No , -1512, -1513, -1514, _Inthe. Jar _e Jtbtral Cirmit

Transcription

1 WEST/CRS No , -1512, -1513, -1514, Unteb Stat nthe Com't Of _gppeals Jar _e Jtbtral Crmt THERASENSE, NC. (now known as Abbott Dabetes Care, nc.) and ABBOTT LABORATORES, Plantffs-Appellants, V. PtL BECTON, DCKNSON AND COMPANY, u.s. COU_To,._ r_t" _pnc_r/4ppeatr _n,_ and NOVA BOMEDCAL CORPORATON,...lcAlCl_C07f u. Def endants-appellee_,ov "- 4 and da[_ho?baly BAYER HEALTHCARE LLC, CLE_,_ Defendant-Appellee. Appeal from the Unted States Dstrct Court for the Northern Dstrct of Calforna n consoldated case nos. 04-CV-2123, 04-CV-3327, 04-CV-3732, and 05-CV-3117, Judge Wllam H. Alsup. BREF OF PLANTFF-APPELLANTS ABBOTT LABORATORES AND ABBOTT DABETES CARE, NC. Roht K. Sngla, Esq. (SBN ) Jason A. Rantanen, Esq. (SBN ) MUNGER, TOLLES & OLSON LLP 560 Msson Street, 27 th Floor San Francsco, CA Telephone: (415) Facsmle: (415) Attorneys for Plantff-Appellants Abbott Dabetes Care, nc. and Abbott Laboratores Bll Ward, Esq. (SBN ) MUNGER, TOLLES & OLSON LLP 355 South Grand Avenue, 35th Floor Los Angeles, CA Telephone: (213) Facsmle: (213)

2 "t UNTED STATES COURT OF APPEALS FOR THE FEDERAL CRCUT Therasense CERTFCATE v. Becton No ,-1512,-1513,-1514,-1595 OF NTEREST Counsel for the (pettoner) Q (respondent) (appellee) (amcu's) (name of party) Abbott D_te_ Caz n, _._a^_ubo_at_a_ certfes the followng (use "None" f aplslcable: use extra sheets f necessary): 1. The full name of every party or amcus repres_nted by me s: Therasense, nc. (now known as Abbott Dabetes Care, nc.) and Abbott Laboratores 2. The name of the real party n nterest (f the party named n the capton s not the real part3,' n nterest) represented by me s: None. 3. A parent corporatons and any publcly held companes tllat own l0 percent or more of the stock of the party or ameus curae represented by me are: None, 4. The names of a[[ law frms and the parmers or assocates that appeared fbr the party or amcus now represented by me n the tral court or agency or are expected to appear n ths court are: See attachment" Please Note: All questons must be answered ec: Bradford Badke, Esq., Morton Amster, Esq. and Rache Krevans, Esq. j. \. %=;So;.,ol Prnted ham,=of counsel

3 CERTFCATE OF NTEREST ATTACHMENT 4. The names of all law frms and the partners or assocates that appeared for the party or amcus now represented by me n the tral court or agency or are expected to appear n ths court are: MUNGER, TOLLES & OLSON LLP Jason Rantanen Jeffrey. Wenberger Roht K. Sngla John Peck (no longer at MTO) Ted G. Dane Bll Ward m_ BAKER BOTTS LLP BNGHAM McCUTCHEN LLP ABBOT/" LABORATORES Davd Wlle James W. Cannon, Jr. Mara W. Boyce Matthew A. Hayenga Shannon H. Hutcheson Steven Mtby Wllam P. Johnson (no longer at BB) Davd Arlngton JeffBaxter Scott Powers Danel Joshua Goldberg Dalton Jose E. Rvera Karen L. Hale

4 TABLE OF CONTENTS Page ntroducton V. V. Jursdctonal Statement... 3 Statement of the ssues... 3 Statement of the Case... 4 Statement of Facts... 5 A. Background of the Technology MedSense TheraSense Electrochemcal sensors... 7 B. The MedSense Patents The '382 patent: A faster medator The contnung need for a membrane The '551 patent: Dsposable test strps for whole blood usng sensors wthout a membrane... 9 C. There s No Evdence That the '382 Patent Taught a Sensor for Whole Blood Wthout a Membrane The '382 embodments all use a membrane for whole blood The '382 nventors dd not themselves know how to test n whole blood wthout a membrane There s no evdence the '382 embodment would work n blood wthout a membrane There are many dfferences between the '382 sensors and the sensors used n the '551 test strps D. The Dstrct Court Reled Only on the "Optonally, but Preferably" Language

5 t V. V. TABLE OF CONTENTS (contnued) Page E. The Dstrct Court Found nequtable Conduct Based on a Purported nconsstency Between Legal Brefs Submtted to the PTO and EPO The statements to the PTO The EPO legal brefs The supposed nconsstency There was no evdence of bad fath F. The TheraSense Patents The '164/'225 patents: Smaller sample sze The '745 patent: Smaller sample sze wth dffusble medators The dstrct court's summary judgment order Summary of the Argument... _ Argument A. '551 Obvousness Standard of revew Dsclosure: The phrase "optonally but preferably" s not a teachng Reasonable Lkelhood of Success: There was no proof of an expectaton that a membraneless '382 sensor would work n blood Enablement: There s no evdence the '382 patent enabled the '551 nventon Motvaton to Combne: There was no reason to combne the pror art references The '551 patent s not just a "deleton of a functon." Objectve ndca: The court gnored commercal success B. '551 nequtable Conduct

6 TABLE OF CONTENTS (contnued) Page 2. o Legal Standard and Standard of Revew Ambgutes and nferences must be taken n Abbott's favor Materalty: MedSense's EPO arguments were not nconsstent wth Abbott's PTO statements Materalty: Lawyer argument about pror art s not nformaton materal to patentablty ntent: Pope's nterpretaton of the EPO brefs was reasonable and n good fath ntent: Dr. Sanghera was enttled to rely on Secton C. The "Non-Flowng Manner" Lmtaton of the ' 164 and '745 Patents The dstrct court gnored the prosecuton hstory ;Fhe dstrct court's constructon volates the laws of physcs The dstrct court msread the specfcaton D. Antcpaton of the '745 Patent The '225 reference dd not dsclose usng dffusble medators The testmony of Defendant's expert created a trable ssue of fact V. 3. The '225 patent dd not dsclose or enable the "background sgnal" lmtaton of the '745 patent Concluson

7 TABLE OF AUTHORTES Page FEDERAL CASES Abbott Labs. v. Sandoz, nc., --- F.3d ---, 2008 WL (Fed. Cr. 2008) Akzo N. K v. U.S. nt'l Trade Comm 'n, 808 F.2d 1471 (Fed. Cr. 1986) Am. Host & Derrck Co. v. Sowa & Sons, nc., 725 F.2d 1350 (Fed. Cr. 1984) B.F. Goodrch Co. v. Arcraft Brakng Sys. Corp., 72 F.3d 1577 (Fed. Cr. 996) Bausch & Lomb, nc. v. Barnes-Hnd/Hydrocurve, nc., 796 F.2d 443 (Fed. Cr. 1986) Beckman nstruments, nc. v. LKB Produkter AB, 892 F.2d 1547 (Fed. Cr. 1989) Board of Regents v. BENQ Am. Corp., 533 F.3d 1362 (Fed. Cr. 2008) Brown & Wllamson Tobacco Corp. v. Phlp Morrs nc., 229 F.3d 1120 (Fed. Cr. 2000) C.R. Bard, nc. v. U.S. Surgcal Corp., 388 F.3d 858 (Fed. Cr. 2004) Celertas Tech., Ltd. v. Rockwell nt'l. Corp., 150 F.3d 1354 (Fed. Cr. 1998)... 61, 62 Demaco Corp. v. t7. Von Langsdorff Lcensng Ltd., 851 F.2d 1387 (Fed. Cr. 1988_ Esa Co. Ltd. v. Dr. Red@'s Labs., Ltd., 533 F.3d 1353 (Fed. Cr. 2008) El Llly & Co. v. Zenth Goldlne Pharms., nc., 471 F.3d 1369 (Fed. Cr. 2006) Hallburton Co. v. Schlumberger Tech. Corp., 925 F.2d 1435 (Fed. Cr. 1991) nnogenetcs, N. K v. Abbott Labs., 512 F.3d 1363 (Fed. Cr. 2008) n re Kumar, 418 F.3d 1361 (Fed. Cr. 2005)... 34, 35 n re Mahurkar, 831 F. Supp (N.D. ll. 1993) n re Shuman, 361 F.2d 1008 (C.C.P.A. 1966) v

8 TABLE OF AUTHORTES (Contnued) Page Judkns v. HT Wndow Fashon Corp., 529 F.3d 1334 (Fed. Cr. 2008) Kngsdown Med. Consultants, Ltd. v. Hollster nc., 863 F.2d 867 (Fed. Cr. 1988) Lfe Teehs., nc. v. Clontech Labs., nc., 224 F.3d 1320 (Fed. Cr. 2000)... : M. Eagles Tool Warehouse, nc. v. Fsher Toolng Co., 439F.3d 1335 (Fed. Cr. 2006) McGnley v. Frankln Sports, nc., 262 F.3d 1339 (Fed. Cr. 2001)... 37, 38 Medchem, S.A. v. Rolabo, S.L., 437 F.3d 1157 (Fed. Cr. 2006)... 32, 33 Motorola v. nterdgtal. Tech Corp., 121 F.3d 1461 (Fed. Cr. 1997) N. Telecom, nc. v. Datapont Corp., 908 F.2d 931 (Fed. Cr. 1990) Newman v.qu_zg, 877 F.2d-1575 (Fed. Cr. 1989) Pandut Corp. v. Dennson Mfg. Co., 810 F.2d 1561 (Fed. Cr. 1987)... 30, 31 pfzer, nc. v. Apotex, nc._ 480 F.3d 1348 (Fed. Cr. 2007) Phllps v. A WH Corp., 415 F.3d 1303 (Fed. Cr. 2005) Purdue Pharma L.P. v. Endo Pharms. nc., 438 F.3d 1123 (Fed. Cr. 2006) Rasmusson v. SmthKlne Beecham Corp., 413 F.3d 1318 (Fed. Cr. 2005)... 34, 35 Rchards v. Chase Elevator Co., 159 U.S. 477 (1895) Rockwell nt'l Corp. v. Unted States, 147 F.3d 1358 (Fed. Cr. 1998) Scanner Teehs. Corp. v. COS Vson Sys. Corp. N. V.., 528 F.3d 1365 (Fed. Cr. 2008)... 41, 46 Smths ndus. Med. Sys., nc. v. Vtal Sgns, nc., 183 F.3d 1347 (Fed. Cr. 1999)... 38

9 1 TABLE OF AUTHORTES (Contnued) Page Star Scentfc, nc. v. R.J. Reynolds Tobacco Co., 537 F.3_ 1357 (Fed. Cr. 2008)... 40, 41, 51, 53 Symbol Techs., nc. v. Optcon, nc., 935 F.2d 1569 (Fed. Cr. 1991) Tate Access Floors, nc. v. nterface Archtectural Res., nc., 279 F.3d 1357 (Fed. Cr. 2002) Upsher-Smth Labs., nc. v. Pamlab, L.L.C., 412 F.3d 1319 (Fed. Cr. 2005) Young v. Lumens, nc., 492 F.3d 1336 (Fed. Cr. 2007) FEDERAL STATUTES 28 U.S.C and U.S.C. 1338(a)... 3 FEDERAL RULES Rule 54(b)... 3 _EDERAL REGULATONS 37 C.F.R. 1.56(d) OTHER AUTHORTES Merram-Webster's Collegate Dctonary (10th ed. 1999) v

10 t o Statement of Related Cases There have been no pror appeals from the dstrct court proceedngs. 2. Abbott has also appealed the dstrct court's September 10, 2008 judgment n favor of BD/Nova regardng Patent Nos. 6,592,745 and 6,143,164, respectvely. Pursuant to ths Court's September 23, 2008 Order, Abbott has addressed the ssues relatng to the '745 (and '164) patents n ths bref, and wll shortly be movng to consoldate the appeal of that Judgment wth the nstant appeal. Abbott has also appealed the dstrct court's August 8, 2008 judgment n favor of BDNova regardng Patent No. 5,628, On August 21, 2008 the dstrct court declared ths an exceptonal case based on ts nequtable conduct fndng. That determnaton has not yet been appealed but any appeal would be related to the nstant appeal.

11 . NTRODUCTON Ths appeal concerns three patents clamng crtcal technology n the feld of dabetes blood glucose test strps: Patent Nos. 5,820,551 ("the '551 patent"), 6,143,164 ("the ' 164 patent") and 6,592,745 ("the '745 patent"). Wth respect to each patent, the dstrct court made the same knd of fundamental error: readng solated statements out of context and n the abstract wthout reference to the patent as a whole. The dstrct court often acted as ts own expert and dsregarded the actual techncal teachngs of the pror art and the testmony of experts, nventors, and other scentfc wtnesses. The '551 patent clams dsposable strps for testng glucose levels n whole blood that use bosensors wthout a membrane. The dstrct court held the patent obvous based on one sentence n a pror art patent, No. 4,545,382 ("the '382 patent"), that descrbes a membrane as "optonal" or "preferred." From ths alone, the dstrct court concluded the '382 rendered obvous a membraneless sensor for blood -- even though (1) every techncal teachng n the '382 patent suggests otherwse, (2) the '382 nventors testfed they had not known how to buld such a sensor, and (3) nothng n the '382 patent would have enabled anyone to practce the '551 nventon. The dstrct court then combned that membraneless sensor wth other pror art references to arrve at the '551 nventon: an elongated

12 dsposable blood-glucose strp desgned for releasable attachment to a meter wth a two-electrode electrochemcal sensorthat s covered by a sngle drop of blood. The dstrct court also held the '551 patent unenforceable -- based on the falure to dsclose not pror art, techncal data, or test results, but legal brefs submtted by German patent counsel regardng the European counterpart of the '382 patent. The court read two passages n these European brefs as nconsstent wth statements made to the PTO n the '551 prosecuton. The dstrct court's nterpretaton of the legal brefs was not only ncorrect - t was rrelevant to the ssue of nequtable conduct, whch nvolves whether Abbott's patent counsel and scentst had at least a reasonable nterpretaton of those documents. They dd. nequtable conduct cannot be premsed on a close readng of at best ambguous legal brefs. The '164 and '745 patents clam advanced electrochemcal bosensors for analyte measurement n very sma[1 blood samples, less than l_tl. The dstrct court construed a shared clam lmtaton -- a sample held n a "non-flowng manner" -- so as to exclude any movement, even the convectve moton present n all lquds. Needless to say, ths unduly restrctve nterpretaton of the clams fnds no support n the specfcatons, prosecuton hstory, or common sense. Fnally, the dstrct court held the '745 antcpated on summary judgment, based (agan) on one sentence n the pror art '225 reference, whch states that non-

13 leachable chemcal medators are "preferred." From ths alone, the dstrct court 1 concluded the '225 reference taught the use of leachable medators -- even though Defendants' expert opned to the contrary.. JURSDCTONAL STATEMENT The dstrct court had subject-matter jursdcton under 28 U.S.C. 1338(a). Ths Court has jursdcton under 28 U.S.C and The dstrct court entered judgment under Rule 54(b) as to the '551 wth respect to Bayer and BDNova on July 2, 2008, as to the '551 and '745 wth respect to Bayer on August 18, 2008, and as to the '745 and '164 wth respect to BDNova on September 10, (JA1-2.3.) Abbott tmely fled notces of appeal on July 21, July 31, September 12, and October 8, (JA7655-JA7664; JA ). STATEMENT OF THE SSUES 1. Whether the '551 nventon was obvous based on a sngle sentence n a pror patent, unsupported by specfc teachngs and absent clear and convncng evdence that the pror patent enabled the '551 nventon. 2. Whether the '551 patent was unenforceable based on the falure to dsclose attorney arguments made to the European Patent Offce n connecton wth a pror art patent, where the lawyer and scentst beleved the brefs were not nconsstent wth the postons taken before the PTO.

14 3. Whether the "non-flowng manner" lmtaton n the '745 and '164 l patents should be construed to render the clams noperatve. 4. Whether the '745 patent could be held antcpated on summary judgment based on pror art that Defendants' own expert testfed dd not dsclose the '745 nventon. V. STATEMENT OF THE CASE n March 2004, Becton Dcknson and Company ("BD") sued TheraSense, now a subsdary of Abbott Laboratores ("Abbott"), seekng a declaratory judgment of non-nfrngement of the' 745 and' 164 patents. (JA ) TheraSense and Abbott subsequently counter-sued BD and ts suppler, Nova Bomedcal, on those patents, as well as the '551 patent and Patent No. 5,628,890 ("the '890 patent"). (Tral Order 1.)1 n August 2005, Abbott sued Bayer Healthcare for nfrngement of the '551 and '745 patents. (JA420.) The Bayer and' BD/Nova cases were subsequently coordnated. (Tral Order 2.) The dstrct court ssued two clam constructon orders, one nvolvng the ' 164 and '890 patents, and one nvolvng the '551 and '745 patents. On Aprl 3, 2008, the dstrct court held numerous clams of the '745 patent to be antcpated on summary judgment. (SJ Order ) The dstrct court also granted BD/Nova 1The Summary Judgment and Tral Orders are attached as Addendums

15 l summary judgment of non-nfrngement of the '745 and '164 patents based on the "non-flowng manner" lmtaton. (d ) From May 27 to June 3, 2008, a bench tral was held on Defendants' defenses wth respect to the '551 patent. (Tral Order 2-3.) On June 24, the court held clams 1-4 of the '551 patent to be obvous. (Tral Order 54.) The dstrct court also held the '551 patent unenforceable due to nequtable conduct. (d.) After ths appeal was taken, the dstrct court held the '551 ltgaton was an exceptonal case and authorzed attorneys' fees to Defendants based on the nequtable conduct determnaton. Defendants' fee clams for approxmately $20 mllon are currently pendng before the dstrct court. V. STATEMENT OF FACTS A. Background of the Technology 1. MedSense Dabetcs today rely on cheap and accurate home glucose sensors to manage ther dsease. Those sensors are the result of poneerng work done at MedSense, formerly known as Genetcs nternatonal and now part of Abbott. (Tral Order 3-5.) At the tme of the events at ssue, the early 1980s, MedSense was a start-up n Oxford, England wth one of the leadng research groups n ths feld n the world. (Tral Order 3-4; JA2610 at 300; JA3726.) Pror to the ground-breakng work at

16 MedSense, dabetc patents reled on "colormetrc" optcal systems n whch patents estmated blood glucose levels by evaluatng the color produced by a chemcal reacton. (JA , 3718; JA2615 at 320.) Such systems were unrelable and nconvenent. (d.) MedSense made several nventons -- ncludng those clamed n the '382 and '551 patents -- that resulted n blood glucose test strps usng electrochemcal bosensors. (Tral Order 4-5.) MedSense's frst commercal product was the Exactech n (d. 50.) Exactech was the frst dsposable electrochemcal test strp on the market, and MedSense sold over one bllon Exactech strps. (JA2640 at 422:21-423:13; JA3725.) t was a revolutonary product, transformng dabetes management and creatng a $3-4 bllon ndustry n the U.S. alone. (JA2614 at 316:19-25.) 2. TheraSense TheraSense was another ground-breakng start-up that nvented n the late 1990's electrochemcal bosensors that requred less than 1_tL of blood. Such a small sample sze permts more frequent testng and less panful blood draws from less senstve areas of the body. (JA ) These advances have been especally benefcal to dabetc chldren who must test ther glucose levels repeatedly everyday. (d.) TheraSense's founder Adam Heller receved the

17 1 l Natonal Medal of Technology and nnovaton for hs work on TheraSense sensors. 2 Abbott acqured TheraSense 3. Electrochemcal sensors n 2004 for $1.12 bllon. n the electrochemcal bosensors of the nventons n sut, glucose n blood reacts wth, and transfers electrons to, an enzyme. (Tral Order 8.) A medator transfers the electrons from the enzyme to a sensor's "actve electrode." (d.) The electrons flow through a meter attached to the sensor, whch calculates the glucose concentraton based on the current flow. (d.) The medator and enzyme are generally coated on the actve electrode tself. (d.) Electrochemcal glucose sensors are not lmted to testng blood. They can measure glucose levels n a varety of lquds, ncludng (a) whole blood (or smply "blood"); (b) blood plasma (the lqud component of blood wthout red blood cells); (c) ntersttal flud (the lqud between cells n the skn or other organs); and (d) buffer soluton. (See, e.g., JA2619 at 337:22-338:3 (blood plasma); JA122 2:40-45 (blood); JA6508 at 3:57-4:2 (ntersttal flud); JA , Example 8 (buffer).) 2 Press Release, (Aug. 25, 2008), at 7

18 B. The MedSense Patents 1. The '382 patent: A faster medator The frst advance by MedSense's scentsts s descrbed n the '382 patent, whch ssued n October 1985, but clams prorty to an October 23, 1981 Unted Kngdom applcaton. (Tral Order 4.) t clams electrochemcal bosensors usng faster ferrocene-based medators. (Tral Order 9-10.) The '382 nventon tself was a tabletop devce usng three electrodes n glass test tubes rather than consumer-ready sensors. (JA at 701:8-705:25; JA6510 7:4-12; JA6505 Fg. 1; JA6509 5:54-56.) The goal of the '382 patent was to develop "an mplantable glucose sensor," not a dsposable test strp. (JA6507 at 1:15-25.) 2. The contnung need for a membrane Before the faster medators of the '382 patent, electrochemcal sensors used dffuson-lmtng membranes. (Tral Order 14.) These membranes, also descrbed as "glucose controllng" or "sempermeable," reduced the rate at whch glucose contacted the sensor chemstry and electrodes because the older medators could not cope wth a rapd nflux of glucose. (JA2742 at 509:23-510:25; JA3002 at 721:14-24; JA6511.) Dffuson-lmtng membranes were dffcult to manufacture and greatly ncreased the tme for measurement. (d.) The faster medators of the '382 patent elmnated the need for dffuson-lmtng membranes. (JA at 721:11-725:2.)

19 The '382 technology dd requre a dfferent knd of membrane -- a "protectve" or "permeable" membrane -- at least n certan crcumstances. Frst, measurements n whole blood (as opposed to blood plasma, ntersttal flud, or buffer) requred protectve membranes to prevent "foulng,".e., red blood cels from stckng to the electrodes and nterferng wth the electron transfer from the medator to the actve electrode. (Tral Order 16-17; JA2733 at 473:3-474:12; JA2740 at 499:18-23.) Second, sensors njected nto the human body (n vvo) used protectve membranes to prevent the chemstry from dssolvng nto the body. (Tral Order ) mportantly, protectve membranes do not nterfere wth the flow of, and are fully permeable to, glucose molecules, whch are much smaller than red blood cells. (JA2733 at 473:3-474:6; JA2738 at 493:8-494:8.) The '382 sensors dd not, however, requre a membrane or buffer, and the '382 patent descrbes membraneless applcatons. (JA6508 3:57-4:2; JA , Example 8.) sensors for such for ntersttal flud 3. The '551 patent: Dsposable test strps for whole blood usng sensors wthout a membrane The orgnal applcaton leadng to the '551 patent was fled n May 1983 n the Unted Kngdom. (JA114.) n the year and a half between the '382 and the '551 applcatons, MedSense's scentsts made sgnfcant advances. The '551 patent descrbes a dsposable electrochemcal test strp, desgned for releasable 9

20 attachment to readout crcutry, for testng glucose n whole blood usng just two electrodes (the' 382 used a three-electrode system). (JA 122.) The '551 nventon was ntended for whole blood: the clams specfcally nvolve whole blood samples and requre the electrodes to be "covered by a sngle drop of whole blood." (JA128 at 13:29-14:47.) Although t s ntended for use wth whole blood, the detaled nstructons for the preferred '551 embodment (Fgures 1-7) do not nclude a membrane. (JA125 8:35-51.) The PTO ultmately found the '551 patent patentably dstngushable from the '382 patent because, unlke the '382 patent, t taught a sensor for whole blood wthout a membrane. (Tral Order 7.) C. There s No Evdence That the '382 Patent Taught a Sensor for Whole Blood Wthout a Membrane. The court's '551 obvousness rulng reled on ts theory that the '382 patent taught PHOSTAs (persons havng ordnary skll n the art) to make and use membraneless sensors for whole blood. (Tral Order 20.) t then combned that membraneless sensor wth other pror art references to arrve at the '551 patent. The dstrct court accepted that pror to the '382 patent, PHOSTAs would hav.e used a membrane for blood. (d. 14.) Thus, the crtcal ssue s whether the '382 taught contrary to that conventonal wsdom. 10

21 1. The '382 embodments all use a membrane for whole blood. The 13 Examples. The dstrct court conceded that although the '382 patent "contaned thrteen workng examples of preferred embodments of the nventon," each "example[] nvolvng blood employed a membrane." (Tral Order (emphass added).) Ths fact bears repeatng: n 13 workng examples, there are none wthout a membrane for blood. The dstrct court dsmssed ths as "happenstance." (d. 16.) Here, the dstrct court began ts habt of actng as ts own expert: No one at tral testfed ths was happenstance. Nor dd the dstrct court or any defense expert explan why, f the nventors ntended such a devce to be part of ther nventon, a membraneless sensor for whole blood was not expressly desc.rbed,. Example 8. The '382 specfcaton actually teaches away from membraneless sensors for whole blood. n Example 8, a sensor s constructed wthout a membrane and tested n buffer soluton,.e., n the absence of red blood cells. (JA ) That sensor was then "modfed" by addng "a cellulose acetate membrane" before testng n buffer and blood. (Tral Order 12; JA6511 9:25-33.) Even though Defendants' expert clamed that testng n blood was "what the whole example s about," the verson of the sensor wthout the membrane was conspcuously never tested n blood. (JA2534 at 250:19-20.) 11

22 The dstrct court got around Example 8 by effectvely actng as ts own expert and clamng that "a membrane was added.., lttle more than [as] a way to nvestgate the tme effect of addng a membrane." (Tral Order 16.) Ths nterpretaton of Example 8 was not supported by the testmony of any wtness or any other evdence. To the contrary: the goal of the Example was testng n blood, the specfc omsson of a test of the membraneless sensor n blood would have been mportant to PHOSTAs. (JA2739 at 497:4-498:21.) Membraneless Embodment for Non-Blood. The dstrct court cted a passage n whch a "dfferent form of the nventon" s descrbed wthout a l membrane. (Tral Order 9-10.) But tellngly, ths membraneless sensor was "envsaged" for "projectng only nto the derms" --.e., for ntersttal flud, not blood. (d. 10; JA2740 at 502:13-18; JA at 691:20-692:2.) 2. The '382 nventors dd not themselves know how to test n whole blood wthout a membrane. The dstrct court gnored the unform testmony from the nventors -- even an nventor pad by Defendants -- that they themselves dd not know how to make a membraneless whole blood sensor at the tme of the '382 applcaton, and only fgured that out durng the research for the '551 nventon. Nether the dstrct court nor Defendants explaned how the '382 patent could dsclose somethng to a 12

23 PHOSTA that the nventors themselves dd not understand, partcularly gven that the MedSense nventors were worldwde leaders n ths feld. Professor Hll, an nventor on both patents, testfed the nventors dd not know whether a membraneless sensor for whole blood was possble n 1981, much less how to buld such a sensor. (JA ) As far as he knew, even after the '382 applcaton "there were no electrodes that could be ntroduced nto blood wthout a membrane." (JA3209.) Even a year later, the nventors merely thought such a sensor theoretcally possble: they "had some experments that suggested that t was possble." (JA3217.) The testmony of Graham Davs, an nventor on the '551 patent, was to the same effect. (JA ) Professor Hggns, an nventor andpad wtness for Defendants, testfed that the experments by the research team n February 1983 were stll usng a membrane for blood. (JA3721; JA3724; JA ) He dd not recall that the nventors knew how to buld a membraneless sensor for whole blood at the tme of the '382 applcaton. They smply thought t "qute concevable that t would be possble." (JA3746 at 119:17-23 (emphass added).) The nventor testmony s supported by the June 1986 applcaton for Patent No. 4,897,173, by Shro Nanka, another "leader n the feld of electrochemcal sensors n the early 1980s." (JA3000 at 712:7-9.) n the ' 173 specfcaton -- submtted nne months after the ssuance of the '382 patent -- a protectve 13

24 membrane s omtted when testng n blood plasma but a membrane s added when testng n whole blood. (JA3000 at 712:17-24; JA , Examples 3-4.) The dstrct court agreed the Nanka patent "dd tend to support the 'conventonal wsdom' argument advanced by Abbott," but nonetheless dsregarded the Nanka reference on unsupported speculaton that Nanka mght not have known about the '382 patent. (Tral Order 19.) o There s no evdence the '382 embodment would work n blood wthout a membrane. The dstrct court based ts concluson that the '382 sensors would have worked n blood wthout a membrane because "the chemstry was fast enough (at least by the tme of the '382 pror-art dsclosure) to obtan acceptable results wthout a membrane." (Tral Order ) Once agan, no evdence supports the court's concluson. Nether the nventors, the expert wtnesses, nor any other wtness so testfed. There s no evdence that the membraneless sensor of Example 8, for example, would have worked f a PHOSTA had even thought to have tred t n whole blood. Defendants' expert, Dr. Turner, Who was part of the MedSense research team, dd not opne that any membraneless sensor from the '382 patent worked n whole blood wthout a membrane. All he couldsay was that the '382 does not expressly state that the membraneless sensor would not work n whole 14

25 blood, but even that gnores Example 8. (JA2533 at 248:25-249:2.) Such an _J opnon certanly does not consttute evdence that the '382 taught membraneless sensors for whole blood, let alone clear and convncng evdence. Moreover, Dr. Turner performed no tests to determne whether the membraneless sensor of Example 8 would work n whole blood. The dstrct court gnored the uncontroverted evdence on ths pont from Dr. Sanghera, another MedSense scentst, who had done "experments that replcated what was n the '382 patent" at Medsense. (JA3007 at 741:9-11.) Dr. Sanghera ponted out that Example 8 reports that the membraneless sensor had a 5% oxygen senstvty when tested n an ar-saturated buffer soluton as compared to one wthout oxygen. (JA3001 at 716:12-717:23; JA6511 at 9:19-21.) Because the oxygen level of whole blood s an "order of magntude hgher" than the buffer, ths 5% dscrepancy ndcated that the membraneless sensor would not work n oxygen-rch blood. (JA3001 at 717:6-23.) Agan, ths evdence of lack of enablement was gnored by the dstrct court. 4. There are many dfferences between the '382 sensors and the sensors used n the '551 test strps. Actng as ts own expert, the dstrct court concluded that the only dfference between the '551 and '382 sensors was the absence of a membrane. But that s not correct and no wtness so testfed. Although t s not clear precsely why the '551 15

26 membraneless sensor works n blood, t s undsputed that the exact electrode confguraton, the materals used, and how the electrodes are prepared can all be crtcal. (JA at ) Professor Hll explaned that as t "turned out," the nventors ultmately learned that "adsorpton,".e., foulng, "depended very much on the structure of the electrode. then there was relatvely lttle" foulng. f an electrode was very well prepared, (JA ) There are many dfferences n the preparaton of the '551 and '382 electrodes, and the dstrct court could not smply presume that the '382 membraneless sensors would have worked n blood lke the '551 sensors. Frst, the '382 patent teaches repeatedly to oxdze the electrode, e.g., "heat[] n an oven for 40 h at 200 C. to gve[] a[n] oxdsed surface." (JA6511 9:1-3; JA6508 3:22-27; JA6510 8:16.) But the nventors later learned that oxdzed electrodes exhbt greater foulng: "f [the electrode] was left n the ar for a long tme to oxdze, then there was much more" foulng. (JA ) The '55 patent pontedly does not oxdze the electrodes. (JA125 8:22-51; JA127 12:31-38.) The dstrct court gnored ths dstncton. Second, the electrode of Example 8 s an ultracarbon rod, whch showed a 5% oxygen senstvty when tested even n a buffer soluton wth lmted oxygen. Supra 15. The '551 patent, by contrast, teaches that electrodes should be constructed from grafol or screen prntng of collodal carbon,.e., a carbon 16

27 suspenson or paste. (JA125 7:15-18; JA127 12:17-40.) When defense expert Dr. Turner worked for MedSense, he reported on ther dscovery that grafol showed much less oxygen senstvty than the ultracarbon rod. (JA6367.) Not concdentally, grafol was the electrode materal used when nventor Davs exclamed n hs 1983 lab notebook -- after the applcaton for the '382 was fled -- that "We can now test [n] whole blood." (JA6437.) Contnung to act as ts own expert on electrochemstry, the dstrct court dsmssed ths dfference because the '382 patent says ts electrodes have "low oxygen senstvty." (Tral Order 21.) That statement must be read n the context of the nventon, whch dscloses a sensor wth a membrane for blood. There s no evdence the '382 sensors' "low oxygen senstvty" was so low that the membrane could be removed n whole blood. The evdence s to the contrary: the membraneless electrode n Example 8 was specfcally not tested n whole blood and showed a sgnfcant oxygen effect. The dstrct court reled also on testmony that the Exactech used carbon paste, "the same materal dsclosed n the '382 patent." (Tral Order 22 n. 10.) But the Exactech was made usng the screen prntng of collodal carbon taught by the '551. (JA3019 at 788:19-25.) ndeed, the ttle of the '551 patent s "Strp Electrode wth Screen Prntng." (JA114.) The '382 does not teach screen prntng. The dstrct court gnored ths dstncton because Abbott had not proven 17

28 that t would affect sensor performance. (Tral Order 22 n. 11.) n so dong, the D. court lost sght of the fact that Defendants bore the burden of provng by clear and convncng evdence that the '382 sensors enabled one to make membraneless sensors for use n blood. The Dstrct Court Reled Only on the "Optonally, but Preferably" Language. Dsregardng all of ths evdence, the dstrct court held that the '382 patent taught membraneless sensors for blood based solely on one sentence n the specfcaton: Optonally, but preferably when beng used on lve blood, a protectve membrane surrounds both the enzyme and medator layers, permeable to water and glucose molecules. (JA6508 4:63-66.) Accordng to the dstrct court, that sentence by tself taught membraneless sensors for blood because t says that the membrane s just "preferred" or "optonal." (Tral Order 15.) The negatve mplcaton from the words "optonal" and "preferred" s the entre bass of the dstrct court's obvousness rulng. The dstrct court, lke Defendants' expert, reled on the supposed "plan language" of that sentence. (JA2531 at 239:20-22.) n dong so, t mproperly focused on a sngle word to the excluson of the rest of the patent dsclosure, whch provdes the necessary context for ths statement. 18

29 Frst, the sentence provdes the key dsclosure that the membrane used n the '382 sensors s "permeable to water and glucose,".e., that t s not the dffusonlmtng membrane prevously used. Ths was the focus of the EPO proceedngs dscussed below. nfra Second, t s undsputed that the protectve membrane was optonal where there are no red blood cells, such as ntersttal flud, blood plasma, and buffer. 1 1 Not concdentally, the '382 patent has examples ofmembraneless sensors but only for use n such fluds. (JA6508 3:57-4:2; JA , Example 8.) Ths resulted from the '382's faster chemstry; glucose-lmtng membranes otherwse would have been requred n all applcatons. n the context of the entre specfcaton and pror art, the sentence does not mean a membrane s always optonal, but rather that a membrane s not necessary n all applcatons. Thrd, by referrng to the use of a membrane n whole blood as "preferable," rather than "necessary," the nventors consdered themselves to have nvented a membraneless sensor n whole blood. Ths concluson was ll-founded. Wtnesses on both sdes agreed that terms lke "optonally" and "preferably" are not always read lterally n patent specfcatons. Abbott's patent prosecutor, Lawrence Pope, explaned that he understood the "optonally, but preferably" phrase, at least as appled to whole blood, to be terms of art, or "patentese." (JA at 627:22-628:20.) Any other nterpretaton would be nconsstent wth the unform techncal 19

30 teachngs throughout the '382 patent that use membranes for whole blood. Pope, a patent prosecutor wth over 35 years of experence, explaned that patent prosecutors often use such words nstead of terms lke "requred" or "needed" to avod lmtatons n the specfcaton from beng mproperly read from the specfcaton nto the clams "one, n draftng a patent applcaton, wants to be t careful not to be unduly restrctve. And so one uses permssve words lke 'preferable' and 'may' to avod beng unduly restrctve." (JA at 627:25, 632:7-11.) Defendants' own expert, Dr. Turner, acknowledged the role ofpatentese. 3 He agreed, for example, that a PHOSTA would understand the sentence, "' [p]referably, the sensor has at least one electrode besdes the actve electrode,'" to mean a second electrode was requred, not merely preferred. (JA2623 at 353:7-17.) PHOSTAs read such language n the context of the whole patent and conventonal wsdom. ndeed, as dscussed nfra 60-62, n challengng the valdty of the ' 164 patent, Turner hmself opned that the sentence "preferably, the redox medators of the present nventon are bound or otherwse mmoblzed" dd not contemplate the use ofnonmmoblzed, leachable medators. Rather he sad the phrase was 3 Defendants' patent law expert, Thomas Smegal, pontedly dd not dspute that "optonal" and "preferred" can be terms of art. 20

31 patent language, or patent tease [sc],... 've read a lot of patents over the years that sort of used ths sort of language. And what t's tryng to tell you s don't use a leachable medator. (JA1571 at 139:15-140:1; JA9738:11-14.) E. The Dstrct Court Found nequtable Conduct Based on a Purported nconsstency Between Legal Brefs Submtted to the PTO and EPO. The dstrct court's nequtable conduct rulng s unprecedented. t s based not on wthholdng of pror art or techncal or scentfc data relevant to patentablty. nstead, the '551 patent was held unenforceable -- and a patent lawyer's unblemshed career was tarnshed -- because the dstrct court nterpreted 1 arguments made by MedSense's German counsel to the European Patent Offce (EPO) regardng the European counterpart of the '382 patent as nconsstent wth statements made to the PTO n the '551 prosecuton. 1. The statements to the PTO n the fall of 1997, after Abbott acqured MedSense, Abbott's patent prosecutor Lawrence Pope assumed responsblty for the fle that resulted n the '551 patent. He sought new clams for a dsposable test strp for measurement n a drop of whole blood wthout a membrane. (JA2976 at 618:1-4.) Pope had prosecuted hundreds of patents, both n prvate practce wth hghly respected law frms as well as n-house for companes such as appellee Bayer and appellant Abbott. (JA at 613:10-616:12) n hs 35 years of practce, Pope had never 21

32 l,, H, g t been even accused of nequtable conduct on any of the hundreds of other patents he has prosecuted. (JA2976 at 616:17-20.) n November 1997, Pope and Examner Shay met to dscuss the "optonally, but preferably" language of the '382 patent. (d. at 619:7-25.) They agreed the '382 patent taught that a protectve permeable membrane was requred for blood,.e., that the "the overall teachng of ths document, ncludng ths language," was that the '382 sensor needed "a protectve membrane whch protected the sensor from some of the components of human blood, lke erythrocytes, but whch allowed free access of glucose and water across the membrane." (d. at 619:7-9; JA2977 at 622:2-623:7.) Shay asked for an affdavt confrmng that a PHOSTA would have agreed wth ther readng of the '382 patent. (JA7639.) Pope submtted a declaraton from Gordon Sanghera, then the research drector of MedSense. Dr. Sanghera revewed the lterature avalable n 1983 and opned that a PHOSTA would have regarded a protectve membrane as necessary, not merely optonal, when measurng n whole blood: a PHOSTA "would not read [the optonally, but preferably language] to teach that the use of a protectve membrane wth a whole blood sample s optonally or merely preferred." (JA2999 at 709:2-13; JA ) Pope submtted legal remarks along the same lnes, explanng that "t]here s no teachng or suggeston of unprotected actve electrodes for use wth whole 22

33 blood." (JA ) The '551 patent then ssued on October 13, (Tral Order 7.) 2. The EPO legal brefs A few years earler, MedSense's German patent counsel had submtted two legal brefs to the EPO n defense of the EP 0,078,636 ("the '636 patent"), whose specfcaton s largely the same as the '382 patent. (Tral Order ) That proceedng addressed a pror art reference known as "D." (JA ) MedSense's German counsel dstngushed the D1 reference because t used the older dffuson-lmtng, "sempermeable membranes," whereas the '382/'636 patents use protectve membranes fully permeable to glucose. (JA6527, JA , JA ) The clearest explanaton of the type of membrane used n the '382/'636 specfcaton happens to be n the "optonally, but preferably" sentence, whch explans that the "protectve membrane" s "permeable to water and glucose molecules." t s not thus surprsng the sentence was quoted n the EPO brefs. 3. The supposed nconsstency The dstrct court concluded that Pope had argued to the PTO that the '382 patent requred a membrane for whole blood, whereas German counsel had argued the opposte to the EPO: "that the 'optonally, but preferably' sentence demonstrated that the '382/'636 nventon dd not need a membrane for measurng 23

34 ,, glucose n blood." (Tral Order 26; see also d ) The dstrct court's readng of the documents s ncorrect. Both the EPO and PTO submssons are consstent wth membranes beng optonal for fluds other than blood, but requred for blood. Pope sad only that a membrane was requred n the '382 patent for whole blood. (See, e.g., passages quoted n Tral Order ) Pope never dsputed that a membrane was optonal for blood plasma, ntersttal flud, and buffer; ndeed, he specfcally ponted out Example 8's testng ofa membraneless sensor n buffer. (Tral Order 25; JA2977 at 623:8-20.) n the EPO, MedSense never argued that a membrane was optonal for whole blood. Rather, German counsel repeatedly ponted out that "For use on human blood the sensor of Example 7 [Example 8 of the '382] was provded wth a protectve membrane." (JA6586 5; see also JA6531.) That drectly refutes the dea that MedSense argued a membrane was unnecessary for blood. To the extent German counsel stated that, unlke the glucose-lmtng membrane of D 1, a membrane was optonal, that statement referred to the undsputed fact that a membrane was ndeed optonal for applcatons such as ntersttal flud, whch was specfcally called out n the '636 clams and specfcaton. (JA ) 24

35 4. There was no evdence of bad fath. Pope testfed at length that he dd not submt the EPO brefs n the '551 prosecuton because he beleved they were rrelevant to the ssue before the PTO,.e., whether the '382 specfcaton taught a membraneless sensor for whole blood. 4 (JA ) He understood the EPO legal arguments to address an entrely dfferent ssue: whether the membrane dsclosed n the '382/'636 specfcaton was the same as the sempermeable membrane of the D1 reference. He even consulted on ths pont wth Dr. Sanghera, who had personally partcpated n the EPO proceedngs and was thus very famlar wth the arguments to the EPO. (JA at 752:24-757:11.) The dstrct court nonetheless concluded that Pope ntended, to deceve the PTO because Pope "knew or should have known" that the EPO submssons contradcted hs statements to the PTO. (Tral Order 33.) But there was no evdence that Pope dd not beleve hs nterpretaton of the EPO documents, no evdence that he adopted that nterpretaton n anythng other than good fath, and no evdence that he beleved hs legal remarks to the PTO were false. The dstrct court smply refused to accept that Pope reasonably could have read the EPO legal brefs dfferently from the court. 4 At the tme, Pope thought "whole blood" and "lve blood" were synonymous. (JA2979 at 628:21-629:15.) 25

36 As for Dr. Sanghera, he testfed that he reled on hs counsel Pope to decde., 1 l 1 t whether to submt the EPO legal brefs to the PTO. (Tral Order 36.) The dstrct court nonetheless concluded Dr. Sanghera ntentonally commtted nequtable conduct merely because he dd not dscuss those legal brefs n the declaraton he provded to the PTO, a declaraton that even Defendants' pad wtness Dr. Hggns sad was "reasonable." (JA3756.) F. The TheraSense Patents 1. The '164/'225 patents: Smaller sample sze The '164/'225 reference descrbes a glucose sensor for blood samples smaller than one mcrolter. 5 (JA8777, Abstract.) Measurement n such a small sample enabled dabetcs to measure ther glucose less panfully and thus more often. (JA8779:11-23.) One dffculty wth developng measurements for smaller samples was that as the electrodes were brought closer together, the "background sgnal" ncreased, degradng accuracy. (JA8787: :2.) The background sgnal s the current generated not by glucose reactons but by factors lke the "shuttlng" of the medator back and forth through the sample between the electrodes. (d.) The ' 164/'225 nventors solved ths problem by employng ether (a) nonleachable l, l 5 The '225 applcaton s the nternatonal counterpart of the '164 patent and shares the same wrtten descrpton as the relevant porton of the '164. (JA ; JA ) 26

37 medators (.e., those that do not substantally dffuse nto the sample); or (b) mmoblzed medators (.e., those that are trapped or bonded to the electrode surface). (JA8786:3-10.) The specfcaton makes abundantly clear that the nventon pertans to nonleachable or mmoblzed medators. (See, e.g., JA8787: :2.) The' 164/'225 specfcatons contan no examples wth dffusble (or "leachable") medators, and there s no evdence that the nventors bult or tested any such sensor n the research for the ' 164/'225 specfcatons. Nor s there evdence that the ' 164/'225 sensors would have worked usng a dffusble medator wthout excessve background sgnal. 2. The '745 patent: Smaller sample sze wth dffusble medators The '745 patent-n-sut descrbes later research at TheraSense. The '745 nventors developed sensors usng dffusble medators wth acceptable background sgnals even n a very small sample. (JA8275, Abstract.) The clams of the '745 patent all nclude a dffusble medator, along wth other elements not dsclosed n the ' 164/'225 reference. (JA ) The PTO ntally rejected the clams of the parent applcaton of the '745 patent based on the 6,120,676 patent that shares (n relevant part) the specfcaton of'164/'225 references. (JA ) TheraSense explaned, however, that the '676 specfcaton does not teach dffusble medators. (d.) The PTO then wthdrew the rejecton, and the rejecton was not repeated durng the examnaton 27

38 of the '745 patent tself, despte explct consderaton of the '676, '164 and '225 references. (JA ) Thus, the PTO rejected the very theory upon whch the dstrct based ts summary judgment nvaldty rulng. 3. The dstrct court's summary judgment order On summary judgment, the dstrct court held that BDNova dd not nfrnge ether the '164 or '745 patents because, the court sad, the "non-flowng manner" lmtaton n both patents' clams requres blood to be completely mmoblzed wthout even convectve flow -- even though t s undsputed that all lquds have some amount of convectve flow. The dstrct court also held the '745 antcpated by the '225 on the theory that a sentence n the ' 164/'225 specfcaton -- "Preferably, there s lttle or no leachng of the redox medator..." -- dsclosed the use of leachable medator. Just as wth the '382 patent, the dstrct court focused on the word "preferably" n solaton and gnored the context, the specfc teachngs, and the embodments of the reference. The dstrct court thought t enough that the "preferably" language "acknowledg[ed] the possblty of usng a leachng or dffusng medator" -- despte testmony from Defendants' own expert that the ' 164 dd not dsclose the use of a leachable medator. (SJ Order 48.) 28

39 V. SUMMARY OF THE ARGUMENT '551 Obvousness: The dstrct court nvaldated clams 1-4 of the '551 patent based on the sngle phrase "optonally, but preferably" n the '382 specfcaton wthout regard to the actual techncal teachngs of the patent, the conventonal wsdom at the tme, or expert and nventor testmony. That phrase does not teach or enable a PHOSTA to make and use the '551 nventon, whch consttuted an mprovement on the '382 nventon. '551 nequtable Conduct: The dstrct court's fxaton on the "optonally, but preferably" sentence carred over to ts fndng that the '551 patent was unenforceable. Ths s not a case of wthheld pror art, techncal data, or test results. Nor s there any separate evdence of an ntent to deceve. Abbott's counsel and scentst smply read the EPO brefs dfferently than the dstrct court. Patents should not be rendered unenforceable, mllons of dollars n sanctons should not be awarded, and reputatons should not be sulled over good fath dsagreements concernng the meanng of lawyer arguments. '164/'745 Nonnfrngement: The dstrct court erred n ts constructon of the "non-flowng manner" lmtaton by gnorng the prosecuton hstory and construng the term so narrowly as to exclude all possble embodments. '745 Antcpaton: On summary judgment, the dstrct court dsregarded the testmony of Defendants' own expert that the ' 164/'225 specfcaton dd not 29

40 dsclose the '745 nventon. That alone created a trable ssue of fact. The '225 specfcaton, moreover, dsclosed not the '745 technology but the opposte. V. ARGUMENT A. '551 Obvousness 1. Standard of revew Defendants had to prove obvousness by clear and convncng evdence. See Am. Host & Derrck Co. v. Sowa & Sons, nc., 725 F.2d 1350, 1360 (Fed. Cr. 1984). "On appeal from a bench tral,... [t]he ultmate concluson of whether a clamed nventon would have been obvous s a queston of law revewed de novo based on underlyng fndngs of fact revewed for clear error." Pfzer, nc. nc., 480 F.3d 1348, 1359 (Fed. Cr. 2007) (ctatons omtted). 2. Dsclosure: The phrase "optonally but preferably" s not a teachng. The dstrct court found a major techncal advance -- a membraneless sensor l for whole blood -- bured n the phrase "optonally, but preferably" n one sentence n the '382 patent. The dstrct court erred n prvlegng that phrase over the rest of the specfcaton and all the other evdence. "[A] pror patent must be consdered n ts entrety,.e., as a whole, ncludng portons that would lead away from the nventon n sut." Pandut Corp. v. Dennson Mfg. Co., 810 F.2d 1561, 30

41 1568 (Fed. Cr. 1987). A "sngle lne out of" a specfcaton cannot be the focus of the analyss: t s mpermssble wthn the framework of secton 103 to pck and choose from any one reference only so much of t as wll support a gven poston to the excluson of other parts necessary to the full apprecaton of what such reference farly suggests to one sklled n the art. Bausch & Lomb, nc. v. Barnes-Hnd/Hydrocurve, nc., 796 F.2d 443,448 (Fed. Cr. 1986) (emphass added). Read as a whole, the '382 patent has no dsclosure of a membraneless sensor for blood. The dstrct court admtted that "the broad teachng" t ascrbed to the "optonally, but preferably" phrase "went beyond the specfcs of the preferred embodments." (Tral Order 16.) But "[g]eneral allegatons, appearng as a departure from the more specfc teachngs.., and devod of detal" are not teachngs. n re Shuman, 361 F.2d 1008, 1012 (C.C.P.A. 1966). The dstrct erred n elevatng the mportance of one sentence, whch, under the court's readng, was a "departure" from the specfc teachngs of the specfcaton, whch all teach away from the '551 nventon. Nether Dr. Turner nor the dstrct court ever explaned why a PHOSTA would have focused on a formalstc constructon of a sngle sentence rather than on the techncal dsclosures of the patent as a whole, ncludng Example 8. Dr. Turner hmself admtted that words lke "preferably" are read by PHOSTAs n the 31