PLAGER, Circuit Judge. ■
Miguel F. Brana,
et al.
(applicants), appeal the March 19, 1993 decision of the United States Patent and Trademark Office (PTO) Board of Patent Appeals and Interferences (Board), in Appeal No. 92-1196. The Board affirmed the examiner’s rejection of claims 10-13 of patent application Serial No. 533,944 under 35 U.S.C. § 112 ¶ 1 (1988).
The examiner’s rejection, upon which the Board relied in rendering its decision, was based specifically on a challenge to the utility of the claimed compounds and the amount of experimentation necessary to use the compounds. We conclude the Board erred, and reverse.
I. BACKGROUND
On June 30, 1988, applicants filed patent application Serial No. 213,690 (the ’690 application)
directed to 5-nitrobenzo[de]isoqui-noline-l,3-dione compounds, for use as anti-tumor substances, having the following formula:
[[Image here]]
where n is 1 or 2, R1 and R2 are identical or different and are each hydrogen, Ci-C6-al-kyl, Ci-C6-hydroxyalkyl, pyrrolidinyl, mor-pholino, piperidinyl or piperacinyl, and R3 and R4 are identical or different and are each hydrogen, Ci-C6-alkyl, Ci-Cg-acyl, C2-C7-alkoxycarbonyl, ureyl, aminocarbonyl or C2-C7-alkylaminoearbonyl. These claimed compounds differ from several prior art ben-zo[de]isoquinoline-l,3-dione compounds due to the presence of a nitro group (02N) at the 5-position and an amino or other amino group (NR3R4) at the 8-position of the iso-quinoline ring.
The specification states that these non-symmetrical substitutions at the 5- and 8-positions produce compounds with
“a
better action and a better action spectrum as antitu-mor substances” than known benzo[de]isoqui-nolines, namely those in K.D. Pauli et al.,
Computer Assisted Structure-Activity Correlations,
Drug Research, 34(11), 1243-46 (1984) (Pauli). Pauli describes a computer-assisted evaluation of benzo[de]isoquinoline-1,3-diones and related compounds which have been screened for antitumor activity by testing their efficacy
in
vivo
against two specific implanted murine (i.e., utilizing mice as test subjects) lymphocytic leukemias, P388 and L1210.
These two
in vivo
tests are
widely used by the National Cancer Institute (NCI) to measure the antitumor properties of a compound. Pauli noted that one compound in particular, benzo[de]isoquinoline-l,3(2H)dione,5-amino-2(2-dimethyl-ammoe-thyl [sic]) (hereinafter “NSC 308847”), was found to show excellent activity against these two specific tumor models. Based on their analysis, compound NSC 308847 was selected for further studies by NCI. In addition to comparing the effectiveness of the claimed compounds with structurally similar compounds in Pauli, applicants’ patent specification illustrates the cytotoxicity of the claimed compounds against human tumor cells,
in vitro,
and concludes that these tests “had a good action.”
The examiner initially rejected applicants’ claims in the ’690 application as obvious under 35 U.S.C. § 103 in light of U.S. Patent No. 4,614,820, issued to and referred to hereafter as Zee-Cheng et al. Zee-Cheng et al. discloses a benzo[de]isoquinoline compound for use as an antitumor agent with symmetrical substitutions on the 5-position and 8-position of the quinoline ring; in both positions the substitution was either an amino or nitro group.
Although not identical to the applicants’ claimed compounds, the examiner noted the similar substitution pattern (i.e., at the same positions on the isoquinoline ring) and concluded that a mixed substitution of the invention therefore would have been obvious in view of Zee-Cheng et al.
In a response dated July 14, 1989, the applicants rebutted the § 103 rejection. Applicants asserted that their mixed disubsti-tuted compounds had unexpectedly better antitumor properties than the symmetrically substituted compounds in Zee-Cheng et al. In support of this assertion applicants attached the declaration of Dr. Gerhard Keil-hauer. In his declaration Dr. Keilhauer reported that his tests indicated that' applicants’ claimed compounds were far more effective as antitumor agents than the compounds disclosed in Zee-Cheng et al. when tested,
in vitro,
against two specific types of human tumor cells, HEp and HCT-29.
Applicants farther noted that, although the differences between the compounds in Zee-Cheng et al. and applicants’ claimed compounds were slight, there was no suggestion in the art that these improved results (over Zee-Cheng et al.) would have been expected. Although the applicants overcame the § 103 rejection, the examiner nevertheless issued a final rejection, on different grounds, on September 5, 1989.
On June 4,1990, applicants filed a continuation application, Serial No. 533,944 (the ’944 application), from the above-mentioned ’690 application. Claims 10-13, the only claims remaining in the continuation application, were rejected in a final office action dated May 1,1991. Applicants appealed the examiner’s final rejection to the Board.
In his answer to the applicants’ appeal brief, the examiner stated that the final rejection was based on 35 U.S.C. § 112 ¶ l.
The examiner first noted that the specification failed to describe any specific disease against which the claimed compounds were active. Furthermore, the examiner concluded that the prior art tests performed in Pauli and the tests disclosed in the specification were not sufficient to establish a reasonable expectation that the claimed compounds had
a practical utility (i.e. antitumor activity in humans).
In a decision dated March 19, 1993, the Board affirmed the examiner’s final rejection. The three-page opinion, which lacked any additional analysis, relied entirely on the examiner’s reasoning. Although noting that it also would have been proper for the examiner to reject the claims under 35 U.S.C. § 101, the Board affirmed solely on the basis of the Examiner’s § 112 ¶ 1 rejection. This appeal followed.
II. DISCUSSION .
At issue in this case is an important question of the legal constraints on patent office examination practice and policy. The question is, with regard to pharmaceutical inventions, what must the applicant prove regarding the practical utility or usefulness of the invention for which patent protection is sought. This is not a new issue; it is one which we would have thought had been settled by case law years ago.
We note the Commissioner has recently addressed this question in his Examiner Guidelines for Bio-tech Applications,
see
60 Fed.Reg. 97 (1995); 49 Pat.Trademark & Copyright J. (BNA) No. 1210, at 234 (Jan. 5, 1995).
The requirement that an invention have utility is found in 35 U.S.C. § 101: “Whoever invents ... any new and
useful ...
composition of matter ... may obtain a patent there-for_” (emphasis added). It is also implicit in § 112 ¶ 1, which reads:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Obviously, if a claimed invention does not have utility, the specification cannot enable one to use it.
As noted, although the examiner and the Board both mentioned § 101, and the rejection appears to be based on the issue of whether the compounds had a practical utility, a § 101 issue, the rejection according to the Board stands on the requirements of § 112 ¶ 1. It is to that provision that we address ourselves.
The Board gives two reasons for the rejection;
we will consider these in turn.
1.
The first basis for the Board’s decision was that the applicants’ specification failed to disclose a specific disease against which the claimed compounds are useful, and therefore, absent undue experimentation, one of ordinary skill in the art was precluded from using the invention.
See Hybritech Inc. v. Monoclonal Antibodies, Inc.,
802 F.2d 1367, 1384, 231 USPQ 81, 94 (Fed.Cir.1986),
cert. denied,
480 U.S. 947, 107 S.Ct. 1606, 94 L.Ed.2d 792 (1987). In support, the Commissioner argues that the disclosed uses in
the ’944 application, namely the “treatment of diseases” and “antitumor substances,” are similar to the nebulous disclosure found insufficient in
In re Kirk,
376 F.2d 936, 153 USPQ 48 (CCPA 1967). This argument is not without merit.
In
Kirk
applicants claimed a new class of steroid compounds. One of the alleged utilities disclosed in the specification was that these compounds possessed “high biological activity.”
Id.
at 938, 153 USPQ at 50. The specification, however, failed to disclose which biological properties made the compounds useful. Moreover, the court found that known specific uses of similar compounds did not cure this defect since there was no disclosure in the specification that the properties of the claimed compounds were the same as those of the known similar compounds.
Id.
at 942, 153 USPQ at 53. Furthermore, it was not alleged that one of skill in the art would have known of any specific uses, and therefore, the court concluded this alleged use was too obscure to enable one of skill in the art to use the claimed invention.
See also Kawai v. Metlesics,
480 F.2d 880, 178 USPQ 158 (CCPA 1973).
Kirk
would potentially be dispositive of this case were the above-mentioned language the only assertion of utility found in the ’944 application. Applicants’ specification, however, also states that the claimed compounds have “a better action and a better action spectrum as antitumor substances” than known compounds, specifically those analyzed in Pauli. As previously noted,
see supra
note 4, Pauli grouped various ben-zo[de]isoquinoline-l,3-diones, which had previously been tested
in vivo
for antitumor activity against two lymphocytic leukemia tumor models (P388 and L1210), into various structural classifications and analyzed the test results of the groups (i.e. what percent of the compounds in the particular group showed success against the tumor models). Since one of the tested compounds, NSC 308847, was found to be highly effective against these two lymphocytic leukemia tumor models,
applicants’ favorable comparison implicitly asserts that their claimed compounds are highly effective (i.e. useful) against lymphocytic leukemia. An alleged use against this particular type of cancer is much more specific than the vaguely intimated uses rejected by the courts in
Kirk
and
Kawai. See, e.g., Cross v. Iizuka,
753 F.2d at 1048, 224 USPQ at 745 (finding the disclosed practical utility for the claimed compounds — the inhibition of thromboxane synthetase in human or bovine platelet mi-crosomes — sufficiently specific to satisfy the threshold requirement in
Kirk
and
Kawai.)
The Commissioner contends, however, that P388 and L1210 are not diseases since the only way an animal can get sick from P388 is by a direct injection of the cell line. The Commissioner therefore concludes that applicants’ reference to Pauli in their specification does not provide a specific disease against which the claimed compounds can be used. We disagree.
As applicants point out, the P388 and L1210 cell lines, though technically labeled tumor models, were originally derived from lymphocytic leukemias in mice. Therefore, the P388 and L1210 cell lines do represent actual specific lymphocytic tumors; these models will produce this particular disease once implanted in mice. If applicants were required to wait until an animal naturally developed this specific tumor before testing the effectiveness of a compound against the tumor
in vivo,
as would be implied from the Commissioner’s argument, there would be no effective way to test compounds
in vivo
on a large scale.
We conclude that these tumor models represent a specific disease against which the claimed compounds are alleged to be effective. Accordingly, in light of the explicit reference to Pauli, applicants’ specification alleges a sufficiently specific use.
2.
The second basis for the Board’s rejection was that, even if the specification did allege a specific use, applicants failed to
prove that the claimed compounds are useful. Citing various references,
the Board found, and the Commissioner now argues, that the tests offered by the applicants to prove utility were inadequate to convince one of ordinary skill in the art that the claimed compounds are useful as antitumor agents.
This court’s predecessor has stated:
[A] specification disclosure which contains a teaching of the manner and process of making and using the invention in terms which correspond in scope to those used in describing and defining the subject matter sought to be patented
must
be taken as in compliance with the enabling requirement of the first paragraph of § 112
unless
there is reason to doubt the objective truth of the statements contained therein which must be relied on for enabling support.
In re Marzocchi,
439 F.2d 220, 223, 169 USPQ 367, 369 (CCPA 1971). From this it follows that the PTO has the initial burden of challenging a presumptively correct assertion of utility in the disclosure.
Id.
at 224, 169 USPQ at 370. Only after the PTO provides evidence showing that one of ordinary skill in the art would reasonably doubt the asserted utility does the burden shift to the applicant to provide rebuttal evidence sufficient to convince such a person of the invention’s asserted utility.
See In re Bundy,
642 F.2d 430, 433, 209 USPQ 48, 51 (CCPA 1981).
The PTO has not met this initial burden. The references cited by the Board, Pazdur and Martin,
do not question the usefulness of any compound as an antitumor agent or provide any other evidence to cause one of skill in the art to question the asserted
utility of applicants’ compounds. Rather, these references merely discuss the therapeutic predictive value of
in vivo
murine tests — relevant only if applicants must prove the ultimate value in humans of their asserted utility. Likewise, we do not find that the nature of applicants’ invention alone would cause one of skill in the art to reasonably doubt the asserted usefulness.
The purpose of treating cancer with chemical compounds does not suggest an inherently unbelievable undertaking or involve implausible scientific principles.
In re Jolles,
628 F.2d at 1327, 206 USPQ at 890. Modem science has previously identified numerous successful chemotherapeutic agents. In addition, the prior art, specifically Zee Cheng et al., discloses structurally similar compounds to those claimed by the applicants which have been proven
in vivo
to be effective as chemotherapeutic agents against various tumor models.
Taking these facts — the nature of the invention and the PTO’s proffered evidence— into consideration we conclude that one skilled in the art would be without basis to reasonably doubt applicants’ asserted utility on its face. The PTO thus has not satisfied its initial burden. Accordingly, applicants should not have been required to substantiate their presumptively correct disclosure to avoid a rejection under the first paragraph of § 112.
See In re Marzocchi,
439 F.2d at 224, 169 USPQ at 370.
We do not rest our decision there, however. Even if one skilled in the art
would have reasonably questioned the asserted utility, i.e., even if the PTO met its initial burden thereby shifting the burden to the applicants to offer rebuttal evidence, applicants proffered sufficient evidence to convince one of skill in the art of the asserted utility. In particular, applicants provided through Dr. Kluge’s declaration
test results showing that several compounds within the scope of the claims exhibited significant antitumor activity against the L1210 standard tumor model
in vivo.
Such evidence alone should have been sufficient to satisfy applicants’ burden.
The prior art further supports the conclusion that one skilled in the art would be convinced of the applicants’ asserted utility. As previously mentioned, prior art — Zee Cheng et al. and Pauli — disclosed structurally similar compounds which were proven
in vivo
against various tumor models to be effective as chemotherapeutic agents. Although it is true that minor changes in chemical compounds can radically alter their effects on the human body,
Kawai,
480 F.2d at 891, 178 USPQ at 167, evidence of success in structurally similar compounds is relevant in determining whether one skilled in the art would believe an asserted utility.
See Rey-Bellet v. Engelhardt,
493 F.2d 1380, 181 USPQ 453 (CCPA 1974);
Kawai,
480 F.2d 880, 178 USPQ 158.
The Commissioner counters that such
in vivo
tests in animals are only preclinieal tests to determine whether a compound is suitable for processing in the second stage of testing, by which he apparently means
in vivo
testing in humans, and therefore are not reasonably predictive of the success of the claimed compounds for treating cancer in humans.
The Commissioner, as did the Board, confuses the requirements under the law for obtaining a patent with the requirements for obtaining government approval to market a particular drug for human consumption.
See Scott v. Finney,
34 F.3d 1058, 1063, 32 USPQ2d 1115, 1120 (Fed.Cir.1994) (“Testing for the full safety and effectiveness of a prosthetic device is more properly left to the Food aiid Drug Administration (FDA). Title 35 does not demand that such human testing occur within the confines of Patent and Trademark Office (PTO) proceedings.”).
Our court’s predecessor has determined that proof of an alleged pharmaceutical property for a compound by statistically significant tests with standard experimental animals is sufficient to establish utility.
In re Krimmel,
292 F.2d 948, 953, 130 USPQ 215, 219 (CCPA 1961);
see also In re Bergel,
292 F.2d 958, 130 USPQ 205 (CCPA 1961). In concluding that similar
in vivo
tests were adequate proof of utility the court in
In re Krimmel
stated:
We hold as we do because it is our firm conviction that one who has taught the public that a compound exhibits some desirable pharmaceutical property in a standard experimental animal has made a significant and useful contribution to the art, even though it may eventually appear that the compound is without value in the treatment in humans.
Krimmel,
292 F.2d at 953, 130 USPQ at 219. Moreover, NCI apparently believes these tests are statistically significant because it has explicitly recognized both the P388 and L1210 murine tumor models as standard screening tests for determining whether new
compounds may be useful as antitumor agents.
In the context of this case the Martin and Pazdur references, on which the Commissioner relies, do not convince us otherwise. Pazdur only questions the reliability of the screening tests against lung cancer; it says nothing regarding other types of tumors. Although the Martin reference does note that some laboratory oncologists are skeptical about the predictive value of
in vivo
murine tumor models for human therapy, Martin recognizes that these tumor models continue to contribute to an increasing human cure rate. In fact, the authors conclude that this perception (i.e. lack of predictive reliability) is not tenable in light of present information.
On the basis of animal studies, and controlled testing in a limited number of humans (referred to as Phase I testing), the Food and Drug Administration may authorize Phase II clinical studies.
See
21 U.S.C. § 355(i)(l); 21 C.F.R. § 312.23(a)(5), (a)(8) (1994). Authorization for a Phase II study means that the drug may be administered to a larger number of humans, but still under strictly supervised conditions. The purpose of the Phase II study is to determine primarily the safety of the drug when administered to a larger human population, as well as its • potential efficacy under different dosage regimes.
See
21 C.F.R. § 312.21(b).
FDA approval, however, is not a prerequisite for finding a compound useful within the meaning of the patent laws.
Scott,
34 F.3d 1058, 1063, 32 USPQ2d 1115, 1120. Usefulness in patent law, and in particular in the context of pharmaceutical inventions, necessarily includes the expectation of further research and development. The stage at which an invention in this field becomes useful is well before it is ready to be administered to humans. Were we to require Phase II testing in order to prove utility, the associated costs would prevent many companies from obtaining patent protection on promising new inventions, thereby eliminating an incentive to pursue, through research and development, potential cures in many crucial areas such as the treatment of cancer.
In view of all the foregoing, we conclude that applicants’ disclosure complies with the requirements of 35 U.S.C. § 112 ¶ 1.
3.
The Commissioner takes this opportunity to raise the question of this court’s standard of review when deciding cases on appeal from the PTO. Traditionally we have recited our standard of review to be, with regard to questions of law, that review is without deference to the views of the Agency,
In re Donaldson,
16 F.3d 1189, 1192, 29 USPQ2d 1845, 1848 (Fed.Cir.1994) (in banc),
In re Caveney,
761 F.2d 671, 674, 226 USPQ 1, 3 (Fed.Cir.1985), and with regard to questions of fact, we defer to the Agency unless its findings are “clearly erroneous.”
See, e.g., In re Baxter Travenol Labs,
952 F.2d 388, 21 USPQ2d 1281 (Fed.Cir.1991);
In re Woodruff,
919 F.2d 1575, 16 USPQ2d 1934 (Fed.Cir.1990);
In re De Blauwe,
736 F.2d 699, 222 USPQ 191 (Fed.Cir.1984).
With regard to judgment calls, those questions that fall “[s]omewhere near the middle of the fact-law spectrum,” this court has recognized “the falseness of the fact-law dichotomy, since the determination at issue, involving as it does the application of a general legal standard to particular facts, is probably most realistically described as neither of fact nor law, but mixed.”
Campbell v. Merit Systems Protection Board,
27 F.3d 1560, 1565 (Fed.Cir.1994). When these questions of judgment are before us, whether we defer, and the extent to which we defer, turns on the nature of the case and the nature of the judgment.
Id.
(“Characterization therefore must follow from an
a priori
decision as to whether deferring ... is sound judicial policy. We would be less than candid to suggest otherwise.”).
The Commissioner contends that the appropriate standard of review for this court regarding questions of law, of fact, and mixed questions of law and fact, coming to us from the PTO is found in the Administrative Procedure Act (APA) at 5 U.S.C. § 706. The standard set out there is that “[t]he reviewing court shall ... hold unlawful and set aside agency action, findings, and conclusions found to be — (A) arbitrary, capricious, an
abuse of discretion, or otherwise not in accordance with law; ... (E) unsupported by substantial evidence_” The Commissioner is of the view that the stated standard we now use, which is the traditional standard of review for matters coming from a trial court, is not appropriate for decisions coining from an agency with presumed expertise in the subject area, and is not in accord with law.
Applicants argue that by custom and tradition, recognized by the law of this court, the standard of review we have applied, even though inconsistent with the standard set forth in the APA, nevertheless is a permissible standard. In our consideration of this issue, there is a reality check: would it matter to the outcome in a given case which formulation of the standard a court articulates in arriving at its decision? The answer no doubt must be that, even though in some eases it might not matter, in others it would, otherwise the lengthy debates about the meaning of these formulations and the circumstances in which they apply would be unnecessary.
A preliminary question, then, is whether this is one of those eases in which a difference in the standard of review would make a difference in the outcome. The ultimate issue is whether the Board correctly applied the § 112 ¶ 1 enablement mandate and its implicit requirement of practical utility, or perhaps more accurately the underlying requirement of § 101, to the facts of this ease. As we have explained, the issue breaks down into two subsidiary issues: (1) whether a person of ordinary skill in the art would conclude that the applicants had sufficiently described particular diseases addressed by the invention, and (2) whether the Patent Act supports a requirement that makes human testing a prerequisite to patentability under the circumstances of this case.
The first subsidiary issue, whether the application adequately described particular diseases, calls for a judgment about what the various representations and discussions contained in the patent application’s specification would say to a person of ordinary skill in the art. We have considered that question carefully, and, for the reasons we explained above in some detail, we conclude that the Board’s judgment on this question was erroneous. Our conclusion rests on our understanding of what a person skilled in the art would gather from the various art cited, and from the statements in the application itself. We consider the Board’s error to be sufficiently clear that it is reversible whether viewed as clear error or as resulting in an arbitrary and capricious decision.
The second subsidiary issue, whether human testing is a prerequisite to patentability, is a pure question of law: what does the practical utility requirement mean in a case of this kind. Under either our traditional standard or under the APA standard no deference is owed the Agency on a question of law, and none was accorded.
If the question concerning the standard of review, raised by the Commissioner, is to be addressed meaningfully, it must arise in a case in which the decision will turn on that question, and, recognizing this, the parties fully brief the issue. This is not that case. We conclude that it is not necessary to the disposition of this case to address the question raised by the Commissioner; accordingly, we decline the invitation to do so.
III. CONCLUSION
The Board erred in affirming the examiner’s rejection under 35 U.S.C. § 112 ¶ 1. The decision is reversed.
REVERSED.