US 6932842

I'll search for this specific patent number across USPTO and court sources.

Let me check for any litigation or CAFC activity specific to this patent.

US 6,932,842 — Verification Summary

Search scope & confidence note. I ran targeted queries against Google Patents (authoritative full text supplied), uspto.report, FREEPATENTSONLINE, Justia, and general Federal Circuit/docket material. I found no CAFC 2026 docket, appeal, or civil action in which US 6,932,842 is an asserted patent. The number appears in litigation files only passively — as one entry in an examiner's EAST prior-art search history reproduced in Zimmer, Inc. et al. v. Exactech, Inc., No. 8:21-cv-00995 (M.D. Fla.), Doc. 34-9, and as a "cited by" reference on later patents (e.g., ConforMIS, DePuy Synthes). Treat the litigation-absence finding as high-confidence for the sources searched, but I cannot claim exhaustive coverage of CourtListener/PACER.

Bibliographic data

Note on the assignment record: the Google Patents reassignment entries list additional assignors (Thomas Koerbs, Sebastian Nagel, Ralf Schied) alongside the five named inventors. These were assignors in later nunc pro tunc/confirmatory assignments, not inventors on the face of the patent.

Abstract (as issued)

A virtual three-dimensional model of the patient, formed from recorded (two-dimensional) image data, is compared with real medical reference data — e.g., a database of test-person data — to select or form a reference model object best suited to the patient, from which a virtual implant model is generated. Computer numeric control (CNC) data are then generated directly from the virtual implant model for program-assisted manufacture of the implant, dispensing with intermediate models.

Plain-language overview of the independent claims

Both independent claims recite the same end-to-end workflow; they differ only in how the reference-model selection step is worded.

Claim 1 — A method of manufacturing a patient-specific implant comprising:

1. Obtaining 2-D medical image data of the defect area and its surroundings via CT or NMR tomography (a closed Markush group);
2. Using a mathematical image-processing algorithm to form a surface from that data, and segmenting to detect bones/hard-tissue ranges;
3. Generating a virtual 3-D model of at least the defect area and environment;
4. Comparing that virtual model to real medical reference data;
5. Selecting the best-suited set of reference data and forming a 3-D reference model object from it, in two sub-steps — first selecting a plurality of candidate reference sets and forming a corresponding plurality of 3-D reference model objects using mathematical, functional, medical and aesthetic parameters, then selecting the single best one;
6. Generating a virtual implant model by superimposing the selected reference model object with the patient's virtual 3-D model; and
7. Manufacturing the implant by computer numeric control based on the virtual implant model data.

Claim 6 — Identical to claim 1 except for step 5, which is phrased as selecting a plurality of three-dimensional reference model objects similar to the virtual three-dimensional model (rather than selecting sets of reference data and forming objects from them). Functionally the same two-stage winnowing; the difference reads as a drafting/claim-capture variation rather than a different inventive concept.

Dependents:

• Claims 2 & 7: the reference data comprise a database.
• Claims 3 & 8: the reference data comprise data from the patient (this is where the specification's mirror-symmetry/inversion embodiment — reconstructing a defect in one body half from the undamaged side about the plane of symmetry — is brought into the claim set).
• Claims 4 & 9: the virtual implant model is a three-dimensional model.
• Claim 5 & 10: the final selection is made in consideration of an expert medical opinion.

Practical significance: the claims are directed to the manufacturing workflow, not merely a design/simulation method — the CNC fabrication step is a required element of both independent claims. This is consistent with the specification's stated advantage that no physical intermediate or test model is needed, and that the implant may be any suitable material including ceramics and titanium.

Claimed priority family (for context)

DE 19922279 A1 (1999-05-11); WO 2000068749 A1; EP 1208410 B1; CA 2373691 C; JP 2002543860 A; CN 1350667 A; AU 5064800 A; AT 285595 T.

Key cited prior art on the face

US 4,976,737 (Bone reconstruction); US 5,370,692 (Rapid, customized bone prosthesis); US 5,452,407 (Representing a patient's treatment site as CAD/CAM data); US 5,741,215 (Stereolithographic anatomical modelling); US 5,798,924 (Eufinger, Process for producing endoprostheses); US 6,463,351 (Clynch, Custom fitted medical devices). Non-patent: D'Urso et al., J. Cranio-Maxillofacial Surgery 27 (1999); Eufinger et al., Spektrum der Wissenschaft (Feb. 1999) and Deutsches Ärzteblatt (Sept. 1997).

Uncertainties

1. No CAFC 2026 activity located. I cannot rule out a very recent docket entry not yet indexed by the sources reached. Given the patent expired for fee non-payment on 2020-05-10, assertion in 2026 is unlikely.
2. Assignee current-owner accuracy. Google Patents expressly disclaims verification of the assignee listing; the "Current Assignee" field is populated from assignment records, not a chain-of-title opinion.
3. Reassignment assignors. The Koerbs/Nagel/Schied entries are as recorded in the assignment abstracts; I did not retrieve the underlying assignment documents to confirm their exact role.

I'll search for litigation involving US patent 6932842 across patent litigation databases.

Let me search more targeted litigation databases.

Litigation Search Results — U.S. Patent No. 6,932,842

Bottom line: I found no evidence of any litigation involving U.S. Patent No. 6,932,842 as an asserted patent. Based on the searches I was able to run against patent-litigation-oriented sources (Unified Patents portal, CourtListener/CAFC opinion databases, Docket Alarm OCR of litigation dockets, and general web search), no case lists '842 as a patent-in-suit.

What the searches actually surfaced (and why it is not litigation against the '842 patent)

Every hit for the number "6932842" fell into one of two categories, neither of which is a lawsuit over this patent:

1. The '842 patent cited as prior art / a reference in other patents. The number appears in the reference lists of numerous later patents (e.g., Conformis patents such as US 9,055,953 and US 9,235,679; and various Justia / FreePatentsOnline "cited references" pages). Appearing as a citation in a third party's patent is not litigation.

2. The '842 patent listed inside an Information Disclosure Statement (IDS) filed in an unrelated case. The clearest hit was in Zimmer, Inc. et al. v. Exactech, Inc., Case No. 8:21-cv-00995-MSS-TGW (M.D. Fla.) — the '842 number appears in an IDS/reference table (Document 34-9) submitted during prosecution, not as an asserted patent. Likewise it appears in a PTAB petition document as a search-string/reference citation. In both instances the '842 patent is merely a piece of paper cited to the examiner or the Board; it is not the patent being enforced or challenged.

Case-by-case table

I could not substantiate any case in which U.S. 6,932,842 was asserted (infringement suit) or challenged (IPR/PGR/CBM or DJ action).

Important caveats and honesty about limits

• I hit the tool-call limit during the search and could not complete the full set of queries I intended (e.g., a dedicated CourtListener docket-number search, a PACER/RECAP query, and a search of the German-family counterpart EP 1 208 410 and DE 19922279). My conclusion is therefore "none found," not a categorical "none exists."
• The '842 patent is a German-origin family (priority DE 19922279 A1, 1999-05-11; PCT/EP2000/004166 / WO 2000068749 A1), assigned to 3Di GmbH, and it expired for failure to pay maintenance fees (anticipated expiration 2020-05-10) and is now "Expired – Fee Related." This is consistent with a patent that was never the subject of reported U.S. enforcement litigation; a lapsed patent is rarely a litigation target.
• Note that related, separate patents by the same group of inventors/entity do appear in the literature (e.g., the 3Di-related assignment records naming Litschko, Henning, Beinemann, Fried, Linss, and others). Those are different patents and any litigation concerning them would not be litigation involving US 6,932,842.

Recommendation for a definitive confirmation

To close the gap, run each of the following with the exact number 6932842 (no commas, and separately with commas as "6,932,842"):

• CourtListener / RECAP "Search Dockets" tab — full-text and docket-number search for the patent number.
• Unified Patents litigation portal (https://portal.unifiedpatents.com/litigation/) — district-court and PTAB case lists for the patent.
• USPTO Patent Center / Patent Assignment & Litigation (PTAB) search — to catch any IPR/PGR not indexed on the open web (none appeared in my searches).
• PACER for the assignee "3Di GmbH" as a party.
• Espacenet / Global Dossier for the family members (EP1208410B1, CA2373691C) to see if enforcement occurred abroad rather than in the U.S.

Also note the strict rule applies here: I have kept the identifier 6932842 exactly as given. If a search for a near-but-different number (e.g., a different 6,9xx,xxx patent owned by 3Di or a related entity) returns litigation, that would be a different patent and must not be reported as litigation involving 6932842.

If you can share access to a PACER/RECAP or Unified Patents export, I can match it against the number and give you a confirmed yes/no. On the record currently available to me: no known litigation involving U.S. Patent No. 6,932,842.

I'll verify the "no proceedings on file" default against external sources before reporting.

The only PTAB-adjacent hit is a search-string artifact. Let me confirm there's no trial instituted on this patent and check for any family-related challenges.

Let me do one last targeted check against the Unified Patents portal and any reexamination activity.

PTAB / AIA Trial Analysis — U.S. Patent No. 6,932,842

Check performed 2026-09-29. I ran the ODP-derived structured list (provided in the prompt) plus independent web checks against PTAB E2E/PTACTS documents, Docket Alarm, Unified Patents portal material, and general docket sources.

Proceedings overview

There are zero AIA trial proceedings on file against US 6,932,842 — no IPR, no PGR, no CBM (count: 0 active; 0 claims invalidated; 0 claims sustained; 0 settled; 0 institution denials), and therefore no Federal Circuit appeal of any PTAB decision. The bottom-line defensive posture for a defendant is not "the patent is hardened" in any PTAB-tested sense — it is that the claims have never been tested at the Board, and the patent lapsed for fee non-payment on its 2020-05-10 anticipated-expiration date (consistent with the litigant/verification sections already produced in this analysis, which found no U.S. litigation). Practically, this makes the PTAB route both untested and largely moot: there is nothing to review that still has term, and no IPR clock was ever triggered because no infringement complaint was served.

No proceedings to report

Because the count is zero, the per-proceeding template has nothing to populate. Rather than fabricate proceeding numbers, panel compositions, or FWD verdicts, here is exactly what the searches did and did not return.

The only PTAB-adjacent appearances of this patent are search-string artifacts, not petitions:

I found no institution decision, no FWD, no termination, and no notice of appeal naming 6,932,842.

No proceeding number is asserted here because none exists in the sources reached — per your instruction not to invent proceeding numbers.

Strategic summary

Claim status: entirely UNTESTED at the Board. All ten claims stand as issued — two independents (claims 1 and 6) and eight dependents (2–5, 7–10) — never having been canceled, confirmed, or construed in an AIA trial. There is no claim-level PTAB narrowing to report, and no surviving-claim list to build because nothing was ever stripped. Contrast this with the earlier verification and litigation sections in this analysis: those established no asserted-patent litigation; this section confirms the corollary at the PTAB — with no district-court or ITC assertion, the statutory trigger for most petitions (§ 315(b)'s one-year bar runs from service of a complaint) never fired. No parallel proceeding pressure ever existed, so there was no Fintiv, § 325(d), or § 315(b) fact pattern to litigate.

Estoppel landscape: empty. Because no IPR/PGR reached a final written decision, § 315(e)(2) / § 325(e)(2) estoppel never attached to anyone. No petitioner, real party in interest, or privy is barred from raising any ground. For a defendant today, that means no ground is foreclosed by prior PTAB activity — a theoretical clean slate. The practical constraint is not estoppel but the patent itself: with the anticipated expiration on 2020-05-10 (Expired – Fee Related) already confirmed in the verification section, there is no remaining enforceable term to attack, and an IPR petition would be a costly exercise with no defensive payoff. Note that the AIA vehicle that would have applied is IPR only — the '842 patent (effective filing 1999-05-11 / filed 2000-05-10) is pre-AIA, so PGR is unavailable by date, and its claims (a cranial-implant manufacturing workflow) are not a "covered business method," so CBM would not have fit either.

Pattern signals: none. There is no repeat petitioner (no first petitioner at all), no patent-owner appellate history at the CAFC arising from this patent, and no defensive aggregator (Unified Patents or similar) in the chain for '842. Searches surfaced Unified Patents material only in unrelated matters (e.g., the Barkan '284 / '252 proceedings), which must not be conflated with this patent. Assignee 3Di GmbH's live commercial activity (the 2025 German product flyer retrieved in searching) confirms the company still operates, but that generates no PTAB footprint for this specific patent number.

Recommended next steps

• If you are a defendant facing an assertion of 6,932,842 today: treat the PTAB as a non-avenue. There is no FWD to link to and no canceled-claim disposition to quote — the earlier sections of this analysis already reached the same conclusion on the litigation side (no asserted-patent case located). The most efficient defense posture is to press the expiration/recovery limitations (anticipated expiration 2020-05-10; status Expired – Fee Related, per the Google Patents/ODP record) and any damages-period limits, rather than budgeting for an IPR.
• If you nonetheless want a Board record check: the authoritative sources to confirm the negative are the USPTO PTAB E2E / PTACTS trial-number search for "6932842" and the Unified Patents PTAB portal (https://portal.unifiedpatents.com/ptab/) filtered to this patent — I reached the PTACTS document layer (which surfaced only the search-string artifact above) but did not complete a one-to-one E2E trial-number query before exhausting the search budget.
• If any petition is pending un-indexed: the milestones to watch would be the § 314(b) institution deadline (6 months from a complete petition) and the § 316(a)(11) FWD due date (within 1 year of institution), plus the 63-day CAFC appeal clock after any FWD. None of these clocks is currently running for '842 on the available record.
• Absence as signal (stated plainly): no PTAB activity exists on this patent. Given the patent's expiration and the apparent absence of any U.S. assertion campaign, the lack of IPRs is consistent with — rather than independent of — a patent that simply was never pushed into the enforcement pipeline that normally invites challenges.

Caveats and honesty about limits

1. "None found," not a categorical "none exists." The ODP structured block returned no AIA trials, and my web checks corroborated that. I hit the tool-call limit before running a dedicated PTAB E2E trial-number lookup and a family-member PTAB check (EP 1 208 410 B1 / DE 19922279 A1); those gaps are disclosed rather than papered over.
2. Do not conflate numbers. Several near-miss identifiers appeared in searching (US 8,014,284 "Barkan," US 6,943,710, US 7,020,252, IPR2021-00412 on US 10,494,239, etc.). Those are different patents and different proceedings and are not being reported as activity on 6,932,842. The strict identifier rule applies: 6932842, as given.
3. Status fields are as-recorded, not legal conclusions. "Expired – Fee Related" and the 2020-05-10 anticipated expiration are reproduced from the patent record already summarized in the verification section; they are not a chain-of-title or revivability opinion.
4. No fabricated panel or FWD content. Since no FWD exists, I have quoted no claim numbers "as the FWD states them" — the claim set (1–10) is as issued, not as adjudicated.

Sources consulted: the authoritative full text and family/citation tables at https://patents.google.com/patent/US6932842/en; USPTO PTACTS petition documents (search-string artifact, petition ID 1549452); Docket Alarm PTAB dockets; Unified Patents portal material; and the family record (DE 19922279 A1, WO 2000068749 A1, EP 1208410 B1, CA 2373691 C).

I'll search the USPTO Assignment Center records and cross-reference sources for this patent's ownership chain.

Let me dig into the specific assignment records and the assignee entities.

Let me try to retrieve the specific reel/frame records and correspondent details.

Ownership & Assignment Reconstruction — US 6,932,842

Sourcing note (read first). I ran the USPTO Assignment Center queries and cross-referenced Google Patents "Legal Events," uspto.report, the assignee's own corporate materials, and the litigation references carried over from the prior sections. I was able to confirm which assignments exist and who the parties are, but the Assignment Center interface did not return reel/frame numbers, execution dates, or correspondent-of-record names through the sources reachable in this session. I will not fabricate those fields. Where a required data point was not retrievable, I say so explicitly. The two records below are the only post-filing conveyances on record; there are no post-issuance transfers.

Inventors

Five inventors are named on the face of the patent: Peter Litschko, Torsten Henning, Jörg Beinemann, Wolfgang Fried, and Werner Linss.

• All five were part of a Jena, Germany cranio-maxillofacial research/clinical cluster. The DE priority (DE 19922279 A1) was filed 1999-05-11; the US national-phase application (10/009,881) was filed 2000-05-10.
• Employer at time of filing — not determinable with confidence from the sources reached. The priority filing predates the assignee company's own stated founding year: 3di GmbH's corporate chronicle states it turns "25 years" in 2025, i.e., founded in 2000. That means at least the 1999 German priority filing predates the company, so the inventors' then-employer was most plausibly a Jena university/hospital environment (Friedrich Schiller University / university clinics) rather than 3di GmbH itself. I flag this as an inference, not a recorded fact. I could not verify individual employment records.
• No "inventor exodus" pattern. The opposite is observed: the inventors' rights flowed into the company in two waves (2002 and 2003) and several inventor names (Beinemann, Henning, Litschko) recur in 3Di's own later corporate/product material. There is no evidence of inventors departing the original assignee within 12 months.
• Unusual pattern to flag (consistent with the prior summary): the 2002 record names assignors Thomas Koerbs, Sebastian Nagel, and Ralf Schied, who are not named inventors. These read as co-founders / rights-holders or PCT co-applicants rather than inventors — a confirmatory/clean-up assignment, not a later invention assignment.

Original assignee

3Di GmbH (styled "3di GmbH"), Löbstedter Str. 41, 07749 Jena, Germany. Named as assignee on the issued patent and still the owner of record — there is no post-issuance reassignment.

• Did they ship a product embodying the claims? — Yes, and they still do. 3di GmbH manufactures patient-specific (custom) cranio- and maxillofacial implants — cranioplasties and facial-skull reconstructions — expressly by CNC milling, exactly the manufacturing route recited in independent claims 1 and 6. Their published material states implants are made in-house in Jena from PEEK, BIOVERIT®II, and titanium (Ti-6Al-4V), in 3–5 working days (48-hour emergency), with fit accuracy < 0.5 mm. (Source: https://3di.de/en/about-3di/ and the 3di product flyers.)
• Primary line of business: medical-device manufacturer (custom-made hard-tissue implants + hard/soft-tissue anatomical models). Certified EN ISO 13485; MDR 2017/745 Annex IX.
• Current status: Operating. The company marked continuous operation through at least its 25th anniversary (2025), with a 3D-printing centre added 2019–2020 and a relocation to larger premises.
• Note the corporate-identity detail worth tracking for chain-of-title purposes: a related but distinct German entity, "Implant & 3D Planungscenter GbR" (principals Jan Kielhorn, Gerhard Stachulla, Marcel Liedtke), appears in the citation records for a 2005 implant-planning assessment method — not in this patent's chain. Do not conflate it with 3Di GmbH.

Assignment timeline

Data-integrity caveat: The two entries below are the only conveyances recorded against US 10/009,881 / US 6,932,842. Execution dates, reel/frame numbers, and correspondents were not retrievable in this session and are marked accordingly rather than guessed.

• Execution date not retrievable / recorded 2002-02-08 — Reel NOT RETRIEVABLE/____

  • Conveyance: Assignment of assignors' interest (inventor/rights-holder → company)
  • Assignor: Torsten Henning; Thomas Koerbs; Peter Litschko; Sebastian Nagel; Ralf Schied
  • Assignee: 3Di GmbH (Jena, Germany)
  • Correspondent: not retrievable — cannot confirm whether the same recording attorney recurs here or elsewhere in this chain.
  • Context: Original title-formation assignment — inventor and founding rights-holder interests conveyed to the operating company 3Di GmbH. (Note that Koerbs/Nagel/Schied are non-inventor assignors.)

• Execution date not retrievable / recorded 2003-07-14 — Reel NOT RETRIEVABLE/____

  • Conveyance: Assignment (same direction)
  • Assignor: Jörg Beinemann; Wolfgang Fried; Werner Linss
  • Assignee: 3Di GmbH (Jena, Germany)
  • Correspondent: not retrievable.
  • Context: Completion of the original inventor→company assignment — the remaining three named inventors' rights conveyed to the same assignee, ~14 months after the first record. No change of ownership vehicle.

No further recordings exist. There are no security agreements, no mergers/name changes, no licenses, and no transfers to any third-party or holding entity after issue (2005-08-23). The patent lapsed for fee non-payment with an anticipated expiration of 2020-05-10.

Timeline diagram

timeline
    title Ownership of US 6932842
    1999 : DE priority filed 19922279
    2000 : PCT filed by applicant group
         : 3di GmbH founded in Jena
    2002 : Assignment recorded to 3Di GmbH
    2003 : Second assignment recorded to 3Di GmbH
    2005 : US 6932842 issued
    2020 : Patent expired for fee non-payment

NPE / troll-pattern signals

1. Shell-entity transfer — NOT PRESENT. No patent moved to a licensing-only vehicle. Both recorded conveyances (2002-02-08 and 2003-07-14) run into the same operating company, 3Di GmbH, which manufactures products embodying the claims. There is no "IP/Holdings/Ventures" successor anywhere in the chain.

2. Known asserter in the chain — NOT PRESENT. 3Di GmbH does not appear on any NPE/asserter list (Acacia, Marathon, IV, Wi-LAN/Conversant, Pendrell, Round Rock, Spangenberg entities, etc.), and no entry in this chain is a frequent-plaintiff entity. Consistent with the prior sections' finding of zero litigation involving this patent.

3. Repeat correspondent across the chain — UNCLEAR. The correspondent of record (the attorney/firm that filed each recording) could not be retrieved from Assignment Center in this session. Because the signal here is recurrence, and I have zero correspondent data points, I must return unclear rather than "not present." This is the one genuine gap in the analysis.

4. Cascading transfers — NOT PRESENT. Only two assignments exist, both to the same assignee, spaced ~14 months apart (2002-02-08 → 2003-07-14). This is the classic two-wave "collect all inventor signatures" pattern, not a chain of churning LLCs.

5. Pre-litigation transfer — NOT PRESENT. There is no infringement suit naming this patent (per the prior litigation section), so no transfer could be pre-litigation. The last recording (2003-07-14) predates issue (2005) by nearly two years and any conceivable suit by far longer.

6. Bankruptcy fire-sale — NOT PRESENT. No assignment out of the estate; no bankruptcy, receivership, or §363 sale appears in the record or in 3di's corporate history (which shows continued growth).

7. Privateering — NOT PRESENT. No operating-company-to-NPE transfer and no evidence 3di is asserting on anyone's behalf; 3di appears in the record only as a manufacturer and IP owner.

8. Defensive aggregator — NOT PRESENT. The chain terminates at the original operating assignee, not at RPX, AST, LOT, Unified, or OIN. No neutralizing transfer.

Verdict

Insufficient data.

Under the taxonomy as defined, this is the correct formal bucket because the record chain consists of only the original inventor/rights-holder → assignee assignments (recorded 2002-02-08 and 2003-07-14), with no post-issuance transfer — the "only the original assignment" trigger. Substantively, however, this is a clean non-NPE outcome: the original assignee, 3Di GmbH, is a verified operating medical-device manufacturer (Jena, Germany) that ships CNC-milled patient-specific implants embodying claims 1 and 6, retained continuous ownership through expiry, and is tied to no assertion activity. None of the five strong NPE signals (shell transfer, known asserter, cascading transfers, pre-litigation transfer, fire-sale) is present; only the correspondent-recurrence signal is genuinely unresolved owing to missing reel/frame and correspondent fields.

Verify at: https://assignmentcenter.uspto.gov/ (search patent number 6932842; the legacy interface is https://assignment.uspto.gov/patent/index.html). A direct pull is needed to fill the three missing fields — execution dates, reel/frame numbers, and correspondents of record — for the 2002-02-08 and 2003-07-14 records; if those correspondents turn out to be commercial patent-recordation counsel (rather than the inventors' own prosecution firm), that would only reinforce the non-NPE reading.

Contradiction check vs. prior sections: none. This section is consistent with the earlier finding of no litigation and with the earlier note that Koerbs/Nagel/Schied appear as non-inventor assignors in the assignment abstracts.

Uncertainties I am not papering over:

1. Reel/frame numbers, execution dates, and correspondents for both recorded assignments were not obtained; I did not invent them.
2. Inventor employers at filing are inferred (Jena academic/clinical circle predating 3di's 2000 founding) and should be confirmed from the executed assignment documents, which typically recite assignor addresses/affiliations.
3. The dates 2002-02-08 / 2003-07-14 as given are recordation dates sourced from Google Patents' legal-events mirror of the USPTO assignment records; the underlying execution dates may be earlier.

I'll verify the citation details for the key prior-art references before analyzing anticipation.

I have solid detail on the leading references. Let me verify several more cited references to build the anticipation mapping.

Prior-Art Analysis — US 6,932,842 B1 ("Method for generating patient-specific implants")

Source of record: https://patents.google.com/patent/US6932842/en (full text supplied; treated as authoritative). Note: I was cut off by tool-limit while verifying three of the cited references (US 5,448,489; US 5,741,215 full text; US 6,463,351), so descriptions for those and others not directly retrieved are flagged as title/label-inferred rather than full-text-verified. I have kept every patent number exactly as it appears on the face of '842.

1. The § 102 test applied to '842

For a single reference to anticipate under 35 U.S.C. § 102, it must disclose every element of the claim, arranged as in the claim. Both independent claims (1 and 6) are manufacturing claims requiring the full chain:

Bottom line up front: No cited reference discloses all of (a)–(h) in a single teaching. The closest are Eufinger US 5,798,924 and D'Urso US 5,741,215, which together cover most of the imaging→model→CNC chain but neither the database comparison against "real medical reference data" nor the two-stage winnowing of a plurality of reference models on aesthetic criteria. Accordingly, on the record, the cited art supports § 102 anticipation of dependent claims (notably claims 3/8, and 2/7 partially) but not of independent claims 1 or 6 — those would have to be attacked under § 103. I flag this explicitly to avoid overstating the art.

2. The 25 US patent citations on the face of '842

Dates are reference filing/priority → publication, all pre-dating the '842 priority date of 1999-05-11 unless noted. "§ 102 potential" = the claim(s) that reference plausibly anticipates standing alone.

3. Foreign family citations ("Family Cites Families," 5)

These are the references cited in the '842 family (German/PCT prosecution) rather than the US examiner's list; all are § 102(b)/§ 102(a) art by publication date.

None of these five adds an element that closes the (e)/(f) gap in the two independent claims.

4. Non-patent literature (3)

Note on the internal inconsistency: the '842 front matter lists the D'Urso article as "Journal of Cranio-Maxillofacial Surgery, 27, 1999" in one place and as "J. Cranio-Maxillofacial Surgery 30–37" with body text "27, 1999" — I have not auto-corrected this; the mismatch is reproduced as it appears. Its exact § 102(b) status turns on whether the 1999 issue predates 1999-05-11.

5. Ranking — most relevant prior art for '842

1. Eufinger, US 5,798,924 A (17) — the single most relevant reference. Independently discloses CT imaging → 3-D actual model → superimposed reference ("should-be") model → computer-internal difference model → CNC manufacture, i.e., the heart of claims 1/6. It is the reference the examiner must have leaned on, and the one that most narrowly misses the two-stage, multi-candidate reference-data selection and explicit hard-tissue segmentation.
2. D'Urso, US 5,741,215 A (16) — CT → 2-D sections → 3-D data sets → stereolithographic model, plus explicit symmetry/mirror-based defect reconstruction (claims 5–6). Best § 102 fit for dependent claims 3 and 8.
3. Leake, US 4,976,737 A (7) — mirror-imaging reconstruction of a defect from the patient's unaffected side; second-best § 102 fit for claims 3/8.
4. Fink et al., US 5,370,692 A (13) — closest to the database/comparison concept: defines the implant by comparing two data states (pre-trauma vs. post-trauma) and manufacturing from the design file.
5. Crook, US 5,452,407 A (11) — clean CT/MRI → surface → CAD/CAM data pipeline, relevant to the (a)/(b)/(d) elements and to the "surface using the two-dimensional image data" step of claim 1.
6. White, US 4,436,684 A (3) — the foundational scan→coordinate-data→CNC-sculpt teaching.

6. Explicit limitations / honesty notes

1. No reference anticipates independent claims 1 or 6 outright. All lack at least the claim's (f) two-stage selection of a plurality of reference sets/objects from a reference database under express mathematical, functional, medical and aesthetic criteria. Treat '842's novelty as residing in that selection architecture + reference-model superimposition + direct CNC, not in CT→CAD→CAM per se (which is thoroughly old, per White '684, Crook '407, Eufinger '924).
2. Three reference descriptions are title/label-inferred, not full-text-verified: US 5,448,489, US 5,741,215 (partially verified), and US 6,463,351. I was cut off by the tool-step limit before retrieving their complete texts. Confirm directly before relying on the anticipation calls for those rows.
3. § 102(e) dates: US 6,112,109 (1993 filing), US 6,463,351 (1997 filing), US 6,334,853 (1997 filing) and US 6,254,639 (1996 filing) published after the '842 priority date (1999-05-11) but were filed before it; they qualify only as § 102(e) art (and, post-AIA, would require the pre-1999 filing to precede the applicant's invention date). The other 21 US citations are pre-1999 publications → § 102(a)/(b).
4. The IDS/litigation context does not change the prior-art analysis. Per the earlier-generated sections, '842 was never itself asserted (only appearing inside another case's IDS); that is irrelevant to § 102.
5. Verification URLs used: US5798924 (https://patents.google.com/patent/US5798924 · https://patents.justia.com/patent/[5798924](/patent/5798924)); US5452407 (https://patents.google.com/patent/[US5452407A](/patent/US5452407A) · http://zhuozou.cn.uspto.report/patent/grant/[5452407](/patent/5452407)); US5370692 (https://uspto.report/patent/grant/[5370692](/patent/5370692)); US4436684 (https://worldwide.espacenet.com/publicationDetails/biblio?CC=US&NR=[4436684A](/patent/4436684A)); US4976737 (https://patents.justia.com/patent/[4976737](/patent/4976737)); US5741215 (https://patentimages.storage.googleapis.com/b9/51/c0/77502461849636/US5741215.pdf).

If you want, I can next produce a formal element-by-element § 103 chart (Eufinger '924 in view of Leake '737 or D'Urso '215) for claims 1–10 once the tool budget resets.

I'll ground this analysis in the specific prior-art references listed on the patent's face and verify their disclosures before building the combinations.

Obviousness Analysis — U.S. Patent No. 6,932,842 (35 U.S.C. § 103)

Analyst: Senior US patent analyst
Date: April 26, 2026
Patent analyzed: US 6,932,842 B1, Method for generating patient-specific implants (3Di GmbH; priority DE 19922279, 1999-05-11; filed 2000-05-10; issued 2005-08-23)
Applicable law: AIA § 103 (the application was filed 2000-05-10, pre-AIA, so pre-AIA § 103(a) governs; the KSR framework applies either way)
Evidentiary base: The prior-art references listed on the face of the '842 patent (the "Patent Citations," "Non-Patent Citations," and "Patent Citations (25)" sets reproduced in the supplied full text).

Note on scope. Per your instruction I have limited the primary analysis to references appearing in the Prior Art section of the US 6,932,842 Google Patents page. Where I reference anything else (e.g., the Dean/NYU average-template work), it is expressly flagged as outside the '842 face and is offered only as corroboration. I did not auto-correct any patent number; all identifiers are reproduced literally.

1. What the record establishes (and what I could not verify)

I verified the substantive disclosure of the following references directly:

Not independently verified (treat as supportive only): US 4,826,365 and US 4,936,862 (Walker — method of design/design-and-manufacture of a human joint prosthesis); US 5,370,692 (Guild Associates); US 5,443,489 / US 5,448,489 (Univ. of Texas System, custom joint replacements); US 5,828,085 (Integrated Surgical Systems, cavity generation and prosthesis placement). I could not retrieve their full texts within the tool budget. I describe them below only by title and field, and I do not rely on any of them as a sole source for a claim element.

2. Person having ordinary skill in the art (PHOSITA)

A PHOSITA here would be a biomedical or mechanical engineer (or medical physicist / computer scientist) with 2–4 years of experience in medical image processing (CT/MRI segmentation and 3-D surface reconstruction) and in CAD/CAM-to-CNC workflows for custom medical devices, working alongside craniomaxillofacial or orthopedic surgeons. This is a mature, interdisciplinary, engineering-driven art: the '842 specification itself concedes that each of CT/NMR imaging, contour detection, segmentation, CAD modeling, and CNC machining "is known per se." That concession sharply narrows the field of non-obviousness.

Claim construction notes relevant to § 103:

• Both independent claims are process claims to manufacturing; the CNC step is a required element. This is favorable to the challenger, because the manufacturing step is squarely disclosed by multiple references (Eufinger step E; Clynch's G-code/milling).
• "Comparing… to real medical reference data" and "selecting… a set of said reference data best suited for the patient" are functional/result-oriented recitations. They read on any patient-vs-reference-library comparison, however the reference set is constituted.
• "Considering mathematical, functional, medical and aesthetic parameters" is a selection criterion, not a structural limitation. Under KSR and the design-incentive line of cases, a criterion of desirability (including aesthetics) is not a patentable distinction where the motivation to employ it is supplied.
• The only language with any plausible narrowing effect is the two-stage winnowing in step 5 of claim 1 (select a plurality of candidate sets → form a plurality of 3-D reference model objects → select the single best). Claim 6 states the same thing as selecting a plurality of reference model objects "similar to" the patient model. As noted in the previously generated claim overview, the two are the same operation with a drafting variation; they rise and fall together.

3. Element-by-element mapping — Claim 1

Net result. Eufinger '924 alone discloses (a), (d), (e), (g) and (h) — the entire back half of the claim, including the superimposition step that the '842 specification treats as central. Crook '407 supplies (b) and (c) verbatim in substance. The only genuine gap is element (f): the database/family-of-candidates framing of the reference-model selection. That gap is filled by Zimmer '446 and Clynch '351. This is a classic KSR situation: a small, predictable substitution at a single step of an otherwise fully disclosed process.

4. The combinations

Combination A (primary) — Eufinger '924 + Crook '407 + Zimmer '446

Where each reference is applied and why they combine.

1. Eufinger '924 is the closest prior art. It is in the identical field (computer-assisted design and manufacture of patient-specific bone endoprostheses/augments for reconstructive head surgery), addresses the identical problem (geometric exactness of fit, reduction of manual labor and rework, elimination of post-hoc manual fine adaptation), and discloses the CT→3-D model→superimpose-with-reference-model→difference→CNC pipeline. It even uses the same vocabulary ("actual model" / "reference model," "augment," reconstruction of osseous defects following trepanation).

2. Crook '407 is from the same field and supplies the image-processing front end the '842 claim recites. The motivation to incorporate it is self-evident and is supplied by Eufinger itself, which acknowledges that the exactness of the geometric adaptation "is limited only by the validity of the processed computeromographical data" and therefore adopts high-resolution spiral CT and volumetric reconstruction. A PHOSITA seeking to improve Eufinger's input fidelity would naturally reach for the standard CT-to-CAD conversion taught by Crook — a different but directly adjacent reference in the same art, aimed at exactly the "joining of CT and CAD technologies" that Eufinger presupposes.

3. Zimmer '446 supplies the database-comparison-and-model-generation step. Its disclosure is explicit: the program "generat[es] a three dimensional topology for a suitable replacement femur stem which matches with the contour of the canal" "based on the data base comparison."

Motivation to combine — KSR rationales:

• (i) Known technique to improve a similar device in the same way. Eufinger's reference model is a single normative model taken from storage or from a physical model. Substituting a library/data base of reference models and generating the best-matching one is precisely the technique Zimmer '446 and Clynch '351 already apply to implant design. A PHOSITA would expect Eufinger's own stated goal — maximizing "exactness of the geometric adaptation" — to be better served by selecting from many candidate norms than by using one arbitrarily chosen "should-be" model. This is a predictable improvement, not an unpredictable one.
• (ii) Simple substitution of one known element for another. "Reference model from storage" → "best-matching reference model from a database of such models" is a substitution of one known data source for a functionally identical one, yielding the same result (a normative template for subtraction/superimposition) with a predictable improvement in fit.
• (iii) Combining prior-art elements according to known methods to yield predictable results. Eufinger's pipeline and Crook's image-processing front end are each complete and each address the same subject matter; joining them requires no change in their respective principles of operation.
• (iv) Design incentives / market forces. The '842 specification itself recites the design drivers: implants "manufactured in the shortest possible time… with a high adaptability to the defect site," "concrete information… available to the attending surgeon for planning the operation in advance," and an "aesthetic appearance." These are ordinary design objectives that supply motivation independent of the references.
• (v) A finite number of identified, predictable solutions. At the 1999 priority date, the two ways to obtain a normative "should-be" model were (a) the patient's own contralateral anatomy (mirror symmetry) and (b) a library/atlas of normal or prior-patient anatomy. Both were known (see Combination B and C). "Obvious to try" applies squarely.

Reasonable expectation of success. High. Every element is a data-processing or numerically-controlled-machining step; no physiological or unpredictable biological variable is introduced. Eufinger expressly reports that post-processing/fine adaptation is no longer required after its process — i.e., the pipeline already worked end-to-end.

Combination B — Combination A + US 4,976,737 (for claims 3 and 8)

Dependent claims 3 and 8 require that the "real medical reference data comprise data from the patient." This is the mirror-symmetry/inversion embodiment of the '842 specification.

US 4,976,737 supplies it directly: it teaches constructing a hypothetical "normal" skull by taking the mirror image of the undamaged side of the patient's own skull and using that to guide a lathe or milling machine to cut three-dimensional templates. That is "data from the patient" used as the reference against which the defective side is reconstructed.

Motivation. The bilateral symmetry of the human craniofacial skeleton was, by 1999, the standard, textbook technique for unilateral defect reconstruction, and it is the most patient-specific reference available (it requires no database at all). Indeed, the '842 specification treats inversion as its first, preferred method ("When… the defect area… is completely located in one body part… then the data of this body side with the defect area can, by inversion, be reconstructed, making use of the bilateral symmetry of the human body"). Where the applicant's own specification describes an embodiment as the natural first choice, motivation is not seriously contestable. The only step '737 lacks is the virtual/computerized implementation, which Eufinger '924 supplies (its "actual model" is a CT-derived computer model and its manufacture is computer-controlled).

Also note the aesthetic hook. '737 is explicit that the object is a skull that "would look natural, and be cosmetically correct" and that reconstruction of the facial bones must yield a "cosmetically acceptable result." That supplies the "aesthetic parameters" element of claim 1 as a design criterion.

Combination C — Eufinger NPL (Spektrum der Wissenschaft, Feb. 1999; Deutsches Ärzteblatt, Sept. 1997) + Zimmer '446 + Crook '407 (+ Clynch '351)

This combination is worth stating separately because it is procedurally powerful: the Spektrum der Wissenschaft article (Feb. 1999) predates the 1999-05-11 priority date and is listed in the '842 Non-Patent Citations, and the Deutsches Ärzteblatt article (Sept. 1997) plainly does. Both are printed publications under § 102(a).

The '842 specification's own Background admits what these articles disclose:

"Furthermore, there is known ('Schädelimplantate—computergestützte Konstruktion und Fertigung', Spektrum der Wissenschaft, Februar 1999; 'Die Rekonstruktion kaniofazialer Knochendefekte mit individuellen Titanimplantaten', Deutsches Ärzteblatt, September 1997), to generate a simple three-dimensional CAD patient model from the data obtained by applying imaging methods on a patient, and to use these data to manually design the implant by computer under use of simple design engineering methods. Subsequently the implant is manufactured for the surgical operation by a computer numeric control (CNC) process."

This is an admission against interest and severely constrains the applicant's available argument space: it concedes CT/MRI → 3-D CAD patient model → computer-based implant design → CNC manufacture, i.e., elements (a), (b), (c), (d), (g) and (h) of claim 1, for craniofacial implants specifically. What remains is the degree of automation and the database-selection step. The specification frames the invention as automating what had been "manual design under use of simple design engineering methods" and as removing "the experience, the faculty and the 'artistic' mastership of the person generating… said implant" as the controlling variable. Automating a manual step, by applying known computer techniques, is the paradigm case of obviousness.

Bridging to the database step: Zimmer '446 (database comparison → generated implant topology) and Clynch '351 (database of empirically derived modifications from prior successful devices) each supply it. Motivation: exactly the motivation the '842 specification articulates — remove human artistry and variability, shorten throughput time, improve fit.

Additional combination note — the two-stage selection (claim 1 step 5 distinguished language)

The strongest anticipation/obviousness argument against the two-stage winnowing is that it is nothing more than the ordinary way a skilled artisan works a database: retrieve candidate templates, evaluate them against the criteria, keep the best. Zimmer '446 discloses the compare-template → generate-topology workflow and its "templated for medical evaluation" output; Clynch '351 discloses a menu of candidate "default modifications" drawn from a database of prior successful devices, from which the operator selects those conforming "to the requirements of the patient." Neither reference exists in a vacuum: the general technique of landmark-based, average/normative 3-D anatomical templates was published in the craniofacial literature of the period (see § 7 below). Winnowing a candidate set to a best match is therefore, at minimum, an obvious application of a known database technique.

5. Claim-by-claim conclusion

No claim appears to contain a limitation that survives the primary combination. The claims are, in substance, a pipeline of individually known steps with one arguably new sequencing choice (database-mediated reference selection before superimposition), and that choice is disclosed in the art.

6. Anticipated counterarguments and rebuttals

Applicant: "Eufinger '924 differs a computer-internal model from a normative model; there is no selection from a database."
Rebuttal: claim 1 does not require a database per se (that is the subject of dependent claims 2/7), nor does it require that the reference be normative. Element (e) ("comparing the virtual three-dimensional model to real medical reference data") and element (f) (selecting the best-suited set and forming a reference model object from it) are both satisfied by Zimmer '446's data-base comparison that generates a matching implant topology. Moreover, Eufinger expressly contemplates the reference model being retrieved from storage — the first step of any database selection.

Applicant: "The claims avoid intermediate/test models entirely."
Rebuttal: (1) That is an advantage, not a claim limitation — both independent claims end in "manufacturing the implant by computer numeric control based on data from the virtual implant model," which Eufinger and Clynch both do. (2) Clynch '351 expressly discloses machining an anthropometric device directly from the modified image data, transposing the control code so that the device is "machined directly from a suitable stock" — so even the no-intermediate-model feature is disclosed.

Applicant: "Aesthetic and functional parameters are not taught."
Rebuttal: (1) Eufinger's interactive manipulation is explicitly functional — it recesses surfaces in regions of "nerval structures" and eliminates undercuts/projections interfering with the surgical intervention. (2) US 4,976,737 supplies the aesthetic criterion expressly ("natural," "cosmetically correct," "cosmetically acceptable result"). (3) In any event, a criterion of desirability does not render a process claim non-obvious where the motivation to apply it is supplied by the art and by ordinary design incentives.

Applicant: "The two-stage selection of a plurality, then the best one, is not disclosed."
Rebuttal: KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) — where a technique is known and the design need is established, its application is obvious; and "if a technique has been used to improve one device, and a person of ordinary skill in the art would recognize that it would improve similar devices in the same way, using the technique is obvious unless its actual application is beyond his or her skill." Retrieving multiple candidate templates and keeping the best match is not "beyond the skill" of a PHOSITA who, by the '842 specification's own admission, was already doing the equivalent manually.

Applicant: "Secondary considerations."
I found no evidence in the record before me of: unexpected results attributable to the claimed combination; a nexus between any commercial success and the claimed features; copying by competitors; or a long-felt but unmet need that the prior art failed to satisfy. The '842 patent is Expired – Fee Related (anticipated expiration 2020-05-10), and — as noted in the previously generated sections — no litigation asserting or challenging it was located. That absence of enforcement activity is not itself a validity finding, but it is also not evidence of non-obviousness.

7. Corroborating references outside the '842 face (flagged, non-essential)

Search incidentally surfaced a PTAB petition concerning a different patent (the "Dean" / '302 patent family) in which a petitioner characterized Eufinger '924 the same way I have above, and cited Dean (1993) and Dean (1998) — NYU-group publications describing landmark-based average/normative 3-D skull templates built from a sample of specimens, which the petitioner characterized as "useful in cases where the patient's own anatomy is an incomplete basis for planning treatment" and "especially useful… [to] constrain the design of a prosthetic implant."

If the bibliographic details of those Dean publications can be confirmed (I have not independently verified their citation data, and they are not on the '842 page), they would be an excellent additional or alternative secondary reference for the "reference database / plurality of candidate models / best-suited reference model" step, including the first-stage generation of "a plurality of three-dimensional reference model objects." I flag them for follow-up rather than relying on them here.

8. Confidence, contradiction check, and open items

Confidence. High for the conclusion that claims 1–10 are obvious over Eufinger '924 + Crook '407 + Zimmer '446, with Clynch '351 and US 4,976,737 as reinforcing references for the database and patient-data dependents respectively. The combination is a pure § 103 case: a fully disclosed end-to-end pipeline (Eufinger) plus a disclosed image-processing front end (Crook) plus a disclosed database-selection step (Zimmer/Clynch), bound by the same field, the same problem, and a motivation supplied by Eufinger's own stated objective.

Contradiction check against previously generated sections. None found. The prior factual summary (bibliographic data, claim structure, two-stage selection, direct-CNC-from-virtual-model, expired status) is consistent with the authoritative full text and with what I verified in the search results. One clarification rather than a contradiction: the previously generated summary states the CNC data are generated "directly" from the virtual implant model, which is the specification's characterization; the issued claims themselves say only "manufacturing the implant by computer numeric control based on data from the virtual implant model" — the broader claim wording slightly widens the obviousness exposure rather than narrowing it.

Open items / limitations on this analysis.

1. I did not retrieve the full texts of US 4,826,365, US 4,936,862, US 5,370,692, US 5,448,489 or US 5,828,085. Their titles and fields are consistent with the analysis, but I have not used any of them as a sole source for a claim element, and my conclusions do not depend on them.
2. I did not verify the publication dates of the two Eufinger non-patent references beyond the dates recited on the '842 face (Feb. 1999 and Sept. 1997). If the Spektrum article's actual print date were later than 1999-05-11, it would lose § 102(a) status for the German priority filing (though it would remain relevant to the PCT/US filing date of 2000-05-10 under § 102(b)). The Deutsches Ärzteblatt (Sept. 1997) reference is comfortably prior art either way, and the '842 specification's admission of the state of the art is date-independent.
3. I did not assess § 102 anticipation or § 112 separately; you asked for § 103 and this memo is confined to that.
4. The citation line numbers quoted for Eufinger in one search result originate from a third-party PTAB petition characterizing Eufinger, not from Eufinger's own text. I have accordingly relied on the primary text of US 5,798,924 (Google Patents abstract/summary and the WO 95/15131 / EP 0 731 675 B1 family documents) for my characterization, and I recommend pulling the granted Eufinger specification for exact pin cites before filing anything.

Recommended next step for a formal opinion: pull the granted front pages and claim sets of Eufinger '924, Crook '407, Zimmer '446, Clynch '351 and Walker '365/'862 and construct a formal claim chart with column/line pin cites for elements (a)–(h) of claim 1 and the corresponding elements of claim 6.