US 11877801

I'll search for authoritative information on this patent, including any 2026 litigation activity.

US Patent 11,877,801 — Analyst Summary

Important caveats up front: I could not locate any Court of Appeals for the Federal Circuit (CAFC) 2026 docket, appeal, or oral-argument entry naming U.S. Patent No. 11,877,801, nor any district-court or PTAB proceeding keyed to that exact number. Searches returned only patent-family, assignee-portfolio, and unrelated documents (including a USPTO paper that contained the different number "11877819" / application 11/877,819 — a different patent entirely, which I have not conflated). If a 2026 CAFC appeal exists, it is not surfacing in the sources I can reach; treat its existence as unverified. I also do not have the verbatim full claim set for the granted patent; the specific claim numbering below is drawn from the pre-grant publication and should be treated as approximate (see "Uncertainties" at the end).

Bibliographic data

Abstract

"The disclosure herein relates to systems, methods, and devices for developing patient-specific spinal implants, treatments, operations, and/or procedures. In some embodiments, systems, methods, and devices described herein can comprise using artificial intelligence, machine learning, and/or predictive modeling to predict the outcome of a spinal surgery, one or more parameters of a spine of a patient after spinal surgery, for example after implantation of a spinal rod which can be patient-specific, and/or one or more parameters of one or more recommended patient-specific spinal rods. Furthermore, in some embodiments, systems, methods, and devices described herein can comprise intraoperative tracking for tracking and/or suggesting improvements during spinal surgery based on a pre-operatively determined surgical plan, for example in real-time or substantially real-time. In addition, in some embodiments, systems, methods, and devices described herein can comprise screw planning prior to spinal surgery."

Independent claims — plain-language overview

This patent appears to contain three (possibly four) independent claims. Two of them share a distinctive technical hook: converting spinal curvature data into the frequency domain, filtering it, running the predictive model there, and converting the result back.

1. Method for generating and assisting a patient-specific spinal treatment (claim 1 family)
A computer receives preoperative medical images of a patient's spine and derives preoperative spinopelvic parameters — specifically one or more of lumbar lordosis (LL), thoracic kyphosis (TK), pelvic incidence (PI), pelvic tilt (PT), or sagittal vertical axis (SVA) for one or more vertebrae. The system then applies a Fourier transformation to those parameters to move them into the frequency domain, and filters out the components whose frequency is above a predetermined threshold (i.e., a low-pass filter that discards noisy high-frequency detail while retaining the informative low-frequency content). A predicted surgical outcome is generated in the frequency domain, and an inverse Fourier transformation converts that prediction back into the spatial domain, where it is used to produce the patient-specific treatment (including, in dependent claims, specifications of a spinal rod to be implanted). The claim recites a conventional computer processor and electronic storage medium.

2. Method of predicting a surgical outcome of spinal surgery (claim 13/14 family)
Preoperative inputs are fed into a computer system: preoperative medical images of the subject's spine plus preoperative non-imaging data. The system determines measurements from the images — including the position of one or more vertebrae — and then derives preoperative spinopelvic parameters (LL, TK, PI, PT, SVA) from those measurements. From these it generates a predictive surgical outcome in the spatial domain. Dependent aspects include: the outcome comprising postoperative spinopelvic parameters and/or spinal rod specifications; automatic determination of the measurements by the computer; use of sagittal and frontal x-rays as the images; the model being a GAN, CNN, or RNN algorithm; and generation of a preoperatively determined surgical plan (again, potentially including spinal rod specifications) from the predicted outcome.

3. Method of training a predictive model for predicting a surgical outcome (claim 21 family)
Both preoperative and postoperative inputs for prior subjects are fed in — preoperative and postoperative medical images of the spine and preoperative and postoperative non-imaging data. The system determines measurements (vertebral positions) from both image sets, derives preoperative and postoperative spinopelvic parameters, compresses them (the disclosure names a Fourier transformation / low-pass filter, and a polynomial function), and filters out compressed parameters having a noise level above a predetermined threshold. It then trains one or more predictive models on the filtered compressed preoperative and postoperative parameters plus non-imaging data, and tests the trained models on a separate set of test subjects distinct from the training subjects. Further dependent claims (as published) add data augmentation: applying a Gaussian process to the measured data, and rotating the preoperative and postoperative images about a vertical axis (up to 180°) to generate artificial training measurements, with the express caveat that augmented data is used for training only, never testing.

Possible additional independent claim (intraoperative tracking hardware). The patent's specification and figures (FIGS. 9–14) describe an "intraoperative tracking module," and the disclosure language recites modules with "notches configured to attach or remove" the module to/from vertebral screws (e.g., tulip screws), a first conduit for a surgical tool, and a second conduit for a spinal rod, with the two conduit longitudinal axes substantially perpendicular and the second conduit having a wider top section than bottom section. I could not confirm from the sources available whether this subject matter is claimed as a separate independent claim in the issued patent or only appears in the specification/figures — flagging this as unresolved.

Underlying architecture (context)

The patent frames its disclosure as an "iterative virtuous cycle" of imaging analysis → case simulation → patient-specific implant production → case support → data collection → machine learning → predictive modeling. It incorporates U.S. Pat. No. 10,292,770 by reference. Trained model types disclosed include GAN, CNN, RNN (including LSTM/GRU), linear models (LARS/LASSO), SVM/SVR, and combinations (e.g., CNN + SVM). It also describes converting a rod into a mathematical object (e.g., Bézier curves, b-splines, segment/arc sequences, or [A;B] endpoint coordinates) so the whole rod shape can be an algorithm input/output.

Uncertainties I want to be explicit about

1. Claim numbering in the granted patent vs. the pre-grant publication. Snippets from the 2020/0315708 publication show "claim 13" as the prediction method, while a Justia page for US 11,877,801 shows "claim 14" as "A computer-implemented method of predicting a surgical outcome a spinal surgery of a subject." I therefore cannot state exact granted-claim numbers with confidence, only the three independent-claim subject-matter families.
2. Total claim count and any additional independent claims (notably the intraoperative-tracking-module subject matter) are unverified from the sources I reached.
3. No 2026 CAFC activity verified. A negative search result is not proof of absence; PACER/CAFC docket records are not fully indexed by the search tools available to me here.
4. Assignee of record ambiguity. Google Patents lists Medicrea International SA (France) as original assignee; portfolio aggregators list "MEDICREA INTERNATIONAL (USA)" for this patent. I have not confirmed the current recorded owner from USPTO assignment records.

Where the patent's own text conflicts with search-derived aggregator data, I have preferred the patent text as authoritative, per the operating rules.

I'll search multiple litigation-focused sources for this specific patent number.

Let me dig into litigation-specific databases and related proceedings.

Bottom line

I found no litigation — district court, ITC, PTAB, or Federal Circuit — in which U.S. Patent No. 11,877,801 is the patent-in-suit, an asserted patent, or an invalidated patent. Every path I searched for the exact number either returned nothing, returned the patent's own bibliographic page, or returned hits for the different identifier 11/877,801 (see the disambiguation warning below). Per your instruction, I am not substituting results for similar numbers.

Because a negative search is not proof of absence, I'm giving you both the negative finding and the specific adjacent items that surfaced, so you can see what was and was not conflated.

1. Litigation naming US 11,877,801

No case in the Unified Patents litigation portal, Docket Alarm, PACERMonitor, CourtListener/RECAP, or general web results keyed to U.S. Patent No. 11,877,801. No IPR/PGR petition naming it appeared either.

Context that may explain the absence: the patent issued January 23, 2024 (app. 16/837,461, filed April 1, 2020, priority April 2, 2019) and is listed as active, adjusted expiration June 20, 2042. It is a relatively recent grant in a portfolio (Medicrea International SA / assignee) whose sibling patents are more likely to be the first asserted.

2. ⚠️ Critical disambiguation — do not read these as hits on your patent

These surfaced repeatedly and are not U.S. Patent No. 11,877,801:

The search engines conflate "11877801" as a patent number with "11/877,801" as an application serial number. This same false-positive was flagged in the earlier section of this analysis; it recurred here and I am not treating it as a litigation hit.

3. Adjacent items found (relevant to the owner/portfolio, but NOT cases involving the '801 patent)

I'm listing these transparently, clearly marked as not '801 litigation:

(a) Medicrea USA Inc. v. K2M Spine Inc. et al. — trade-secret/"unlawful acts" suit

• Plaintiff: Medicrea International USA Inc. (UNiD ASI technology)
• Defendants: K2M Spine, Inc. and several other individuals/entities
• Jurisdiction: U.S. District Court for the Southern District of New York (Judge Analisa Torres)
• Filing date: November 8, 2017
• Case number: not confirmed in the sources I reached (I will not invent one)
• Status: not determined from available sources
• Why it is not an answer: it was filed in 2017, more than a year before the '801 patent's April 2, 2019 priority date, and the reported cause was "unlawful acts," not infringement of the '801 patent.
• Sources: https://www.odtmag.com/breaking-news/medicrea-files-lawsuit-against-k2m-spine-others/ ; https://www.combourse.com/News/MEDICREA_Medicrea_intente_une_action_en_justice_contre_K2M__2250918.html

(b) Neuropro Spinal Jaxx, Inc. v. Medtronic PLC et al.

• Plaintiff: Neuropro Spinal Jaxx, Inc. d/b/a Spinal Jaxx
• Defendants: Medtronic PLC, Medtronic, Inc., Medtronic USA, Inc., Medtronic Sofamor Danek USA, Inc.
• Jurisdiction: W.D. Tenn. (Chief Judge Sheryl H. Lipman)
• Case No.: 2:25-cv-02893
• Filed: September 22, 2025
• Asserted patents: per the Patent/Trademark Report, 7,727,280; 10,213,321; 11,141,289 (and others) — not the '801 patent
• Status: pending (summonses served Sept. 24, 2025; notice of appearance Sept. 30, 2025)
• Why it is worth noting: Medtronic is the post-acquisition parent of Medicrea, and one exhibit is a "unid-spine-clinical-brief-data" document (UNiD being the Medicrea product line). But the '801 patent is not asserted. This is competitor-side litigation against the Medtronic/Medicrea family, not enforcement of the '801 patent.
• Source: https://cdn.pacermonitor.com/public/case/60265007/Neuropro_Spinal_Jaxx,_Inc_v_Medtronic_PLC_et_al

(c) A 2017-in-era EPO opposition where Medicrea International was the opponent on EP1926444 (verified owner Zimmer Spine) — an administrative proceeding, not U.S. litigation, and unrelated to the '801 patent.

4. What I could not verify (explicitly)

1. No 2026 Federal Circuit activity naming U.S. Patent No. 11,877,801 was located. The CAFC docket is not exhaustively indexed by the search tools available here. A "not found" is not proof of non-existence.
2. No PTAB proceeding (IPR/PGR) naming the '801 patent was located. I could not retrieve a caselist result directly from https://portal.unifiedpatents.com/litigation/caselist filtered on this patent number.
3. No ITC Section 337 investigation naming the '801 patent was located.
4. The 2017 K2M case number and its ultimate disposition remain unconfirmed from the sources reached.

5. Recommended verification steps (for the file)

To convert this negative finding into a documented clearance, run the exact number (with and without commas) against:

• PACER / PACERMonitor and CourtListener RECAP docket searches — "U.S. Patent 11,877,801" and "11877801";
• USPTO PTAB (PTAB Center / Patent Trial & Appeal Board End-to-End) — search by patent number for IPR/PGR;
• Unified Patents Portal — https://portal.unifiedpatents.com/litigation/caselist?patents=11877801, and confirm the result isn't silently returning a fuzzy/near-number match;
• CAFC docket search and Docket Alarm;
• The patent's own file wrapper / assignment record to confirm the current recorded owner (Google Patents lists Medicrea International SA (France) as original assignee, while some aggregators list "MEDICREA INTERNATIONAL (USA)"), since the plaintiff-of-record matters for any future enforcement.

Also note a date inconsistency in the task framing: the prompt states "Current Date: April 26, 2026," whereas the session date is October 1, 2026. Nothing in what I found turns on the difference, but for a litigation-clearance memo the search date should be stated precisely.

Conclusion: as of the search performed, no known litigation involves U.S. Patent No. 11,877,801. That conclusion is based on sources reachable here and should be re-confirmed against PACER/PTAB directly before being relied on as an absence-of-litigation opinion.

PTAB / AIA Trial Review — U.S. Patent No. 11,877,801

Sourcing note. The canonical list in the structured "PTAB proceedings on file" block (USPTO Open Data Portal ingest) reports zero AIA trial proceedings for this patent. Per the operating instructions, that block controls. I nonetheless ran independent web searches to try to surface anything the ODP ingest had missed; the results are discussed under "What the searches did and did not return" below. I did not find any proceeding number for this patent, and I will not invent one.

Proceedings overview

There are zero AIA trial proceedings on file against U.S. Patent No. 11,877,801 — 0 active, 0 with claims invalidated, 0 with claims sustained, 0 settled, 0 institution denials — which means no claim of this patent has been tested at the PTAB and, for a defendant, the patent is at full, unamended-claim scope and entirely unchallenged: there is no FWD to exploit, no cancellation order to point to, and no § 315(e)(2) estoppel to lean on, but also no reason to assume the claims are robust — they have simply never been litigated.

No proceedings to enumerate

Because the set is empty, the per-proceeding template (petitioner, panel, grounds, institution decision, FWD, settlement, appeal, defensive value) has no entries. Recorded plainly so the negative is auditable:

• Type: N/A — no IPR, no PGR, no CBM.
• Filed: N/A.
• Status: N/A. The ODP list returns no records for US 11,877,801.
• Judge panel: N/A — no panel has been assigned.
• Petition grounds: N/A — no petition has been filed or served on the patent owner as far as I can determine.
• Institution decision: N/A.
• Final Written Decision: N/A.
• Settlement / termination: N/A.
• Appeal: No PTAB FWD exists, so there is nothing to have been appealed to the Federal Circuit from the PTAB. (This is distinct from the separate question of district-court appeals; see below.)
• Defensive value: Minimal for a PTAB-based defense — a defendant today must build the invalidity case from scratch rather than leveraging an existing FWD. That is a real cost, but it is also unclaimed ground: nobody has burned the best art yet.

What the searches did and did not return

• No PTAB hit. Searches combining the patent number, the title, the assignee (Medicrea International), and "IPR / inter partes review / PTAB" returned only Google Patents family listings, FDA 510(k) clearances, and Medicrea investor documents. Nothing pointed to a petition, institution decision, FWD, or appeal on this patent.
• A trap number to avoid. One search hit was a PTAB documents download (https://ptacts.uspto.gov/ptacts/public-informations/petitions/1510855/...) that contains the string "11877801." It is not this patent. It is an IDS acknowledgement for application 11/877,819, "SYSTEMS AND METHODS FOR PROCESSING DATA FLOWS," first named inventor Harsh Kapoor, filed 2007-10-24, attorney docket CRSB-0003-P08. That is a different patent family entirely (the "11877819 / 11/877,819" collision flagged in the patent summary). Do not let a keyword search on "11877801" pull this into a brief.
• No CBM pathway exists. Even leaving aside that no CBM petition was filed, CBM review is unavailable here: the patent issued from an application filed 2020-04-01, is post-AIA, and CBM for newly filed petitions is closed (Sunset under AIA § 18; the transition period ended 2020-09-16).
• PGR is time-barred. PGR is available only within 9 months of grant. The patent granted 2024-01-23; the PGR window closed 2024-10-23. PGR is therefore off the table as of today, 2026-10-01.
• IPR remains available to any petitioner whose § 315(b) clock has not run out (see estoppel/timing section below).

Strategic summary

Claim status: everything is UNTESTED. No claim of US 11,877,801 has been canceled, confirmed, or amended in an AIA trial. If the patent summary's reading is right — three independent-claim families built around (1) Fourier-transforming preoperative spinopelvic parameters into the frequency domain, filtering, predicting there, and inverse-transforming back; (2) predicting surgical outcome from preoperative images plus non-imaging data with vertebral-position extraction; and (3) training a predictive model on Fourier- or polynomial-compressed and noise-filtered pre/post-operative parameters with Gaussian-process and image-rotation augmentation — then all three families stand untouched, including the dependent claims adding spinal-rod specifications, GAN/CNN/RNN model types, and automatic image measurement. Note the open item from the prior summary: whether the FIGS. 9–14 intraoperative tracking module subject matter (notches, first conduit for a tool, second perpendicular conduit for a rod, wider top than bottom) is a separate independent claim in the granted patent was not resolved in the earlier section and remains unresolved here. A defendant should get the granted claim set from the face of the patent before mapping product features to claims — this is a prerequisite to any IPR, since you cannot petition on claim numbers you have not confirmed.

Estoppel landscape. Because there has been no IPR or PGR, no § 315(e)(2) estoppel attaches to anyone. That cuts both ways: there is no adverse estoppel binding a defendant, but there is also no prior petitioner's work product or FWD claim constructions to free-ride on. The relevant timing rule for a new petitioner is § 315(b): a petition must be filed within one year of service of a complaint alleging infringement of this patent. If a demand letter has arrived but no complaint has been served, the clock has not started — that is the window to file. Given the granted 2024-01-23 issue date, any party served in the initial 2024–2026 assertion wave may already be time-barred and should check its service date immediately rather than assuming IPR is open. There is no defensive-aggregator footprint on this patent that I could find; nothing suggests Unified Patents, RPX, or a similar entity has petitioned on the Medicrea spinal-predictive-model family.

Pattern signals. Medicrea historically ran an offensive posture, not a defensive one: a 2017-11-08 suit by Medicrea USA, Inc. against K2M Spine, Inc. and others in the U.S. District Court for the Southern District of New York (Judge Analisa Torres), reported publicly as a trade-secret/confidential-information and employee-poaching action rather than a patent-infringement action, and a 2019 announcement that the USPTO had issued key UNiD ASI patents (US 10,318,655; 10,314,657; 10,292,770) as part of a stated "aggressive patent strategy" spanning "10 independent families." Medicrea was subsequently absorbed into the Medtronic organization, so any future assertion would likely be driven by a large, well-funded patent owner — the kind of party that defends IPRs vigorously and appeals adverse FWDs. The family is also still growing — continuation patents including US 11,925,417, US 11,944,385, US 12,251,165, US 12,274,511, and US 12,564,447 appear as priorities/related filings on the Google Patents family record — so even a successful IPR against 11,877,801 would not clear the field; a defendant should map the whole family before committing to a single-patent strategy. That same expanding family is why the absence of IPRs is notable: this is precisely the profile (patient-specific spinal rods, AI-driven surgical planning, a deep-pocketed owner, a growing continuation chain) that normally attracts challenger petitions. The absence so far is a real signal, but a weak one — it does not establish that the claims are strong, and the patent is only a little over two years past grant as of 2026-10-01.

Recommended next steps

1. Say it plainly: there is no PTAB activity on this patent. Do not represent otherwise in a client memo or a litigation hold. There is no FWD to quote and no cancellation order to cite; the claim-level disposition is N/A across the board.

2. Verify the negative yourself before relying on it. I am working from an ODP ingest that can lag, plus web searches that are not a substitute for the docket. Confirm on the primary sources:

• PTAB E2E (trial proceeding search): https://ptacts.uspto.gov/ptabweb/
• USPTO Open Data Portal / PTAB API: https://api.uspto.gov
• Federal Circuit opinions and dockets (for any district-court appeal, as distinct from a PTAB appeal): https://www.courtlistener.com
  If E2E shows a recently filed petition the ODP ingest missed, everything below changes — the institution decision would be due within 6 months of the petition's filing and the FWD within 12 months of institution (35 U.S.C. § 316(a)(11)).

3. Fix your § 315(b) clock. If you have been served with a complaint asserting 11,877,801, count one year forward from the service date. If you are inside the window, a petition drafted now is your only path to PTAB. If you are outside it, IPR is foreclosed to you and your invalidity case moves to the district court (and to any DJ action), where the same art is available but without the PTAB's specialized treatment.

4. Do not lead with art you have not validated against the granted claims. Since no FWD exists to establish any ground, the art-of-record question is wide open. Two structural targets worth early attention: (a) Medicrea's own earlier family patents, notably US 10,292,770, which this patent expressly incorporates by reference — self-prior-art framing is often a soft spot when a specification is carried forward; and (b) the § 112 exposure of the frequency-domain claims (Fourier transform, low-pass filter with a "predetermined threshold," inverse transform, prediction "in the frequency domain") — the written-description and enablement questions around predicting in the frequency domain rather than merely filtering there are worth a dedicated read of the granted specification. § 101 is also a live lever for a claim set that reads as image-analysis/statistical-modeling applied to diagnostic data; that argument is unreachable in an IPR and is a reason to keep a parallel district-court strategy.

5. Escalate to the family, not the patent. Before budgeting a single-patent challenge, pull the claims of the continuation chain (US 11,925,417; 11,944,385; 12,251,165; 12,274,511; 12,564,447 and any newly issued members) and identify which claims actually read on your product. Challenging 11,877,801 alone risks spending the invalidity budget on a patent the plaintiff can substitute out of.

6. Treat the absence of IPRs as a timing fact, not a validity finding. Well-asserted patents eventually attract IPRs; this one simply has not been asserted hard enough (in public, indexed sources) to draw them yet. The defensive read for today is: no one has handed you a roadmap, but no one has taken away your best art either.

I'll research the assignment chain for this patent using authoritative sources.

I was able to retrieve the patent's own legal-events data and the assignee's corporate history, but I could not reach the USPTO Assignment Center record set in this session. Below is what is documented versus what remains unverified.

Inventors

Pattern note (important, and not an NPE tell here): all five inventors are the assignors on a single assignment executed to their employer (recorded 2020-07-22, per the patent's Google Patents legal-events entry). This is the ordinary employee-to-employer assignment, not a mass inventor exodus. The relevant corporate event is not an inventor departure but a change of control of the assignee itself: Medicrea — including its founder/CEO Sournac — was acquired by Medtronic via a friendly tender offer in which "Medicrea's largest shareholders, including founder, president and CEO Denys Sournac as well as certain other key managers, employees and directors," committed to tender. I found no evidence of any inventor leaving Medicrea within 12 months of the 2019-04-02 priority filing; the triggering event (19 months later) was the Medtronic tender offer, not attrition.

Original assignee

Medicrea International SA (also rendered as Medicrea International S.A. / S.A.S.), Lyon/Rillieux-la-Pape, France — matching both the patent's "Original Assignee" field and its "Current Assignee" field on Google Patents.

• Primary line of business: surgical implants for complex and personalized spinal surgery; specifically the UNiD ASI™ (Adaptive Spine Intelligence) platform — AI-driven preoperative planning plus patient-specific 3D-printed titanium spinal rods. Medicrea obtained FDA clearance for custom pre-bent rods in late 2014 and described itself as the first to commercialize made-to-measure spinal implants (Medicrea 2016 AGM presentation; Medtronic 2020 release citing "30+ 510(k) cleared or CE Marked implant technologies" and "a database of over 6,000 surgical cases").
• Did they ship a product embodying the claims? Substantially yes. The claims are directed to generating patient-specific spinal treatments/rod specifications from preoperative imaging using predictive modeling. The UNiD ASI platform is precisely that commercial workflow (planning + personalized rod manufacture), and the Medtronic release describes it as "powered by predictive modeling and sophisticated algorithms that measure and digitally reconstruct a patient's spine to its optimal profile."
• Current status: Operating, as an acquired subsidiary. Medtronic plc (NYSE: MDT) completed the friendly tender offer on 2020-11-16, owning >90% of Medicrea's capital and voting rights and requesting a French squeeze-out so that Medicrea became a wholly-owned Medtronic subsidiary. The entity remains legally alive: a 2024 Vietnamese device-registration filing lists the owner as MEDICREA INTERNATIONAL S.A.S., with the owner explaining that the trailing "S" reflects "the change in legal status after Medtronic bought Medicrea." Not dissolved, not in bankruptcy.

Assignment timeline

⚠️ Limitation up front: I could not obtain reel/frame numbers. The USPTO Assignment Center (https://assignmentcenter.uspto.gov/; mirror at https://assignment.uspto.gov/patent/index.html) is the primary source required for this section, and it did not return an indexed record for US 11,877,801 in the searches available to me. I did not locate any reel/frame citation, any correspondent of record, or any post-issuance record. I am therefore reporting only the assignment documented in the patent's own legal-events data and the documented corporate acquisition. I have not invented reel/frame numbers.

• 2020-07-22 (recorded) — Reel not retrieved / not located — Conveyance: Assignment (reassignment)

  • Assignor: SOURNAC, DENYS; Javelot, Christophe Xavier Guillaume; AUGAGNEUR, CELINE; MOSNIER, THOMAS; Senac, Agathe (all five named inventors)
  • Assignee: MEDICREA INTERNATIONAL
  • Correspondent: not determinable from sources reached — the Assignment Center record, which is the only place the correspondent of record is exposed, was not retrievable. I cannot assess the "repeat correspondent" signal for this chain.
  • Context: Routine employee/inventor-to-employer assignment of the application (filed 2020-04-01) to the operating company — the ordinary first link, not an acquisition or fire-sale.

• 2020-11-16 — no USPTO assignment located (transaction is a share purchase, not a patent transfer)

  • Conveyance: Change of control via friendly tender offer (Medtronic plc acquires >90% of Medicrea's share capital; squeeze-out to follow). Medtronic news release, 2020-11-16.
  • Assignor/Assignee (patent-record sense): none — because this was an equity acquisition, title to the patent remained with Medicrea International and no patent-level assignment would ordinarily be recorded. A French-law squeeze-out would typically produce, at most, a Change of Name or no recording at all. I could not confirm whether such a name-change record exists.
  • Context: Acquisition (change of control), not a transfer to an asserting entity.

No further assignments located. No security agreements, no mergers, no license recordals, no releases, and no transfers to any LLC were found for this patent.

Cross-references attempted: Google Patents legal events (only the 2020-07-22 inventor assignment appears); Medtronic newsroom / 8-K-level coverage of the acquisition; Medicrea investor-relations materials. SEC filing caveat: Medicrea International was a Euronext Growth Paris issuer (ALMED), not an SEC registrant, so there are no Medicrea 10-K/8-K filings to mine for assignment context. Medtronic plc is SEC-registered, but no Medtronic filing I found treats this specific patent as a transferred asset — consistent with a share-purchase structure. RPX / Unified Patents asserter directories: I found no listing of Medicrea International or any related entity as a high-frequency plaintiff; my search budget was exhausted before I could do an exhaustive check, so treat this as "not found," not "confirmed absent."

Timeline diagram

timeline
    title Ownership of US 11877801
    2019 : Priority application filed
    2020 : Nonprovisional filed by Medicrea
         : Inventors assign rights to Medicrea
         : Assignment recorded 22 Jul 2020
         : Medtronic completes Medicrea buyout
    2024 : US 11877801 issues
         : Continuation family members issue
    2025 : Family continues to mature

NPE / troll-pattern signals

1. Shell-entity transfer — Not present. The only recorded assignment runs to the operating company (Medicrea International), not away from it. The current assignee of record is a French implant manufacturer that shipped FDA-cleared product (UNiD ASI). No "IP / Holdings / Ventures / Licensing" entity appears anywhere in the chain.

2. Known asserter in the chain — Not present. No assignee in the chain matches Acacia, Marathon/IPNav, Intellectual Ventures, Wi-LAN, Mosaid/Conversant, Vringo, Pendrell, Innovatio, MPHJ, Lumen View, Round Rock, Document Generation Corp, or any Spangenberg-linked entity. The only entities are Medicrea International and (by equity) Medtronic plc.

3. Repeat correspondent across the chain — Unclear. This is the one signal I genuinely cannot reach. The correspondent of record is exposed only on the Assignment Center entry, which I could not retrieve; with a single documented link there is no recurrence to test anyway. Flagging this as a data gap rather than clearing it — if you can pull the 2020-07-22 record, the correspondent name is the one field worth capturing.

4. Cascading transfers — Not present. One assignment, in 2020. No chain of LLCs, no consecutive transfers within 24 months, no shared correspondent addresses.

5. Pre-litigation transfer — Not present. I found no infringement suit naming US 11,877,801 at all, so there is no pre-suit transfer to time against. The chain contains no assignment to an asserting party in any window.

6. Bankruptcy fire-sale — Not present. The opposite occurred: Medicrea was sold in a friendly, all-cash voluntary tender offer at €7.00/share, a 22% premium over the prior close and 56% over the three-month VWAP (Medtronic/Medicrea joint release, 2020-07-15). That is a strategic premium sale, not a distressed disposition. (For completeness: the AMF later sanctioned a Mr. Reynouard / SR Capital for trading on inside information about the unannounced Medtronic approach — an insider-trading matter concerning the buyer's offer, not a bankruptcy.)

7. Privateering — Not present on the current record, but the structural setup warrants a watch note. The patent now sits inside Medtronic, an operating company, and there is no NPE interposed between Medtronic and the patent. However, Medtronic's own release frames the portfolio as competitively differentiating ("Medtronic will become the first company to be able to offer an integrated solution including AI driven surgical planning, personalized spinal implants and robotic assisted surgical delivery"), and the July 2020 announcement openly cites head-to-head competition with Globus Medical, NuVasive, Stryker, Alphatec, and DePuy Synthes (framed in the Medicrea materials as rival UNiD/"custom rod" programs). That is ordinary operating-company competition, not privateering. No evidence of assertion by a proxy entity.

8. Defensive aggregator (anti-NPE) — Not present. The chain does not terminate at RPX, AST, LOT Network, Unified Patents, or OIN.

Verdict

Insufficient data — specifically, "only the original assignment," per the rubric.

Justification: the only assignment I can document is the 2020-07-22 inventor-to-employer assignment of all five named inventors to Medicrea International, plus the 2020-11-16 Medtronic equity acquisition that left Medicrea as a wholly-owned, still-operating subsidiary holding title. Because I could not retrieve the USPTO Assignment Center record set (no reel/frame, no correspondent of record), I cannot fully reconstruct the chain, and I will not characterize a chain I have not read. That said, every NPE signal I could evaluate is negative: no shell entity, no known asserter, no cascading LLC transfers, no pre-suit transfer, no distressed sale, and no defensive-aggregator termination — with signal #3 marked unclear rather than cleared. On the evidence reached, this reads as an operating-company asset retained through a change of control, not an NPE acquisition.

Verify at: https://assignmentcenter.uspto.gov/ (search US 11,877,801) and https://assignment.uspto.gov/patent/index.html — and specifically capture the correspondent of record on the 2020-07-22 record, which is the open item in this analysis.

Contradictions / open items flagged

• Assignee name form: previous section noted aggregators listing "MEDICREA INTERNATIONAL (USA)." The patent's own data (Google Patents original and current assignee fields) shows Medicrea International SA, and third-party regulatory filings confirm the French entity MEDICREA INTERNATIONAL S.A. → S.A.S. I treat the French entity as authoritative and the "(USA)" rendering as an aggregator artifact (Medicrea did operate a US distribution subsidiary from 2007). Consistent with prior section; no correction needed.
• Claim-numbering caveat from the prior section still stands and is unaffected by anything found here.

I'll search for authoritative prior-art and citation data on US 11,877,801.

I need the specific "References Cited" list for this patent. Let me probe several patent databases.

Let me try to locate the granted patent's front page (which carries the "(56) References Cited" list) and sibling-family listings.

Prior-Art Analysis — US 11,877,801 B2

1. Retrieval status — read this first

I could not obtain the patent's (56) References Cited list (the front-page citation table) with the tools available to me. Specifically:

• The authoritative full text I hold (the Google Patents fetch of https://patents.google.com/patent/US11877801B2/en, dated 2026-09-24) renders the front page and classification sections but omits the "Citations / Cited By / References Cited" tables — every such table in that document came back as blank whitespace.
• My searches for the search-report and bibliography pages returned unrelated documents. Notably, the "patentimages…WO2020172395A1.pdf" hit that appeared to match an application-number query was actually a mixed PDF whose contents (US 2005/0087359 A1, CN 106104933 A, JP 2003‑208928 A) are electrical-connector art and cannot belong to this spinal-implant family.
• I did not reach Patent Public Search / Patent Center / PTAB-PACER raw records before exhausting my search budget.

Therefore: I will not fabricate a citation list. Everything below is either (a) verified from the patent's own text, or (b) expressly labeled as unconfirmed candidate/comparative art. Per the operating rules, where I don't know something with high confidence, I'm flagging it.

2. Critical number-disambiguation finding (do not auto-correct this)

A literal USPTO search for 11877801 surfaced USPTO papers for a different proceeding keyed to the serial number 11/877,801 — application 11/877,801, filed 2007‑10‑24, first named inventor Harsh Kapoor, titled "Systems and methods for processing data flows" (attorney docket CRSB‑0003‑P04/P08, Art Unit 2433, Examiner Michael D. Anderson). See e.g. the IDS/acknowledgement receipt bundle at https://ptacts.uspto.gov/ptacts/public-informations/petitions/1510855/download-documents.

• 11/877,801 (serial, 2007) ≠ 11,877,801 (patent, 2024).
• This is the same trap the prior summary flagged with the number "11877819." Two independent collisions now, and neither has been conflated into the analysis.
• Practical consequence: any prior-art or litigation search keyed loosely to "11877801" will pull in 2007 data-flow software art. Filter on application 16/837,461 or patent 11,877,801, never on the bare digit string.

3. References actually verifiable from the record

Only item 1 is a "citation" in the § 102 sense that I can confirm against this patent. Note that an incorporation-by-reference is not evidence the document was applied as prior art by the examiner — a distinction that matters for the § 102 analysis you asked for.

4. Why I cannot responsibly do the requested per-reference § 102 mapping yet

You asked for, per cited reference: full citation, publication/filing date, description, and the claim(s) it potentially anticipates under 35 U.S.C. § 102. That requires the Form PTO‑892 (examiner's Notice of References Cited) and/or the PTO‑1449/IDS plus the reference texts themselves. I have neither for this patent. Producing a per-reference anticipation chart from memory would be exactly the fabrication the operating rules forbid.

Where the authoritative list lives (for a follow-up run with page-fetch capability):

1. USPTO Patent Public Search → application 16/837,461 → "References Cited" and the "Examiner's search strategy / 892" documents.
2. Patent Center → application 16/837,461 → "Documents & Transactions" → the Non-Final/Notice of Allowance and any PTO‑892.
3. The granted patent PDF front page, field (56) — not the HTML abstract page I was served.
4. EP 2020059254 / WO 2020/201353 search report (the EPO's "DOCUMENTS CONSIDERED TO BE RELEVANT" table with X/Y/A categories), which is a good proxy for the material the family's examiners considered.

5. § 102 framework — what a reference must show to anticipate this patent

Because this is an AIA patent (filed 2020‑04‑01, priority 2019‑04‑02), § 102(a)(1)/(a)(2) govern, with § 102(b)(2)(C) common-ownership potentially in play for Medicrea's own documents. Anticipation requires every element of the claim in a single reference. Applied to the three independent-claim families identified in the prior summary:

• Frequency-domain claim family (Fourier transform → filter out above-threshold components → predict → inverse transform back to spatial domain). A § 102 reference must disclose, in one document, (i) derivation of spinopelvic parameters (LL/TK/PI/PT/SVA) from preoperative images, and (ii) a transform-to-frequency-domain step applied to those parameters, and (iii) threshold filtering there, and (iv) prediction in the frequency domain, and (v) inverse transformation. Art that merely does image-domain low-pass filtering, or that applies a Fourier descriptor to an implant/rod shape rather than to the spinopelvic parameters, would fall short. This is almost certainly the point of novelty, so a genuine § 102 hit is unlikely from generic ML-spine-planning art — that art is more naturally § 103 fodder.
• Prediction claim family (images + non-imaging data → vertebral positions → spinopelvic parameters → predicted outcome in spatial domain, GAN/CNN/RNN). Here a § 102 reference needs the full ladder from vertebral-position extraction to parameter derivation to model output. Broad "predict postoperative alignment from preoperative radiographs using a neural network" disclosures are the strongest candidates and should be charted element-by-element.
• Training claim family (paired pre/post images + non-imaging data → measurements → compression [Fourier or polynomial] → noise-threshold filtering → train → test on disjoint subjects; plus Gaussian-process and vertical-axis-rotation augmentation). The augmentation dependents map onto generic GAN/CNN data-augmentation art; the compression-then-filter-then-train combination is again the discriminator. Note the specification's own express caveat that augmented data is never used for testing — a reference that augments the test set would not anticipate those dependents.
• Possible intraoperative-tracking-module claim (unresolved per the prior section). If it is in fact claimed, the § 102 art would be mechanical: tulip/polyaxial screw-head attachment structures and rod-reduction instruments in A61B17/70xx (the classification string I retrieved from the patent confirms A61B17/7001, 17/7032, 17/7074–17/7082, A61B34/20). I have not verified that this subject matter is claimed, so I have not charted it.

6. Candidate / comparative art — expressly NOT confirmed as cited of record

These surfaced in my searches as documents in the same citation neighborhood or same assignee portfolio. I am labeling them "candidate" because I could not confirm they appear on any 892/ISR for 11,877,801. Do not treat the claim mapping as an examiner position.

I am not listing any ML-spine-planning or Fourier-descriptor references by name, because I did not retrieve any with enough confidence to reproduce a citation accurately, and an inaccurate citation is worse than an acknowledged gap.

7. Bottom line

1. The requested deliverable — a per-citation § 102 anticipation chart — cannot be produced from the sources I reached, because the patent's (56) list is not present in the authoritative full text I was given and I could not retrieve the PTO‑892/ISR before exhausting searches. This is a retrieval gap, not a finding that no citations exist.
2. One citation is verified from the document itself: U.S. Pat. No. 10,292,770, incorporated by reference under 37 C.F.R. § 1.57. I could not confirm its title or issue date from a reliable source, so I have not stated them.
3. A number-collision hazard is confirmed and documented: USPTO serial 11/877,801 (Kapoor, "Systems and methods for processing data flows," 2007) is a different matter entirely from patent 11,877,801. Together with the earlier "11877819" collision, any automated search on this ID needs a hard filter on application 16/837,461.
4. Most likely § 102 outcome: given the transform→filter→predict→inverse-transform loop and the compression/noise-filter training recitations, straight anticipation of the independent claims is improbable from the generic spinal-ML art; the realistic contest is § 103. The mechanical intraoperative-tracking subject matter (if claimed) is the part most exposed to § 102 on A61B17/70xx art.
5. To close the task, re-run with page-fetch access to: (a) the granted PDF's field (56); (b) Patent Center documents for 16/837,461 (PTO‑892); (c) the EP 2020059254 / WO 2020/201353 search report. Those three will yield the exact reference list and the examiner's own claim mapping.

I'll start by retrieving the prior-art references cited on this patent's face, since the copied page text stripped out the Citations/Prior Art tables.

I need the actual cited-reference list. Let me pull more of it.

I have partial front-page citations. Let me hunt for the technically relevant cited art (frequency-domain / ML-diagnostic references) and the foreign/NPL citations.

Obviousness analysis — U.S. Patent No. 11,877,801 (35 U.S.C. § 103)

Framework note / disclosure of limitation. The Google Patents page for US 11,877,801 that was supplied to me had its "Citations," "Cited By," "Similar Documents" and "Prior Art" tables stripped out of the copied text. I therefore reconstructed what I could from (a) the granted patent's own front-page "References Cited" block as it appears in the issued PDF, and (b) references that cite US 11,877,801 back. That reconstruction is partial and imperfect (details below). Everything I could not verify, I say so explicitly rather than filling in. I do not repeat the bibliographic/abstract material already generated above.

1. What is actually in the "Prior Art" section (as retrievable)

From the issued front page (US11877801 PDF, front page):

• U.S. Patent Documents — a very long list (~200 entries) that I could only partially recover. The recoverable run begins 4,382,438 A (Jacobs), 5,006,984 A (Steele), 5,163,440 A (DeLuca), 5,209,752 A (Ashman), 5,224,035 A (Yamashita), 5,521,127 A (Raab), 5,291,001 A (Graf), 5,305,203 A (Raab), 5,312,403 A (Korotko), 5,413,116 A (Radke), 5,561,180 A (Heggenss), 5,567,506 A (Sutterlin), 5,748,767 A (Raab), 5,785,663 A (Sarvazyan), 6,015,409 A (Jackson), 6,013,958 B (Winder), 6,282,437 B1 (Franck), 6,302,888 B1 (Mellinger), 6,364,849 B1 (Wilcox), 6,385,475 B1 (Cinquin), 6,490,684 B1 (Wilk), 6,443,953 B1 (Perra), 6,499,488 B1 (Hunter), 6,565,519 B2 (Benesh), 6,585,666 B2 (Suh), 6,711,432 B1 (Krause), 6,715,213 B2 (Richter), 6,716,213 B2 (Shitou), 6,746,449 B2 (Jones), 6,786,930 B2 (Biscup), 7,066,938 B2 (Slivka), 7,538,526 B2 (Steinberg), 7,509,183 B2 (Liu), 7,534,263 B2 (Burdulis), 7,542,791 B2 (Mire), 7,57… (truncated).
• Foreign Patent Documents — a run of WO publications: WO 17001851, WO 16137347, WO 16148675, WO 16165030, WO 17059596, WO 17064719, WO 17066513, WO 17077356, WO 17170965, WO 17172838, WO 17151949, WO 17172237, WO 18045086, WO 18055494, WO 18055183, WO 18078536, WO 18078758, WO 18131044, WO 18130154, WO 18183314, WO 181185755, WO 18193316.
• Incorporated by reference (specification, not the §102 table): U.S. Pat. No. 10,292,770.

Two cautions I want on the record. First, the two-letter OCR strings above are as printed in the source I retrieved; several are internally irregular (e.g., 5,521,127 A is printed with a 10/1993 date, and WO 181185755 is malformed). Per my operating rules I have not auto-corrected them; their true identities need verification against the paper/PACER-quality copy. Second, the top of this US list is overwhelmingly spinal-fixation hardware and image-guided-navigation art (Raab is the foundational surgical-navigation family: 5,521,127, 5,305,203, 5,748,767; 6,385,475 Cinquin is computer-assisted/robotic surgery; 7,538,526 Steinberg is surgical navigation). That distribution matters for the analysis below: the hardware/navigation half of the claims has deep, old art; the frequency-domain half is where any §103 fight will be won or lost.

For the technically material side I could verify these additional references exist and what they say:

Important date arithmetic. Google lists the priority date as 2019-04-02, but the specification also claims benefit of provisionals filed 2019-11-08 and 2019-12-23. Whether the frequency-domain claim elements enjoy the April date or only the December date is a claim-by-claim question that materially changes the §103 field of art. This should be resolved before any invalidity position is finalized.

2. Level of ordinary skill and the controlling standard

For these claims the POSITA is a hybrid: a computational engineer/data scientist (medical image processing, statistical modeling, neural networks) working with spine-surgery domain input, or a spine engineer with equivalent programming/ML skill. That person, as of 2019, knew: landmark extraction from sagittal/frontal radiographs to get LL/TK/PI/PT/SVA; regression and neural-network prediction of alignment; Fourier analysis, low-pass filtering and their use as compression/denoising; train/test splits and cross-validation; Gaussian-process synthetic sampling; and image augmentation by rotation/mirroring. Under KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007), a combination is obvious where the references, the nature of the problem, or common sense supply the motivation, and where the combination is "a predictable use of prior-art elements according to their established functions." In re Boesch / In re Kuhle also make clear that selecting a threshold or parameter value absent demonstrated criticality is within ordinary skill.

3. The obviousness theories

Combination 1 — Claim 1 family (Fourier/hook), over predictive-planning art + frequency-domain signal processing

What the primary art teaches. The Medicrea platform as publicly disclosed before April 2019 (UNiD ASI: measure spinopelvic parameters → surgeon sets a target → predict compensatory mechanisms → output a rod). US 10,292,770 (incorporated into the very specification at issue) frames the identical "virtuous cycle," including "machine learning, and/or predictive modeling." US 10,413,365 B1 shows deriving the corrective geometry from a sagittal radiograph. US 11,903,655 shows predicting a postoperative correction from digitized vertebral positions. These supply every element of the claim except the frequency-domain step.

What the secondary art teaches. A POSITA's own field supplied the rest: Fourier transformation followed by low-pass filtering with a defined cutoff, and inverse transformation to return to the spatial domain, is a textbook compression/denoising pipeline. The specification concedes this framing. From the FIG. 6 and FIG. 7 description:

"the system can be configured to utilize a low-pass filter for data compression, which can be advantageous for machine learning purposes, while preserving most of the relevant information."

"the system can be configured to utilize one or more data compression techniques, such as Fourier transformation and/or a polynomial function."

Motivation, expressly supplied by the specification itself (usable as an admission of what the art would recognize as desirable): (i) reduce the number of model input parameters; (ii) remove noisy high-frequency content that "can comprise a large number of noise or unclear data, thereby reducing the quality of data"; (iii) make the model "more complex"/more trainable. These are recognized, pre-existing engineering needs. Combining a known predictive-modeling pipeline with a known compression/denoising front end is the paradigm "predictable use of a prior-art element according to its established function."

Predictable variation, not invention, in the remaining limitations:

• "one or more predictive models … GAN … CNN … RNN" — the specification itself lists these as interchangeable options and even as combinations (CNN+SVM). Choosing among a finite, identified set of known ML architectures to fit a known data type (sequence → RNN; image/curve array → CNN; generative output → GAN) is exactly the KSR "finite number of identified, predictable solutions" scenario.
• "frequency level above a predetermined threshold" — a filter cutoff is a design parameter; absent evidence of criticality, In re Kuhle applies.
• "inverse Fourier transformation" — the necessary and only sensible way to return a frequency-domain prediction to a physically usable rod geometry; no separate inventive weight.

Realistic counter-argument (and its weakness). Applicant will argue that the art teaches predicting in the spatial domain, and that moving the model into the frequency domain is not suggested. The counter is that the stated purpose of the transform (fewer parameters, less noise) is itself the motivation, and the specification states the benefit rather than demonstrating surprise. This is a fight over "motivation" and "predictable result," not over missing elements.

Combination 2 — Dependent claims reciting the intraoperative tracking module (if claimed)

Verified art: the face-of-patent navigation family (Raab 5,521,127/5,305,203/5,748,767; Cinquin 6,385,475; Steinberg 7,538,526), plus EP 3431032 A1 (Mazor, published 2019-01-23) which expressly stores a screw's angular orientation and location in a surgical-plan data structure. The patent is itself classified in A61B17/7076–7082, the classes for screw-driving/positioning/anchor-moving tools — i.e., the sleeve/extension instrument class.

The added structure is a battery pull-tab plus a screw-head-mounted sleeve: two substantially perpendicular conduits (tool channel + rod channel), a wider top than bottom for rod capture, and notches that clip to a tulip-screw head. Each feature performs its conventional function (guide a tool, guide a rod, retain on the screw head, prevent accidental discharge of a battery). Activation by removing a blocking strip is a ubiquitous, decades-old convention in battery-powered medical devices. Motivation to combine: intraoperative verification of the preoperatively planned alignment in real time — the stated purpose of the whole platform. This claim family looks the weakest of the set.

Note: the strip/power-circuit language appears verbatim in the 2025 continuation US 2025/0152249 A1 claim 1 (FPO copy). I could not confirm that the granted US 11,877,801 carries this element — the prior section flagged this as unresolved, and my retrieval does not resolve it. Treat Combination 2 as conditional.

Combination 3 — Claim 13/14 family (spatial-domain prediction)

Elements: images + non-imaging data in; measurements out, including vertebral position; spinopelvic parameters derived from the measurements; predictive outcome out; rod specifications out; sagittal + frontal x-rays; GAN/CNN/RNN. This is a straightforward KSR combination of (a) x-ray landmark/endplate extraction art (the Surgimap/UNiD workflow, plainly public and commercialized pre-2019) with (b) supervised-learning regression of postoperative alignment on preoperative spinopelvic parameters. The specification's own justification for extracting vertebral position before computing angles ("some information can be lost if measurements are taken directly from x-ray images") is stated as an advantage but is not shown to be critical. Motivation: every prior-art planner needs input parameters, and deriving more of them from fewer measurements is an ordinary optimization.

Combination 4 — Claim 21 family (training method)

The train/test partition on subjects distinct from the training set is the definition of held-out validation, described in the specification itself as conventional ("testing an algorithm on the same data it was trained on may not provide accurate testing results"). The Gaussian process augmentation is a known synthetic-data technique, and the specification admits its purpose ("to increase the learning dataset for training the predictive model"). 180° rotation/mirroring of spine images to manufacture training examples is a stock item of medical-image augmentation; the specification recites the purpose (more data), not an unexpected result. The only claim-specific grace note — that augmented data is used for training and never for testing — is a methodological hygiene rule, not a technical advance; it is the kind of limitation a POSITA applies by rote to avoid data leakage.

4. Secondary considerations

There is a genuine commercial story: UNiD ASI, FDA clearance in 2014, "over 4000 cases," competitor entry. But under In re GPAC and WBIP, the nexus must run to the claimed feature. The Fourier/low-pass limitation and the Gaussian-process augmentation are not what 4,000 surgeons bought; the platform's pre-existing planning-and-rod-manufacturing workflow is. Absent evidence that the frequency-domain step drove adoption or solved a problem others had failed at, the secondary considerations do not rescue these claims.

5. Where the patent is strongest (and what to test)

1. The frequency-domain predictive step is the only genuinely uncommon limitation. A rigorous §103 position needs a concrete secondary reference teaching low-pass filtering of spinal curvature/alignment data, not merely general signal-processing knowledge. Prior-art families worth targeting: harmonic/Fourier analysis of scoliotic and sagittal spinal curves, and use of low-order polynomial/harmonic descriptors of the spine as a compact feature vector. I searched for these but could not retrieve a specific qualifying reference within my step budget — I am not asserting one exists.
2. Priority-date exposure. If the frequency-domain limitations are supported only by the November/December 2019 provisionals, the field of art widens considerably for that limitation.
3. Common-ownership complication. US 10,292,770 is incorporated by reference and shares inventors with the patent-in-suit. As a §102(a)(2) reference it would likely be disqualified by the §102(b)(2)(C) common-ownership exception; the pre-2019 WO publications in the foreign-documents list, by contrast, are §102(a)(1) printed publications and are not disqualified by common ownership. Those WO documents are the most promising statutory-prior-art candidates on the page and I could not verify their contents.

6. Conclusion

• Claim 1 family: Likely obvious under §103 as a predictable combination — a known spinal-outcome predictive pipeline (Medicrea's own disclosures, US 10,292,770, US 10,413,365, US 11,903,655, EP 3431032) plus the admitted, conventional use of Fourier transformation with low-pass filtering for compression/denoising, with the model architecture selected from a finite list the specification itself enumerates. Confidence: moderate-to-high, contingent on producing a specific secondary reference for low-pass filtering of spinal alignment data.
• Claim 13/14 family: Likely obvious; the elements are individually and collectively conventional. Confidence: high.
• Claim 21 family (training): Likely obvious; held-out testing, Gaussian-process augmentation, and rotational augmentation are routine. Confidence: high.
• Intraoperative tracking module claims (if present in the granted claims): Likely obvious over the navigation art on the patent's own face plus EP 3431032. Confidence: moderate-to-high, conditional on the claims actually issuing in that form.

Explicit non-findings. I did not verify the substance of the ~200 U.S. references or the ~22 WO references printed on the front page; I did not locate the non-patent-literature citations; and I did not confirm the granted claim set of US 11,877,801 beyond the language carried over from US 2020/0315708 A1 and US 2025/0152249 A1. Any statement above that depends on those gaps is flagged as such rather than asserted. Nothing here contradicts the earlier-generated sections; the one item it advances is the intraoperative-tracking question the earlier section left open — the strip/power-circuit limitation is confirmed in the 2025 continuation, but still not confirmed in the granted patent.