T 2153/14 () of 2.7.2020

Summary of Facts and Submissions

I. The appeal is against the decision of the examining division dated 23 June 2014 refusing European patent application No. 08762935.8 pursuant to Article 97(2) EPC. The application was published as international application WO 2009/066136 A2.

II. The documents cited in the decision under appeal included the following:

D1: WO 2007/058632 A1;

D8: H.A.F. Gratama van Andel et al: "Removal of bone in CT angiography by multiscale matched mask bone elimination", 31 August 2007, XP12103149.

III. The application was refused on the following grounds.

- Claim 1 of the main request and the first auxiliary request then on file did not meet the requirements of Article 84 EPC (see decision, points 15 and 30 of the Reasons).

- Claim 1 of the main request and the first auxiliary request then on file did not specify processing multiple series of biological images obtained from a patient affected by an ischemic stroke. Therefore, D8 was an appropriate starting point for the assessment of inventive step (see decision, points 16 to 18 and 30 of the Reasons). The subject-matter of claim 1 of the main request and the first auxiliary request did not meet the requirements of Article 56 EPC because it lacked inventive step over the disclosure of D8 combined with the common general knowledge of the person skilled in the art (see decision, points 19 to 24 and 31 to 35 of the Reasons).

- The subject-matter of claim 1 of the second auxiliary request then on file did not meet the requirements of Article 56 EPC because it lacked inventive step over the disclosure of D1 combined with the common general knowledge of the person skilled in the art (see decision, points 37 to 46 of the Reasons).

IV. The applicant (hereinafter: appellant) filed notice of appeal. With the statement of grounds of appeal, the appellant filed claims according to a main request and an auxiliary request and submitted that this main request corresponded to the second auxiliary request forming the basis for the decision under appeal (see statement of grounds, page 1, last paragraph). It requested that the decision under appeal be set aside and that a European patent be granted on the basis of the claims of the main request or the auxiliary request filed with the statement of grounds. The appellant indicated a basis for the claims in the application as filed and provided arguments as to why the claims met the requirements of Articles 54 and 56 EPC.

V. The board issued a summons to oral proceedings. In a communication under Article 15(1) RPBA (Rules of Procedure of the Boards of Appeal, OJ 2007, 536), annexed to the summons, the board introduced documents D9 (US 2003/0045791 A1) and D10 (EP 1 860 453 A1) into the appeal proceedings and gave the following preliminary opinion.

- Claim 1 of the main request and the auxiliary request did not validly claim the priority of the previous application (Article 87(1) EPC). Therefore, document D10 was prior art under Article 54(2) EPC.

- The subject-matter of claim 1 of the main request and the auxiliary request lacked inventive step over the combined disclosures of documents D1 and D9 and the common general knowledge of the person skilled in the art (Article 56 EPC).

- The subject-matter of claim 1 of the main request and the auxiliary request lacked inventive step over the combined disclosures of documents D1 and D10 and the common general knowledge of the person skilled in the art (Article 56 EPC).

VI. With its reply dated 14 April 2020, the appellant filed amended claims according to a main request and an auxiliary request replacing the previous requests on file. It indicated a basis for the amendments in the application as filed and submitted arguments as to why the claimed priority was valid (Article 87 EPC) and the amended claims met the requirements of Articles 54 and 56 EPC. The appellant announced that it would not be attending the oral proceedings scheduled for 14 May 2020.

VII. The board notified the appellant that the oral proceedings to be held on 14 May 2020 had been cancelled. In a communication dated 23 April 2020, the appellant was asked whether, in view of its announcement that it would not be attending the oral proceedings, it wished to withdraw its request for oral proceedings.

VIII. The appellant informed the board that it did not wish to have new oral proceedings scheduled.

IX. Claim 1 of the main request and claim 1 of the auxiliary request read as follows (the additional features of claim 1 of the auxiliary request are in italics):

"A method for identifying a region representing an area affected by an ischemic stroke, performed by a processing entity (102) of a computing apparatus (100), said apparatus comprising input means (108) and output means (110) communicatively coupled to said processing entity (102), said method comprising obtaining a plurality of series (306) of biological images (3041, 3042, 3043, 3044), including a diffusion-weighted image series and/or a perfusion-weighted image series, obtained from a patient affected by an ischemic stroke, all series evidencing a plurality of associated parameters, wherein certain images in any series are corresponding to certain images in other series, wherein corresponding images consist in an array of pixels or voxels having respective values, all corresponding images in any series having same pixel or voxel spacing, origin and orientation, characterized in that it further comprises (208):

- creating a respective segmentation mask (604, 622, 624) for a selected image (602, 612, 614) in a first series, including a diffusion-weighted image series and/or a perfusion-weighted image series, previously obtained from a patient affected by an ischemic stroke, said first series being formed by a set of images evidencing a first parameter, defining a set of pixels to be eliminated from said image and wherein creating a segmentation mask is based on an input from a user (114) of said computing apparatus (100) received by said input means (108);

- applying said segmentation mask to a corresponding image in a second series, including a diffusion-weighted image series and/or a perfusion-weighted image series, previously obtained from a patient affected by an ischemic stroke, said second series being formed by a set of images evidencing a second parameter, said second parameter being different than said first parameter, wherein applying a segmentation mask to said corresponding image comprises creating a filtered image for which each pixel or voxel value is set to the value of its corresponding pixel or voxel of said corresponding image, excluding all pixel or voxel values that are set to a baseline value when said corresponding pixels or voxels are captured by said segmentation mask;

- outputting by said output means (110) a filtered image created from the corresponding image on which a segmentation mask was applied."

X. The examining division considered D1 to be the closest prior art for the assessment of inventive step (see decision under appeal, point 37 of the Reasons).

XI. The appellant's arguments, where relevant to the present decision, may be summarised as follows.

(a) D1 was the closest prior art for the assessment of inventive step (see statement of grounds of appeal, page 13, second paragraph, and letter dated 14 April 2020, page 7, fourth paragraph).

(b) D1 did not disclose creating a segmentation mask in an image series "evidencing" a first parameter and applying the mask to images of a second series evidencing a second parameter to filter the images (see statement of grounds of appeal, the paragraph bridging pages 11 and 12, and the letter dated 14 April 2020, page 7, fifth paragraph).

(c) The problem to be solved might be identified as that of extracting information useful for diagnostics from a plurality of images (see statement of grounds of appeal, page 13, last paragraph, and the letter dated 14 April 2020, page 8, first full paragraph).

(d) D9 did not clearly disclose using two series of images. D9 did not disclose creating a segmentation mask in an image series "evidencing" a first parameter and applying the mask to images of a second series evidencing a second parameter to filter the images (see letter dated 14 April 2020, page 9, first paragraph). The steps of calculating and filtering mentioned in D9, paragraphs [0062] to [0065] simply consisted of applying a threshold in a single series of images (see the letter dated 14 April 2020, the paragraph bridging pages 8 and 9 and page 9, first full paragraph).

(e) D9 did not require a user input to create the mask (see the letter dated 14 April 2020, page 9, second full paragraph).

Reasons for the Decision

1. Withdrawal of the request for oral proceedings

1.1 By letter dated 18 May 2020, the appellant informed the board that it did not wish to have a new date for oral proceedings appointed ("Nous vous informons que la demanderesse, pour des raisons economiques explicitées dans notre courrier du 14 avril 2020, ne souhaite pas de nouvelle date de convocation à la procedure orale") but instead wished to have its submissions of 14 April 2020 taken into account in the written proceedings ("En revanche, la demanderesse souhaite que les amendements déposés le 14 avril 2020 soient pris en compte dans le cadre de la procedure d'examen de ladite demande citée en objet").

1.2 The appellant's statements referred to in point 1.1 above are an unambiguous expression of its wish to withdraw its request for oral proceedings.

2. Main request and auxiliary request - inventive step over D1 and D9 (Article 56 EPC)

2.1 D1 is the closest prior art for the assessment of inventive step (see also points X and XI(a) above).

2.2 D1 discloses (see also decision under appeal, points 37 to 46 of the Reasons, and statement of grounds of appeal, page 11) a method for identifying a region representing an area affected by an ischemic stroke (see paragraphs [0056] and [0057]: "If the segmented GRE region is larger than the segmented FSE region, then the stroke is determined to be hemorrhagic [...] Otherwise, the stroke is determined to be ischemic [...] the brain images may be [...] visually inspected and manually processed to determine if the stroke is hemorrhagic or ischemic"), performed by a computer comprising input means and output means coupled to a processing entity (see Figure 1). The method comprises the following steps:

- obtaining a plurality of series of images including a diffusion-weighted image series and a perfusion-weighted image series, all series evidencing a plurality of associated parameters (see paragraph [0035]: "the captured brain images may include two or more of [...] diffusion weighted imaging (DWI), perfusion weighted imaging (PWI) [...] ADC, CBF, CBV, MTT, TTP, and PKHT maps. In a particular embodiment, the superposition may include a DWI image and a PWI image [...] a DWI image or an ADC MAP may show infarct regions in the brain, a PWI image or maps derived from the PWI image may show penumbra regions in the brain");

- processing corresponding images in the series such that they have the same pixel or voxel spacing, origin and orientation (see paragraph [0079]: "the brain images, with the contours for delineating the actual infarct and penumbra regions, are co-registered to a co-ordinate system [such as] a 2D or 3D Cartesian or Talairach co-ordinate system" and paragraph [0090]: "two sets of images or atlas images may be considered to correspond to each other [...] when they satisfy the following conditions: 1. Both sets of images have the same image size and resolution, and the same number of associated atlas images; 2. Both sets of atlas images have the same shape of the cortex and ventricular system; and 3. Both sets of atlas images are located in the same (e.g. Talairach) coordinate system");

- removing noise or background from the captured DWI and PWI images (see paragraph [0051]: "A captured brain image may be subject to post-capture processing or alteration, such as removal of noise or background, intensity correction including field inhomogeneity correction" and paragraph [0072]: "one or more perfusion maps may be used for determining the penumbra regions and generating the penumbra contours [...] it may be necessary or desirable to remove any artifacts and/or geometric distortions from the brain images before segmentation");

- creating respective segmentation masks for the DWI and the PWI images, based on a threshold which is either automatically set or set by the user and applying said mask to segment the DWI and PWI images (see paragraph [0060]: "CAD application 208 may have a routine for automatically segmenting an infracted [sic] region from the DWI image, such as by using an adaptive intensity threshold"; paragraph [0062]: "For instance, an [sic] region of interest may be selected, such as interactively, and the selected region, such as a local volume or a slice, may be segmented using thresholds determined either automatically or interactively for the region of interest"; paragraph [0063]: "A user may dynamically adjust the value of the threshold to adjust the contours of the infarct regions and inspect the modified view to see if any improvement has been made to the segmented image"; and paragraph [0072] "The penumbra regions may be segmented either automatically or interactively, or both. For instance, the PWI image may be automatically and globally segmented using a suitable thresholding technique"); and

- displaying a filtered image created from the corresponding image on which a segmentation mask was applied (see paragraph [0064]: "The displayed image may be enlarged, or zoomed, and panned to allow more precise adjustment of the contours. For instance, FIG. 5 shows another screenshot 500 of GUI 402, where a pull-down menu window 502 is shown. As can be seen, the menu provides menu items for modifying the contour lines 406 that enclose the infarct regions 408").

2.3 The board agrees with the appellant that D1 does not disclose creating a segmentation mask in an image series "evidencing" a first parameter and applying the mask to images of a second series evidencing a second parameter to filter the images (see point XI(b) above). Thus, the subject-matter of claim 1 of both requests differs from the disclosure of D1 in that artefacts, noise and background in a second image (such as a PWI image) are filtered based on a mask created for a first image (such as a DWI image).

2.4 The board shares the appellant's view that the problem to be solved may be identified as that of extracting information useful for diagnostics from a plurality of images (see point XI(c) above).

2.5 D9 discloses a method for generating images from calculated hemodynamic parameters acquired from NMR data after contrast agent injection (see paragraph [0060]). In a pre-contrast phase (no acquisition of hemodynamic parameters), the mean value of each voxel is calculated and voxels whose pre-contrast mean value does not exceed a threshold are removed (see paragraphs [0062] to [0065]). Thus, D9 illustrates that it is advantageous to create a mask for filtering noise, background and artefacts in an image in which these can be easily identified and to then apply the mask to other images. Therefore, the person skilled in the art would create the masks known from D1 in an image in which the artefacts to be removed can be easily identified and then apply the masks to other images.

2.6 The board has not been persuaded that D9 does not disclose the use of two series of images (see point XI(d) above). D9, paragraph [0062] discloses that the mean value of each voxel is calculated during a pre-contrast phase by averaging over a first series of N frames during this phase. A mask is created by determining those voxels whose mean value does not exceed a threshold. This mask is applied to subsequent frames (a second series of images) to filter out ("remove from consideration") voxels which are located in the air outside the patient's head (see paragraph [0065]). Thus, the mask is generated in a first series of images during a pre-contrast phase and applied to a second series of images including the frames following the contrast arrival time.

2.7 The board agrees with the appellant that D9 discloses a method for automatically calculating an arterial input function (AIF) from the acquired NMR data (see point XI(e) above). However, D1, paragraph [0051] discloses that post-capture processing such as noise or background removal may be performed automatically by a computer or interactively with input from a user. The person skilled in the art using the background removal known from D9 in the method known from D1 would consider allowing the user to, for instance, manually input the threshold.

2.8 In summary, the subject-matter of claim 1 of neither request meets the requirements of Article 56 EPC because it lacks inventive step over the combined disclosures of documents D1 and D9 and the common general knowledge of the person skilled in the art.

3. In view of the board's conclusion set out in point 2.8 above, the questions of the validity of the priority and inventive step over the combined disclosures of D1 and D10 need not be considered.

4. Since neither request is allowable, the appeal is to be dismissed.

Order

For these reasons it is decided that:

The appeal is dismissed.