Issue Preclusion in Relation to a Larger Patent Portfolio

By Dennis Crouch

One risk of claim construction is potential for inconsistent judgments between different district court judges who each are required to construe disputed claim terms.  In e.Digital v. Futurwei (Huawei), the district court cut this knot by applying the doctrine of collateral estoppel (issue preclusion) to preclude the patentee from seeking a construction that varied from that of a prior court.  On appeal, the Federal Circuit has largely affirmed — however, the appellate panel rejected the notion that collateral estoppel applies to the construction of identical terms found in a second asserted patent since that second-patent was not at issue in the prior case and had material differences. [Read the Court Decision]

Due process generally requires Courts to allow parties a chance to make their case. However, the doctrine of collateral estoppel blocks parties from re-litigating issues that had already been finally decided.  Generally, the issue being precluded must have been actually litigated by the party being precluded and necessarily decided as part of a valid final judgment.  Because claim construction is an interlocutory decision and subject to modification, preclusion does not kick-in until a final judgment is awarded.

Here, a prior Colorado Court construed the terms of e.Digital’s U.S. Patent Nos. 5,491,774. The ‘774 patent covers a handheld audio recorder/player. All of the asserted claims require  “a flash memory module which operates as the sole memory of the received processed sound electrical signals” (sole memory limitation).  The patentee argued that this sole-memory-limitation should not rule-out the use of microprocessors that also require RAM to operate.  The Colorado district court disagreed and held that the claims excluded any device with RAM. Under the arguably narrow construction it was clear that the accused infringers did not actually infringe and the parties stipulated to a dismissal of the case.

In the present action against Huawei, e.Digital asserted both the ‘774 patent and Patent No. 5,839,108.  Although the ‘108 patent is not an official ‘family member’ (e.g., not a continuation, CIP, or divisional), it does cover the same basic device and includes the identical sole memory limitation.

On appeal, the Federal Circuit found that, despite the identical limitation, it was improper for the district court to rely on issue preclusion to bar the patentee from litigating construction of the sole memory limitation term with respect to the ‘108 patent.

The ’108 patent . . . presents a separate claim construction issue. The ’108 patent is not related to the ’774 patent, but does disclose a purported improvement to the ’774 patent. While the ’108 patent may incorporate by reference the ’774 patent as prior art, it does not change the fact that the patents are not related. The ’108 patent discloses a separate invention, includes a distinct prosecution history, and is supported by a different written description—including Figures 3 and 4 which clearly depict RAM. These distinctions reinforce the well understood notion that claims of unrelated patents must be construed separately. Texas Digital Sys., Inc. v. Telegenix, Inc., 308 F.3d 1193, 1211 (Fed. Cir. 2002) (citing Abbott Labs. v. Dey, L.P., 287 F.3d 1097, 1104 (Fed. Cir. 2002)) (explaining that a claim of an unrelated patent “sheds no light on” the claims of the patent in suit). Because the asserted patents are not related, the ’108 patent requires a new claim construction inquiry and the court therefore erred in applying collateral estoppel to the ‘108 patent.

The holding here is partially based upon the fact that the two patents are not formally related according to the rules of priority.  However, the court added – in dicta – that issue preclusion might not apply even if they had been related.

To be clear, our decision that collateral estoppel cannot apply to the construction of a claim in one patent based on a previous claim construction of an unrelated patent is not an invitation to assume the opposite is always justified. That is, a court cannot impose collateral estoppel to bar a claim construction dispute solely because the patents are related. Each case requires a determination that each of the requirements for collateral estoppel are met, including that the issue previously decided is
identical to the one sought to be litigated. A continuation-in-part, for instance, may disclose new matter that could materially impact the interpretation of a claim, and therefore require a new claim construction inquiry.

The point here is that issues involving a related but previously un-asserted patent ‘might’ be subject to issue preclusion based upon whether the new patent requires consideration of arguments and facts that are distinct from the already issued patent.

Going forward, this case offers some considerations for patent portfolio management both at the filing/prosecution stages as well as the assertion stage.  This case, coupled with other prior Federal Circuit decisions provide value in splitting-up larger patents into smaller component patents and in potentially holding-back some patents from initial rounds of litigation.

41 thoughts on “Issue Preclusion in Relation to a Larger Patent Portfolio

  1. 5

    What a wonderful system we have, where clearly erroneous decisions cannot be corrected:

    “Here, a prior Colorado Court construed the terms of e.Digital’s U.S. Patent Nos. 5,491,774. The ‘774 patent covers a handheld audio recorder/player. All of the asserted claims require “a flash memory module which operates as the sole memory of the received processed sound electrical signals” (sole memory limitation). The patentee argued that this sole-memory-limitation should not rule-out the use of microprocessors that also require RAM to operate. The Colorado district court disagreed.”

    Well the Colorado distri9ct court was wrong. Indicating that a flash memory modules is the sole memory OF RECEIVED PROCESSED SOUND ELECTRICAL SIGNALS” does not mean that there is no other memory for holding program instructions or for doing calculations!

    OMG save us from the poly-sci majors.

      1. 5.1.1

        Read the decision. The patentee disclaimed RAM in order to overcome prior art.

        From the decision:
        Turning first to the intrinsic record, the notion that Mr. Norris was disclaiming the use of RAM (or other non-flash memory) during the analog-to-digital conversion and data compression phases of his device is consistent with the reasons given by the USPTO for denying the patent and Mr. Norris’ concessions made to overcome that denial. The USPTO’s position was that Mr. Norris’ initial application was anticipated by the Schroder patent because the distinguishing characteristic of Mr. Norris’ invention — the use of flash memory as the final storage device — would have been obvious to a person with skill in the art contemplating alternatives to the Schroder device’s use of a hard drive or floppy disk for storage purposes. The Plaintiff’s present position — that the claimed device differs from the Schroder device insofar as the Schroder device uses a hard drive for storage whereas the Norris device uses flash memory — presents precisely the same contention that the USPTO considered and rejected in ruling on Mr. Norris’ initial application.


          PA, but what does the final storage have to do with whether the controlling microprocessor uses RAM for is program memory?

    1. 5.2

      If the Colorado district court was wrong, the patentee should have appealed the decision. They did not, and thus it’s immaterial if it was wrong. They can’t retroactively try to appeal a decision they failed to appeal earlier.

      1. 5.2.1

        PA, of course they are bound in court.

        Had they fixed the problem in the PTO by claiming a microprocessor with RAM, the story here would have been different.

      2. 5.2.2

        Yeah, what Ned said. The claim was allowed, so apparently the Examiner came to the conclusion that Flash was not an obvious, under 103, substitute.

  2. 4

    Is this really well thought out?

    Suppose patentee P sues defendant D and gets an unwanted claim construction C1. Then, like the generic pharmaceutical settlement cases from last month where defendants help plaintiffs, P sues defendant E and stipulates to settle on the condition that E helps P get new construction C2 before settlement is final. E is not bound by estoppel and can seek a new construction that benefits P in some other case easily enough.

    Then does P have a choice of construction C1 and C2 to be estopped by?

    And what happens — for the law profs out there — when P sells the patent to D who then sues F and they get to Markman?

    1. 4.1

      Owen, it happens like this. P gets favorable in case 1. Infringer in case 2 gets favorable in case 2. P is bound by 2.

      All P has to do is lose just once.

    1. 3.1

      What’s favorable to a patentee in one case might not be favorable in the next, depending on the accused infringer’s product or the prior art they manage to dig up.

    2. 3.2

      Does this mean that claim constructions favorable to patentees can’t be re-litigated

      You mean challenged by different defendants? It definitely does not mean that.

    3. 3.3

      Issue Preclusion only binds the parties involved in the original litigation. So, the construction (or validity finding) in favor of the patentee can be re-litigated if the patentee sues a different entity for infringement.

      1. 3.3.1

        Dennis, I think you meant “claim” preclusion.

        Issue preclusion binds a party where there is non mutuality of parties.

        Claim preclusion bind parties to a litigation.

      2. 3.3.2

        Dennis, in other words, if the parties are the same, positive results for a party will bind the other party. If the parties are not the same, a party to the former litigation is bound but the non party is not. Thus a non party can ignore rulings it does not like. The party cannot. I don’t think the whether the issue is favorable or not to a party has any bearing on this.

    4. 3.4

      Issue preclusion can’t be used against third parties. A favorable claim construction maybe can’t be relitigated against the same defendant in a subsequent suit, but it certainly can be against the rest of the world.

      1. 3.4.1

        Anonymous, just to be clear, a claim construction favorable to the patent owner is binding on a privy of the party infringer.

  3. 2

    CAFC: The ’108 patent is not related to the ’774 patent, but does disclose a purported improvement to the ’774 patent. While the ’108 patent may incorporate by reference the ’774 patent as prior art, it does not change the fact that the patents are not related.

    Those facts don’t also don’t “change the fact” that the sky is blue. But they are certainly facts that are relevant to the construction of a claim term that appears in both patents, where that claim term isn’t expressly defined by the applicant.

    More specifically, the the incorporation by reference of an earlier document will always be relevant to the construction of a term that appears in both the patent and the earlier document (at least in those cases where that term is not expressly defined by the patentee to exclude a definition given to the term in the earlier document). Likewise, a statement in the specification that the claimed invention is an improvement to an invention described in a prior document will always inform the construction of a term that appears in both the patent claim and in the description of the improved-upon embodiment in the prior document (at least in those cases where that term is not expressly defined by the patentee to exclude a definition given to the term in the earlier document).

    There’s no way around that sort of common sense unless the CAFC is trying to erect some sort of wall around what facts can be considered in construing a claim term that has not been defined expressly or specifically by an applicant/patentee.

    The ’108 patent discloses a separate invention, includes a distinct prosecution history, and is supported by a different written description—including Figures 3 and 4 which clearly depict RAM. These distinctions reinforce the well understood notion that claims of unrelated patents must be construed separately.

    Those distinctions only “reinforce” the trivial point that differences between documents can lead to different claim constructions. When the same applicant uses the same term in the same technological field to describe some aspect of an invention, there is every reason to believe that the applicant is defining that term in the same way, regardless of the “relationship” between the documents. If there is evidence (other than self-serving boilerplate) to suggest that the applicant is not using the term in the same way then by all means it should be considered.

    The key point here is that a requirement that claim terms in “separate” patent be construed “separately” doesn’t mean that the construction of the claim term in the prior patent isn’t relevant or can’t be considered by judges, jurists or potential defendants. That would make no sense whatsoever.

    With respect to the collateral estoppel issue, the same considerations apply. Given two patents in “separate families”, collateral estoppel might very well apply to preclude re-litigation of claim construction of a particular term. It depends on the facts (e.g., the similarity between the cases and what the patentee argued in the first case).

    1. 2.1

      I wonder if the court should reconsider its interpretation of Abbott Labs. v. Dey, L.P., 287 F.3d 1097, 1104 (Fed. Cir. 2002)) (explaining that a claim of an unrelated patent “sheds no light on” the claims of the patent in suit).

      Perhaps the focus should be on what it means to be ‘related.’ One meaning involves priority claims or terminal disclaimer that expressly links the inventions. Here, we had other signs of relation: an express incorporation by reference; patents filed by the same applicant and inventor using identical claim language. I suspect that the applications were also filed by the same patent attorney and perhaps examined by the same examiner. These elements are likely enough to at least create a prima facie case that identical claim terms are to be given identical meaning.

      1. 2.1.1

        I wonder if the court should reconsider its interpretation of Abbott Labs. v. Dey, L.P., 287 F.3d 1097, 1104 (Fed. Cir. 2002)) (explaining that a claim of an unrelated patent “sheds no light on” the claims of the patent in suit).

        I don’t remember what else might be in that opinion but the assertion in parentheses doesn’t seem reasonable without some important caveats. At the very least, a claim in an “unrelated” patent (where “unrelated” means “unrelated by priority claim) is extrinsic evidence that could (and should) “shed light” on how a skilled artisan would understand the meaning of the term. And, as you noted, the amount of light that would be shed would seem to be significantly greater when it’s the same inventor using the term for the same reason that any non-patent documents where the inventor uses the term in a particular way are always going to be very relevant to the construction of an otherwise undefined claim term. How could it be otherwise?

        Patentees are entitled to be their own lexicographers. That is all the flexibility that is needed for the patentee. If the patentee wants a term to have a particular and unambiguous meaning in a particular application, then the patentee define that term expressly and unambiguously in the specification. Otherwise the patentee should be prepared to have that term defined in a reasonable manner based on the evidence that was set forth in Phillips, without any other per se restrictions. Neither the patent system nor the public is served by a claim construction paradigm where a single assignee is entitled to have identical but otherwise undefined terms construed in contradicting ways — in the assignee’s favor — by excluding evidence of how that term was construed in other “unrelated” patents/applications owned (or previously owned) by the same assignee.

    2. 2.2

      MM, let us put it this way, there is a very strong presumption that collateral estoppel does not apply to unrelated patents even though the claim terms may be the same.

      The issues have to be identical.

      1. 2.2.1

        Collateral estoppel doesn’t — and shouldn’t — require 100% “identicalness” of all facts.

        The mere fact that two patent applications are not “related” by priority claim is a fact to consider like any other fact but by itself it shouldn’t create any special “presumption” that must be overcome to find collateral estoppel. I’ll grant you that that the likelihood of the criteria for finding collateral estoppel may be lowered when the applications are not “related” by priority claim, but that’s equally true of a lot of other facts.


          MM, there is also the point that even when the specs are the same the claims may not be so that even though the same claim term can be in two different claims the issues might not be identical.

          Consider two independent claims with the same claim term. Take the present case as an example where two claims have the “sole memory” element. A first claim has a microprocessor. The second claim has a microprocessor with RAM. A construction that “sole memory” would exclude microprocessors with RAM would clearly not apply to the second claim that expressly included a microprocessor with RAM.

          This is why subsequent reexamination my alter claim construction. Assume the patent owner had added a RAM limitation to the microprocessor during reexamination. That claim could not longer justly be construed to exclude a microprocessor with RAM.


            The patentee couldn’t have added material so as to capture a microprocessor with RAM in re-exam. That would be an impermissible broadening.

            The original claim was construed to be “Microprocessor WITHOUT USING ram” (a negative limitation),thus a product that used ram wouldn’t infringe. If the patentee in reexam tried to claim “microprocessor WITH using ram”, the claim would have been impermissibly broadened and would not be allowed, as it would capture material that would not have infringed pre-reexam. See In re Freeman, 30 F.3d 1459 (1994) (“A claim is enlarged if it includes within its scope any subject matter that would not have infringed the original patent.”).


              PA, so….. I add “RAM” as a limitation to the claim in reexamination and I broaden the claim?

              No, that is why the initial claim construction of the court was wrong or has to be reconsidered.

              People here appear to be brainwashed.


                ? Your argument doesn’t make sense – it’s a negative limitation, so yes, adding the thing that was specifically determined to not be present in the claim broadens it.

                For example:

                I claim a chair with “no more than 3 legs”. Court says that is construed as meaning “all chairs with 1, 2, or 3 legs BUT NO MORE”

                If in reexam I change my claim to be “no more than 3 legs and one more leg” I’ve broadened the claim. I now capture 4-legged chairs, something that was specifically NOT part of my previous claim. Prior to reexam, a four legged chair would not have infringed. After, it does. Therefore, violation of In re Freeman.

                Did you look at the original reasons why the court decided the claim specifically excluded microprocessors with RAM? otherwise, not sure how you can say it was wrong. Assuming it was right, you have to accept that now capturing material post reexam that would not have infringed pre-reexam would be a broadening of the claim. Read Freeman, it explains all this. Point is: Just because you add something to a claim doesn’t necessarily mean it is narrower.

                1. PA, but the limitation did not say microprocessor without RAM. That was a gloss put on it by the court and it was clearly wrong. We are not talking about limitations in the claims.

                  We have a flash memory module recited in the claim. Assume it has a microprocessor. Who in the world would think that that microprocessor would have to have flash as its program memory?

                  You also assume the court construction is somehow binding on the patent office. It is not. They are free to ignore it, and come to a different construction. The PTO may not view the addition of RAM to a microprocessor limitation as broadening the scope of the claim because they might view the court construction as clearly erroneous.

                  Again, we are not talking about positively recited limitations here.

                2. That was a gloss put on it by the court and it was clearly wrong.

                  Here’s the Colorado district court’s claim construction decision. The argument you were making was specifically decided by that court, and e.Digital (as far as I can tell) did not appeal the decision. For all intents and purposes it became final. So regardless of whether or not it was wrong, that’s irrelevant. They should have appealed then.

                3. PA, what you fail to understand is that collateral estoppel does not apply the patent office, and collateral estoppel does not apply after proceedings in the patent office and completed if the claim term under consideration has been amended.

                4. Ned, you’re missing the point. Collateral estoppel applies to the patentee. The PTO is immaterial. The patentee was bound to a narrower construction, and couldn’t try to change that construction at the PTO without violating In re Freeman.

                  Keep in mind that the patentee, in the present case, cited nothing to indicate it had advocated for a different construction at the PTO. They should not get to get a do-over because of the decision they made to not appeal the Colorado decision, just because they put a patent in reexam. The incentives that creates to try to expand claim scope are inconsistent with In re Freeman.

                5. Also, the claim term wasn’t amended at the PTO. Other stuff was added, but the particular term at issue was not amended. Collateral estoppel.

                6. PA, regarding Freeman, it is an odd case. The Board disagreed with the district court, but felt it was bound by its claim construction. The Federal Circuit (Rich) said that whether the Board was bound by the court claim construction was the issue it had to decide. Then it did not decide that issue, but a different issue, whether the applicant, Freeman was bound by issue preclusion.

                  Under any Supreme Court case you may want to review, a third party is NOT BOUND, unless they are in privity. Certainly Freeman is way, way out of step with controlling authority on the issue that Rich said that the Federal Circuit had to decide and apparently did decide adversely to Freeman. The PTO simply is not bound by district court rulings to which they were not a party.

                  The exacerbating problem here also is that Freeman attempted to amend the claims to restore the original, intended meaning, a meaning that the Board agreed was always the intended meaning.

                  Again, both the examiner and the Board were not bound by the district court ruling and were within their rights to have ruled differently. However, if they agreed with and then adopted the district court construction, and then rejected the claims based on that construction, I daresay that I would then agree that the applicant might be estopped under the principles of collateral estoppel to argue that the PTO’s construction was wrong.

                  Rich’s thinking in this case is obtuse.

                7. From:
                  In re Trans Texas Holdings Corp., 498 F.3d 1290 (Fed. Cir. 2007).

                  Contrary to Trans Texas’s argument, our decision in Freeman does not require the application of issue preclusion. In Freeman, we held that a patentee in a PTO proceeding was barred by issue preclusion from asserting a claim construction already rejected in a district court infringement action brought by the patentee against a third party. 30 F.3d at 1469. Nothing in Freeman suggests that issue preclusion could be applied against the PTO or another non-party to the infringement proceeding…

                  In other words, Rich railroaded Freeman, gave him the old knife in the back. The Board disagreed with the district court, but Rich did not let the Board’s opinion get in the way of skewering Freeman.


          Yes let’s just have judges sit in their chambers and pretend they are experts like in Alice and decide whether collateral estoppel should apply. They could come up with a new test Alice where they just consider the gist of the claims and then wash away all the differences. We can have opinions like, “well, both of these have to do with information processing, so maybe we should just adapt the claim construction from IBM for the Microsoft patent. Save us all a lot of time.”

  4. 1

    I thought the most interesting issue was that the claim asserted and for which collateral estoppel was applied, claim 33, was added in a reexamination conducted after the prior litigation. The term at issue in the prior litigation and in litigation at bar was not at issue in the reexamination. The implication of this is that if that limitation was the subject of the reexamination, that collateral estoppel may not apply.

      1. 1.1.1

        Dennis, I think this might apply to any post grant proceeding where the Board gives the term at issue a construction at variance from the court’s construction as opposed to simply adopting the court construction.

        Petitioners should be aware not to argue a different construction of a term for which there is a prior favorable court construction. Petitioners might be inclined to do this to read a claim on prior art.

Comments are closed.