The Ghost of Lemelson: PTA Consequences of Exelixis v. Kappos

As shown, applications A and B have identical prosecutions with one subtle difference.  Applicant A files a first RCE just before the three year deadline, whereas applicant B files a first RCE just after the three year deadline.  Although that difference is subtle and slight, it results in 12 years of additional patent term (see purple line above)!

Applicants can also delay the prosecution of their applications to reach the 3 year deadline.  Any PTA reduction due to that delay will pale in comparison to the potential benefit.  35 U.S.C. 154(b)(2)(C).  In the above example, applicant A might reach the 3 year deadline simply by filing a one month extension of time.  That extension of time will result in a one month reduction of PTA.  Id.  Yet that single month of reduction effectively purchases about 12 years of additional patent term (12 years minus one month = 11 years and 11 months).

Of course, the doctrine of laches, and PTO regulations, provide disincentives against applicants causing “unnecessary delay.”  37 C.F.R. 11.18(b)(2)(i).  But these are unlikely to prevent applicants from delaying examination to reach the three year deadline.  Applicants can argue that the delay is necessary to optimize patent term.  Notably, the PTO recently proposed enabling applicants to purposefully delay examination for up to 30 months—without suggesting that doing so violates rule 11.18.  See PTO Proposes Major New Patent Application Processing Rules.

Despite these anomalous consequences, the district court defended its interpretation based on several reasons.  First, the court stated that the plain language of the statute compels the court’s interpretation.  I’m sympathetic to the court’s argument here.  Yet the statute is so complex, if not positively vague, that it defies any attempt to give it a “plain” meaning.  The court refers to “tolling” and a “clock.”  But the statute’s plain language does not include those terms.  The statute includes the adjacent phrases (A), (B), and (C) without stating, explicitly, whether phrase (B) modifies phrase (A) or (C).  If the statute’s meaning is so plain, how did the PTO get the interpretation wrong?

Moreover, if the PTO’s interpretation is plainly wrong, how did the public fail to detect the error until the district court’s decision?  As far as I know, the district court is the first to ever advance its interpretation of the statute.  The plaintiff, Exelixis, made a far more modest request of the district court.  See the discussion of Exelixis’s original request at the PharmaPatants blog.

The district court based its interpretation on two further reasons related to the statute’s overall “structure and purpose”: (1) that the title of the disputed subsection is “Guarantee of no more than 3-year application pendency,” and (2) that continued examination is not considered “applicant delay” under 154(b)(2)(C).  Both of these reasons are far from persuasive.

The district court’s statement that the goal of the statute is “Guarantee of no more than 3-year application pendency” is incorrect.  That is the title of the relevant subsection, but the title is a misnomer.  The subsection is directed to adjusting patent term, not application pendency.  Nothing in the statute guarantees that applications will remain pending for less than three years.  One wonders how the statute can make that guarantee, when applications routinely remain pending beyond three years.  If the statute guarantees pendency of less than three years, then what are these applications that remain pending for longer than three years—outlaws?

If one strained the interpretation, one might speculate that the statute is designed to incentivize the PTO to prevent the three year clock from expiring.  Otherwise the patent may receive a perpetually extended term, like application B in the above example.

There are two problems with this strained interpretation.  First, the potential for a longer, Lemelson-type patent term is hardly a threat to the PTO.  The PTO generally has no interest in shaping the lengths of patent terms, as long as the PTO fulfills its mandate in accordance with law.  Second, even if the PTO makes this three year deadline (e.g. by issuing a final rejection early enough), applicants still have the option to file an RCE.  That RCE may still extend pendency beyond three years, despite the PTO’s efforts.  Nothing in the statute guarantees that applicants will not do that.

If anything, the statute guarantees, with major exceptions, “no less than 17 years of patent term.”  That guarantee is very different than the title’s “[g]uarantee of no more than 3-year pendency.”  In fact, the district court’s interpretation gives applicants motivation to keep their applications pending.  Like Lemelson, and like applicant B above, these applicants can extend their patent term indefinitely into the future.

The district court’s reliance on the statute’s reference to applicant delay under 154(b)(2)(C) is also unavailing.  The court reasons that (A) adopting the PTO’s interpretation would construe RCE’s as “punishment,” (B) punishment is reserved for “applicant delay” under 154(b)(2)(C), and (C) RCEs cannot constitute applicant delay under 154(b)(2)(C) because they are not currently listed in 37 C.F.R. 1.704.  Each of these inferences is controversial.

The statute does not state that “applicant delay” under 154(b)(2)(C) is the only potential justification for denying applicants PTA.  On the contrary, every aspect of the statute denies applicants PTA, by defining the limits of PTA.  Defining the limits of PTA excludes otherwise infinite PTA.

The district court’s interpretation itself denies PTA to applicants by tolling the running of the 3 year clock.  Without the tolling, applicants would begin receiving type B PTA sooner.  So the district court’s concept of tolling “punishes” applicants for RCE’s filed before the 3 year clock expires.  If RCEs are a “valuable tool in the patent prosecution process,” as the district court repeatedly states, then why would Congress deny applicants PTA by tolling the 3 year clock for RCE’s?

Similarly, the district court bases its decision on the fact that the PTO does not list an RCE as one of the causes of applicant delay under rule 704.  But this reasoning is circular.  The court already considers the PTO’s interpretation of 154(b)(1)(B) as “punishing” applicants.  So the PTO had no reason to further punish applicants by listing RCEs as applicant delay.  To do so would doubly punish the applicants.  If the PTO had known that the court would overrule its interpretation, then the PTO may have categorized RCE’s as applicant delay (in appropriate circumstances).  But the PTO did not have the benefit of the district court’s interpretation when drafting rule 704.

Contrary to the district court’s statements otherwise, I believe that the statute’s overall “structure and purpose” supports the PTO’s interpretation.  The reasons are simple.  First, the purpose of the PTA statute is to convert patent terms from 17-years-from-issuance to 20-years-from-filing.  But the court’s interpretation reverses that change whenever applicants wait long enough to file an RCE.  Second, the court’s interpretation results in huge amounts of overlapping type B and type C PTA.  But these subsections appear to be mutually exclusive.  I’ll discuss these in turn.

The district court’s interpretation has the following profound consequence: the PTO’s failing to meet the three year deadline (e.g. by issuing a final rejection early enough) converts the application into a pre-1994 17-years-from-issuance application.  We can call these “Lemelson-type” applications.  You can see this in the example above: application B effectively has a 17-year-from-issuance patent term.  By filing the first RCE after the three year deadline, applicant B converted the application into a Lemelson-type application.  In contrast, applicant A “missed the boat,” and received a post-1994 20-years-from-filing term.  That term happened to be 12 years shorter than applicant B received.

Of course, Congress’s entire goal in 1994 was converting the patent term from 17-years-from-issuance to 20-years-from-filing.  According to the district court, Congress achieved that goal with one giant exception.  If the PTO fails to finally dispose of the application within three years, then the patent’s term snaps back to 17-years-from-issuance.  There is no evidence that Congress intended to create a giant exception to its overall goal of creating the 20-years-from-filing patent term.

In fact, RCEs are the functional equivalents of the continuations that Lemelson used to indefinitely extend his patent terms.  To the patent prosecutor, the differences between RCEs and continuations are relatively slight.  RCEs are simpler to prepare, which creates lower service fees.  Also, the government fee for an RCE is slightly less than for filing a continuation.  So the district court’s interpretation enables applicants to do with RCE’s what Lemelson famously did with continuations.  I doubt that Congress intended to create this loophole.

Second, the court’s interpretation results in huge amounts of overlapping type B and type C PTA.  According to the court, the RCE provision of subsection (b)(1)(B)(i) does not prevent the accumulation of type B PTA after the three year deadline.  Yet, by the same logic, the interference proceeding, secrecy order, and appeal provisions of (b)(1)(B)(ii) and (iii) would also not prevent type B PTA from accumulating after that deadline.  Section 154 reads in relevant part:

(B) Guarantee of no more than 3-year application pendency.— Subject to the limitations under paragraph (2), if the issue of an original patent is delayed due to the failure of the United States Patent and Trademark Office to issue a patent within 3 years after the actual filing date of the application in the United States, not including—

(i) any time consumed by continued examination of the application requested by the applicant under section 132 (b);

(ii) any time consumed by a proceeding under section 135 (a), any time consumed by the imposition of an order under section 181, or any time consumed by appellate review by the Board of Patent Appeals and Interferences or by a Federal court; or

(iii) any delay in the processing of the application by the United States Patent and Trademark Office requested by the applicant except as permitted by paragraph (3)(C),

The statute’s structure requires the court to treat subsections (ii) and (iii) (interference proceedings, secrecy orders, and appeals) the same as subsection (i) (RCEs).  But doing so apparently conflicts with 35 U.S.C. 154(b)(1)(C), which is specifically directed to interference proceedings, secrecy orders, and appeals.

Consider, again, the above example of the district court’s interpretation.  Application B survived the 3 year period without filing an RCE, thereby becoming a Lemelson-type application.  After the end of the 3 year period, the application will receive type B PTA for every day until the patent issues.  These days may include interference proceedings, secrecy orders, and appeals.  Yet the application will also receive type C PTA for any interference proceeding, secrecy order, or successful appeal.  35 U.S.C. 154(b)(1)(C).

In fact, subsection (b)(1)(C) specifies, among other things, that appeals must be successful in order to receive type C PTA.  Subsection (b)(1)(B) has no such qualifier.  So, in the above example of the court’s interpretation, applicant B receives PTA for unsuccessful, or even frivolous, appeals!  In general, it does not make sense to grant applicants PTA for unsuccessful appeals.

The better interpretation is that Congress designed subsections (b)(1)(B) and (b)(1)(C) to be mutually exclusive.  Subsection (b)(1)(B) excludes interference proceedings, secrecy orders, and appeals because subsection (b)(1)(C) addresses those situations.  Subsection (b)(1)(B) does not specify that appeals must be successful to receive PTA because subsection (b)(1)(C) specifies that limitation.  And subsection (b)(1)(C) governs appeals—all of them—regardless of whether the three year deadline has passed.

Of course, the statute addresses the potential for overlap when granting PTA.  35 U.S.C. 154(b)(2)(A).  But that provision is apparently directed to addressing potential overlap between PTA types A and B or between types A and C (or between grants within type A, B, or C).  The fact that (b)(2)(A) addresses other potential overlap does not change the fact that subsection (b)(1)(B) excludes certain situations because subsection (b)(1)(C) addresses those situations.  Although other portions of section 154 may overlap, Congress designed subsections (b)(1)(B) and (b)(1)(C) so that they did not.

At the time of my writing, the PTO still has the option to appeal the district court’s order.  If the PTO does, I think it has good odds of success.  The meaning of the statute is not as plain as the district court holds, as explained above.  Moreover, the district court’s interpretation has anomalous consequences, as the above example shows.  The PTO called these anomalous consequences “absurd,” and I tend to agree.  Lastly, the statute’s overall “structure and purpose” support the PTO’s interpretation, as explained above.

Of course, if the statute plainly supports the court’s interpretation, then the court should not alter that meaning to improve public policy.  That is the duty of Congress.    Meanwhile, in that case, clever inventors can convert their applications back to receiving 17-years-from-issuance patent terms.  So long as the court’s interpretation prevails, the specter of Lemelson’s ghost will continue to haunt accused infringers.