Ex Parte Gunderson et al

Patent Trial and Appeal Board

Nov 26, 2013

11564126 (P.T.A.B. Nov. 26, 2013)

UNITED STATES PATENT AND TRADEMARK OFFICE
__________

BEFORE THE PATENT TRIAL AND APPEAL BOARD
__________

Ex parte BRUCE D. GUNDERSON, MARK L. BROWN, and
AMISHA SOMABHAI PATEL1
__________

Appeal 2011-010827
Application 11/564,126
Technology Center 3700
__________

Before TONI R. SCHEINER, ERIC GRIMES, and JEFFREY N.
FREDMAN, Administrative Patent Judges.

GRIMES, Administrative Patent Judge.

DECISION ON APPEAL
This is an appeal under 35 U.S.C. § 134 involving claims to a method
of processing data generated by an implantable medical device, which have
been rejected for obviousness. We have jurisdiction under 35 U.S.C. § 6(b).
We reverse.

1 Appellants identify the Real Party in Interest as Medtronic, Inc. (Appeal
Br. 3).
Appeal 2011-010827
Application 11/564,126

2
STATEMENT OF THE CASE
The Specification states that increasing amounts of time are required
to review data generated by implantable medical devices, such as
implantable cardioverter-defibrillators (ICDs), and therefore “an algorithm
that post-processes and automatically reviews each previously detected
episode upon interrogation could address these concerns by accurately
classifying episodes and potentially suggesting ICD parameter changes
and/or medical therapy, such as changes in medication” (Spec. 2, ¶ 4).
The Specification provides “a flowchart of classifying of a cardiac
event” (id. at 3, ¶ 10; Fig. 3). One of the steps in the algorithm is to
“Generate/Store Template” (Fig. 3). The Specification states that the
templates are generated when a cardiac event is classified as a VT/VF event
(Spec. 15, ¶ 45), and are used to determine whether a subsequent cardiac
event should be re-classified as a ventricular tachycardia event (id. at 14,
¶ 44). The Specification provides an exemplary method for generating a
plurality of templates and selecting one of them for storage (see id. at 15-17,
¶¶ 46-50; Fig. 7).
Claims 1-10 are on appeal. Claim 1 is illustrative and reads as follows
(emphasis added):
1. A method of generating a template during post-processing of
sensing data generated by and stored within an implantable medical device,
comprising:
receiving the sensing data generated by the implantable medical
device with an external access device, wherein the stored sensing data
includes sensed atrial events and sensed ventricular events;
determining, with the external access device, in response to the
received data, instances where the implantable medical device identified a
cardiac event being detected in response to the sensing data;
Appeal 2011-010827
Application 11/564,126

3
generating, with the external access device, a plurality of
templates in response to correlated morphologies of a plurality of adjacent
intervals prior to a detection interval corresponding to the cardiac event
being identified as the cardiac event and a morphology of the detection
interval, wherein each of the templates is associated with one or more of the
detection interval and the adjacent intervals based on the correlated
morphologies of the detection interval and the plurality of adjacent intervals;
and
selecting, with the external access device, one of the plurality of
templates associated with a greatest number of the intervals for arrhythmia
type classification of other detected events.

Claim 6, the only other independent claim, is directed to a system that
includes “means for selecting one of the plurality of templates associated
with a greatest number of the intervals.”
DISCUSSION
The Examiner has rejected claims 1, 2, 4-7, 9, and 10 under 35 U.S.C.
§ 103(a) as obvious based on Gunderson2 and Kim3 (Answer 3). The
Examiner has rejected claims 3 and 8 under 35 U.S.C. § 103(a) as obvious
based on Gunderson, Kim, and Duffin4 (Answer 5). The same issue is
dispositive for both rejections.
The Examiner finds that Gunderson teaches the first three steps of
claim 1, including “generating a plurality of templates in intervals adjacent
to the detection interval (e.g. 606) and determining a correlated morphology
(e.g. 608)” (id. at 4), but does not teach “selecting the template with the
greatest number of interval[s] for use as arrhythmia classification” (id.). The

2 Gunderson et al., US 2003/0204215 A1, Oct. 30, 2003.
3 Kim et al., US 2005/0192506 A1, Sept. 1, 2005.
4 Duffin, 5,193,550, Mar. 16, 1993.
Appeal 2011-010827
Application 11/564,126

4
Examiner finds that “Kim discloses matching intervals to multiple templates
and then recording the intervals that matched each template (e.g. Fig. 14A).
Then Kim compares the templates and chooses the template that matched the
most criteri[a] for the current template (e.g. Fig. 13)” (id.). The Examiner
reasons that this disclosure meets the disputed limitation:
Gunderson teaches creating the templates from the plurality of
intervals and Kim disclosing choosing the template that
matches the criteria the most (fig. 13). The template that
matches the most beats will have the highest correlation.
Therefore, the template that matches the most beats (the one
with the highest correlation) will be selected.
(Id. at 6.) The Examiner concludes that it would have been obvious “to
modify the system as taught by Gunderson, with selecting the template that
matched the most criteri[a] as taught by Kim, since such a modification
would provide the predictable results of selecting a template based on it
matching the most waveforms to provide the predictable results of selecting
the most accurate template” (id. at 4).
Appellants argue that “Kim simply describes comparing each of a set
of templates to each representative cardiac beat, calculating the correlation
coefficient for each template, and selecting the template that has the highest
average correlation coefficient across the entire set of cardiac beats”
(Appeal Br. 9). Appellants argue, therefore, that “Gunderson in view of
Kim fails to disclose or suggest selecting, with the external access device,
one of the plurality of templates associated with a greatest number of the
intervals for arrhythmia type classification of other detected events, as
recited in Appellant’s claim 1” (id.).
Appeal 2011-010827
Application 11/564,126

5
We agree with Appellants that the Examiner has not provided a
persuasive basis for concluding that it would have been obvious, based on
Kim, to modify Gunderson’s method by selecting a template associated with
the greatest number of intervals, as claimed. The Specification describes a
method of generating a plurality of templates and selecting one of them for
storage (Spec. 15-17; Fig. 7). The exemplified method begins with defining
an interval associated with detection of a cardiac event (id. at 15, ¶ 46). The
morphology of the cardiac waveform in that interval is defined as a first
template (e.g., template A).
Then the morphology of the interval preceding the detection interval
is compared to the first template: if the morphology of the interval has a
correlation with the first template above a certain threshold (e.g., 0.82), the
interval is determined to correlate with the first template; if its morphology
has a correlation with the first template below the threshold, it is defined as a
new template (e.g., template B) (id. at 15, ¶ 47). This process is repeated for
additional intervals preceding the detection interval, and a determination is
made whether their morphologies correlate with each of the previously
defined templates (id. at 16-17, ¶¶ 48-50). After the desired number of
intervals have been analyzed, “the generated template that is correlated with
the greatest number of intervals (template B in the example of FIG. 7) is set
as the template associated with the detected event [ ] and is stored in
memory” (id. at 17, ¶ 50).
Gunderson discloses a method for determining whether an
“arrhythmia detection is due to cardiac oversensing and not an appropriate
VF detection . . . [by] compar[ing] consecutively sensed signal
Appeal 2011-010827
Application 11/564,126

6
morphologies” (Gunderson 8, ¶ 74). “If alternating morphologies are
occurring, T-wave oversensing is diagnosed as the cause of the VF detection
. . . , and the episode is identified as an inappropriate detection” (id.).
In the method described, the “morphology of the sensed event
occurring at VF detection, referred to as R(I), is stored as a first template,
TEMPLATE(1)” (id. at 8, ¶ 75). Then “[t]he morphology of the sensed
event prior to R(I), referred to as R(I-1), is compared to the stored template,
TEMPLATE(1). . . . If the morphology of R(I-1) approximately equals the
TEMPLATE(1) . . . then R(I-1) is labeled as a TEMPLATE(1) match” (id.).
“If the morphology of R(I-1) is different than TEMPLATE(1), it is stored as
a second template, TEMPLATE(2)” (id.).
“[T]he morphology analysis continues by . . . compar[ing] the next
previous template [sic, event], R(I-N) to TEMPLATE(1). . . . If the
morphology of R(I-N) does not match TEMPLATE(1), the method [ ]
determines if any other morphology templates have been stored.” (Id. at 8,
¶ 76.) “If other stored templates do exist, . . . the morphology of R(I-N) is
compared to the other stored templates” (id. at 8, ¶ 77). Gunderson states
that “[a]fter completing the morphology analysis [ ], the method . . . can
determine . . . if alternating signal morphologies are occurring that would be
evidence of T-wave oversensing” (id. at 8, ¶ 78).
Kim discloses “a method and system for selecting a template
representative of a particular type of cardiac event” (Kim 1, ¶ 9). More
specifically, “[w]hen a pace pulse produces a contractile response in heart
tissue, the contractile response is typically referred to as capture. . . . Capture
verification has been accomplished by examining the cardiac waveform
Appeal 2011-010827
Application 11/564,126

7
following a pacing pulse. Such an examination may involve, for example,
comparison of the cardiac waveform following the pace pulse to a
morphology template.” (Id. at 1, ¶¶ 4, 6.) “Morphology templates may also
be used to detect and/or classify types of cardiac tachyarrhythmias” (id. at 1,
¶ 7).
Kim states that its Figure 13, cited by the Examiner, “illustrates a
method of selecting capture-related templates” (id. at 8, ¶ 96). “A plurality
of templates characterizing a particular type of capture-related event is
provided [ ]. For example, the templates may characterize an evoked
response, a pacing artifact, a captured response,” etc. (Id.)
Next, “cardiac signals representative of the type of capture-related
event are detected . . . [and] used to . . . evaluat[e] the correlation between
the cardiac signals representative of the type of capture-related event and a
template characterizing the type of capture-related event” (id. at 8, ¶ 97).
“A template is selected [ ] as a current capture-related template based on the
comparison. For example, a template having a highest average correlation
with the cardiac signals representative of the capture-related event may be
selected and stored as the current template.” (Id.)
Kim states that its Figure 14A, also cited by the Examiner, illustrates
one method of selecting a template (id. at 8, ¶ 99). Kim states that the
“correlation coefficient of a representative cardiac beat is calculated [ ] for
each template. The process continues [ ] until a preset number of correlation
coefficients have been calculated for each template.” (Id. at 8, ¶ 100.) The
appropriate template can then be selected by various methods (id. at 8,
¶ 102). “For example, a template having a highest average correlation with
Appeal 2011-010827
Application 11/564,126

8
the cardiac signals representative of the capture-related event may be
selected” (id. at 8, ¶ 97).
The Examiner has not pointed to a disclosure, in either Gunderson or
Kim, of selecting a template “associated with a greatest number of the
intervals for arrhythmia type classification,” as required by the claims. In
addition, the Examiner has not provided a persuasive reason for concluding
that, even if the required selection method was disclosed, it would have been
obvious to modify Gunderson’s method by selecting a template associated
with the greatest number of intervals.
As discussed above, the relevant disclosure in Gunderson relates to
detecting T-wave oversensing, which indicates that a detected VF event is an
inappropriate detection. Gunderson’s method is based on “compar[ing]
consecutively sensed signal morphologies” (Gunderson 8, ¶ 74) to each
other. Gunderson discloses that “if alternating signal morphologies are
occurring that would be evidence of T-wave oversensing” (Gunderson 8,
¶ 78). Thus, Gunderson’s method does not rely on matching interval/event
morphologies to a template, but only on whether the signals alternate
between different morphologies. Given Gunderson’s focus on changes in
signal morphology, rather than matching to any specific morphology per se,
the Examiner has not persuasively explained why it would have been
obvious to modify Gunderson’s method by “selecting . . . one of the plurality
of templates associated with a greatest number of the intervals,” as required
by the claims on appeal.
Appeal 2011-010827
Application 11/564,126

9
SUMMARY
We reverse both of the rejections on appeal.

REVERSED

lp