Harmless Error at the Federal CircuitFebruary 28, 2010

A recent decision by the Court of Appeals for the Federal Circuit addressed the issue of how much deference should be given to a decision by the Board of Patent Appeals and Interferences when the Board makes an error in ascertaining the teachings of references.

The appeal concerns the status of U.S. Patent Application number 09/719,045 which had been rejected as either anticipated by or obvious under one or two references cited by the examiner at the U.S. Patent and Trademark Office. The technology of the appeal concerns "divalent antibody fragments," which are pieces of an antibody which are used for their advantages over full antibodies in certain diagnostic and theraputic uses. The prior art had described some elements of the claimed invention, but the Board made several factual findings during the appeal that were later determined to be false.

In its decision, the CAFC noted that the USPTO is subject to the Administrative Procedures Act, the caselaw from this act requires that the decisions of an administrative proceeding can only be reversed for error if the error is not deemed to be "harmless." The burden of establishing that the error was not harmless rests with the party attacking the decision of the agency. In reviewing the art and the decisions made by the Board, the CAFC determined that the errors made were not harmless, and therefore remand was necessary. The CAFC did not address the applicant's arguments that the claimed invention was nonobvious.

As previously described, the present application involves divalent antibody fragments. These are produced by digesting antibodies using specific enzymes, in this case pepsin, to create antibody fragments. Antibodies are generally Y-shaped with two "arms," each arm formed by one heavy and one light "chain." The chains are linked at the base of the arm by a disulphide bridge (formed by a covalent bond between two sulphur atoms) and at their other end bind to antigens on specific foreign bodies. The two arms are joined to the stem and one another by another disulphide bridge. The pepsin enzyme acts to remove the stem of the antibody, leaving the two arms joined together in a dumbell shape.

The inventors, Andrew P. Chapman and David J. King (together "Chapman"), in their application teach removing the stem from the antibody, removing the light chains from each of the resulting arms, and coupling two heavy chains through the use of an "interchain bridge" rather than the disulphide bridge. This interchain bridge is accomplished through a polymer molecule attached to the heavy chains. The polymer also increases the "circulating half life" of the resulting antibody fragment.

The examiner cited two references against Chapman's application, Gonzalez (U.S. Patent No. 6,025,158) and Barbanti (U.S. Patent No. 5,436,154). The examiner rejected claim 1, either anticipated by or, in the alternative, obvious over Gonzalez. During the appeal to the Board, the examiner made, and the Board adopted, a number of factual findings regarding the scope of Gonzales. It was acknowledged by the government on appeal to the CAFC that these factual findings were incorrect.

On appeal, Chapman argued for reversal based on Gonzales teaching away from his invention and the invention being non-obvious as a matter of law. The court found that as a factual matter, Gonzalez does not teach away from the Chapman invention, and that the question of obviousness should be determined on remand to the Board when it considers the question of obviousness without the incorrect factual findings of the examiner.

The court's decision therefore was centered on whether the error in the three factual findings was harmless, under the Administrative Procedures Act. On the first error, the court noted that the board did not rely on the information stated in the first factual finding, and therefore the error could be considered harmless. The second and third errors were used, in part, by the Board to support its finding of obviousness. Had the errors not been made, then the question of obviousness may have been made differently, and therefore the error was not harmless. In order for error to be harmless, the court must be able to "say with confidence that the Board would have reached the same conclusion in the absense of these errors." Slip Op. at 16 (citing Kotteakos v. United States 328 U.S. 750, 765 (1946)). However, on remand, the board only needs to review its finding of obviousness with respect to the correct interpretation of Gonzalez.

The case of In re Chapman is available here.

← Return to Filewrapper

Stay in Touch

Receive the latest news and updates from us and our attorneys.

Sign Up