In dissenting from the conclusion of the majority, I do so because it is my belief that the facts of record require a reversal of the decision of the board.
Appellant’s position here is predicated upon his discovery that an admittedly known compound, designated NF-246, could be effectively used, when orally administered in dosage unit form to a host, to combat bacterial urinary tract infections. His position is thus based upon 35 U.S.C. § 100(b) which provides that a process which involves the new use of a known composition of matter may be patented provided the other conditions for patentability are satisfied.
The position of the Patent Office is that, in view of the close relationship of the chemical structures involved, it would be obvious that appellant’s compound, NF-246, would also possess the same properties as nitrofurantoin, NF-253, a well-known urinary tract antibacterial. Considering this position as prima facie establishing obviousness, it seems to me this prima facie case is rebutted by the record developed by the appellant. It is on this basis that I would reverse the decision of the board.
Structural similarities in chemical compounds, like homology, provide the chemist with a ready tool for classification. This fact provides a continuing temptation to convert a chemical classification into a legal presumption, i. e., that since all chemicals of a given class have certain common properties there is a “legal presumption” that one would be legally obvious in view of the other. This fallacy has been before the court in different guises and frequently is predicated on a doctrinal extension of our decision in In re Henze, 181 F.2d 196, 37 CCPA 1009 (1950). However, as we pointed out in In re Mills, 281 F.2d 218, 47 CCPA 1185 (1960), the Henze case is not authority which supports such a presumption. As we pointed out in Mills, 281 at 222, id. at 1190:
The term “presumption of unpatentability,” as it is used in the Henze case, refers to an inference of fact. [Footnote omitted.]
******
In the Henze case, this court went to a considerable length to safeguard applicants against the observed tendency of the Patent Office to freeze into legal rules of general application what, at best, are statements applicable to particular fact situations. Thus, the “presumption of unpatentability” referred to in the Henze case, was limited to a claim directed to a composition of matter (a new compound), the *650adjacent homologue of which was old in the art.
After pointing out how this “presumption” placed a burden on the applicant, the court [in Henze] even suggested certain types of evidence which applicant could produce to overcome this evidentiary burden and further concerned itself with limiting the types of proofs which could be called for by the Patent Office when it said:
This does not mean that in every case the Patent Office would be justified in exacting empirical data from an applicant respecting the properties of the prior art homologue tested under the same conditions and limitations as are set forth for the claimed compound where the reaction of the old compound under those conditions is a well established fact of general cognizance in the art. * * * [37 CCPA at 1015, 181 F.2d at 201, 85 USPQ at 265.]
The court [in Henze] further stated:
The appellant was not refused a patent in limine but was only placed under a reasonable requirement to overcome a presumption reasonably raised and which he could reasonably be expected to meet. *. * * The Patent Office as the public’s representative has the right to require such evidence of invention as is suitable to dissolve a presumption of unpatentability arising out of the nature of the subject matter where the criteria raising the presumption, as here, are of universal acceptance by those skilled in the art involved. * * * [37 CCPA at 1016, 181 F.2d at 201, 85 USPQ at 265.]
Thus considered, the structural similarities of the compounds NF-246 and NF-253 give rise to the same legal issues as did the existence of homologous compounds in the Mills case. As we there stated, 281 F.2d at 224, 47 CCPA at 1191:
* * * Homology provides for the chemist a convenient system of structural classification. Inherent in that system are differences as well as similarities in the properties and reactions of the members of any given homologous series.
A chemist, and it is from the standpoint of a chemist skilled in this art that the question of obviousness must be resolved, would consider the differences as well as the similarities in the properties and reactions of the members of any given homologous series. * * * Homology per se should, therefore, be treated as a chemist would treat it, being nothing more than a fact which must be considered with all other relevant facts before arriving at the conclusion of “obviousness” specified in 35 U.S.C. § 103.
Appellant argues that reference to the compound NF-246 in Gever et al., which admittedly teaches that NF-246 has a wide spectrum of anti-microbial activity and can be orally administered, does not suggest urinary tract activity for the compound. He relies for factual support of this position on the affidavit of Dr. Mary Paul as rebutting the examiner’s position, and asserts error in the treatment accorded the Paul affidavit in the Patent Office. The Paul affidavit contains a relevant showing of a factual basis from which we can draw a legal conclusion. See In re Weber, 341 F.2d 143, 52 CCPA 1015 (1965); In re Chilowsky, 306 F.2d 908, 50 CCPA 806 (1962).
It is noted that neither the examiner nor the board challenged Dr. Paul’s qualifications in this art. She states that she was familiar with each of the references relied upon to support the rejection. In particular, she commented on her publication (a “prior art” reference against appellant’s claims) as follows:
The Paul et al. publication, some years subsequent to the Gever et al. patent dealing with new chemical compounds including NF-246 and the Mintzer et al. publication dealing with an investigation of nitrofurantoin, presents data on the physiological disposition and distribution of a variety of nitrofurans when administered to *651laboratory animals. These data are tabulated in a series of Tables (I-VII). They represent a digest of then available determinations of blood level, urinary excretion, fecal excretion, plasma protein binding, excretion in milk, biliary excretion, presence in cerebrospinal fluid, and tissue degradation of certain nitrofurans, affording to those interested an appreciation of their physiological disposition and distribution.
In the commentary on the urinary excretion of the nitrofurans involved at pages 293 and 294 and referring to Table I it is noted that among the nitrofurans urinary excretion varied from little or none to 75% of the dose administered. The influence of species is also noted, nitrofurantoin maintaining its urinary excretion in rat, dog and man but another nitrofuran, nidroxyzone, dwindles from 25% in rats to 16% in dogs to 6% in man. Other remarks to particular nitrofurans include reference to secretion of one of them, NF-189, by renal tubules and the absence of any substantial amount of furazolidone in the urine. There is no further reference to any particular nitrofuran.
Table I lists a series of nitrofurans and the determination of plasmas levels, urinary excretion and fecal concentration thereof when administered orally to rats. From the data presented there I would not expect, with a high degree of probability, that a trial of NF-246 as a urinary tract agent would be successful. In comparison to nitrofurantoin, its water solubility is lower (88 vs. 190); its plasma level somewhat greater, (lf.7 vs. 2.6); its urinary excretion considerably lower (10 vs. 52) and its fecal concentration considerably less (0.1 vs. 2.0). On their face these values would not convince me that a trial of NF-246 as a urinary tract agent could have any better than hope of success. [Emphasis added.]
While some of Dr. Paul’s affidavit is a statement of her opinion, it is entitled to consideration as the opinion of an expert predicated on specific factual data as there indicated in those portions of the affidavit which have been underlined. The factual differences referred to by Dr. Paul indicate that the water solubility and urinary excretion properties of NF-246 are inconsistent with the comparable properties of NF-153. Those inconsistent properties provide a factual rebuttal to the conclusion reached by the examiner and the board from inferences predicated on similarities in chemical structures of the two compounds. Thus, I feel the board and the examiner erred in not giving proper weight to the Paul affidavit as rebutting the inferential prima facie case of obviousness upon which they relied.
Appellant properly points out that the significance of the structural similarities between compounds is not conclusive on the issue of obviousness. He states:
* * * That seemingly slight variations in chemical structures can result in marked differences in biological properties is illustrated in the present case by a comparison of the sensitivity of certain strains or organisms to NF-246 while lacking sensitivity to NF-153 (Table 4 of Attachments filed with Affidavit of Regan * * *) and the marked difference in urinary excretion rate, 52% for NF-153, while only 10% for NF-246 * * *.
The affidavit of Dr. John Ward Regan also was submitted by appellant to show the effectiveness of NF-246 in treating urinary tract infection. Appellant argues that this affidavit, together with the supportive studies, shows NF-246 to be effective against a greater range of bacterial organisms than NF-153 and to exhibit a lesser incidence of toxic manifestations.
The examiner dismissed the affidavit for its failing to show such improvement over NF-153 as to overcome the presumption of obviousness. He dismissed the showing of effectiveness against a greater range of bacterial organisms as based on an in vitro test rather than on an in vivo test and there*652fore found it not definitive. He failed-to note that while the test was in vitro, it was nevertheless conducted against organisms found in the urine of those patients participating in the clinical study.
Further, the in vitro tests seem to me to have been a direct comparison of the 'effects of both NF-153 and NF-246 on organisms found in the urine of the clinical patients. Twenty of the organisms were sensitive to NF-246 and insensitive to NF-153 while only 3 were sensitive to NF-153 and insensitive to NF-246. This test, as noted by the author, is the standard and only way to determine which agent would be most likely to eradicate the offending organism.
The examiner and the board seem to have ignored the in vitro tests and factual conclusions which Dr. Regan draws from them. As pointed out in In re Isaacs, 347 F.2d 887, 52 CCPA 1791, (1965):
* * * One thing seems clear: both the examiner and the board felt that appellants should have submitted evidence of in vivo tests. No authority has been cited and we have been able to find none which requires that in order to secure a patent, utility of a pharmacologically active substance must be proved by in vivo testing. * * *
Appellant avers that the board erred in failing to recognize that the clinical studies reported in the Regan affidavit were run, as are most initial investigations on new drugs, at dosages far below the recommended levels. Appellant urges that the results reported show good results, i. e., sterility of the urine in 18 out of the 30 patients (60%) receiving the drug. This is considered by appellant to be outstanding, especially if one notes that the 12 patients (40%) who failed to show improvement were receiving only 200 mg. per day (one received 300 mg.). This compares with the recommended dosage of the prior art compound NF-153 at 400 mg. per day for treatment of severe infections. Table 4 of the Regan affidavit shows that the 12 patients who failed to respond had an average age of about 75 and most had undergone prostatectomys.
Appellant strongly urges that the Regan affidavit does show properties for NF-246 which are both surprising and unexpected in view of the teachings in the prior art. One skilled in the art would not expect NF-246 to be useful as a urinary tract antibacterial in view of its low urinary excretion. However, even assuming that one would be led to experiment with NF-246 in this area,1 he would not expect to find a 60% cure where the majority of dosage levels ranged from about % to y2 that recommended for the leading urinary tract antibacterial NF-153. These results are even more surprising if one takes into consideration the decided disadvantage of a 10% urinary excretion as compared with a 50% urinary excretion for the urinary tract antibacterial NF-153.
Similarly, Mintzer et al. is cited by the examiner as describing four characteristics of a good urinary tact antibacterial:
(1) 4% or less of the amount administered orally is found in the feces; (2) high excretion of the compound in the urine (45% of administered dose); (3) low blood level; and (4) a broad antibacterial spectrum.
Mintzer et al., in my view, did not disclose that these four properties were criteria of a good urinary tract antibacterial. Rather, these criteria were a mere statement of properties previously disclosed by Eaton Laboratories investigations of NF-153 in animals. The following paragraph of that reference, however, as appellant urges, seems to me to indicate more aptly the property of NF-*653153 which interested Mintzer et al. in investigating the use of this compound as a urinary tract antibacterial wherein he states :
Since a concentration of Furadantin in the urine can reach at least 40 mg. per 100 cc. during therapy, in spite of its apparent maximum solubility in urine of about 20 mg. per 100 cc. at pH 5 in vitro, these results seem paradoxical. This is due to the fact that when large amount of Furadantin are dissolved in urine in vitro at the blood and glomerular pH of 7.5, reduction of the pH to 5.5 does not produce precipitation. A state of stable supersaturation exists. This phenomenon, coupled with the extremely low blood levels of the drug, suggests that no crystalluria need be feared with the use of this preparation. This conclusion is supported by our clinical observations.
Thus, it seems Mintzer et al. were primarily interested in the high urinary concentrations of Furadantin (NF-153), i. e. concentrations which exceeded its solubility in urine but still existed in a state of stable supersaturation. Thus, appellant argues that it was this phenomenon which was of prime interest to these investigators and it is in exactly this property in which NF-246 differs so appreciably from NF-153, 52% vs. 10% as evidenced by Table I of the Paul et al. reference. It is appellant’s contention that if one skilled in the art had envisioned possible urinary tract activity for NF-246, a reading of Mintzer et al. would discourage such an envisioned use of the compound. It is also appellant’s position that the discovery of urinary tract activity for NF-246 is both unexpected and surprising in view of its relatively low urinary excretion.
The factual base which appellant developed below is in my opinion of such strength as to overcome the prima facie case of the Patent Office based on inferences drawn by the examiner from the naked teachings of the art. In my opinion the reasons are sufficient to reverse the decision of the board.
. The court has often stated its basic disagreement -with an “obvious to try analysis” under 35 U.S.C. § 103. See In re Lindell, 385 F.2d 453, 55 CCPA - (1967) ; In re Tomlinson, 363 F.2d 928, 53 CCPA 1421 (1966) ; In re Huellmantel, 324 F.2d 998, 51 CCPA 845 (1963).