US 8735428

I'll verify the patent record and check for CAFC 2026 litigation activity.

Let me check the specific litigation and Orange Book/regulatory record.

I have the bibliographic and litigation record. Now let me try to retrieve the actual claim text, since the supplied full text cuts off before the claims.

US 8,735,428 B2 — Verification Summary

Search scope: Google Patents record for US8735428 (authoritative full text supplied), DrugPatentWatch litigation records for patent 8,735,428, Orange Book listings for pomalidomide/POMALYST (NDA 204026), Federal Circuit docket records, and district-court filings from D.N.J. and S.D.N.Y. that quote the patent's claims and file history.

Important caveat up front: I could not retrieve the patent's verbatim claim set from an authoritative source. The full text you supplied is truncated mid-specification (it ends at "…4-(amino)-2-(2,6-dioxo" in the Actimid dosing passage) and does not include the claims. My claim description below is reconstructed from court filings and is flagged accordingly. I also found no CAFC 2026 docket activity for this patent.

1. Bibliographic Record

The active moiety is pomalidomide (Celgene code name "ACTIMID"), designated in the specification as 4-(amino)-2-(2,6-dioxo(3-piperidyl))-isoindoline-1,3-dione, PubChem CID 134780 / CAS-linked formula C₁₃H₁₁N₃O₄.

Family position (per S.D.N.Y. 1:25-cv-05237 complaint, ¶¶176–191): provisional 60/380,842 (May 17, 2002) → provisional 60/424,600 (Nov 6, 2002) → application 10/438,213 (May 15, 2003) → division 12/229,074 → continuation 12/640,702 → continuation 13/488,888 → continuations 13/782,728 and 13/782,612 (both filed Mar 1, 2013). The '428's May 15, 2023 expiry is the 20-year term measured from the May 15, 2003 parent filing. I note the Cigna complaint asserts the parties treated November 6, 2002 as the operative priority date during litigation, not the May 2002 date Google Patents displays — a discrepancy worth flagging.

2. Abstract (as published)

Methods of treating, preventing and/or managing cancer as well as and diseases and disorders associated with, or characterized by, undesired angiogenesis are disclosed. Specific methods encompass the administration of an immunomodulatory compound alone or in combination with a second active ingredient. The invention further relates to methods of reducing or avoiding adverse side effects associated with chemotherapy, radiation therapy, hormonal therapy, biological therapy or immunotherapy which comprise the administration of an immunomodulatory compound. Pharmaceutical compositions, single unit dosage forms, and kits suitable for use in methods of the invention are also disclosed.

Note this is a boilerplate, genus-level abstract that reads onto the whole IMiD family and does not mention multiple myeloma or pomalidomide by name. The actual inventive contribution is in the claims, not the abstract.

3. Plain-Language Overview of the Independent Claims

⚠️ Reconstruction, not verbatim. No independent claim text was available from the authoritative fetch. The following is drawn from claim-construction and infringement filings in Celgene Corp. v. Apotex Inc. / Hetero (D.N.J. 2:17-cv-03387, Dkt. 211) and Dkt. 251, which quote the '428 claims and its file history.

• Claim 1 (independent) — Method of treating multiple myeloma. In plain terms: A method of treating multiple myeloma by administering pomalidomide (or a pharmaceutically acceptable salt, solvate, or stereoisomer) to a patient who has multiple myeloma, with a patient-population limitation — the filings indicate claim 1 is directed to patients who have undergone prior therapy, e.g., with thalidomide, lenalidomide, or a proteasome inhibitor. The full preamble as quoted in the D.N.J. briefing reads: "A method of treating multiple myeloma, which comprises administering to a patient having multiple myeloma, and which patient…" (text continues beyond the available snippet).

• A further independent claim — Dosage-limited method. Claims recite administration of "about 1 mg to about 5 mg per day" of the pomalidomide compound (or a salt/solvate/stereoisomer thereof). Related family filings reference a 1 to 4 mg/day range. This limitation was the crux of the §112/§103 disputes over whether salt, solvate, and stereoisomer weights count toward the recited dose.

• The patent's independent claims appear to be method-of-use claims closed over the prior art by the combination of (a) the specific compound, (b) the specific daily dose, and (c) the prior-therapy patient population — precisely the three elements Celgene argued were not suggested by Kyle (2001), Davies (2001), Corral (1999), Muller (1999), and US 6,555,554.

I cannot confirm: the total claim count, the exact number of independent claims (at least two), the precise claim 1 wording, or whether combination-with-dexamethasone and 21-day/7-day cycling limitations appear in the '428 itself or only in its sibling patents '262 and '3939. The Dex/cycling language I encountered in the record appears tied to the '262 reexamination, not the '428.

4. Prosecution History Highlights

• The examiner initially rejected the claims for double patenting over the '262 and for obviousness over Kyle (2001), Davies (2001), Corral (1999), Muller (1999), and the '554 patent.
• On October 4, 2013, Celgene conducted an examiner interview urging that pomalidomide unexpectedly treats myeloma resistant to lenalidomide despite the structural similarity of the two compounds.
• On October 8–9, 2013, Celgene filed terminal disclaimers over the '262 and '3939 (and any patents issuing from related applications) to clear the double-patenting rejection.
• The Thakurta Declaration (filed October 9, 2013) supplied the evidentiary basis for the unexpected-results argument; the claims issued May 27, 2014.
• Materiality alert: the S.D.N.Y. antitrust complaints (Cigna, 1:25-cv-05237; Hagens Berman Pomalyst action) allege that this unexpected-results showing was false and procured by inequitable conduct, asserting that pomalidomide's myeloma activity was already disclosed by Lentzsch (2001/2002), Schey (April and October 2002), and D'Amato (2001) — references the complaints allege were concealed or not disclosed with their import. These are allegations in pending litigation, not adjudicated findings, and I have no information that the '428 has been held invalid or unenforceable.

5. Litigation and CAFC Docket Check

Confirmed dockets naming patent 8,735,428 (per DrugPatentWatch, last updated March 2026):

• Celgene Corp. v. Par Pharmaceutical, Inc., D.N.J. 2:17-cv-03387 (filed May 4, 2017; terminated Feb. 6, 2019), §271 ANDA infringement
• Celgene Corp. v. Mylan Pharmaceuticals Inc., D.N.J., aff'd, No. 2021-1154 (Fed. Cir. Nov. 5, 2021) — venue dismissed as improper in New Jersey; dismissal of Mylan N.V. affirmed. This is the only Federal Circuit appeal I found touching this patent, and it is a 2021 decision on venue, not a merits ruling on the '428.
• M.D.N.C. 1:18-cv-00540 (filed June 21, 2018), §271 infringement
• In re Motion to Compel, D. Mass. 1:20-mc-91045 (Jan.–Feb. 2020)
• Antitrust follow-on actions: Louisiana Health Service & Indemnity Co. v. Celgene, S.D.N.Y. 1:23-cv-07871; New York Hotel Trades Council v. Celgene, S.D.N.Y. (2024); CenterWell Pharmacy, Inc. v. Celgene, S.D.N.Y. 1:24-cv-06924; The Cigna Group v. Celgene, S.D.N.Y. 1:25-cv-05237 (filed June 24, 2025)
• Hagens Berman Pomalyst Pomalidomide action (2023–2025 filings)

On your specific request for CAFC 2026 dockets: I searched for 2026 Federal Circuit activity naming 8,735,428 and found none. The Google Patents litigation panel lists only Federal Circuit case 21-1154. The 2026 Federal Circuit items my searches surfaced (e.g., REGENXBIO v. Sarepta, the VirtaMove/Google mandamus petition) involve different patents entirely and should not be attributed to 8,735,428. Note also that the "542" hits in my first search (Genzyme v. Novartis Gene Therapies, Sarepta) are U.S. Patent No. 9,051,542 — a different patent, not 8,735,428. Per your instruction on literal identifier interpretation, I did not treat those as results for this patent.

6. Explicit Uncertainty Flags

1. Claim text is not verified against the patent's own claim set — the supplied full text is truncated before the claims, and I could not obtain verbatim claims from a primary source within this search. Treat the §3 overview as a well-grounded reconstruction from litigation quotes, not a quotation.
2. Number of claims and independent claims is unconfirmed.
3. Priority date discrepancy (Google Patents: May 17, 2002 vs. 20-year term running from the May 15, 2003 parent filing; litigation treats Nov. 6, 2002 as operative).
4. Inventorship: the single named inventor per the record is Jerome B. Zeldis. The complaints characterize Zeldis as a Celgene senior executive and the sole inventor on the pomalidomide method-of-use patents — I did not independently verify inventorship of the '428 beyond the assignment record.
5. PTAB activity: I found no IPR/PGR specific to 8,735,428. The "IPR2023-00608" that appeared in one search hit belongs to the Sarepta/Genzyme '542 patent matter, not this patent.
6. All inequitable-conduct, fraud-on-the-PTO, and sham-litigation assertions are allegations in pending or settled civil actions and are not findings of fact or law.

Primary sources: https://patents.google.com/patent/US8735428/en · https://www.drugpatentwatch.com/p/litigation/patent/8735428 · https://www.courtlistener.com/opinion/[5303687](/patent/5303687)/celgene-corporation-v-mylan-pharmaceuticals-inc/ · https://business.cch.com/ald/TheCignaGroupvCelgeneCorporationComplaint.pdf · Orange Book listings for POMALYST (NDA 204026).

I'll research litigation involving US Patent 8,735,428. Let me run several searches.

Good initial results. Let me dig into the specific cases and dockets.

Litigation Involving US Patent 8,735,428 ("the '428 patent")

Patent identified: US 8,735,428 B2, "Methods for treating multiple myeloma with 4-(amino)-2-(2,6-dioxo(3-piperidyl))-isoindoline-1,3-dione" (pomalidomide), Celgene Corporation, issued May 27, 2014, application 13/782,612, priority May 17, 2002, expired/expiring May 15, 2023. Orange Book listed against POMALYST (NDA 204026) under use codes U-1360 and U-2254. (Terminal disclaimer links it to the '262 and '3939 method-of-treatment patents.)

A. Hatch-Waxman / ANDA patent-infringement cases (the '428 patent asserted)

Note on the Markman ruling: In Celgene Corp. v. Hetero Labs Ltd., the court construed "treating multiple myeloma" in the '262, '3,939 and '428 patents (common specification) and declined to read in an efficacy limitation — a ruling that substantially undercut Celgene's method-of-treatment patents.

B. Additional cases listed in the patent record (Unified Patents litigation data on the Google Patents page)

The Google Patents "Family has litigation" record for this patent lists these additional matters. Party names could not be confirmed for most of them:

C. Antitrust cases in which the '428 patent is cited as part of the challenged Celgene patent portfolio

These are not infringement actions on the '428 patent; the patent is pleaded as part of the patent estate alleged to have unlawfully maintained the Revlimid/Thalomid/Pomalyst monopoly (e.g., In re Revlimid & Thalomid Purchaser Antitrust Litigation, D.N.J. lead No. 2:19-cv-07532). The '428 patent's expiration (May 15, 2023) is cited as relevant to the alleged no-generic-entry-until-2026 settlement structure.

Status in these matters is generally pending/consolidated; the court has already resolved Noerr-Pennington/"sham litigation" motions in the consolidated Revlimid/Thalomid litigation (see the June 2024 opinion in 2:19-cv-07532).

D. Post-grant proceedings

No IPR/PGR directly challenging the '428 patent was found. The only PTAB matter surfacing in the search (IPR2015-01103, CFAD VI v. Celgene) challenged a different patent (U.S. 6,315,720) and merely used the Pomalyst Orange Book patent listing — including the '428 patent — as an exhibit.

Important caveats

1. One-stop confirmation was not possible. The Unified Patents litigation portal (portal.unifiedpatents.com) entries for this patent were not directly retrievable; I relied on the litigation data reproduced on the Google Patents page for US8735428, the Stanford NPE Litigation Database page for patent 8735428, and DrugPatentWatch's "Litigation for Patent 8,735,428" page.
2. Google Patents' "family litigation" list can include suits on family members, not only on the '428 patent itself. Treat Section B as leads requiring docket-level verification rather than confirmed '428 assertions.
3. Case numbers/dates for the Section C antitrust matters were only partially retrieved; the docket numbers shown for the S.D.N.Y. cases were not confirmed.
4. No litigation was found in which the '428 patent was the sole or primary asserted patent outside the Pomalyst ANDA cases (Celgene v. Par; Celgene v. Hetero).

If you need the exact docket-level detail (filing dates, termination orders, settlement terms) for Section B or the S.D.N.Y. antitrust matters, that requires pulling the PACER dockets or the Unified Patents case pages directly.

I'll start by verifying whether any AIA trial proceedings exist on this patent, since the structured block reports none.

Let me run more targeted searches on PTAB proceeding numbers and the litigation chain.

Let me check the drugpatentwatch PTAB listing for pomalidomide and search PTAB dockets directly.

AIA Trial Proceedings — US 8,735,428 B2

Proceedings overview

Zero. US 8,735,428 has no AIA trial proceedings on file — no IPR, PGR, or CBM, instituted or denied — per the USPTO Open Data Portal structured block supplied for this patent, and no such proceeding surfaced in web searching. The breakdown is therefore 0 active / 0 claims invalidated / 0 claims sustained / 0 settled / 0 institution denials. Bottom line for a defendant: the patent is not "hardened" by a PTAB win — it has simply never been tested at the Board at all. Every invalidity theory you would want to raise against the '428 is un-adjudicated and, correspondingly, no PTAB estoppel exists that could block anyone. The defensive constraint here is not estoppel but the calendar: the '428 is listed in the structured data as Expired – Lifetime, with an anticipated expiration of 2023-05-15, which materially changes what an IPR is worth.

Proceedings

None to report. There are no proceeding numbers to list. To be explicit about the scope of that finding:

• The structured PTAB block for US 8,735,428 is empty (ODP returns no AIA trials).
• Targeted searching for an IPR/PGR on the '428 produced no petition, no institution decision, and no Final Written Decision. The Celgene pomalidomide PTAB activity that does exist is on other patents in the Pomalyst Orange Book family — I flag these below only as adjacent context, not as proceedings against the '428.

Adjacent Celgene PTAB activity — NOT proceedings on the '428 (listed to prevent misattribution)

These are different patents. Do not cite any of these as a proceeding against US 8,735,428.

Sources: Celgene 10-Q/10-K Pomalyst and USPTO-notes disclosures (https://www.getfilings.com/sec-filings/[180726](/patent/180726)/CELGENE-CORP-DE-_10-Q/); Jones Day experience page for the Bass REMS IPRs (https://www.jonesday.com/en/practices/experience/2015/04/celgene-defends-rems-patents-against-kyle-bass-ipr-petitions); Managing IP on the CAD institution denial (https://www.managingip.com/article/2a5bugryp18sj8vowx3i8/celgene-win-leaves-bass-even-in-ptab-institution-game); Celgene Preliminary Response in IPR2018-01504 (https://www.docketalarm.com/cases/PTAB/IPR2018-01504/Inter_Partes_Review_of_U.S._Pat._9056120/).

Why the '428 appears to have escaped the Board. The '428 was asserted in the 2017 Pomalyst Hatch-Waxman wave (Teva, Apotex, Hetero, Aurobindo, Mylan, Breckenridge, later Synthon; consolidated in D.N.J., e.g., Celgene Corp. v. Hetero, No. 2:17-cv-03387), where the generics chose the district court counterclaim route rather than the PTAB — answers and DJ counterclaims of invalidity were filed from July 2017 onward. The '428 never got a PTAB petition; it got a district-court claim construction instead: a 2020-06-16 ruling by Judge Salas construing the method-of-treatment patents (per plaintiffs' later antitrust pleadings describing that decision — treat the date and substance as second-hand until you pull the order yourself). The portfolio then resolved through settlements rather than judgments, which is why nothing was ever taken to the Board on the '428.

Strategic summary

Claim status: all claims UNTESTED. US 8,735,428 has one independent claim directed to a method of treating multiple myeloma with pomalidomide with a specified cycle structure (21 days on / 7 days rest), for relapsed/refractory MM with demonstrated progression after prior therapy, plus a dependent claim adding dexamethasone. None of those claims has been canceled, confirmed, or even construed by the PTAB. Independent prior-art challenges exist on paper (pre-2002 D'Amato/Children's Hospital/EntreMed art, Davies 2001, Kyle 2001, and the examiner's own pre-allowance obviousness rejections over Kyle (2001), Davies (2001), Corral (1999), Muller (1999)), and the generic defendants catalogued more than 100 references in their 2017-12-17 invalidity contentions — but none of that has been adjudicated by the Board. If you are being asserted on this patent, you are the first party with a shot at the PTAB, which means zero § 315(e)(2) estoppel runs against you and no prior petitioner's expert record exists to leverage.

Two hard practical limits on an IPR. First, expiry: the structured record shows Expired – Lifetime, anticipated expiration 2023-05-15. An IPR against an expired patent is of little operational value — there is no prospective injunction to defeat, and the realistic upside is pressure on a past-damages claim within the § 286 six-year lookback. Second, the § 315(b) one-year bar: every would-be petitioner who was served with a complaint in the 2017–2022 D.N.J. MDL-style Pomalyst cases, or in the related N.C. M.D. action, has long been time-barred. In practice the "free swing" at the Board has effectively closed for the parties who were actually sued.

Pattern signals worth noting. (1) There is no serial petitioner on this patent — no Unified Patents, no IPR aggregator, no repeat challenger. Unified's litigation portal indexes the underlying district-court cases but that is case tracking, not a Unified-filed IPR. (2) Celgene does get challenged aggressively at the PTAB on its other thalidomide-analog patents — it lost the '501 and '720 REMS patents to Final Written Decisions in 2016 — so the absence of any '428 IPR is a signal about petitioner strategy (they litigated Hatch-Waxman instead) rather than about patent strength. (3) The structured family data flags Federal Circuit case 21-1154, but that is recorded as a Court of Appeals for the Federal Circuit litigation entry for the family — I have not verified it as an appeal of any PTAB FWD, and I would not describe it as one. Treat it as an unverified district-court-family appeal until the docket is pulled.

Recommended next steps

1. Confirm the negative directly before relying on it. My search coverage was not exhaustive on this run. Verify with two primary sources: the PTAB E2E / Patent Trial and Appeal Board End-to-End system and USPTO Patent Center for US 8,735,428 (application 13/782,612). If both return no trial numbers, the "no PTAB activity" conclusion is solid.

2. Run a § 315(b) and expiration analysis first, before any merits work. With the '428 expired as of 2023-05-15 and the mass ANDA litigation dating to 2017, ask: (a) has your client or a privy already been served with a complaint on the '428 (one-year bar)? and (b) is there any live past-damages exposure that would justify IPR spend on an expired claim? If the answer to (b) is no, do not file an IPR — attack validity as a district-court defense, where the 2017 defendants already built the record.

3. Mine the existing district-court invalidity record rather than building from scratch. The 2017-12-17 invalidity contentions against the '262, '939, '428 and '427 (D.N.J., ECF No. 922, Ex. D per the antitrust complaints) plus the 2020-06-16 claim construction are the highest-value assets here — they were never tested by the PTAB, so they are unencumbered by estoppel and fully available.

4. Do not overstate the claim-construction point. The assertion that the method-of-treatment patents claim no efficacy (and are therefore invalid on the patentee's own concession) comes from plaintiffs' pleadings in the subsequent antitrust cases. Pull the 2020-06-16 order in D.N.J. yourself and read the actual holding before citing it in a brief.

5. If you are drafting a demand-letter response: cite the absence of any PTAB adjudication accurately — you can say the claims are untested and un-estopped, but you cannot say any claim has been canceled, because none has been. Fabricating a PTAB outcome here would be the single most damaging error available in this file.

I'll research the assignment record and litigation history for this patent before writing up. Let me run several searches.

Let me dig into the assignment records and the litigation specifics.

Assignment & Ownership Analysis — US 8,735,428 B2

Patent: Methods for treating multiple myeloma with 4-(amino)-2-(2,6-dioxo(3-piperidyl))-isoindoline-1,3-dione
App.: 13/782,612 (filed 2013-03-01) · Issued: 2014-05-27 · Earliest priority: 2002-05-17 · Term: expired 2023-05-15
Verify at: https://assignmentcenter.uspto.gov/ (search "8735428") and https://assignment.uspto.gov/patent/index.html

Inventors

• Sole named inventor. Zeldis was a Celgene officer — contemporaneous court filings describe him as "Celgene's then‑Vice President and Chief Medical Officer." He is a repeat named inventor across the Celgene IMiD portfolio (e.g., US 7,468,363; US 7,189,740). This is a captive/insider inventor pattern, the normal configuration for a brand-pharma method‑of‑use portfolio.
• No unusual departure pattern. The "all inventors gone within 12 months" distress tell is absent: Zeldis remained with Celgene for well over a decade after the 2002 priority filing (later President, Celgene Global Health). There is no inventor-side abandonment preceding a portfolio sale.
• Note: the '428 patent's claims were allowed after a 2013-10-09 declaration by Anjan Thakurta (a Celgene employee, not a named inventor). Plaintiff-side antitrust complaints (Cigna v. Celgene; Louisiana Health Service v. Celgene, 1:23‑cv‑07871 S.D.N.Y.) allege that declaration was fraudulent. That is an inequitable-conduct/validity theory, not an ownership issue — flagged only because it is the source of the later "sham litigation" narrative.

Original assignee

Celgene Corporation, a Delaware corporation, 86 Morris Avenue, Summit, NJ 07901 (at the time of recordation; earlier filings list 7 Powder Horn Drive, Warren, NJ 07059).

• Primary line of business: commercial biopharmaceutical — immunomodulatory drugs (IMiDs). Celgene marketed REVLIMID (lenalidomide) and POMALYST (pomalidomide).
• Did it ship a product embodying the claims? Yes. POMALYST was approved 2013‑02‑08 under NDA 204026 (pomalidomide capsules, 1/2/3/4 mg). US 8,735,428 is Orange Book–listed against POMALYST with use codes U‑1360 and U‑2254 and a listed expiry of 2023‑05‑15. The '428 is one of the three Celgene "MOT" (method‑of‑treatment) pomalidomide patents — the '262, '939 and '428 — that were front-and-center in the ANDA litigation.
• Current status: acquired — Bristol Myers Squibb's ~$74B acquisition of Celgene closed in November 2019. Celgene Corporation survived the merger as a wholly owned BMS subsidiary, so no assignment of the '428 patent was required or recorded to effect the deal; Google Patents still lists "Celgene Corp" as current assignee, consistent with that. Not in bankruptcy, not dissolved.

Assignment timeline

Important sourcing caveat: this analysis is built from Google Patents legal events, USPTO recordation notices reproduced in FDA citizen‑petition dockets and Federal court exhibits, the Stanford NPE Litigation Database, and IPR mandatory notices. I could not directly execute the Assignment Center query in this session, so the exact reel/frame for the '428 recordation itself was not retrieved. Everything below is either directly documented or explicitly marked as not retrieved. No reel/frame has been inferred or invented.

The chain contains one substantive assignment link — inventor → Celgene — re-recorded against each successive application in the family. Celgene's practice was to record the same 2003‑era inventor assignment on the filing date of each new continuation.

• 2003-09-03 (executed) / recorded 2013-03-01 — Reel not retrieved for 13/782,612 (see family analogues below)
  • Conveyance: Assignment of Assignors Interest ("see document for details")
  • Assignor: ZELDIS, JEROME B.
  • Assignee: CELGENE CORPORATION, 86 Morris Avenue, Summit, NJ 07901
  • Correspondent: not retrieved for this recordation — flagged as "unclear" in signal 3 below rather than guessed. For context on Celgene's recording practice, the adjacent REMS-patent family (App. 10/383,275, Reel 013982/0697, recorded 2003‑04‑23) was recorded through Woodcock Washburn LLP, Stephen C. Timmins, One Liberty Place, 46th Floor, Philadelphia, PA 19103‑7301 — a mainstream operating-company patent firm (later merged into BakerHostetler), not an NPE recording mill. That is a single adjacent data point, not proof for this reel.
  • Context: Original inventor → employer assignment. Ordinary corporate IP-capture, executed 2003‑09‑03 as part of the 2002 pomalidomide filing program; re-recorded per-application (family analogues below).
• No post-issuance assignment. There is no recorded transfer to any IP/holdings/licensing LLC, no security interest recorded against the patent, no merger deed of assignment, and no defensive-aggregator acquisition.

Family recordation analogues (same Zeldis→Celgene instrument, different reel/frame per application) — useful for confirming the pattern but these reels do not apply to the '428:

If Assignment Center returns records for "8735428," the expected single entry is the Zeldis→Celgene assignment of assignor's interest recorded 2013‑03‑01. There is no downstream chain to reconstruct.

Timeline diagram

timeline
    title Ownership of US 8735428
    2002 : Priority application filed
    2003 : Zeldis assigns rights to Celgene
    2013 : Continuation filed as 13/782612
    2014 : Patent 8735428 issues
    2017 : Celgene sues generic ANDA filers
    2019 : Celgene acquired by Bristol Myers Squibb
    2023 : Patent term expires

NPE / troll-pattern signals

1. Shell-entity transfer — not present. The sole documented link is Zeldis → Celgene Corporation, a large operating pharma with a physical R&D/manufacturing footprint and an approved NDA. There is no transfer to any entity bearing an "IP / Holdings / Licensing / Ventures" suffix, no registered-agent-only address, and no single-purpose LLC anywhere in the record. No such reel exists to cite, because no such transfer occurred.

2. Known asserter in the chain — not present. The assignee is Celgene (now BMS) — an operating company, on no public NPE list (Acacia, Marathon, IV, IPNav, Wi‑LAN, Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Spangenberg entities, etc.). Caution on a false positive: the Stanford NPE Litigation Database returns three hits for this patent — CELGENE CORPORATION v. HETERO LABS LIMITED, 2:17‑cv‑03387 and 2:20‑cv‑02601 (D.N.J.), and CELGENE CORPORATION v. PAR PHARMACEUTICAL, 2:17‑cv‑03159 (D.N.J.). Those are indexed because the database crawls broad patent-litigation dockets, not because the plaintiff is an NPE. In all three, Celgene is the plaintiff asserting against generic ANDA filers — brand-vs-generic Hatch-Waxman litigation, the textbook operating-company posture.

3. Repeat correspondent across the chain — unclear. There is only one substantive link, so "recurrence" is structurally impossible to observe here. I could not retrieve the correspondent of record on the '428 recordation, and I will not name one. The only correspondent I can document in this patent's immediate neighborhood is Woodcock Washburn LLP / Stephen C. Timmins on Reel 013982/0697 (a different Celgene family) — an operating-company firm, and a single appearance is not a finding under your own criterion.

4. Cascading transfers — not present. One link only; no chained LLCs, no shared correspondent address across successive assignees, nothing within any 24-month window.

5. Pre-litigation transfer — not present. The assignment was executed 2003‑09‑03 and recorded 2013‑03‑01; the first infringement suits over the '428 were filed in May 2017 (Celgene v. Par, 2:17‑cv‑03159, filed 2017‑05‑04; Celgene v. Hetero, 2:17‑cv‑03387, filed 2017‑05‑11). The gap is ~4 years from recordation and ~14 years from execution — the opposite of a venue/standing-arranging transfer.

6. Bankruptcy fire-sale — not present. Celgene never filed Chapter 7/11. It was acquired at a massive premium by Bristol Myers Squibb in November 2019, at which point POMALYST was generating roughly $2B+/yr. No §363 sale, no distressed patent auction.

7. Privateering — not present. Celgene is the asserting party against competitors' generic filers. There is no transfer to an NPE asserting on Celgene's behalf, and no EFF/Patent Progress "privateering" coverage. (Celgene's conduct was attacked on other theories — sham litigation, reverse payment, inequitable conduct — but none of those is a privateering transfer.)

8. Defensive aggregator — not present. No RPX, AST, LOT, Unified Patents, or OIN acquisition; the chain terminates at Celgene/BMS, which continues to hold the expired patent.

Additional non-signal worth recording: the '428 was filed 2013‑10‑08 with terminal disclaimers as to the '262 patent, and it expired naturally on 2023‑05‑15. Generic pomalidomide entry occurred via litigation settlements (e.g., the Apotex consent judgment of 2021‑04‑20 listing the '428 among the Patents-in-Suit), not via any ownership transfer.

Verdict

Operating-company assertion.

The chain is a single inventor→employer assignment (Zeldis to Celgene Corporation, executed 2003‑09‑03, recorded against this application 2013‑03‑01), with no downstream transfer of any kind — no LLC, no asserter, no aggregator. Celgene is a commercial biopharma that ships a product embodying the claims (POMALYST, NDA 204026, pomalidomide; '428 listed in the Orange Book with use codes U‑1360/U‑2254) and asserted the patent itself in Celgene v. Par Pharmaceutical, 2:17‑cv‑03159 and Celgene v. Hetero Labs, 2:17‑cv‑03387 (both D.N.J., filed May 2017) against actual generic competitors, before being acquired by Bristol Myers Squibb in November 2019. The three Stanford "NPE database" hits for this patent are plaintiff-side Celgene dockets — a database artifact, not evidence of NPE status.

Confidence note: the verdict rests on well-documented facts (identity of plaintiff/assignee, product, litigation). The one genuinely open item is the correspondent and reel/frame on the 2013‑03‑01 recordation, which I could not retrieve and have deliberately left blank rather than estimated. A direct Assignment Center pull by patent number should close that gap in a single query.

I'll start by pulling the patent's own "References Cited" record and verifying the cited patents one at a time.

Let me try to retrieve the front-page "References Cited" list for US 8,735,428 directly.

I found a substantial portion of the '428 front page in a court filing. Let me retrieve the full "References Cited" list.

Scope note before the analysis

I could not query a USPTO database directly with the tools available — I have web search only. What follows is grounded in (a) the authoritative full text you supplied for US 8,735,428 (which includes the specification-body citations but is truncated before the claims and before the front page), and (b) the actual front page of US 8,735,428 B2 as reproduced in D.N.J. litigation exhibits (the Hetero case docket, Celgene Corp. v. Hetero Labs Ltd., No. 2:17-cv-03387), plus prosecution-history material quoted in the Cigna/Hagens Berman complaints and the Celgene v. Hetero briefing. I did not retrieve a direct USPTO Patent Center / Open Data Portal page in this session.

One hard limit, stated up front: I recovered only a partial front-page "References Cited" list. The excerpt I captured begins mid-list at 6,335,349 B1 (implicitly preceded by earlier entries) and ends mid-entry at 2003/0144325 A1. I flag every place where the list is visibly incomplete. I did not obtain the verbatim claim set; claim references below are drawn from litigation quotes and are labeled as such.

1. Confirmed bibliographic anchors (updates to earlier sections)

Governing law note: The '428 was filed March 1, 2013 — before the March 16, 2013 AIA cutoff — and claims 2002 priority. Pre-AIA § 102/§ 103 governs, so the relevant anticipation categories are § 102(a), (b), and (e). This matters for how the cited references are analyzed.

2. The patent's own citation record — what was cited, and where

The '428's citation record has three distinct layers, and conflating them is the single most common analytical error on this patent:

Critical distinction for § 102 purposes: the inequitable-conduct allegations in S.D.N.Y. (Cigna 1:25-cv-05237; Hagens Berman) assert that D'Amato (2001), Lentzsch (2001), Lentzsch (2002), Schey (April 2002), and Schey (October 2002) were NOT cited by the examiner and that their import was concealed. Whether or not those allegations are credited, they are — not the front-page list — the references with real § 102 exposure. I separate them accordingly in § 3.2 below.

3. Reference-by-reference § 102 analysis

3.1 Front-page U.S. Patent Documents (captured portion)

Dates are the printed front-page dates. "§102 bucket" reflects the analysis under a May 17, 2002 / Nov. 6, 2002 priority.

(The list is visibly truncated at the front: earlier entries — almost certainly including 5,635,517; 5,874,448; 5,929,117; 6,045,501; 6,281,230; 6,316,471 — were not captured. Those are analyzed in § 3.3 below from the authoritative specification text, which cites them expressly.)

3.2 Front-page U.S. Published Applications (captured portion)

Net: every entry in the captured application list is dated on or after 2001 and mostly on or after the May 2002 priority, and none discloses the four-element combination. None anticipates. Most were cited for § 103 or as background.

3.3 Specification-body citations (from the authoritative full text) — and why they are not § 102 art

These were cited by the applicant in the specification, not by the examiner as prior art:

• US 6,281,230 (Muller et al.) and US 6,316,471 (Muller et al.) — the IMiD compound/use patents. The '471 (issued 11/13/2001) teaches pomalidomide's use for cancers and autoimmune disease, oral capsules/tablets of 1 to 100 mg, and combination with steroids including dexamethasone (per Cigna ¶134). This is the best patent-side § 103 reference against the dexamethasone-combination claims, but it does not anticipate: a disclosed 1–100 mg range does not, without more, describe the claimed "about 1 to about 4 mg per day" sub-range, and it contains no MM-specific dosing cycle.
• US 5,635,517 (Muller et al.) — 1-oxo/1,3-dioxo isoindolines amino-substituted on the benzo ring; discloses pomalidomide and TNF-α reduction (the "primary reference" in the '262 prosecution). No anticipation of the '428's MM claims; it reaches cancer only via the TNF-α mechanism.
• US 5,874,448 (3-fluoropiperidinyl isoindolines) and US 5,929,117 (cyano/carboxy substituted styrenes) — disclosed as alternative IMiD genera. No pomalidomide-MM disclosure.
• Boilerplate formulation/delivery citations — 3,845,770; 3,916,899; 3,536,809; 3,598,123; 4,008,719; 5,674,533; 5,059,595; 5,591,767; 5,120,548; 5,073,543; 5,639,476; 5,354,556; 5,733,566 (controlled release); 5,134,127 (cyclodextrins); 5,391,485 / 5,393,870 / 5,229,496 (GM-CSF); 4,810,643 / 4,999,291 / 5,528,823 / 5,580,755 (G-CSF). Zero § 102 relevance to the method claims.
• US Ser. No. 09/972,487 (filed Oct. 5, 2001) and US Prov. 60/372,348 (filed Apr. 12, 2002, Hariri et al., stem cells) — cited as co-pending/related work; the Oct. 5, 2001 application is § 102(e)-capable in principle but discloses isoindole-imide chemistry, not the MM dosing method.

3.4 The references with real § 102 exposure (third-party, MM-specific)

These are the most relevant prior art to the '428 in substance, and the references that the S.D.N.Y. plaintiffs allege were not cited by the examiner:

4. Ranked "most relevant prior art" for US 8,735,428

5. Bottom-line § 102 conclusions

1. No U.S. patent or published application cited on the '428 front page anticipates any of the 27 claims. All are compound-genus, anti-angiogenic, formulation-boilerplate, or the patentee's own family art. The Zeldis-family patents cannot even qualify as § 102(e) art because they are not "by another."
2. The only genuine § 102 candidates are the 2001–2002 third-party journal publications — principally Schey (April 2002) and D'Amato (2001).
3. The anticipation question turns entirely on claim scope, which I could not verify verbatim. If the independent claim requires the 21-consecutive-days-then-7-days-rest cycle, Schey (April 2002) does not literally anticipate (though it is severe § 103 art). If a claim requires only pomalidomide + MM (+ prior therapy), Schey and D'Amato are strong § 102(b) references.
4. The May 17, 2002 vs. Nov. 6, 2002 priority question is dispositive for Schey (October 2002) and is therefore a first-order issue, not a footnote.

6. Corrections to earlier sections of this analysis

• Claim count: previously flagged as "unconfirmed." The front page shows 27 claims, 1 drawing sheet. Resolved.
• 21/7 cycle limitation: the earlier summary said the Dex/cycling language "appears tied to the '262 reexamination, not the '428." That should be corrected. The Cigna complaint ¶186 and the HBS brief both attribute to application 13/782,612 (→ the '428) claims to "21 consecutive days followed by seven consecutive days of rest in a 28 day cycle," and defendant Hetero counterclaimed on claims 12–13, 15, 25–26. The cycle limitation does appear in the '428's claims; the parallel '3939 instead recites the generic "period of time followed by a period of rest."
• Dosage range: the earlier "about 1 mg to about 5 mg per day" reconstruction is not corroborated; the litigation quotes consistently read 1 to 4 mg/day. Treat 1–4 mg/day as the better-supported figure.
• Independent-claim count: the earlier sections said "at least two" and then "one." The Cigna complaint describes an independent claim plus a dependent dexamethasone claim, which is consistent with a single independent claim; I still cannot confirm this from the patent itself.

7. Explicit uncertainty flags

1. Front-page citation list is incomplete — retrieved from 6,335,349 B1 through 2003/0144325 A1 only. Earlier U.S. patents (probable: 5,635,517; 5,874,448; 5,929,117; 6,045,501; 6,281,230; 6,316,471) and all foreign patent documents and the full "Other Publications" list were not captured.
2. Titles/descriptions marked "[not verified]"? I declined to supply them from memory. Pull the front page or Patent Center to confirm.
3. Claim text remains unverified verbatim. Claim numbers 12–13, 15, 25–26 are taken from Hetero's counterclaim; claim 1's preamble from the Cigna/HBS briefing.
4. Kyle (2001) full citation (venue, volume, page) was not verified this session; only "page 584" is quoted in the record.
5. All inequitable-conduct / concealment assertions (including that D'Amato 2001, Lentzsch, and Schey were withheld) are pending-litigation allegations, not adjudicated findings. I have no information that the '428 has been held invalid or unenforceable.
6. No foreign counterpart citations were retrieved; if you need the Art. 54(3)/EP equivalents, that is a separate pull.

Would you like me to try again to retrieve the complete front-page citation list (including foreign documents and the "Other Publications" block) so each entry can be individually analyzed? That is the one gap that materially limits the completeness of § 3.1–3.2 above.

I'll ground the prior-art references before building the § 103 analysis.

Let me pull the remaining prior-art references and attempt to locate verbatim claim text.

Obviousness Analysis — US 8,735,428 B2 under 35 U.S.C. § 103 (pre-AIA)

Analyst note on scope. You asked me to work from the Prior Art section of this page. Before analyzing, I must be explicit about what that section actually contains here, because it is unusually thin and that fact matters to the analysis.

0. What the "Prior Art" section of this page actually contains

The supplied Google Patents record for US 8,735,428 exposes only four prior-art data points in its structured panel:

The page's citation tables ("Patent Citations" / "Non-Patent Citations") were not captured in the text supplied to me — the record jumps from the classification tree to the Definitions section. I therefore cannot inventory the face-of-the-patent references from this page alone, and I am not going to pretend otherwise.

Two things follow, and both are analytically significant:

1. The keyword set is itself evidence about how the invention was categorized. The five prior-art keywords are administered, patients, compound, cancer, day. Conspicuously absent are "multiple myeloma," "amino," "isoindoline," "piperidyl," "TNF," "dexamethasone," "relapsed," or "refractory." The examiner's art-classification signal is that this is an administration/dosing-and-scheduling invention in a generic cancer patient population — not a new-compound invention and not a new-mechanism invention. That framing is the single most important input to the § 103 analysis below.

2. The prior-art date is 2002-05-17, which matches the priority date Google displays. As my previously generated sections flagged, this is contradicted by the litigation record, which treats November 6, 2002 as the operative priority date, and by the 20-year term, which runs from the May 15, 2003 parent filing (US 10/438,213). I carry that discrepancy forward because it changes which references qualify. See § 6.

For the reference set, I therefore rely on (a) the two structured fields above, (b) the references Celgene itself identified as the IMiD art in the '428 specification, and (c) the references the examiner actually applied and Celgene actually distinguished during prosecution — all of which are documented in the file wrapper materials quoted in the litigation record, most directly in Celgene's own appeal/response brief (https://storage.courtlistener.com/recap/gov.uscourts.njd.[348812](/patent/348812)/gov.uscourts.njd.348812.250.4.pdf).

1. The limitations a § 103 case must meet

⚠️ Contradiction flag — the prior sections of this analysis disagree on claim 1. I must surface this rather than paper over it:

These cannot both be right, and I cannot resolve them — the supplied full text is truncated mid-specification and contains no claims. Per the operating rule on literal interpretation, I do not auto-resolve in favor of either. Instead I analyze each limitation independently, so the analysis holds under either reading. The four candidate limitations are:

• (A) the compound pomalidomide (4-amino-2-(2,6-dioxo(3-piperidyl))isoindoline-1,3-dione, "ACTIMID," CC-4047);
• (B) the patient population — MM patients previously treated (relapsed/refractory; prior thalidomide, lenalidomide, or proteasome inhibitor);
• (C) the dose — about 1–5 mg/day (a 1–4 mg/day variant also appears in family filings);
• (D) the schedule — a cyclic 21-day/7-day (or 4-week) regimen, and/or co-administration with dexamethasone.

Any § 103 combination must supply every limitation that actually appears in the claim as issued. Because (D) is contested, I build the case in layers: A → A+B → A+B+C → A+B+C+D.

2. Reference-by-reference disclosure map

Important § 103(c) note. The '517, '230, '471 and '554 are commonly owned Celgene patents. Pre-AIA § 103(c) disqualifies common ownership art only where it qualifies "only under § 102(e), (f) or (g)." The '517 issued July 1, 1997 and the '554 issued April 29, 2003 — both are also § 102(b) printed publications relative to a 2002/2003 effective filing date. So § 103(c) does not remove them, and they remain available as § 103 art. Only the co-pending application art would be at risk. This is a point on which a careless analysis would err.

3. The § 103 case — four layered combinations

Combination α — the examiner's actual rejection (the baseline)

US 5,635,517 (primary) + Muller 1999 + Corral 1999 + Kyle 2001 + Davies 2001 + US 6,555,554

This is not hypothetical; it is what the PTO did. Celgene's own brief recites the rejection "under § 103(a) as being unpatentable over Kyle, in view of Davies, Corral, Muller and U.S. [6,555,554]." I reproduce it because the fact of the rejection fixes the starting position and identifies the limitations the art was held to supply versus those Celgene was forced to argue.

• '517 + Muller 1999 → the compound. Muller expressly discloses "4-amino substituted analogs of thalidomide" as potent TNF-α inhibitors; the '517 (per the pleadings) claims the TNF-α-reduction method wherein the compound is pomalidomide. Together they place pomalidomide squarely in the art, structurally and functionally.
• Corral 1999 → the amino-substituted analogue class and the express clinic-use motivation.
• Hideshima 2000 / Davies 2001 → IMiDs act on MM, including drug-resistant MM, and in patient MM cells.
• Kyle 2001 → MM is a thalidomide-responsive disease; thalidomide/Dex combination used clinically; dose titration to response and toxicity is routine practice (the examiner relied on Kyle p. 584 for exactly this: "obvious to adjust the amount of the drug depending on its efficacy and side effects").
• '554 → the closely related 4-aminoisoindoline analogue, i.e., the genus in which the POSA would work.

Why a POSHITA would combine them. All references are in one field (IMiD/thalidomide analogues for hematologic malignancy, principally MM). The '517 and Muller references come from the same research group (Muller/Celgene) that generated the compounds, and Corral is the in-house immunology characterization of them — KSR Int'l v. Teleflex "familiar elements arranged according to known methods," and an express "try the known class members in the known indication"} rationale. The motivation is not mere hindsight: Corral expressly points to clinic use, and Hideshima 2000 expressly points to drug resistance, which is the prior-therapy limitation.

What defeated it — and why that matters. Celgene's counter was that the references did not teach pomalidomide by name: the '517 "did not expressly teach ACTIMID," and Davies "does not teach ACTIMID." Per Celgene's brief, the argument was that without teaching the specific compound, the combination "does not direct the skilled person to use the recited compound." Two observations:

1. That is a species-disclosure argument, not an absence-of-motivation argument. It is a weaker position than a true teaching-away; under In re Gorman / Merck v. Biocraft, a single species within a disclosed genus may be obvious where the art gives structural guidance — and Muller 1999 arguably does more than that by naming the 4-amino substitution.
2. The argument's factual premise is directly challenged in the antitrust complaints, which allege that the '517 did disclose and claim pomalidomide, that Davies did study pomalidomide (as one of "IMiD1/2/3," left unnamed by chemical structure), and that D'Amato (2001) did teach pomalidomide for MM — and that these facts were affirmatively misrepresented. (https://www.courtlistener.com/docket/69159985/1/centerwell-pharmacy-incv-celgene-corporation/; https://business.cch.com/ald/TheCignaGroupvCelgeneCorporationComplaint.pdf)

⚠️ Those are allegations in pending civil actions, not adjudicated findings. I have no information that the '428 has been held invalid or unenforceable, and the examiner's rejections were overcome.

Combination β — compound + refractory population (limitations A + B)

Muller 1999 (or '517) + Hideshima 2000 + Lentzsch 2001 + Davies 2001 + Kyle 2001

This is the strongest combination on paper and it is stronger than α because it does not depend on the '517's disclosure of pomalidomide by name.

Motivation. The chain is compact and non-hindsight: (i) thalidomide is clinically active in relapsed/refractory MM (Kyle 2001; Singhal 1999, cited in the patent's own Lentzsch-recited reference list); (ii) thalidomide's analogues are more potent and act directly on MM cells and on drug-resistant MM (Hideshima 2000; Corral 1999; Lentzsch 2000); (iii) Muller 1999 identifies the 4-amino-substituted subgenus as the potent species; (iv) therefore administer the 4-amino analogue to the refractory MM patient. The expected benefit — a more potent, orally available thalidomide analogue with direct anti-MM activity — is the ordinary expectation of medicinal chemists, not an unexpected one. This is "obvious to try" in the KSR sense (a finite, identified class of analogues; predictable direction of modification).

Combination γ — adding the dose limitation (A + B + C)

Combination β + Schey Apr. 2002 + US 6,316,471 + Kyle 2001 (dose titration)

If claim 1 (or a separate independent claim) recites "about 1 mg to about 5 mg per day," this combination meets it head-on:

• Schey (Apr. 2002) discloses CC-4047 (= pomalidomide) orally in relapsed/refractory MM at cohorts of 1, 2, 5 and 10 mg/day, establishing MTD = 5 mg/day — i.e., the entire recited range, disclosed as the working clinical range, in the recited patient population.
• US 6,316,471 (per pleadings) teaches oral pomalidomide at 1–100 mg per unit dose.
• Kyle 2001 supplies the ordinary clinical practice of titrating to efficacy/toxicity — the rationale for arriving at any particular sub-range within the disclosed 1–10 mg spread.

Motivation. Dose selection is the paradigmatic In re Aller / In re Boesch situation: where the prior art discloses a range and the claimed range is a routine optimization within it, the burden shifts to the applicant to show criticality — that the claimed sub-range produces a different result in kind, not degree. Schey's MTD of 5 mg/day is the pre-claimed anchor. Nothing in the reference set points to the 1 mg/day end as uniquely efficacious in a way not already captured by "MTD-driven dose selection."

Combination δ — the "D'Amato" combination (A + B, via a different primary)

D'Amato (2001) + the D'Amato/Children's Hospital patents + Lentzsch 2001 + Schey Apr. 2002

If the plaintiffs' pleading is correct that D'Amato (2001) taught pomalidomide to treat multiple myeloma and that a D'Amato patent claims pomalidomide for MM, then the compound-plus-indication limitation is met by a single reference, and the § 103 inquiry collapses to the dose/schedule limitations alone. That is the strongest version of the invalidity case, and it is also the least verifiable from the material I can access — I have only the pleadings' characterization of D'Amato (2001), not the reference itself.

Combination ε — adding schedule/dexamethasone (A + B + C + D)

γ + Kyle 2001 (Thal/Dex) + US 6,316,471 (IMiD + dexamethasone) + Schey Apr. 2002 (4-week oral cycle)

If the PTAB-section description of claim 1 is correct (21-on/7-off + Dex dependent claim):

• Dexamethasone co-administration: Kyle 2001 reports thalidomide + dexamethasone dosing schedules in detail (40 mg/d on days 1–4, 9–12, 17–20 in odd cycles; 20/26 response); the '471 (per pleadings) teaches IMiD + steroids including dexamethasone. In oncology the combination of an anti-MM agent with dexamethasone was standard; treating the combination as inventive would require evidence that the IMiD/Dex pairing produced an unexpected result.
• Cyclic 21/7 or 4-week schedule: Schey (Apr. 2002) already administers pomalidomide orally for 4 weeks in relapsed/refractory MM; the patent's own specification recites the 4-to-6-week cycle with 1–2 weeks' rest as the invention's convention (spec: "administered daily in a single or divided doses in a four to six week cycle with a rest period of about a week or two weeks"). Where the specification itself frames the cycle as conventional, the cycle limitation is a routine scheduling choice and a weak § 103 bulwark.

4. Motivation to combine — consolidated

Reduced to the KSR factors:

1. Same field of endeavor. Every reference is thalidomide-analogue ("IMiD") immunomodulation in hematologic malignancy, predominantly MM. No cross-field leap is required.
2. Same problem, known solution class. Kyle/Davies/Hideshima pose the problem as "thalidomide works in MM but is toxic and resistance emerges." Corral/Muller/Hideshima answer it: more potent amino-substituted analogues that act directly on resistant MM cells.
3. Express pointers in the references. Corral 1999 expressly concludes the analogues are candidates for clinical use. Hideshima 2000 expressly says the analogs overcome drug resistance. These are not implicit; they are the references' own stated conclusions.
4. Predictable structural modification. The relevant modification (4-amino substitution) was already made and reported in Muller 1999, with potency data. Selecting it is not "obvious to try" over an unexplored field — it is the known best member of a disclosed subgenus.
5. Routine optimization for dose/schedule. Kyle 2001 (titrate to response and toxicity), Schey Apr. 2002 (dose-escalation cohorts, MTD), the '471 (1–100 mg unit dose), and the '428's own spec (4–6-week cycles) all frame the remaining limitations as ordinary clinical development.
6. Reasonable expectation of success. In vivo anti-MM activity in a drug-resistant model (Lentzsch 2001) plus human MM cell killing (Davies 2001) supply the expectation of success that KSR requires — no unpredictable mechanism was being bet on.

5. Where the § 103 case is genuinely contestable

An honest analysis must give the patentee its best ground. The strengths are not in the compound but in the clinical limitations:

1. Unexpected results / non-cross-resistance. Celgene's October 4, 2013 examiner interview argued that pomalidomide unexpectedly treats myeloma resistant to lenalidomide despite the structural similarity of the two compounds — and the Thakurta Declaration (Oct. 9, 2013) was submitted to evidence it. Structurally, lenalidomide (4-amino-1-oxo-isoindoline, i.e. the '554 species) and pomalidomide (4-amino-1,3-dioxo-isoindoline) are near-identical; the patentee's best § 103 argument is that the near-identical structure produced a different resistance profile. This is a § 103 "unexpected results" argument, and it must be met with evidence (comparative data) rather than assertion. The plaintiffs directly attack this showing as false and inequitably procured (citing Lentzsch 2001/2002, Schey Apr./Oct. 2002, D'Amato 2001 as art that rendered the result expected) — again, allegations, not findings.
2. Criticality of the low dose. In re Aller cuts both ways: if the patentee can show the 1–5 mg/day range gives a result different in kind from the '471's 1–100 mg disclosure (e.g., efficacy preserved with materially reduced myelosuppression/toxicity, which is the real-world Pomalyst dose profile), the dose limitation could survive even though the range numerically overlaps Schey's 1/2/5/10 mg cohorts. Numeric overlap alone does not defeat criticality.
3. Species-selection hurdle (In re Baird / In re Jones). If the primary reference is a genus that does not name pomalidomide — as Celgene successfully argued for the '517 and Davies — then the combination α fails, and the case must be re-founded on Muller 1999's express "4-amino" disclosure or on D'Amato. The antitrust complaints assert the '517 does name pomalidomide, but that is the contested factual core.
4. Teaching-away argument (weak). Celgene tried to convert Kyle's own toxicity data into a teaching-away: the brief notes Kyle p. 586 records "significant toxicities" for thalidomide/dexamethasone. This is a poor teaching-away argument — toxicity constrains dosing, it does not direct the POSA away from the analogue class. Indeed Kyle's toxicity data is the motivation to use a more potent analogue at lower dose.
5. The unrebuttable weakness. Note that the patent's own specification contains the POSA's motivation. It recites the IMiD class (Muller '230/'471), recites that IMiDs are "potent co-stimulators of T cells," recites the TNF-α IC₅₀ data, recites Actimid/Revimid dosing, and recites the 4–6-week cycling convention. A specification that supplies the class, the mechanism, the analogues and the dosing conventions is an unhelpful place to look for an inventive step.

6. § 102/§ 103 date qualification (the gating issue)

The '428 was filed March 1, 2013 — before the AIA's March 16, 2013 first-to-file date — and claims § 120 benefit through 10/438,213 (filed May 15, 2003). Pre-AIA § 102/§ 103 therefore applies.

This is the crux of the priority-date discrepancy I flagged earlier, and it is outcome-determinative. The dose limitation (C) is supported principally by Schey (Apr. 2002) — the single reference that discloses the whole 1–5 mg/day range in relapsed/refractory MM. If the operative priority is May 17, 2002, or if the patentee can antedate an April 2002 publication under Rule 131, then limitation (C) is left supported only by Kyle 2001's generic titration teaching and the '471's 1–100 mg unit-dose disclosure — neither of which discloses the claimed sub-range in the claimed population. Limitation (C) is therefore the most defensible limitation in the patent, and the whole § 103 case may turn on whether Schey (Apr. 2002) and Lentzsch (Apr. 2002) are properly § 103 art.

Conversely, under the litigation's November 6, 2002 operative date, Schey (Apr. 2002) and Lentzsch (Apr. 2002) both qualify as § 102(a) art (subject only to a Rule 131 swearing-behind, which requires the applicant to prove an earlier completion of the invention — a factual burden), and the case for (C) is materially stronger.

7. Bottom line

Combination most likely to render the claims obvious — Combination β + γ:

Muller 1999 (4-amino thalidomide analogues, potent TNF-α inhibitors) + Hideshima 2000 (thalidomide analogs overcome drug resistance in MM) + Davies 2001 (IMiDs act on patient MM cells and augment anti-MM NK cytotoxicity) + Lentzsch 2001 (S-3APG = pomalidomide; in vivo anti-MM activity in drug-resistant MM) + Kyle 2001 (thalidomide/Dex in MM; routine dose titration) + Schey Apr. 2002 (pomalidomide, 1/2/5/10 mg/day, MTD 5 mg/day, relapsed/refractory MM)

This combination supplies all four candidate limitations and rests on express reference teachings at each step rather than on hindsight. If a cycle/dexamethasone limitation is in the claims, US 6,316,471 (pomalidomide + dexamethasone; 1–100 mg unit dose) and Schey Apr. 2002 (4-week oral dosing) supply them as routine scheduling choices.

My confidence levels, stated honestly:

Two limitations on this analysis you should carry forward:

1. I could not read the '428's claims. The supplied text is truncated mid-specification. Everything above is conditioned on the reconstructed limitations in § 1, which the prior sections of this analysis describe inconsistently. Before relying on this in any filing, pull the issued claims from USPTO Patent Center (application 13/782,612).
2. The prior-art panel on this page was incomplete. The structured data furnished only the five keywords and the 2002-05-17 prior-art date; the citation tables were not captured. I have reconstructed the reference set from the patent family front pages, the file-wrapper materials quoted in litigation, and the references themselves — and I have labelled which propositions are drawn from pleadings rather than from adjudicated findings or from the references directly.

Primary sources: https://patents.google.com/patent/US8735428/en · https://pubmed.ncbi.nlm.nih.gov/11740813/ (Kyle) · https://pubmed.ncbi.nlm.nih.gov/11418482/ (Davies) · https://pubmed.ncbi.nlm.nih.gov/10384139/ (Corral) · https://pubmed.ncbi.nlm.nih.gov/11049970/ (Hideshima) · https://storage.courtlistener.com/recap/gov.uscourts.njd.348812/gov.uscourts.njd.348812.250.4.pdf · https://www.courtlistener.com/docket/69159985/1/centerwell-pharmacy-incv-celgene-corporation/ · https://business.cch.com/ald/TheCignaGroupvCelgeneCorporationComplaint.pdf · https://www.drugpatentwatch.com/p/patent/6555554