. 3. Appeal 2021-001353 Application 15/968,836 5 We agree with Appellant that the Examiner errs. Recitation [1] is directed to predicting which regions cannot have accurate hemodynamic index values computed unless user feedback is provided. This is consistent with the disclosure that “[e]mbodiments of the present invention utilize one or more learning models to indicate/predict certain parts of the arterial geometry where user feedback is required for obtaining accurate computed FFR (cFFR) results.” Spec. ¶ 30. Although not recited in claim 1, such predictions can be used to request user feedback for the identified regions before final cFFR values are computed. Spec. ¶ 47, Fig. 2; Appeal Br. 20 (Claims App., claim 10). The Examiner’s interpretation of the limitation “for which user feedback is required” in recitation [1] as merely being directed to an intended use is not a reasonably broad interpretation in light of the Specification. Moreover, the Examiner does not show that recitation [1] would have been obvious in light of the claims of Itu ’336. Accordingly, we do not sustain the Examiner’s obviousness-type double-patenting rejection of claim 1, and claims 2-23, based on Itu ’336. Anticipation and Obviousness The Examiner rejects claim 1 as anticipated by Itu ’336. Final Act. 5 (citing Itu ’336 17:1-25). The Examiner also rejects claim 1 as anticipated by Itu ’209, citing a disclosure in Itu ’209 that is similar to the disclosure in Itu ’336 with respect to recitation [1]. Id. at 6 (citing Itu ’209 ¶ 69); compare Itu ’209 ¶ 69 with Itu ’336 17:1-25. In particular, the Examiner notes that Itu ’209 teaches that “features are determined for the original patient-specific anatomical model, and they are adapted, either automatically or by using information input by the user, and finally the machine learning based trained surrogate model is applied to compute the post-treatment Appeal 2021-001353 Application 15/968,836 6 hemodynamic index.” Final Act. 6; Ans. 7 (“a user provides input in Itu [’336] and the machine learning based model computes a hemodynamic metric, while claim 1 provides for predicting where user input would be required for accurate computation of a hemodynamic index, which in some embodiments can be used to request user input”); Ans. 10-11. Appellant contends the Examiner errs because “the machine learning based model in the cited portions of Itu [’336] ‘is applied to compute the post-treatment hemodynamic metric,’ . . . and is not applied for ‘predicting regions in the automatically generated patient-specific anatomical model for which user feedback is required.’” Appeal Br. 8. Appellant acknowledges that Itu ’336 describes “receiving user input,” but Appellant argues that Itu ’336 nonetheless fails to “teach or suggest ‘predicting regions . . . for which user feedback is required.’” Id. at 9; Reply Br. 4-5. Appellant makes similar arguments with respect to the Examiner’s reliance on Itu ’209 because the cited portions of Itu ’209 “disclose substantially similar subject matter as disclosed” in the cited portion of Itu ’337. Appeal Br. 12. We agree with Appellant the Examiner errs. Itu ’336 discloses that “first the features are determined for the original patient-specific anatomical model, next they are adapted, either automatically or by using information input by the user . . . , and finally the machine learning based training surrogate model is applied to compute the post-treatment hemodynamic metric.” Itu ’336 17:5-13 (emphases added); see also Itu ’209 ¶ 69. That is, the machine learning based model cited in Itu ’336 is used after information is input from the user. Thus, rather than predicting where user input is required, the cited machine learning based model of Itu ’336, at best, performs computations that incorporate user input that has already been supplied. Therefore, we agree with Appellant that the Examiner’s findings do not show that either Itu ’336 or Itu ’209 discloses disputed Appeal 2021-001353 Application 15/968,836 7 recitation [1]. The Examiner also does not show that Gulsun cures the noted deficiency of Itu ’209. Accordingly, we do not sustain the Examiner’s 35 U.S.C. § 102(a)(1) rejections of claim 1, and claims 2-10 and 12-23, which contain similar recitations. We also do not sustain the Examiner’s 35 U.S.C. § 103 rejection of claim 11. CONCLUSION In summary: Claim(s) Rejected 35 U.S.C. § Reference(s)/Basis Affirmed Reversed 1-23 Obviousness-type double-patenting (Itu ’336) 1-23 1, 14, 19 102(a)(1) Itu ’336 1, 14, 19 1-10, 12-23 102(a)(1) Itu ’209 1-10, 12-23 11 103 Itu ’209, Gulsun 11 Overall Outcome 1-23 REVERSED