US 7465800

I'll search for authoritative information on this patent, including the USPTO record and any 2026 CAFC docket activity.

Let me check for any 2026 Federal Circuit (CAFC) docket activity involving this patent.

Let me verify the exact claim set of the '800 patent, since independent claim structure matters for your request.

US Patent 7,465,800 — Analyst Summary

Sourcing note: I was able to confirm the bibliographic data and full claim text across multiple sources (Google Patents, Espacenet, DrugPatentWatch, and District Court filings). I did not find any 2026 CAFC docket in which the '800 patent itself is on appeal — see the litigation/uncertainty section below.

1. Bibliographic data (confirmed)

The compound 3-(4-amino-1-oxo-1,3 dihydro-isoindol-2-yl)-piperidine-2,6-dione is lenalidomide, the active ingredient in REVLIMID®. Note the patent's own text characterizes this as a polymorph patent built on the compound disclosed in U.S. Pat. Nos. 5,635,517 and 6,281,230 (both to Muller et al.).

2. Abstract (verbatim, as printed)

"Polymorphic forms of 3-(4-amino-1-oxo-1,3 dihydro-isoindol-2-yl)-piperidine-2,6-dione are disclosed. Compositions comprising the polymorphic forms, methods of making the polymorphic forms and methods of their use are also disclosed."

3. Plain-language overview of the claims

The specification discloses eight crystalline forms (Forms A–H), but the claims of the '800 patent are directed to only one of them: the hemihydrate (Form B). There is a single independent claim;

Independent Claim 1 (the only independent claim; claims 2–14 all depend directly on it):

"Crystalline 3-(4-amino-1-oxo-1,3 dihydro-isoindol-2-yl)-piperidine-2,6-dione hemihydrate."

In plain terms: a crystalline solid of lenalidomide that contains roughly half a mole of water per mole of drug (a "hemihydrate"). The claim recites the compound and the hydration state but does not, on its face, name "Form B."

The dependent claims then add identifying characteristics, all tied back to claim 1:

• 2–3: XRPD peaks at ~16, 22 and 27° 2θ (claim 3 adds ~18°)
• 4–5: XRPD peaks at 15.8, 22.2 and 26.7° 2θ (claim 5 adds 18.2°)
• 6: XRPD pattern corresponding to FIGS. 6, 32, 33 and 34
• 7–8: DSC thermogram with an endotherm maximum at ~268 °C (claim 8 adds a dehydration endotherm)
• 9: DSC thermogram corresponding to FIGS. 9, 42, 43 and 44
• 10: approximately 0.46–0.59 moles of water per mole of compound (this is the numeric definition of the "hemihydrate" range)
• 11: TGA weight loss of ~3.1%–4.0% when heated from ~30 °C to ~175 °C
• 12–13: IR peaks at ~3513 and 1960 cm⁻¹ (claim 13: IR spectrum of FIG. 7)
• 14: Raman spectrum corresponding to FIG. 8

Important caveat: The specification's written description covers eight forms and compositions/methods of use, but none of that breadth is captured in the claims. All fourteen claims are limited to the hemihydrate/Form B.

4. Litigation and 2026 docket check

Confirmed historic litigation (Google Patents "Family has litigation" list plus court records):

• Celgene v. Natco Pharma, D.N.J. No. 2:10-cv-05197 — the leading case. On May 27, 2014, Judge Susan Wigenton construed "hemihydrate" (claims 1–14 of the '800 patent) as "a hydrate containing approximately half a mole of water to one mole of the compound forming the hydrate" — adopting Celgene's construction and rejecting Natco's narrower 1:2-ratio/Form-B-only reading (see Casetext summary of Celgene Corp. v. Natco Pharma Ltd.; CourtListener Dkt. 252).
• Numerous other D.N.J. actions against generics: Natco, Lotus/Alvogen, Dr. Reddy's, Cipla, Zydus, Sun, Hetero, Apotex, Mylan (N.D. W. Va. 1:20-cv-00003), plus a D.D.C. case (1:09-cv-01080).
• The '800 patent has been repeatedly listed in Orange Book paragraph IV certifications (e.g., Dr. Reddy's notice letter, Sept. 9, 2016, certified against '800 among others).
• Pending 2026 antitrust litigation (BMS/Celgene antitrust complaints filed in D.N.J., e.g., a May 2026 complaint) continues to attack the polymorph patents as improperly obtained/invalid and as the basis for output-restriction settlements with generics.

On CAFC 2026 dockets specifically: I did not locate any 2026 Federal Circuit appeal docketed for U.S. Patent 7,465,800. The 2026 appellate activity I found in the Revlimid ecosystem is not this patent or this court — it is the UMB Bank (former Celgene shareholders) v. Bristol Myers Squibb CVR dispute in the Second Circuit (revived August 2026), concerning the Breyanzi/ozanimod/ide-cel contingent value rights, entirely unrelated to the '800 patent. I state this as a limitation of my search, not as proof no such docket exists — a definitive negative would require a direct PACER/CAFC docket query.

5. Points of uncertainty

1. Priority basis: The specification claims benefit of provisional 60/499,723 (Sept. 4, 2003), consistent with the listed 2003-09-04 priority date; I did not independently verify the provisional filing.
2. Assignee today: Celgene Corporation was acquired by Bristol Myers Squibb in 2019; the '800 patent is asserted in post-acquisition litigation by BMS/Celgene. The patent record still names Celgene Corp.
3. Expiration: April 27, 2027 reflects patent term adjustment. Different filings in the record describe the family's nominal expiry as Sept. 3, 2024, with the '800, '217 and '357 patents receiving PTA — the April 27, 2027 date is the operative adjusted date.
4. No CAFC 2026 docket found for the '800 patent (see above).

Primary sources: https://patents.google.com/patent/US7465800/en ; https://www.drugpatentwatch.com/p/patent/7465800 ; Espacenet claim set (US7465800B2); Celgene Corp. v. Natco Pharma Ltd., D.N.J. 2:10-cv-05197, Markman opinion (May 27, 2014) and Dkt. 249/252.

I'll research litigation involving US patent 7465800. Let me run several targeted searches.

I have strong leads. Let me pull details on specific cases and verify with litigation databases.

Litigation Involving U.S. Patent No. 7,465,800

Short answer: Yes — the '800 patent has been extensively litigated. It is one of the two "polymorph" (crystal form) patents that Celgene asserted against essentially every lenalidomide ANDA filer, and it is the second-to-last listed Orange Book patent on REVLIMID (NDA 021880), expiring April 27, 2027. Nearly all of the infringement actions were resolved by settlement/consent judgment (permanent injunctions plus volume-limited licenses) rather than by trial verdicts.

1. The patent at issue

(Claims: the widely-discussed limitation is "hemihydrate," construed in the Natco Markman opinion as "a hydrate containing approximately half a mole of water to one mole of the compound forming the hydrate." Celgene also asserted the sister patent US 7,855,217 — "the '217 patent," Form A — alongside the '800 patent in these cases.)

2. Direct Hatch-Waxman / ANDA infringement actions asserting the '800 patent

Per the consolidated antitrust complaint filed by Hagens Berman, "Celgene asserted that the '800 and '217 patents covered Revlimid and asserted these patents in litigations against every ANDA filer" (Table 7, "The '800 Patent Family: Patents asserted in Litigation"), naming eleven generics' ANDAs. The following are the specific cases I was able to confirm as asserting the '800 patent, with the strongest documentation.

Confirmed docket links:

• Zydus: https://npe.law.stanford.edu/case/188700 (Civil Action 2:17-cv-02528, D.N.J., filed Apr. 12, 2017; patents 7465800, 7855217, 7968569, 8530498, 8648095, 9101621, 9101622)
• Hikma: https://dockets.justia.com/docket/new-jersey/njdce/2:2021cv10398/[471118](/patent/471118) (filed Apr. 28, 2021)
• Amneal: https://dockets.justia.com/docket/new-jersey/njdce/2:2024cv00500/[539555](/patent/539555) (filed Jan. 26, 2024; consent judgment Feb. 5, 2024)
• Celgene v. Amneal consent-judgment analysis: https://www.patsnap.com/de/resources/blog/litigation/celgene-v-amneal-pharmaceuticals-lenalidomide-patent-infringement-patsnap/

3. Litigation history of the '800 patent in the Natco case (the most substantive merits proceeding)

From the D.N.J. record (2:10-cv-05197 and the later antitrust recitals):

• Markman (May 27, 2014): The court rejected Natco's proposed construction of "hemihydrate" and adopted Celgene's approximation-based definition. However, that ruling exposed the '800 patent to new invalidity defenses.
• June–July 2014: Natco moved to amend to add invalidity of the '800 patent for indefiniteness, lack of written description, and lack of enablement (arguing the specification only discloses Form B, not other hemihydrates). Over Celgene's opposition, Magistrate Judge Arleo granted leave; Judge Wigenton affirmed July 9, 2015.
• Celgene settled with Natco (announced Dec. 22, 2015) before any infringement or invalidity of the '800 patent was adjudicated, "avoiding scrutiny of the infringement or invalidity of any of its patents."
• In parallel, Dr. Reddy's (2:16-cv-07704) argued its ANDA product was amorphous, not "crystalline," and Celgene ultimately acquiesced in that construction (March 23, 2018), effectively conceding a non-infringement path for amorphous products.

Bottom line on the merits: the '800 patent has never been the subject of a final judgment of infringement or invalidity. Every case ended in settlement or consent judgment.

4. PTAB / IPR — no IPR against the '800 patent found

I found no inter partes review petition directed at the '800 patent. The PTAB proceedings cited by the House Oversight Committee (IPR2018-00685 Apotex; IPR2018-01504 & -01509 Dr. Reddy's; IPR2018-01714 Alvogen) concerned Celgene's use patents ('569, '740 dosing/treatment patents), not the '800 polymorph patent. Alvogen's IPR was instituted but the PTAB refused to invalidate the '569 patent.

Foreign counterpart: the European counterpart of the polymorph family, EP 1,667,682, was opposed by Mylan (Jan. 8, 2012) and Teva and was revoked by the EPO (relevant EPO opposition register: https://register.epo.org/application?number=EP04783095). This is not US litigation but is frequently referenced in the US antitrust pleadings as evidence bearing on the '800 family's validity.

5. Follow-on antitrust litigation in which the '800 patent features

These are not infringement actions on the '800 patent, but the '800 is a central subject of them:

• In re Revlimid and Thalomid Purchaser Antitrust Litigation, D.N.J. (2:14-cv-06997 and related) — insurers/direct purchasers allege the '800 and '217 patent suits were part of a "sham litigation"/reverse-payment scheme. Court opinion: http://business.cch.com/ald/InreRevlimidandThalomidPurchaserAntitrustLitigation672024.pdf
• Molina Healthcare, Inc. v. Celgene Corp., D.N.J., 2:22-cv-04561; and The Cigna Group v. Celgene Corp., D.N.J., 1:25-cv-05237 (filed June 24, 2025) — Section 2 Sherman Act claims referencing the crystal patents.
• Bristol-Myers Squibb/Celgene antitrust complaint, D.N.J. (2026) — https://www.courthousenews.com/wp-content/uploads/2026/05/bristol-myers-squibb-celgene-antitrust-complaint-new-jersey.pdf

6. Additional case numbers from Google Patents' litigation panel (parties not fully verified)

Google Patents' "Family has litigation" panel for US 7,465,800 lists these docket numbers. I was able to confirm parties for some but not all within this research pass, so I list them as unverified leads rather than firm entries:

Confirmed: 2:10-cv-05197 (Natco), 2:16-cv-07704 (Dr. Reddy's), 2:17-cv-02528 (Zydus), 2:18-cv-11630 (Sun), 2:21-cv-10398 (Hikma), 2:24-cv-00500 (Amneal).

Unverified/party not yet identified: 1:09-cv-01080 (D.D.C.); 1:20-cv-00003 (N.D. W. Va.); 2:17-cv-06163; 2:17-cv-06842; 2:18-cv-08964; 2:18-cv-11518; 2:18-cv-17463; 2:19-cv-14731; 2:20-cv-00315; 2:20-cv-07759; 2:20-cv-08570; 2:20-cv-14389; 2:21-cv-11261; 2:21-cv-12927; 2:21-cv-20099; 2:21-cv-20459 (Hikma companion, per Robbins Kaplan); 2:22-cv-02952; 2:22-cv-03753; 2:23-cv-02992; 2:23-cv-21189.

Source: https://patents.google.com/patent/US7465800/en (litigation panel) and Darts-IP family 34272860 (https://patents.darts-ip.com/?family=34272860).

7. Caveats and confidence level

• High confidence: the '800 patent is Celgene's Form B polymorph patent; it was asserted against Natco, Dr. Reddy's, Zydus, Hikma, Amneal, Sun, Alvogen/Lotus, Cipla, and the 2021 wave of generics; the Amneal consent judgment (Feb. 5, 2024) and Hikma consent judgment (Jan. 13, 2022) name/apply to the '800 patent; the '800 patent is listed on REVLIMID through April 27, 2027.
• Moderate confidence: exact filing dates and settlement dates for the 2021 generic wave (Cipla, Aurobindo, Torrent, Biocon, Lupin, Alembic, Mylan) and the date each consent judgment was entered — these are drawn from secondary antitrust recitals, not from each docket directly.
• Low confidence / could not verify: the identity of the parties in the ~20 unlisted New Jersey docket numbers in Section 6. I did not confirm that each of those specific docket numbers actually asserts the '800 patent (as opposed to another patent in the same family, e.g., '217 or '357). No auto-correction of any patent number has been applied; all numbers are reproduced as found.
• I found no trial verdict, no final judgment of infringement, no final judgment of invalidity, and no PTAB IPR specifically adjudicating the '800 patent. Its practical effect has been achieved through settlements/consent judgments.

I'll verify against live sources before writing anything.

Let me search specifically for AIA trials naming this patent.

PTAB / AIA Trial Report — U.S. Patent No. 7,465,800

Patent: 7,465,800 B2 — Polymorphic forms of 3-(4-amino-1-oxo-1,3 dihydro-isoindol-2-yl)-piperidine-2,6-dione (lenalidomide polymorphs; Revlimid®)
Patent owner: Celgene Corporation (now Bristol-Myers Squibb subsidiary) · Inventors: Jaworsky, Chen, Muller
Priority: 2003-09-04 (provisional 60/499,723) · Filed: 2004-09-03 · Granted: 2008-12-16 · Adjusted expiration: 2027-04-27

Proceedings overview

Total AIA trial proceedings on file for US 7,465,800: zero. The USPTO Open Data Portal structured block reports no IPR, PGR, or CBM naming this patent, and my independent web checks surfaced no petition, institution decision, FWD, or appeal referencing the '800 patent.

That means the breakdown is: active 0 / claims invalidated 0 / claims sustained 0 / settled 0 / institution denied 0. There is no FWD for this patent to link to, and no claim of the '800 patent has been canceled or even tested at the Board.

Bottom-line defensive posture: the usual "claims 1–5 are dead, the troll has no case" narrative does not apply here. All claims of the '800 patent — which the D.N.J. claim-construction record refers to as claims 1 through 14 — stand as issued and have never been challenged in an AIA trial. Equally, the patent owner has no PTAB win to hide behind. A defendant gets no free estoppel benefit, but also faces a fully intact patent whose only real-world stress test has been district court Markman briefing and European oppositions.

Proceedings detail

Because the canonical list is empty, there is no proceeding to break out claim-by-claim. What follows is (a) confirmation of the negative, and (b) the adjacent Revlimid-family AIA activity that defendants routinely confuse with this patent.

No proceeding — verified negative

• Type/Status: N/A
• Verification: Structured ODP "PTAB proceedings on file" block = empty. Google Patents record for US7465800B2 (https://patents.google.com/patent/US7465800/en, fetched 2026-10-01) lists litigation links and a worldwide-family litigation link but no PTAB proceedings section.
• Web checks performed: searches for IPR/petition activity keyed to "7,465,800," "7,855,217," and "lenalidomide polymorph" returned only district court Markman briefing (D.N.J. 2:10-cv-05197), ANDA complaints (Cipla, Sun, Amneal, Lotus/Alvogen, Apotex), and antitrust pleadings (BCBSA/Humana-vintage complaints) — no Board proceeding on either polymorph patent.
• Caveat: I could not in this session query PTAB E2E or Docket Alarm directly for a case-by-case docket dump on the '800. The negative is strongly supported but is a "no evidence of" finding, not a docket-level certification.

Adjacent Revlimid-family AIA proceedings — NOT on the '800 patent (context only)

These are the PTAB proceedings everyone conflates with the polymorph patents. None of them names US 7,465,800. Source for the list: House Committee on Oversight and Reform staff report, 2020-09-30, n.81 (https://oversightdemocrats.house.gov/sites/evo-subsites/democrats-oversight.house.gov/files/Celgene%20BMS%20Staff%20Report%2009-30-2020.pdf).

Sanctions side-note with precedent value: Celgene moved to sanction the CFAD petitioner for abuse of process (hedge-fund short-selling strategy). The Board denied sanctions on 2015-09-25, holding that "profit is at the heart of nearly every patent and nearly every inter partes review," and that the AIA imposes no competitive-interest or Article III standing requirement. (IPR2015-01092 et al., Paper 19; https://www.finnegan.com/en/insights/blogs/at-the-ptab-blog/bass-escapes-sanctions-in-celgene-ipr.html.)

Strategic summary

Claim-by-claim status of the '800 patent: all claims untested, all claims presumed valid and enforceable.

• CANCELED: none. There is no IPR certificate for this patent.
• SUSTAINED (adjudicated): none. The Board has never ruled on the '800 patent. Its claim set (claims 1 through 14 per the D.N.J. Markman record) is intact, including the hemihydrate claim 1 and claim 10's "between approximately 0.46 and approximately 0.59 moles of water per mole of 3-(4-amino-1-oxo-1,3 dihydro-isoindol-2-yl)-piperidine-2,6-dione" limitation.
• UNTESTED: everything. That is the whole story. The family's PTAB losses were on the REMS/distribution patents and the use patents — the commercially decisive but structurally different patents. The polymorph patents were never taken to the Board.

Estoppel landscape. With no instituted IPR, § 315(e)(2) estoppel attaches to no one. Every defendant — including Apotex, Dr. Reddy's, Alvogen, Natco, Sun, Cipla, Lotus, Amneal — remains free to raise any § 102/§ 103/§ 112 ground against the '800 in district court, and remains free to file a first IPR petition (subject to the § 315(b) one-year bar from service of a complaint, and to § 325(d) / § 314(a) discretionary-denial risk, whose contours have shifted repeatedly under recent USPTO guidance — verify current practice before filing). The absence of estoppel cuts both ways: the patent owner cannot point to a PTAB affirmance for res judicata or "patent survived IPR" rhetoric, but a would-be petitioner has no roadmap either.

Pattern signals.

1. The polymorph family was defended by covenant, not by PTAB win. Celgene issued express covenants not to sue on sibling patent 7,855,217 — one to Lotus Pharmaceutical/Alvogen Pine Brook (dated 2018-08-08, D.N.J. 2:17-cv-06842) and one shielding the '217 from Natco's declaratory-judgment counterclaims. That is the opposite of a hardened, Board-validated patent; it is a patent kept out of harm's way.
2. No defensive aggregator in the chain. Unified Patents does not appear in any Revlimid polymorph challenge. The ODP block's "family has litigation" links are Unified Patents litigation records (D.N.J. ANDA cases), not Unified Patents IPR filings. The AIA attack came from a hedge fund (CFAD) on different patents, not from an aggregator on the '800.
3. Foreign counterparts fared worse. The European counterpart in this family drew opposition proceedings (EP 1667682, per the House Oversight report n.78), and a Celgene lenalidomide use patent (EP 1505973) was revoked 2013-02-25. European invalidation of the polymorph/use patents is a recurring theme in the antitrust pleadings; the U.S. '800 has simply never faced the equivalent test.
4. Assertion history is heavy, litigation-driven. The '800 has been asserted against essentially every ANDA filer (11+ generics per the antitrust complaints), and as recently as 2024-02-05 was the subject of a consent judgment and permanent injunction against Amneal (D.N.J. 2:24-cv-00500, Patents-in-Suit = 7,465,800 and 7,855,217). Its expiry is 2027-04-27 per the patent record, and it is Orange Book–listed with orphan-drug exclusivity (ODE-245) running to 2026-05-28. The commercial leverage today comes from settlement entry dates and the non-patent exclusivity, not from an adjudicated patent win.

Recommended next steps

1. Do not represent to a court that any claim of the '800 has been canceled. There is no FWD to cite. Any argument built on "the PTAB killed this patent" would be a misrepresentation — the invalidated Celgene claims belong to 6,045,501 and 6,315,720 (Celgene Corp. v. Peter, 931 F.3d 1342 (Fed. Cir. 2019); FWD IPR2015-01092, Paper 73, 2016-10-26). Verify any adverse assertion against the actual patent number in the demand letter before building a defense.
2. Assume no estoppel shield and budget accordingly. Because no IPR was instituted, no petitioner is estopped, and the prior art space — including art the examiner considered during the 2004–2008 prosecution (see the '800 file-history references catalogued in Celgene v. Apotex, D.N.J. 2:19-cv-13994, Dkt. 1 authorities: Bernstein 2002/2004, Brittain 1999/2002/2007, Caira 1998, Grant 1999, Threlfall 2000, Knapman 2000, plus the 2007-10-10 Office Action) — is fully open. Note § 325(d) risk is real if you re-run art already of record, so lead with art the examiner did not consider.
3. The live defense is § 112, not § 102/103. After Judge Wigenton adopted Celgene's broad construction of "hemihydrate" (approximately half a mole of water per mole of compound) on 2014-05-27 (Celgene Corp. v. Natco Pharma Ltd., No. 10-5197 (SDW), D.N.J.; https://g.casetext.com/case/celgene-corp-v-natco-pharma-ltd), the D.N.J. permitted defendants to amend contentions to plead indefiniteness, lack of enablement, and lack of written description (see https://cases.justia.com/federal/district-courts/new-jersey/njdce/2:2010cv05197/[247596/366](https://assignmentcenter.uspto.gov/search/patent/reelFrameDetail?reelFrame=247596-0366)/0.pdf). That is the road the generics actually took. Note the asymmetry: the same 2014 opinion construed "Form A" narrowly (adopting Natco's construction, requiring all specification-attributed characteristics) for the sibling '357 and '598 patents, which largely gutted Celgene's infringement case on those forms — but that construction does not apply to the '800's hemihydrate claims.
4. If you file an IPR, file early and pick forum-conflict strategy deliberately. The § 315(b) clock runs from service of a complaint, the § 315(a)(1) civil-action bar can be triggered if you serve an invalidity contention first in some fact patterns (confirm your local practice), and discretionary denial under §§ 314(a)/325(d) is the single biggest procedural risk to any first-ever petition against a heavily litigated Orange Book patent. Current USPTO guidance on discretionary denial has changed since the 2022-06-21 Interim Procedure memo — confirm the operative guidance as of your filing date.
5. Check the license/settlement posture before investing. Multiple generics hold negotiated entry dates in the 2022–2026 window; if your client is a licensee, a settling party, or a privy, the practical assertion risk is different from that of a new at-risk launcher. The BMS/Celgene Revlimid antitrust litigation (and the 2024-11-05 corrected amended complaint) is worth reviewing for admissions about the patent family's strength.

Verification trail and confidence

• High confidence: no AIA trial proceeding exists on file for US 7,465,800 as of 2026-10-01 (ODP block empty; Google Patents record shows litigation only; targeted web searches returned nothing).
• High confidence: the CFAD IPRs (IPR2015-01092, -01096, -01102, -01103, -01169) targeted the '501, '720, and '517 patents — not the '800. The Board's sanctions decision in that set of cases lists those patents expressly.
• High confidence: the 2018–2019 IPR wave (Apotex, DRL, Alvogen) targeted Revlimid use/method patents ('120, '569 and relatives), not the polymorphs.
• Not fully verified in this session: whether any lower-profile, unpublished, or recently filed petition naming the '800 exists outside indexed sources; the exact PTAB disposition (institution denial vs. institution-then-petitioner-loss) for each individual 2018–2019 Revlimid-family IPR; and the Supreme Court's disposition of the Celgene Corp. v. Peter cert petition (No. 19-1074). Treat those three items as open and confirm against PTAB E2E / CourtListener / the Supreme Court docket before relying on them.

I'll research the assignment record and litigation context before assembling the analysis.

Let me pin down the recorded reel/frame and check for any post-issuance transfers.

Ownership & Assignment Analysis — US 7465800 B2

Patent: US 7465800 B2 — "Polymorphic forms of 3-(4-amino-1-oxo-1,3 dihydro-isoindol-2-yl)-piperidine-2,6-dione"
Application: US 10/934,863 · filed 2004-09-03 · issued 2008-12-16 · priority 2003-09-04 (prov. 60/499,723)
Compound: lenalidomide (REVLIMID® active ingredient); patent listed in the Orange Book as a drug-product/drug-substance patent, adjusted expiry 2027-04-27.

Sourcing note (read first): The USPTO Assignment Center / Assignment Search record for this patent shows exactly one recorded assignment — the original inventor-to-Celgene transfer — with no subsequent post-issuance assignments. The reel/frame number and recorded correspondent for that single entry were not retrievable from the indexed sources (Google Patents legal events, CourtListener exhibits) I could access in this session, and I will not invent them. That one field is the highest-value follow-up: pull it directly at USPTO Assignment Center (search Patent No. 7465800) or assignment.uspto.gov.

Inventors

Named on the issued patent (3):

Discrepancy to flag (material): The recorded assignment of 2005-01-04 (below) lists five assignors — CAMREON, LOUISE M.; CHEN, ROGER SHEN-CHU; JAWORSKY, MARKIAN S.; MULLER, GEORGE W.; and SAINDANE, MANOHAR T. — whereas the granted patent front page names only three inventors. Louise M. Camreon and Manohar T. Saindane therefore appear in the chain of title but not on the '800 patent face.

• This is most likely explained by a single assignment instrument covering a family of co-filed applications rather than this one application alone (the '800 family includes US 7,977,357; US 7,855,217; US 8,058,443; US 8,143,286, several of which were later-filed continuations). I could not verify that explanation from an authoritative source — treat it as a hypothesis, not a finding.
• It is not evidence of an unusual inventor-departure pattern. There is no record in the sources reviewed of all inventors exiting Celgene within 12 months of filing, and Muller remained a Celgene scientist well into the 2010s.

Unusual-pattern check: No "all inventors depart, then fire-sale" signal. The inventors assigned to their employer and the patent stayed with Celgene/BMS for its entire life.

Original assignee

Celgene Corporation.

• Legal form: Delaware corporation; at filing, offices at 86 Morris Avenue, Summit, NJ 07901 (earlier Celgene assignments in this program list 7 Powder Horn Drive, Warren, NJ 07059).
• Primary business: Commercial pharmaceutical manufacturer — an operating company, not a licensing vehicle.
• Product embodying the claims: Yes. Celgene holds NDA No. 021880 for lenalidomide capsules and sells them as REVLIMID® (approved 2005-12-27). The '800 patent is an Orange Book–listed patent for that product.
• Current status: Operating as a wholly owned subsidiary. Celgene was acquired by Bristol-Myers Squibb in a ~$74 billion all-stock transaction completed November 2019. Celgene survived as a BMS subsidiary, which is why no assignment was required or recorded — this was a stock acquisition, not an asset/patent transfer. BMS/Celgene has continued to assert the '800 patent post-acquisition (see 2024 Amneal consent judgment).

Assignment timeline

One recorded assignment exists; there are no post-issuance transfers.

• 2005-01-04 (executed date not separately shown; recording/assignment date on Google Patents legal events) / recorded 2005-01-04 — Reel not retrieved; verify at Assignment Center
  • Conveyance: Assignment of assignors' interest (brief text: "ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS)")
  • Assignor: Camreon, Louise M.; Chen, Roger Shen-Chu; Jaworsky, Markian S.; Muller, George W.; Saindane, Manohar T. (five assignors — see inventorship discrepancy above)
  • Assignee: Celgene Corporation
  • Correspondent: Not retrieved. I could not access the Assignment Center's correspondent field for this entry. Flag: this is the single data point most worth pulling manually — for a pre-issuance, in-house-style assignment to an operating pharma, the correspondent is normally Celgene's own IP department or its regular outside prosecution firm, and a "repeat correspondent" signal would only exist if Celgene used the same firm across the whole patent family.
  • Context: Ordinary employee/consultant invention assignment to the operating employer — not a sale, securitization, or reorg.
• 2008-12-16 — Patent granted. No assignment accompanies issue.
• 2019-11 — Celgene acquired by Bristol-Myers Squibb (stock purchase). No assignment recorded, because Celgene remains the record owner as a BMS subsidiary. Context: parent-level M&A, not a patent transfer.
• No further recorded conveyances (no security agreements, no releases, no change of name, no transfer to any IP-holding or licensing entity).

Bottom line: For practical purposes the chain of title is a single hop: inventors → Celgene Corporation → (still Celgene Corporation, now a BMS subsidiary).

Timeline diagram

timeline
    title Ownership of US 7465800
    2003 : Provisional application filed
    2004 : Utility application filed
    2005 : Inventors assign to Celgene
    2008 : Patent US 7465800 issues
    2010 : Celgene sues Natco over 800 patent
    2019 : Celgene acquired by Bristol-Myers Squibb
    2024 : Consent judgment against Amneal

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT.
The only recorded assignee is Celgene Corporation, a named operating pharmaceutical manufacturer with an FDA-approved product (REVLIMID®, NDA 021880), not a "Holdings/Licensing/Ventures" LLC. No registered-agent-service address, no single-purpose Delaware/Texas LLC appears anywhere in the chain. (Record: single assignment dated 2005-01-04.)

2. Known asserter in the chain — NOT PRESENT.
Celgene does not appear on any of the NPE directories named in the brief. To the contrary, the Stanford NPE Litigation Database classifies Celgene as a "Practicing Entity" / "Product company" in Celgene Corp. v. Zydus Pharmaceuticals (USA) Inc., D.N.J. 2:17-cv-02528 — a case asserting this patent (7465800) alongside 7855217, 7968569, 8530498, 8648095, 9101621, 9101622. Source: https://npe.law.stanford.edu/case/[188700](/patent/188700). No Acacia / Marathon / IV / Wi-LAN / Conversant / Vringo / Pendrell entity appears at any point.

3. Repeat correspondent across the chain — UNCLEAR / NOT ASSESSABLE.
There is only one recorded conveyance, so "recurrence" cannot be established even in principle. The correspondent of record surfaced for this entry was not retrievable in this session. Not a finding.

4. Cascading transfers — NOT PRESENT.
Zero consecutive assignments, let alone multiple LLC-to-LLC hops within 24 months. The chain is static for ~16 years.

5. Pre-litigation transfer — NOT PRESENT.
The single assignment executed/recorded 2005-01-04, roughly four years before grant (2008-12-16) and ~5–9 years before any assertion of the '800 patent (the earliest Celgene v. Natco action on this patent family is D.N.J. 2:10-cv-05197, filed 2010; the patent was asserted against the ANDA filer cohort from 2010 onward). Nothing resembling a transfer timed to set venue or standing.

6. Bankruptcy fire-sale — NOT PRESENT.
No Chapter 7/11 proceedings involving the assignee. Celgene's exit was a ~$74B all-stock acquisition by Bristol-Myers Squibb (Nov 2019), the opposite of a distressed sale.

7. Privateering — NOT PRESENT.
Celgene/BMS asserts the '800 patent in its own name against actual generic competitors (Natco, Dr. Reddy's, Cipla, Sun, Amneal, Alvogen/Lotus, Zydus, Apotex, Hetero, etc.). No transfer to an NPE to assert on Celgene's behalf. Note: several commentators and antitrust complaints characterize Celgene's Revlimid settlements as reverse-payment / output-restriction arrangements (e.g., In re Revlimid & Thalomid Purchaser Antitrust Litig.; the Hagens Berman amended complaint at https://www.hbsslaw.com/sites/default/files/case-downloads/revlimid/2024-11-05-corrected-amended-complaint-redacted.pdf). That is an antitrust characterization of the brand owner's settlement conduct — not an NPE/privateering signal and it involves no assignment.

8. Defensive aggregator (anti-NPE) — NOT PRESENT.
The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN. The patent remains held by the operating originator and is actively asserted, not neutralized.

Verdict

Operating-company assertion.

Celgene Corporation is the sole recorded assignee (assignment dated 2005-01-04, executed four years pre-issuance) and remains record owner today as a wholly owned Bristol-Myers Squibb subsidiary; the chain contains no IP-holding LLC, no security interest, no change of name, and no post-issuance transfer of any kind. Every assertion of US 7465800 in the record is Celgene suing actual generic competitors in ANDA litigation (e.g., Natco 2:10-cv-05197; Amneal 2:24-cv-00500 resulting in a 2024 consent judgment and permanent injunction), and Stanford's NPE database classifies Celgene as a "Practicing Entity / Product company" — the case is the inverse of a troll pattern. The only unresolved item is the reel/frame and correspondent of record for the single 2005-01-04 assignment, which should be pulled manually for completeness; it does not change the verdict.

Verification links:

• USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (search Patent No. 7465800)
• USPTO Assignment Search (legacy) — https://assignment.uspto.gov/patent/index.html
• Google Patents legal events — https://patents.google.com/patent/US7465800/en
• Stanford NPE Litigation Database, Celgene v. Zydus — https://npe.law.stanford.edu/case/188700

Limitations of this analysis: (a) The reel/frame and correspondent fields for the 2005-01-04 assignment were not obtained and are reported as unknown rather than estimated. (b) I found no Assignment Center records indicating any post-issuance conveyance; if a record exists that is not indexed on the sources searched, this section should be updated after a direct Assignment Center query. (c) The five-assignor vs. three-inventor discrepancy is flagged as unresolved and should be checked against the assignment document itself and the sibling '800-family patents.

I'll research this systematically. Let me start with searches on the patent and its cited references.

I now have the claim set confirmed. Let me dig deeper into the specific "Patent Citations" and "Non-Patent Citations" listed on the face of US7465800.

Let me try to access the specific citation tables on the patent's Google Patents page and the USPTO record.

Preliminary note on scope and sourcing

I want to be explicit about two things before the substantive analysis:

1. The identification is confirmed and literal. US 7,465,800 B2 is a real, distinct grant. I did not return results for look-alike numbers (e.g., US 7,465,801, US 7,485,330, US 7,465,363), per your instruction.
2. A limitation on the citation record. The "Patent Citations"/"Non-Patent Citations" tables that normally appear on the front page of a US patent did not surface in the sources I was able to retrieve (Google Patents returned the description-based "Definitions" section rather than the examiner's reference list). What I can ground with specificity is the set of references the patent itself discusses and distinguishes. I flag below where I am relying on the patent's own text versus external data, and I do not invent citation rows I could not retrieve.

1. Patent identification (verified)

Source: https://patents.google.com/patent/US7465800/en

Critical fact for the anticipation analysis: the granted claims are narrowed to a single polymorph — Form B, the hemihydrate. They are not generic to the compound and not generic to "polymorphs A–H." The claim set (per Espacenet claim text, https://sk.espacenet.com/publicationDetails/claims?CC=US&NR=[7465800B2](/patent/7465800B2)&KC=B2) is:

• Claim 1: Crystalline 3-(4-amino-1-oxo-1,3 dihydro-isoindol-2-yl)-piperidine-2,6-dione hemihydrate.
• Claims 2–3: XRPD peaks at ~16, 22, 27 (and further ~18) °2θ.
• Claims 4–5: XRPD peaks at 15.8, 22.2, 26.7 (and further 18.2) °2θ.
• Claim 6: XRPD patterns of FIG. 6/32/33/34.
• Claims 7–9: DSC endotherm with maximum at ~268 °C (and dehydration endotherm; FIG. 9/42/43/44).
• Claim 10: 0.46–0.59 moles water per mole compound.
• Claim 11: TGA weight loss 3.1–4.0% from ~30 to ~175 °C.
• Claims 12–14: IR peaks at ~3513 and 1960 cm⁻¹ (FIG. 7); Raman of FIG. 8.

2. Prior art analysis

2.1 References cited within the patent (the only ones I could confirm)

The patent's Background section expressly identifies two references and incorporates them by reference:

Reference A — US 5,635,517 (Muller et al.)

• Citation: U.S. Pat. No. 5,635,517, Muller et al., "Method of reducing TNFα levels with amino substituted 2-(2,6-dioxopiperidin-3-yl)-1-oxo- and 1,3-dioxoisoindolines," assigned to Celgene.
• Date: granted 1997 (pre-dates the 2003/2004 priority by ~7 years). — Source: US7465800 text ("U.S. Pat. Nos. 5,635,517 and 6,281,230 both to Muller et al.") and https://pharsight.greyb.com/ingredient/lenalidomide
• Description: Discloses the genus of amino-substituted 2-(2,6-dioxopiperidin-3-yl)-isoindolines and their use to reduce TNFα, which encompasses the 4-amino-1-oxo (lenalidomide) species.
• § 102 analysis: This is the closest compound-disclosure art, but it is a compound/utility reference. It discloses the molecular entity; it does not disclose the crystalline hemihydrate, the 0.46–0.59 mol water stoichiometry, the ~268 °C DSC endotherm, or the 3513/1960 cm⁻¹ IR doublet. Under § 102, a reference anticipates only if it discloses every limitation arranged as in the claim. It therefore does not anticipate claims 1–14. Its realistic role is (a) § 102 for the compound per se if a claim recited the compound alone — but no claim here does — and (b) § 103 as the base compound for an obviousness attack on the hemihydrate (a theory the patent would rebut with the unexpected stability/hygroscopicity data at §5.2.2). This reference was also the subject of an IPR (Coalition for Affordable Drugs VI LLC; institution denied 16 Nov 2015 — https://pharsight.greyb.com/drug/revlimid).

Reference B — US 6,281,230 (Muller et al.)

• Citation: U.S. Pat. No. 6,281,230, Muller et al., "Isoindolines, method of use, and pharmaceutical compositions," Celgene.
• Date: granted 2001-08-28 (family filing ~1996; the greyb record lists a 2016-07-24 expiry consistent with a mid-1996 priority). — Sources: US7465800 text; https://pharsight.greyb.com/ingredient/lenalidomide
• Description: Discloses isoindoline compounds including 3-(4-amino-1-oxo-1,3-dihydro-isoindol-2-yl)-piperidine-2,6-dione, pharmaceutical compositions thereof, and methods of use for inflammatory/autoimmune disease and cancer. The patent's own §5.2 calls it out as the disclosure of the compound.
• § 102 analysis: Same structure of analysis as Reference A but closer, because it expressly claims the specific compound and pharmaceutical compositions. Still, it is silent on solid-state form. It does not anticipate claims 1–14. It would be the primary § 103 reference (isolated compound + obviousness of a stable hydrate), and also the anticipatory reference for a hypothetical claim drawn to the compound or a composition thereof — but US 7,465,800 contains no such claim.

Note on the "both to Muller et al." phrasing: the patent says both references "disclose 3-(4-amino-1-oxo-1,3 dihydro-isoindol-2-yl)-piperidine-2,6-dione." Read literally (your instruction), these two Muller references are the applicant's own identified anticipatory-level compound disclosures — i.e., the applicant concedes the compound is old and stakes the invention solely on the polymorph.

2.2 Narrower cross-references named in the specification (not § 102 art against the polymorph claims)

The specification also cites, and incorporates by reference, related Zeldis applications — application Ser. Nos. 10/411,649 (MDS), 10/438,213 (cancer), 10/411,656 (myeloproliferative diseases), 10/693,794 (pain), 10/699,154 (macular degeneration) — plus US 6,235,756 and US 6,114,335 to D'Amato. These are method-of-treatment/utility references only; they contain no solid-state characterization and cannot anticipate a claim requiring a crystalline hemihydrate with defined XRPD/DSC/IR/Raman signatures.

2.3 Non-patent literature cited in the patent

• *P. DiMartino, et al., J. Thermal Anal., 48:447–458 (1997)* — general polymorph physical-property paper (solubility/flow/compression). Background/§ 103-adjacent only.
• Knapman, K., Modern Drug Discoveries, 2000, 53 — general statement that polymorphs affect solubility, stability, flowability, fractability, compressibility, safety, efficacy.

Both are general-principle references, cited to establish the state of the art of polymorph screening; neither discloses lenalidomide Form B. They cannot anticipate any claim, and alone they do not render the isolated hemihydrate obvious, because the law does not treat "try to make a polymorph" as sufficient without a reason to expect the specific form.

3. Bottom line on § 102 anticipation

No reference in the record I could retrieve anticipates any of claims 1–14. The reason is structural: every claim requires a specific crystalline hemihydrate of lenalidomide with defined spectral/thermal fingerprints. The Muller references (the strongest art) are limited to the compound and its therapeutic use and are silent on hydration state, XRPD, DSC, TGA, IR, and Raman. Their genuine significance is as § 103 base references, and the patent's § 5.2.2 data (Form B hemihydrate's aqueous stability, non-hygroscopicity, ~3.1% volatile loss, ~146 °C dehydration / ~268 °C melt) is the applicant's evidence of non-obviousness. Note also the later same-family grants US 7,855,217, US 7,977,357, US 8,058,443, US 8,143,286 (all continuations of this disclosure), which confirms the family's strategy was to claim additional Forms rather than broader compound claims.

4. Honesty caveats (per your operating rules)

• I could not retrieve the actual examiner-cited front-page reference table for US 7,465,800. My analysis is anchored to the references the patent's own text identifies (US 5,635,517 and US 6,281,230), the incorporated method-of-use applications, and the two NPL items (DiMartino 1997; Knapman 2000). If a complete front-page "Patent Citations" list exists containing additional US/foreign documents, it is not reflected here and I have not fabricated entries for it.
• Precise grant/filing dates for US 5,635,517 and US 6,281,230 are given with the sourcing shown; I am highly confident of the US 6,281,230 → 2001-08-28 grant and the US 5,635,517 → 1997 grant, but I could not independently confirm the exact 5,635,517 grant day in this session and have not asserted one.
• The KR family member (KR 10-1005991 B1 / KR 2006-0087543 A) and UA 83504 C2 are same-family counterparts, not prior art, and are excluded from the anticipation table.

Primary source: https://patents.google.com/patent/US7465800/en · claims: https://sk.espacenet.com/publicationDetails/claims?CC=US&NR=7465800B2&KC=B2 · litigation/docket corroboration: https://www.courtlistener.com/docket/[6917120](/patent/6917120)/1/celgene-corporation-v-zydus-pharmaceuticals-usa-inc/

I'll research the prior art landscape and relevant obviousness doctrine for this polymorph patent before building the analysis.

Let me pull the specific claim language and any examiner-cited prior art from the file history.

Obviousness Analysis — U.S. Patent 7,465,800 (35 U.S.C. § 103, pre‑AIA)

Patent: US 7,465,800 B2, "Polymorphic forms of 3‑(4‑amino‑1‑oxo‑1,3 dihydro‑isoindol‑2‑yl)‑piperidine‑2,6‑dione" (Celgene Corp.; inventors Jaworsky, Chen, Muller; app. 10/934,863 filed 2004‑09‑03; priority to prov. 60/499,723 filed 2003‑09‑04). Source: https://patents.google.com/patent/US7465800/en
Governing law: Pre‑AIA § 103(a) (priority 2003‑09‑04), Graham v. John Deere, 383 U.S. 1 (1966); KSR Int'l v. Teleflex, 550 U.S. 398 (2007).

1. The claims at issue

Claim text (from the published claims): https://sk.espacenet.com/publicationDetails/claims?CC=US&NR=[7465800B2](/patent/7465800B2)

Note the breadth point litigated in Celgene v. Natco: the court construed "hemihydrate" as "a hydrate containing approximately half a mole of water to one mole of the compound forming the hydrate" — i.e., approximation, not an exact 1:2 ratio (Markman Opinion, D.N.J. 2:10‑cv‑05197, ECF 312, at 6‑7; discussed at https://storage.courtlistener.com/recap/gov.uscourts.njd.[247596](/patent/247596).249.0.pdf). That construction expands claim 1 (and claims 7–14, which all depend from it) well beyond Form B.

Hypothetical POSA: a formulation/process chemist or solid‑state scientist with an advanced degree and 2–5 years of pharmaceutical development experience, familiar with XRPD, DSC, TGA, KF, IR/Raman, and with routine crystallization/slurry/desolvation techniques.

2. Prior art on the face of the patent

The Google Patents page's prior‑art keywords are polymorph, representative, isoindol, dione, piperidine; the enforceable prior art for § 103 purposes is the set the specification itself identifies and characterizes:

Critical admissions in the specification itself (usable as evidence of the state of the art, not as prior art):

• "Polymorphs of a molecule can be obtained by a number of methods known in the art … melt recrystallization, melt cooling, solvent recrystallization, desolvation, rapid evaporation, rapid cooling, slow cooling, vapor diffusion and sublimation."
• Polymorphs "can be detected, identified, classified and characterized using well‑known techniques such as … DSC, TGA, XRPD … IR … Raman …"
• "drugs are approved by agencies such as the U.S. FDA only if they meet exacting purity and characterization standards" (regulatory driver).
• Form B "can be obtained from many solvents, including, but not limited to, hexane, toluene, and water."
• Form E "was obtained most readily by slurrying … in water"; and the "laboratory procedure … to exclusively produce polymorph B" is a "stirred 10 volume water slurry at ~75 °C for 6‑24 hours."

3. Combination 1 — Muller '517 + '230 + Knapman + DiMartino → claim 1

The combination. '517/'230 disclose the compound (a small molecule bearing an imide N–H and two lactam/imide carbonyls — a strong hydrogen‑bond donor/acceptor array) and its pharmaceutical utility. Knapman and DiMartino teach that (a) drug substances commonly exist as polymorphs, (b) polymorphs have materially different solubility, stability, compressibility and processability, and (c) thermal/vibrational methods exist to detect and distinguish them.

Motivation to combine (KSR rationales).

1. Known compound + known technique yielding a predictable result. Once a compound is identified as a development candidate, producing a crystalline, characterizable, pharmaceutically acceptable solid form is a design requirement, not an invention. The specification admits the requisite techniques were "known in the art."
2. Finite number of identified, predictable solutions. The available pharmaceutically acceptable recrystallization solvents/solvent systems — including water, and the water‑miscible systems (acetone/water, acetonitrile/water) — constitute a small, known set. The dihydrate (Form E) and hemihydrate (Form B) are precisely the hydration products a POSA would expect from aqueous systems. Under KSR, "obvious to try" suffices where the prior art identifies a finite set of predictable solutions.
3. Market/regulatory design incentive. Knapman/DiMartino supply the express reason: solubility, stability, flowability, and compressibility drive form selection, and the specification itself concedes the FDA's characterization requirements.
4. Same utility. The hemihydrate has the identical utility as the '517/'230 compound; under In re Papesch, 315 F.2d 381 (CCPA 1963), a claimed form cannot be rescued by properties inherent in the known molecule.

Why claim 1 is the most exposed claim. It recites no XRPD, DSC, TGA or water‑content limitation. To invalidate it, a challenger need only show that some crystalline hemihydrate of lenalidomide was an obvious result of routine aqueous recrystallization/slurrying — not that Form B was predictable. The KSR "predictable solutions" and "known techniques" rationales map directly onto a genus this broad.

4. Combination 2 — + DiMartino (DSC/TGA) and the admitted "well‑known" characterization methods → claims 2–5, 7–8, 10–12

Claims 2–5 (XRPD °2θ positions), 7–8 (DSC endotherms at ~268 °C and a dehydration endotherm), 10 (0.46–0.59 mol water) and 11 (TGA 3.1–4.0 % mass loss) recite measured, inherent properties of the same crystalline hemihydrate. DiMartino expressly uses thermal analysis to characterize polymorphs, and the specification concedes XRPD/DSC/TGA/IR/Raman are "well‑known techniques." Under Papesch and Sanofi‑Synthelabo v. Apotex, 550 F.3d 1075, 1079 (Fed. Cir. 2008) ("a chemical compound and its properties are inseparable"), adding a characterization datum does not confer patentability on an otherwise obvious form. Salix v. Norwich, 98 F.4th 1056, 1066 (Fed. Cir. 2024), is the closest analogue: where the claimed crystal "can be readily produced" by known conditions and characterization methods are "well known and readily available," the difference over the art is "nothing more than the performance of routine characterization."

5. Combination 3 — + the specification's admitted route to Form B → claims 4–6, 9, 13, 14

For the picture claims (FIGS. 6/32/33/34, 7, 8, 9/42/43/44), the challenger's theory is the same but must reach the specific Form B. Here the specification's own admissions supply the missing teaching: Form B "can be obtained from … hexane, toluene, and water," and the production procedure is a water slurry at 75 °C. That is a routine, conventional crystallization/slurry operation using a solvent Knapman/DiMartino‑type art identifies as a standard screening solvent. A POSA seeking a stable hemihydrate for capsule formulation would have had every reason to run a water slurry of the known compound.

Related inherency/anticipation theory (record‑grounded). Practitioners have argued that practicing the '517 process inevitably yields a specific form: the EPO revoked Celgene's EP '682, the Form A polymorph patent, on the stated rationale that Form A was anticipated by the '517 patent because following '517 necessarily produces it. See In re Revlimid & Thalomid Purchaser Antitrust Litig., No. 2:19‑cv‑07532 (D.N.J.), slip op. at 233–35, http://business.cch.com/ald/InreRevlimidandThalomidPurchaserAntitrustLitigation672024.pdf. The same plaintiffs alleged that the '800 and '217 patents "are also obvious variants of the composition of matter patent." If a water‑slurry/recrystallization step practiced on '517‑made material inherently gives the hemihydrate, the inherency bridge to claim 1 is short — though the Federal Circuit requires that the result be "the natural result flowing from the operation as taught," Millennium Pharms. v. Sandoz, 862 F.3d 1356, 1365 (Fed. Cir. 2017).

6. Combination 4 — Muller + D'Amato + Zeldis (Apr–May 2003) → motivation/prima facie glue

D'Amato ('756/'335) and the Apr–May 2003 Zeldis applications show the compound was being developed for chronic oral dosing across multiple indications. That supplies the "market forces/design incentive" KSR factor: a chronic oral product needs a reproducibly manufacturable, non‑hygroscopic, storage‑stable crystalline form. Caveat: the Zeldis filings are § 102(e) art, commonly owned by Celgene, and therefore cannot be used in a § 103 combination under pre‑AIA § 103(c). Their proper use is only to establish the state of the art/background, not as § 103 references.

7. Counterarguments (and why they may defeat some claims)

1. Unpredictability/unreasonable expectation of success. Grünenthal v. Alkem, 919 F.3d 1333, 1341‑43 (Fed. Cir. 2019) (polymorph screening disclosure, even a Byrn‑type flowchart of variables, is insufficient); Pharmacyclics v. Alvogen (N.D. Cal.), aff'd (Fed. Cir. 2022) (no reasonable expectation of producing a specific form). Only ~30–35 % of compounds are polymorphic; the prior art here does not disclose that lenalidomide is polymorphic at all, let alone identify any form. See https://www.finnegan.com/en/insights/articles/ip-law-in-defense-of-polymorphs.html.
2. Unexpected properties. The specification and figures establish Form B is uniquely stable: no significant weight change 5→95 % RH; no form conversion after 85 days at RT/0 % RH and 40 °C/93 % RH, whereas Forms A, C, D convert to Form B and Form E converts to Form H under stress. Form B was selected as the API form and used in clinical product. That is classic unexpected‑results evidence and distinguishes Salix, where the polymorph patent "did not identify any advantageous properties" and the claimed form could be "readily produced" from disclosed conditions.
3. Claim‑scope mismatch. The unexpected‑stability evidence supports only Form B. It is commensurate with claims 4–6, 9, 13–14, but not with claim 1, which sweeps in every hemihydrate (per the Markman construction). This asymmetry is the strongest invalidity wedge.
4. Commercial success (Revlimid) has weak nexus because the '517 compound patent barred competition — Sanofi, 550 F.3d at 1090.

8. Opinion and vulnerability ranking

Bottom line: Claim 1 is the claim most likely to fall under § 103 on the combination of U.S. 5,635,517 and 6,281,230 with Knapman (2000) and DiMartino (1997), with the specification's own admissions that recrystallization/slurry/desolvation and DSC/TGA/XRPD/Raman characterization were "known/well‑known in the art" supplying the motivation and reasonable expectation of success. The narrower Form B picture claims present a materially closer question, and are supportable on the basis of Form B's demonstrated anomalous aqueous‑phase stability and stress resistance.

Stated uncertainties: I do not have the prosecution history of 10/934,863 (actual examiner rejections, prior‑art statements, terminal disclaimers), the full text of Knapman or DiMartino beyond the single sentences the specification quotes, or the complete "References Cited" list from the patent front page. The combinations above are built from the references cited on the patent page and the specification's quoted characterizations of them; any additional of‑record references (e.g., an examiner‑cited amorphous‑lenalidomide application noted in Natco's prosecution) would need to be verified before filing an invalidity contention.