US 9372134

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

Let me run targeted searches for litigation and USPTO ownership records.

Let me check the granted claims text directly, since I want to describe the independent claims accurately.

US Patent 9,372,134 B2 — Analyst Summary

Scope note / search disclaimer: I was asked to search USPTO and CAFC 2026 dockets. My searches surfaced the Google Patents record and third‑party patent databases, but I found no CAFC 2026 docket, appeal, or district‑court litigation involving US 9,372,134. I cannot claim an exhaustive docket check, so treat "no litigation found" as absence of evidence in my searches, not proof of none. One search result also displayed the title "Fluid interface cartridge for a microfluidic chip" next to "US9372134B2" — that is an artifact of a Google Patents similar‑documents/citation table (the page was WO2010118427A1). Per your strict rule I did not adopt that title; the record title below is the one from the '134 patent itself.

---

Bibliographic data (from the patent record)

Representative figure listing: FIGS. 1–10 (first embodiment), 11–17 (second), 18–24 (third), 25–28 (fourth).

---

Abstract (verbatim paraphrase of record)

A micro-fluidic device includes a filter case, a first capture array, and a second capture array. The filter case has an inlet for a sample containing different kinds of targets, an outlet for discharging the sample, and a channel between them. The first capture array sits in an upstream portion of the channel and has a plurality of first forward funnels arranged orthogonally to the sample flow direction to capture the targets. The second capture array sits in a downstream portion of the channel and has a plurality of second forward funnels likewise arranged orthogonally to capture the targets.

---

Plain‑language overview of the independent claims

The record provides verbatim claim‑style language for two independent claims (device and method). Note: the granted patent's dependent claims are described extensively in the specification (reverse funnels, dispersing‑unit posts, obstruction bodies, partition plates, discharge path) but I do not have an independently verified verbatim dependent‑claim set, so I limit the claim overview to the independent claims below.

Independent Claim 1 — Micro‑fluidic device (apparatus)
A device built around a filter case containing:

1. an inlet that introduces a sample carrying different kinds of targets, an outlet for discharging sample, and a channel running between them;
2. a first capture array in an upstream part of the channel, made of multiple first forward funnels lined up across the channel (perpendicular to flow) to catch the different target types; and
3. a second capture array in a downstream part of the channel, made of multiple second forward funnels likewise arranged across the channel to catch the different target types.

In substance: size‑based capture is achieved by two (or more) stages of tapering funnel structures spanning the channel, rather than by antibody/binding‑moiety‑coated posts.

Independent Claim 2 — Target separating method
A multi‑step method using such a device:

1. Provide a micro‑fluidic device whose filter case has a channel for the target‑bearing sample and multiple capture arrays in multiple stages along the flow direction. Each array has forward funnels across the channel and a "reverse flow guide means" that lets captured targets flow opposite to the sample flow direction when the filter case is mounted overturned.
2. Supply the sample into the channel so the forward funnels capture the different kinds of targets.
3. Turn the filter case upside down so the forward funnels are inverted (their tapered direction reverses).
4. Supply a carrier fluid into the channel to flush the captured targets out of the filter case.
5. Collect the discharged targets.

In substance: capture in one flow direction, then flip the cartridge and back‑flush to harvest — which is the stated advance over binding‑moiety capture, whose collected cells are hard to release.

---

Supporting structure described in the specification (context for the claims)

• Dispersing units 20‑1/‑2/‑3: staggered post arrays (posts 22 narrowing from ~300–400 µm → 200–300 µm → 100–200 µm) that spread the sample transversely to raise capture rate.
• Forward funnels 40 (entrance 42 upstream → exit 44 downstream, tapering down) and reverse funnels 50 (entrance downstream → exit upstream) arranged alternately across the channel; the reverse funnels constitute the claimed "reverse flow guide means."
• Exit widths described as ~15–20 µm (capture ≥15 µm targets), ~10–15 µm (capture ≥10 µm targets), minimum ~5 µm; red blood cells (6–8 µm) pass because cytoplasm deforms, while less‑deformable target cells are retained.
• Obstruction bodies 64 / 66 assist capture / guide reverse flow, and may be omitted (FIG. 7).
• Stand 70/170 with slant table and holder, sample supply syringe 82, carrier‑fluid syringe 92, and (third/fourth embodiments) stacked partition plates 200 dividing the channel into parallel sub‑channels for throughput.

---

Points of uncertainty (stated honestly)

1. Claim set: I verified the two independent claims as reproduced in the patent, but I have not independently confirmed the exact number/text of dependent claims in the granted '134 patent.
2. Litigation: No CAFC 2026 docket entry for 9,372,134 appeared in my searches, nor did any district‑court case. I cannot certify that none exists in PTAB/PACER/CAFC records I could not fully query.
3. Assignee naming: The record shows three related names — "Cytogen Inc," "CYTOGEN CO., LTD." (2012 assignment), and "Cytogen Co., Ltd." (citation tables). These are legal‑entity naming variations in the record, which I am reporting literally rather than normalizing.
4. Family: The '134 patent shares a 2010‑04‑15 KR priority date with KR101254680B1, KR101226515B1, and KR101254679B1, and has a US pre‑grant publication US20130074613A1. I did not confirm the full international family.

I'll search for litigation involving US patent 9372134 across patent litigation databases and court resources.

Let me search more specific litigation databases and the patent's assignee.

Let me search a few more targeted sources directly.

Litigation Search Results — U.S. Patent No. 9,372,134

Bottom line

I was unable to identify any litigation involving U.S. Patent No. 9,372,134 (US 9,372,134 B2). Based on the public sources I was able to search, there is no known district court, Federal Circuit, ITC, or PTAB proceeding asserting or challenging this specific patent.

What I searched and what I found

I ran targeted searches using the exact number 9372134 (and the forms 9,372,134 and US9372134B2) against sources that index patent litigation. The results did not surface any case asserting this patent. The hits fell into these categories, none of which concern the '134 patent:

Important caveats

1. Sources I could not fully query. My access to live databases was limited. The Unified Patents litigation portal (https://portal.unifiedpatents.com/litigation/caselist), PACER, CourtListener/RECAP, and the ITC EDIS system are the authoritative places to confirm a "no known litigation" result, and I was not able to exhaustively query them for this specific patent number. A definitive negative requires a direct query of those systems (e.g., CourtListener's docket search filtered on the patent number, and PACER's patent docket reports).

2. Scope of "exhaustive." Public web search can miss older, settled, or sealed cases, and cases that named the patent only in an exhibit rather than the complaint.

3. The related microfluidic patents in the Inguran/ST disputes show that this technology space (microfluidic cell sorting/isolation) does generate litigation — but none of those cases involved the '134 patent specifically, and I found no evidence tying Cytogen (the '134 assignee) to any asserted case.

Recommendation

To obtain an authoritative answer, run these checks directly:

• CourtListener / RECAP: docket search for "9,372,134" or "9372134"
• Unified Patents Portal: https://portal.unifiedpatents.com/litigation/caselist filtered by patent number
• USPTO Patent Trial & Appeal Board (PTAB) API / PatentCenter for any IPR/PGR/CBM naming the patent
• Google Patents "Litigation" tab on the US9372134B2 page (it shows a litigation section when litigation exists)

Unless one of those systems returns a hit, the appropriate conclusion is that no litigation is known for U.S. Patent No. 9,372,134.

I'll verify the structured data with targeted searches before drafting.

PTAB proceedings

No AIA trial proceedings found. I checked the canonical structured list first and then ran independent web searches; I found no IPR, PGR, or CBM petition ever filed against US 9,372,134.

---

Proceedings overview

Total proceedings on file: 0. The USPTO Open Data Portal returns no AIA trial proceedings for US 9,372,134, and my independent searches (patent number, patent title, "Cytogen" + IPR/PTAB, and assignee-based queries) surfaced nothing further. That means there is no claims invalidated / claims sustained / settled / institution denied breakdown to report — every claim of the patent is untested at the Board.

Defensive posture: the opposite of a hardened patent by PTAB metrics — but also the opposite of a paper tiger. Nothing has been canceled, so every claim you might be accused of infringing is still live and carries a statutory presumption of validity. The real signal here is the absence of activity: this is a thinly asserted patent with no IPR track record, which usually means either (a) the patent owner has pursued licensing quietly without triggering district-court suits, or (b) assertion is recent enough that no defendant has yet cleared the 35 U.S.C. § 315(b) one-year window to petition. For a defendant, that is an open field, not a hardened patent.

No proceedings to itemize

There is no proceeding number, petitioner, panel, institution decision, Final Written Decision, settlement, or Federal Circuit appeal to report for US 9,372,134. Per the constraints of this task, I will not manufacture one.

Two things I did confirm while searching, which are worth flagging because they are common sources of false positives:

---

Strategic summary

Claim status. Every claim of US 9,372,134 is UNTESTED at the PTAB. Nothing is canceled, nothing is sustained on the merits, nothing is narrowed by certificate. Statutorily, no claim of the '134 has been invalidated by the Board. Independent claim 1 recites the microfluidic device — a filter case with an inlet, an outlet and a channel, a first capture array of forward funnels arranged orthogonally to sample flow in an upstream portion of the channel, and a second capture array of second forward funnels likewise arranged in a downstream portion. The specification also describes a target-separating method aspect (supplying the sample, turning the filter case upside down so the forward funnels are overturned, supplying a carrier fluid, and collecting the targets), which appears as method claims in the patent. Because no FWD exists, I cannot state with confidence which claim numbers carry the apparatus vs. method subject matter beyond what the specification's summary of the invention describes — do not treat any claim-number mapping as verified unless you pull the printed claims.

Estoppel landscape. § 315(e)(2) estoppel is not triggered at all — estoppel attaches only after a Final Written Decision, and there is none. No petitioner and no privy is barred from any ground. Practically, a defendant today can raise any ground available under § 311(b): anticipation and obviousness under § 102/§ 103 based on patents and printed publications. Note the statutory ceiling on an IPR — no § 112 grounds, and no public-use, on-sale, or prior-knowledge art in an IPR. If your best invalidity theory is a § 112 written-description or indefiniteness attack on the "funnels," "forward funnels"/"reverse funnels," or "staggering pattern" geometry, the IPR vehicle will not carry it; that theory lives in district court or in an ITC/§ 282 defense. Also note this is a pre-AIA patent (priority 2010-04-15, filed 2011-04-15), so IPR grounds are governed by pre-AIA § 102/§ 103, and the § 102(e) treatment of certain pre-2013 published applications can matter.

Vehicle availability. Because the effective filing date predates 2013-03-16, PGR is unavailable. CBM is likewise unavailable and, in any event, the covered-business-method program sunset. IPR is your only AIA trial vehicle on this patent.

Pattern signals. No petitioner has filed anything — so there is no serial-petitioner pattern, no follow-on petitions, and no defensive aggregator (Unified Patents or similar) in the chain on this patent so far. Equally, there is no PTAB appeal history, which means the patent owner has never had a claim confirmed at the Board and has never had to defend one. That is a two-edged fact: the claims have never been tested against a well-funded invalidity attack and the owner has never demonstrated they can survive one. Given the patent's adjusted expiration of 2033-03-01, there is a long royalty tail, which is exactly the fact pattern that usually eventually attracts an IPR — either from a district-court defendant or from a defensive aggregator filing preemptively to protect the sector.

---

Recommended next steps

1. Confirm the negative with primary sources tied to your own case, not just this report. Pull the prosecution history and file wrapper for US 9,372,134 via USPTO Patent Center, and check PTAB E2E / the USPTO Open Data Portal for the patent number directly. Absence of PTAB activity is a real finding, but it should be verified against the docket for the specific litigation or demand you are facing.

2. Since no claims are invalidated, there is no FWD to link to and no claim-level disposition to quote. The opportunity here runs the other way: if you decide to petition, you will be the first mover, you will not face § 315(e) estoppel, and you will not be met with a Patent Owner who has already litigated the same art before the Board.

3. Watch the § 315(b) clock. If you have been served with a complaint asserting the '134, you have one year from service to file an IPR or you are barred entirely. There is no prior petition to piggyback on or coordinate with, so the entire invalidity budget is yours to allocate.

4. Evaluate the § 112 theories early, in parallel with the IPR art search. Because IPR cannot reach § 112, and because claim 1's funnel-geometry limitations ("arranged along a direction orthogonal to a flow direction," "first forward funnels," "second forward funnels," the staggering relationship between arrays) are the kind of relative-orientation language that generates indefiniteness and written-description fights, decide up front whether your strongest theory is an IPR-eligible art attack or a district-court § 112 attack. Splitting the theories across fora without a clear lead theory is how defendants end up with an instituted IPR and no stay.

5. If your dispute is really with the broader Cytogen family rather than the '134 alone, run this same check on each asserted sibling (US 2013/0071304 and the "Cell collecting device" family, plus EP2560000B1 for European exposure). PTAB coverage across a family is frequently uneven, and the '134 being clean says nothing about whether a sibling has already been narrowed.

Confidence and gaps: I have high confidence that no PTAB proceeding exists on US 9,372,134, based on the structured ODP list plus multiple independent searches. I have no confident knowledge of any U.S. district-court litigation asserting this specific patent, and I did not locate one — if a complaint exists, it did not surface in my searches, so treat the litigation-activity inference in the "Pattern signals" paragraph as an inference from silence rather than a verified negative. I also did not independently verify the patent's printed claim set (numbering and scope of each claim), so do not rely on claim-number characterizations here without pulling the patent text.

I'll search for the assignment records and ownership chain for this patent.

Ownership & Assignment Chain — US Patent 9,372,134 B2

Analyst caveat up front (read before using): I could retrieve the parties and dates of the recorded assignment from the Google Patents legal‑events table, but I was unable to retrieve the USPTO reel/frame numbers or the correspondent of record for this patent — the Assignment Center search at https://assignmentcenter.uspto.gov/ was not directly queryable in this session, and no third‑party index surfaced the reel/frame for US 9,372,134. Every reel/frame slot below is therefore marked [NOT RETRIEVED]. Do not treat that as "no record" — the record exists (Google Patents legal events shows one assignment), I simply could not read its frame number. This is a material gap for one of the signals (repeat correspondent) and I flag it rather than guess.

This builds on the prior sections' noted uncertainty #3 (assignee naming variants) — that naming issue is resolved below and is not a separate assignment.

---

Inventors

Pattern note — not a fire‑sale signal. The classic red flag you asked me to watch for ("all inventors departing the original assignee within 12 months of filing, often preceding a portfolio fire‑sale") is absent here, and in an unusual way: the sole inventor is the controlling principal of the assignee. Byung‑Hee Jeon has served continuously as CEO & Director of Cytogen (Yahoo Finance / Investing.com / StockAnalysis.com profiles, all listing Jeon as CEO of KOSDAQ:217330). There is no inventor‑departure signal to find — the inventor and the assignee are the same economic actor, which is a founder‑inventor/owner pattern, not an employee‑inventor-abandonment pattern. Jeon's patent leaderboard entry shows 10 granted US patents, all tied to Cytogen (first 2009, most recent Oct 2023), i.e. an anchor inventor who stayed.

---

Original assignee

Entity on the issued patent: Cytogen Inc. (Google Patents "Original Assignee" field), which the patent record's assignment event renames in Korean corporate form as CYTOGEN CO., LTD. (㈜싸이토젠).

• Primary line of business: an operating in‑vitro diagnostics / liquid‑biopsy company. It commercializes a CTC (circulating tumor cell) platform that enriches CTCs from blood using high‑density microporous chips — i.e., size‑based capture, the very mechanism claimed in the '134 patent. Products: Smart Biopsy™ Cell Isolator, IF Stainer, Cell Image Analyzer, Isolation/Culture Kits (StockAnalysis.com company profile; Crunchbase; Cytogen's own IR deck at cytogenlab.com, which lists Cell Isolator/IF Stainer/Cell Image Analyzer among platform hardware).
• Product‑embodying the claims? Yes, credibly. Cytogen's own filings describe a "Cell Isolator" that "enrich[es] CTCs from human blood … using high density microporous chips," and the '134 specification expressly targets "white blood cells having a diameter of from 12 µm to 25 µm from the human blood" via staged funnel exits (15–20 µm / 10–15 µm / min 5 µm). The claimed subject matter maps onto the shipped instrument, not to a bare license.
• Current status: Operating and publicly listed — KOSDAQ: 217330, KOSDAQ listing Nov 22, 2018 (company history slide, kitox.re.kr); HQ Seoul, ~45–47 employees; FY2025 revenue ≈ $18.3M, TTM 30‑Jun‑2026 ≈ $18.9M; net loss ≈ –$8.7M FY2025 (PitchBook/Morningstar data). Not acquired, not dissolved, not in bankruptcy. Note the entity trades in a financially strained state (negative net income, convertible‑bond issuance per thevc.kr), which is a going‑concern stress factor — but stress is not a recorded event and does not by itself create an assignment.

---

Assignment timeline

Only one recorded post‑filing assignment appeared in the patent's legal events. It is the routine founder→company transfer, not a post‑issuance monetization event.

• 2012‑12‑11 (executed date not disclosed in the record I could read; this is the recording date) — Reel [NOT RETRIEVED]/[NOT RETRIEVED]
  • Conveyance: Assignment (Patent Assignment Abstract of Title — "ASSIGNMENT OF ASSIGNORS INTEREST; SEE DOCUMENT FOR DETAILS")
  • Assignor: JEON, BYUNG HEE
  • Assignee: CYTOGEN CO., LTD. (Korea)
  • Correspondent: [NOT RETRIEVED] — could not read the correspondent/recording attorney from the Assignment Center. No recurrence analysis possible; this is the one signal I cannot evaluate.
  • Context: Original inventor→assignee perfection. The application (13/641,094, national stage of WO2011/129651) was already filed 2011‑04‑15 naming Cytogen as applicant; this 2012 recordation formally perfects Cytogen's title of record for the US national‑stage case. It is not an acquisition, fire‑sale, reorg, securitization, or transfer‑to‑asserter.

No other assignments, security agreements, mergers, name changes, or releases appear in the legal‑events record. The two assignee spellings — "Cytogen Inc" (Google Patents current/original assignee field) and "CYTOGEN CO., LTD." (assignment event + the EP family member EP2560000/ES2715828, which lists proprietor "Cytogen Co., Ltd., 100%") — are romanization variants of the same Korean entity ㈜싸이토젠, not a change‑of‑name conveyance. Recording a Korean‑form name in an assignment while a US‑form name sits in the assignee field is normal and is not a finding.

Cross‑reference to earlier sections: the prior "Patent summary" flagged the three assignee name variants (Cytogen Inc / CYTOGEN CO., LTD. / Cytogen Co., Ltd.) as uncertainty #3. That is now resolved: single entity, three romanizations, one assignment.

---

Timeline diagram

timeline
    title Ownership of US 9372134
    2010 : Korean priority filed
         : Cytogen Inc founded in March
    2011 : PCT application filed by Cytogen Inc
    2012 : Inventor assigns to Cytogen Co Ltd
    2016 : US patent 9372134 issued
    2018 : Cytogen lists on KOSDAQ

---

NPE / troll‑pattern signals

Opposite‑direction signal worth noting (not in your list): the inventor holds the CEO role at the assignee and the patent is asserted on the company's own product family (Smart Biopsy CTC platform). That combination — inventor‑principal + product‑practicing owner + zero serial‑assignment history — is a negative indicator for NPE behavior.

---

Verdict

Operating‑company assertion — with an explicit qualification on the word "assertion."

Justification: the entire chain is one recorded transfer, 2012‑12‑11, Jeon → CYTOGEN CO., LTD. (Reel [NOT RETRIEVED]), an inventor‑to‑company title perfection; there is no subsequent assignment, no LLC intermediary, no known NPE in the chain, and the current/sole owner is Cytogen, Inc. (KOSDAQ: 217330), a commercially operating diagnostics company whose Smart Biopsy Cell Isolator practices the size‑based funnel capture claimed in the '134 patent. Because the prior litigation section found no suit asserting this patent, the "assertion" half of the category is unverified — what is verified is that the owner is a practicing operating company, not an NPE; if your taxonomy requires a documented lawsuit to earn "assertion," the correct call would be operating‑company‑owned, non‑litigating, and the closest pre‑defined label is still this one (the alternatives — NPE, defensive aggregator, insufficient data — are each affirmatively contradicted by the single assignment and the KOSDAQ product line). The two residual gaps driving any residual uncertainty are the unretrieved reel/frame + correspondent and the fact that Cytogen files with DART/KRX, not SEC/EDGAR, so US filing cross‑checks are incomplete.

Verify at: USPTO Assignment Center — https://assignmentcenter.uspto.gov/ (patent number 9372134; also the legacy https://assignment.uspto.gov/patent/index.html), plus https://patents.google.com/patent/US9372134B2/en → Legal Events.

I'll research the prior-art references cited for US 9,372,134. Let me run targeted searches.

Let me retrieve the full citation lists from the patent record pages.

I found the cited-reference list. Let me verify the details (titles, dates) of each cited reference.

Scope and method note (read first): I retrieved the citation list for US 9,372,134 from the Justia patent record (which reproduces the "Referenced Cited" table), plus the Google Patents full text you supplied, plus one search-report hit that bears on one of the entries. I hit my tool-call limit before I could pull the full text of each cited reference, so the "brief description / claim-mapping" below is tentative for every reference whose text I did not open. Where I could not confirm a reference's subject matter, I say so rather than guess.

---

Prior Art Cited on the Face of US 9,372,134

Source for the list: https://patents.justia.com/patent/9372134 (Referenced Cited → U.S. Patent Documents / Foreign Patent Documents), cross-checked against the Google Patents record.

0. Correction to the earlier generated section (flag)

The earlier "Patent summary" in this thread stated the patent has "two independent claims" and rendered the method as Claim 2. That is not what the record shows. The Justia rendering of the granted claims shows:

• Claim 1 — device (independent)
• Claim 11–13 — dependent device claims
• Claim 14 — device (independent; recites the dispersing unit with posts)
• Claim 15–16 — dependent
• Claim 17 — "A target separating method, comprising the steps of…" (independent)
• Claim 18–19 — dependent

So there are three independent claims (1, 14, 17), and the method is claim 17, not claim 2. The earlier text was reproducing the "Aspects of the Invention" bullet from the Summary section, not the granted claim numbering. Treat the claim numbers above as authoritative for the prior-art mapping that follows.

---

1. U.S. Patent Documents cited

§102 analysis per reference

Preliminary — statutory categories. The '134 has a KR priority date of 2010‑04‑15 and a US/PCT filing date of 2011‑04‑15. Therefore:

• References published before 2010‑04‑15 can be §102(a)/102(b) or §102(e) art.
• References published between 2010‑04‑15 and 2011‑04‑15 can only be art if their own effective filing date precedes the '134's priority (i.e., §102(e) with an earlier priority).
• References #11 (2013) and #12 (2015) postdate the 2011‑04‑15 filing and cannot be §102 prior art for the '134 as published — unless they carry an earlier priority/filing date, which would have to be checked on their faces. Their presence in the list is anomalous and suggests either (a) they were cited against a related later application, or (b) the Justia table has merged a "Cited By" section into "Referenced Cited." Flagging as unresolved.

#5 — US 2007/0259424 A1 (Toner et al.). This is the only cited reference I can describe with confidence, because the '134 patent itself describes it verbatim in its Background. It discloses a microfluidic device with a top layer, a bottom layer and a plurality of obstacles (micro‑posts) whose surfaces are coated with binding moieties (antibodies, charged polymers, cell‑coupled molecules); blood enters an inlet, flows along channels, exits an outlet, and cells are captured by the binding moieties.

• Anticipation of claims 1/14/17? No. It lacks (i) the claimed "forward funnels" arranged orthogonally across the channel, (ii) the second capture array of funnels, (iii) the reverse flow guide means, and (iv) the overturn/back‑flush step of claim 17. Capture is affinity‑based, not size/geometry‑based. It is the closest§103 reference** (the patent frames its whole problem statement as an improvement over it), not a §102 reference.

#1 — US 5,624,815. ⚠️ Number/inventor discrepancy. The Justia table lists it as "Grant et al.," but a WIPO search report (WO 2000/045950 A1) cites "US 5 624 815 A (Evans Christopher T et al.) 29 April 1997." Under your strict rule I am reporting both literally and not auto‑correcting. I could not open the document to resolve it. Because I cannot confirm its subject matter, I cannot responsibly map it to a claim. (Two possibilities to check: an early cell/particle separation or assay patent, or an apparatus case — unverified.)

#2 — US 5,800,784 (Horn, 1998). Listed in the same "Referenced Cited" row as #1. Title and subject matter not verified in this session. A 1998 patent is comfortably §102(b) art by date; claim mapping not determinable without text.

#3 — US 7,064,321 (Franzen, 2006). Pre‑priority‑date art by date (2006 < 2010). Subject matter not verified. Franzen‑family patents in this space generally concern filtration/particle handling, but I will not assert that without the text. Claim mapping not determinable.

#4 — US 2004/0141880 A1 (Handler et al., 2004). Pre‑priority art by date. Subject matter not verified. Claim mapping not determinable.

#6 — US 2009/0100810 A1 (Smith, 2009). Pre‑priority art by date. Subject matter not verified. Claim mapping not determinable.

#7 — US 2009/0226957 A1 (Paterlini‑Brechot, 2009). This is the reference I would flag as the second‑most likely to be substantively important, because the Paterlini‑Brechot work is associated with size‑based isolation of rare cells from blood by filtration (the "ISET" approach). If, on its face, it discloses a filter case with inlet/outlet/channel and size‑based trapping of target cells, it could potentially reach claim 1's preamble and the concept of capture by size. However, to anticipate claim 1 it must disclose the "first/second forward funnels arranged orthogonal to the flow direction" — a tapered‑funnel array geometry. I did not verify that it does, so I will not assert anticipation. Unverified — do not rely on this mapping. It is a strong §103 candidate if the funnel geometry is absent.

#8 — US 2009/0314725 A1 (Parro, 2009). Pre‑priority art by date. Subject matter not verified. Claim mapping not determinable.

#9 — US 2011/0008223 A1 (Tsao et al., Jan 13, 2011). Published after the '134 priority date (2010‑04‑15) but before the US filing date (2011‑04‑15). It can only be art under §102(e) if its own effective filing date predates 2010‑04‑15. Not verified; claim mapping not determinable.

#10 — US 2011/0083964 A1 (Ulmanella, Apr 14, 2011). Published one day before the '134 US filing date and after the '134 priority date. Same §102(e) caveat as #9: relevant only if it has an earlier effective filing date. Not verified; claim mapping not determinable.

#11 — US 2013/0244906 A1 (Collins, 2013) and #12 — US 2015/0118739 A1 (Kobayashi, 2015). Both postdate the '134 filing date. They are not §102 art against the '134 unless they carry pre‑2010‑04‑15 priority (not verified). Treat their presence in the list as a data‑provenance anomaly (see Preliminary above).

---

2. Foreign Patent Documents cited

§102 note: the four pre‑2010 entries (CN 1628907, EP 12740, JP 2003‑102710, JP 2005‑140790, JP 2008‑526251, JP 2009‑109232) are all §102(b)‑eligible by date. The two 2011 CN publications (Jan and Feb 2011) fall in the priority‑to‑filing window and require an earlier §102(e)‑equivalent date to count.

Most notable item: JP 2009‑109232 A. The '134 record specifically lists, as separately submitted references:

"Derwent abstract of JP 2009‑109232 A, ACC‑No. 2009‑J61096, and JPO machine translation of JP 2009‑1092232 A (JP 2009‑109232 A published May 21, 2009)."

The fact that the applicant filed a Derwent abstract and a JPO machine translation of JP 2009‑109232 indicates this reference was treated as material during prosecution (this is the classic pattern of a reference the examiner raised and the applicant overcame or distinguished). I could not retrieve its text, so I cannot state what it discloses. This is the single reference I would prioritize opening first in any follow‑up, alongside US 2007/0259424 A1.

All other foreign documents (CN 1628907, CN 101947516, CN 101963552, EP 12740, JP 2003‑102710, JP 2005‑140790, JP 2008‑526251): subject matter not verified; no claim mapping asserted.

---

3. Other (non‑patent) cited material

Suk Ho Kang, Particle Technology (1995, p. 34) is the only printed publication in the non‑patent list that could itself be §102(b) art. Its cited page (p. 34) is presumably a figure/text on particle classification or filtration. Unverified — could only anticipate if that single page discloses every element of a claim, which is unlikely for claims 1/14/17 but would have to be read.

---

4. Bottom line — "most relevant prior art" for US 9,372,134

1. US 2007/0259424 A1 (Toner et al.) — the pre‑eminent reference; expressly discussed in the '134 Background as the device the invention improves upon. Not anticipatory of claims 1, 14 or 17 (no funnels; no reverse flow guide means; no overturn/back‑flush step). Best framed as a §103 starting reference.
2. JP 2009‑109232 A (with Derwent abstract + JPO translation in the file) — highest‑priority unresolved item; its having been submitted in translated form signals examiner reliance. Text not retrieved.
3. US 2009/0226957 A1 (Paterlini‑Brechot) — the cited reference most likely to be a size‑based cell‑filtration disclosure; potential §102/§103 candidate against claim 1 if it shows a tapered‑funnel capture geometry (unverified).
4. US 5,624,815, US 5,800,784, US 7,064,321, US 2004/0141880 A1, US 2009/0100810 A1, US 2009/0314725 A1 — date‑eligible but unverified; cannot be mapped.
5. US 2011/0008223 A1, US 2011/0083964 A1, CN 101947516, CN 101963552 — in the priority‑to‑filing window; art status depends on earlier effective dates (unverified).
6. US 2013/0244906 A1, US 2015/0118739 A1 — post‑date the '134 filing; not §102 art absent earlier priority. List placement is anomalous.

No cited reference can presently be asserted to anticipate claims 1, 14 or 17, because (a) the three independent claims share the distinctive forward‑funnel arrays orthogonal to flow, and claim 17 additionally requires the reverse flow guide means plus the "turn the filter case upside down" step; and (b) none of the cited references that I could characterize discloses funnel arrays rather than cylindrical posts (Toner) — and I could not verify the remaining references. Any anticipation theory would have to turn on one of the unverified references (most plausibly JP 2009‑109232 or Paterlini‑Brechot).

---

5. Explicit uncertainties (do not treat as settled)

1. I could not open the full text of any cited reference except US 2007/0259424 A1 (described by the '134 itself). All §102 claim mappings for the others are withheld rather than guessed.
2. Assignee/inventor naming in the citation table is literal only; I did not normalize and did not verify each reference's bibliographic data independently.
3. The US 5,624,815 inventor discrepancy ("Grant et al." in the Justia table vs. "Evans Christopher T et al." in WO 2000/045950's search report) is reported, not resolved.
4. The 2013/2015 entries are, on their face, incapable of being §102 art for this patent; the list provenance is unverified.
5. Interrogating the actual reels: the authoritative source for the true "References Cited" is the USPTO PatentCenter / IFW for application 13/641,094 (and the PTO‑892/1449 forms and the examiner's search notes), plus the PCT/KR2011/002705 ISR dated Feb 6, 2012. I did not query those directly; the list above is from the Justia mirror.

I'll research the prior art record for this patent before drafting the obviousness analysis.

Let me dig into the specific prior-art references and closest analogous art.

Let me look at the specific prior-art citation record and the closest size-separation art.

Obviousness Analysis — U.S. Patent No. 9,372,134 B2 under 35 U.S.C. § 103

Scope note. This builds on the previously generated sections (bibliographic record, independent-claim overview, litigation and PTAB "empty field" findings) and does not repeat them. One correction/addition to the record from this pass: the reference cited on the face of the '134 patent — U.S. Patent Publication No. 2007/0259424 A1 (Toner et al., published 2007-11-08) — is not merely background; it is an admitted prior-art reference and is the correct starting point for the § 103 analysis. I could not retrieve the patent's own Google Patents "Patent Citations" table in this session, so the combinations below are built from (i) the admitted prior art, (ii) the Toner/Massachusetts General Hospital family that dominates the same classification space, and (iii) the reverse-flow recovery art. Where a reference's prior-art date is questionable, I flag it rather than assume it.

---

1. Legal framework

The '134 patent has a 2010-04-15 Korean priority date and a 2011-04-15 U.S. filing date. It is therefore a pre-AIA patent, and § 103 is applied under pre-AIA § 102(a)/(b)/(e) art plus the Graham v. John Deere factors as refined by KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007). Because the patent is pre-AIA, "prior art" includes § 102(e) U.S. application publications as of their U.S. filing dates, and a foreign priority claim under § 119 can be used to antedate a § 102(e) reference (but a foreign priority date of a reference cannot be used to give that reference an earlier § 102(e) date — In re Hilmer). This date arithmetic matters here, because the closest recovery art in this space sits within weeks of the '134 priority date (see § 3, "Date risk").

---

2. What must be shown

Per the previously generated claim overview (which I adopt, with the stated caveat that the printed dependent-claim set was not independently verified):

Everything else in the specification (reverse funnels 50, obstruction bodies 64/66, dispersing posts 22, staggered rows, graded exit widths 15–20 / 10–15 / 5 µm, partition plates 200) reads as dependent-claim / embodiment matter. I analyze it separately in § 6 rather than treating it as part of the independent claims.

---

3. The prior art — identified, with dates and statutory basis

Date risk on the NUS reference (important)

The NUS publication US 2014/0087456 A1 is the most on-point teaching of the claimed method (reverse-flow dislodgement of mechanically trapped cells). But its lineage is:

• CIP filed 2013-09-18;
• parent U.S. 12/765,576, filed 2010-04-22;
• provisional 61/172,250, filed 2009-04-24.

Its § 102(e) date is therefore 2013-09-18 for new matter, 2010-04-22 for matter carried from the parent, and 2009-04-24 only for matter supported by the provisional. The '134's 2010-04-15 priority date is one week before the parent's 2010-04-22 filing, so any subject matter first added in the parent (but not the provisional) is antedated by the '134's § 119 priority. Whether the reverse-flow-displacement disclosure is supported in the 2009 provisional is a verification step I could not complete. If it is, this reference is § 102(e) art as of 2009-04-24 and is the single most damaging reference on claim 2. If it is not, the reverse-flow teaching must be sourced elsewhere (general backwash art) and the claim-2 case weakens. This is the highest-value factual question in the whole § 103 analysis and should be resolved first.

---

4. Level of ordinary skill

A POSITA here would be a person with a B.S./M.S. in mechanical, biomedical, or chemical engineering (or a Ph.D. in a microfluidics-adjacent field) and 2–5 years' experience designing microfabricated cell-separation devices — i.e., familiar with soft lithography, micron-scale channel constrictions, DLD/bump arrays, weir and membrane microfilters, and blood-cell mechanics (RBC deformability, WBC 8–25 µm). That person would treat Toner's obstacle arrays, DLD arrays, weir filters, and membrane backwash as a single, well-developed toolbox.

---

5. Claim 1 — the device claim

Element-by-element mapping

Combination 1 (primary): Toner '424 + Toner '718 + Toner '809

Take Toner '424 as the base microfluidic cartridge (admitted prior art, and the very device the '134 characterizes as the state of the art). Replace its antibody-coated micropost array with the size-selective 2-D obstacle arrays of Toner '718/'809 — which are expressly described as "capture modules… that selectively filters or captures cells … having a hydrodynamic size greater than a particular gap size" — and arrange two of them in series with the downstream one having smaller gaps. Toner '809 itself teaches running the same sample through successive size cut-offs (8 µm, then 3–6 µm), which is precisely the "first array upstream / second array downstream, both capturing the different kinds of targets" structure of claim 1.

Claim 1 does not require funnels to be anything more than converging constrictions that capture by size. Because the claim's capture mechanism is purely geometric ("so as to capture the different kinds of targets") and imposes no binding moiety, no particular taper angle, and no numeric exit width, the funnel element reads on any array of converging, size-grading constrictions spanning the channel. The Toner wedge-shaped obstacles plus the ordinary weir-filter geometry (an upstanding wall with a tapered gap, ubiquitous in microfabricated blood filtration) supply it.

Combination 2 (alternative): weir/tapered-constriction filter (primary) + Toner '809 (secondary)

If one insists that the primary reference must itself disclose the funnel shape, invert the order: start from a microfabricated weir/tapered-gap filter (the "microfiltration apparatus … for separating cells whilst minimising mechanical trauma," US 2009/0188864 A1; and the membrane/backwash filters of WO 2009/087375), which captures cells larger than the gap while passing smaller, deformable cells, and combine with Toner '809 for the multi-stage, graded-size, in-series arrangement and with Toner '781 for row-wise obstacle arrays spanning the channel.

Combination 3 (size-based deflection): DLD

Huang et al. 2004 (and the U.S. 5,837,115 / WO 2004/113877 arrays cited in '809) supply a 2-D post/bump array spanning the whole channel width, sized to sort by particle diameter, with staggered rows so that each row is offset from the last to maximize particle–obstacle interactions. That supplies the "arranged along a direction orthogonal to the flow direction" limitation and the staggered geometry if it is later read into the claims.

---

6. What the specification adds as dependent matter (if claim 1 is too broad, these are the fallbacks)

The specification describes, and the dependent claims (unverified numbering) likely recite:

• Reverse funnels 50-1/50-2 alternating with the forward funnels — but the '134 itself states each reverse funnel "is formed of the first guide 46 of each of the … forward funnels … arranged at the right side thereof and the second guide 48 of each of the … forward funnels … arranged at the left side thereof." In other words, the reverse funnels are not a separate structural addition — they are the negative space between adjacent tapered guides. Any array of alternating wedge/funnel profiles in a channel inherently defines such inter-funnel passages. This is a strong § 103 point: the "reverse flow guide means" of claim 2 appears to be inherent in the funnel geometry the art already teaches, not an added component.
• Exit areas equal as between forward and reverse funnels — a mathematical consequence of the funnels sharing the same guides.
• Graded exit widths (15–20 µm, 10–15 µm, ≥5 µm) — the identical numeric scheme (15–20 µm / 10–15 µm / 5–10 µm in three stages, chosen to collect WBCs of 12–25 µm) appears in the same-family Cytogen application US 2013/0071304 A1 for its mesh-filter stages. That is a same-family disclosure and is not usable as prior art, but it confirms that the numbers are a design choice keyed to known blood-cell dimensions — the kind of "finite number of identified, predictable solutions" that KSR makes obvious.
• Obstruction bodies 64/66 — the specification concedes a FIG. 7 modified example in which both are omitted and the targets are simply captured within the forward funnels. A limitation that the patent expressly discloses as omissible cannot supply the inventive weight.
• Dispersing units 20-1/-2/-3 (staggered post arrays, 300–400 → 200–300 → 100–200 µm) — staggered, axially-graded post arrays to spread flow transversely are the defining feature of Toner '809's "staggered … each subsequent row … offset" arrays and of every DLD device. Purely a known flow-uniformity expedient, motivated by the admitted problem that "human blood … may flow through the channel in a biased state depending on the viscosity thereof."
• Partition plates 200 stacking to form parallel sub-channels — parallelization for throughput. US 2007/0059718 A1 FIG. 4 already shows "a plurality of size-based separation modules in parallel," and Toner '781 expressly contemplates scaling to "more than 2, 10, 20, 30, … 1000 rows." Stacking plates is a routine mechanical scaling choice.

---

7. Claim 2 — the method claim

Claim 2's steps and the corresponding art:

The obviousness argument for claim 2

The '134's own BACKGROUND supplies the motivation: the admitted Toner '424 device "suffers from a problem in that the capture rate of targets is very low" and "has a difficulty in collecting the targets captured by the binding moieties… not suitable for use in separating the targets from a large quantity of blood." Faced with a known device having two expressly identified deficiencies, a POSITA has a clear reason to (a) swap affinity capture for size-based capture (removing the release problem entirely, since mechanically trapped cells are not covalently bound), and (b) add a reverse-flow flush — a technique that is the standard way to harvest retained particles from a filter. The NUS reference independently diagnoses the same problem and independently prescribes the same cure: "the isolated cells are likely to be difficult to retrieve due to the binding of the tumor associated antigens to the device. Retrieving these cells may require high mechanical forces or biochemical agents and the integrity of these cells might be affected" — followed by reverse-flow retrieval.

That is a textbook KSR fact pattern: a known problem, a finite number of known solutions, a predictable result, and a reference that articulates the very rationale.

---

8. Motivations to combine — the KSR rationales, stated explicitly

1. Known problem, known cure. The '134 BACKGROUND concedes the low-capture-rate and hard-to-harvest problems of the Toner '424 device. Reason to improve = express.
2. Simple substitution of one known element for another. Replacing antibody-coated microposts with size-graded constrictions is a substitution of a known capture mechanism for a known capture mechanism, to obtain the predictable benefit of size-based separation without biochemical release. KSR (predictable results).
3. Known technique to improve a similar device in the same way. Multi-stage, series-connected, size-graded separation modules are expressly disclosed in the same Toner family (US 2008/0138809 A1: 8 µm then 3–6 µm; US 2007/0059718 A1 FIG. 4; US 2007/0059781 A1: many rows). Adding a second, finer capture array downstream of a first is the obviousness of doing again what you already did, with a smaller gap.
4. Design incentive / market pressure. The BACKGROUND itself identifies the commercial driver: "research and development have been extensively made on the collection of live cells from the human blood"; and the specification and the sibling Cytogen filings show the target performance metric (recovering 12–25 µm WBCs from whole blood at useful throughput). This is the "design incentive… in the field" rationale.
5. Turning a known technique into a method. Back-flushing a filter to release retained matter is ancient in filtration and present in microfluidics (NUS; WO 2009/087375). Applying it, and choosing to do so by inversion rather than by valve re-plumbing, is a predictable mechanical choice with no asserted unexpected result. The '134 claims no performance data showing the flip produces a result that flow reversal alone would not.
6. "Obvious to try." Gap widths of 15–20 / 10–15 µm to catch ≥15 µm and ≥10 µm cells, and a 5 µm floor to pass 6–8 µm deformable RBCs, are among a finite, identified set of solutions dictated by the known dimensions of blood cells (Toner '718 itself tabulates cell types and sizes in the 4–6 µm RBC / >10 µm WBC range). Selecting them is not inventive.

---

9. Where the obviousness case is weakest — and the counterarguments to expect

I flag these because a rigorous § 103 opinion has to anticipate the Patent Owner's response:

1. Claim 1 may be anticipated, not merely obvious. If "funnel" is construed broadly, Combination 3 (DLD, Huang 2004) alone or Toner '809 alone may be § 102 art for claim 1. Test this first — it is cheaper than an obviousness case.
2. The "funnel" limitation is the real battleground. Patent Owner will argue that a funnel (smooth, converging, four-slant-surface guide with rounded corners 46e/48e) is structurally distinct from a post/micropost, a bump, or a weir, and that no reference teaches an array of paired tapered guides forming both a forward funnel and an adjacent reverse funnel. Expect an argument that this geometry is not merely a design choice but is what enables the flip-and-flush to work without wedging or cell damage ("round corners … serve to assure smooth flow of the targets and to prevent the targets from getting damaged"). This is the strongest non-obviousness argument in the case, and it should be met with: (a) the round-corner feature is a dependent matter, not in claim 1; (b) reduced-shear rounded constrictions are a known microfabrication expedient; (c) the specification's own FIG. 7 shows the structure works with the obstruction bodies omitted, undermining the claim that the geometry is critical.
3. The physical-inversion step. A Patent Owner will argue that no art teaches turning the case upside down, and that this is what distinguishes over a mere valve-based flow reversal. The response is the specification's own admission that overturning merely "oriented [the funnels] in the direction opposite to the filtering direction" — the claim recites no gravity-dependent or hydrostatic effect that differs from reversing the pressure gradient. Consider whether this limitation is even entitled to patentable weight under the printed-publication and In re line of cases on functional/orientation language (as flagged in the earlier report's § 112 discussion — § 112 theories are IPR-ineligible, so this fork matters).
4. No secondary considerations of record. From the previous sections: no litigation and zero PTAB proceedings on the '134. That leaves an evidentiary vacuum — there is no evidence of commercial success, licensing, copying, industry praise, or unexpected results that could rebut a prima facie § 103 case, and equally no adjudicated validity finding to rely on. The adjusted expiry of 2033-03-01 means the analysis has a long commercial tail, which raises the stakes on the funnel-geometry question.
5. Same-family contamination risk. If an obviousness chart is built using US 2013/0071304 A1, KR101254679B1, KR101226515B1, KR101254680B1, or WO 2011/129649 A2, the chart is defective — those are the applicant's own same-family documents, not third-party prior art. My search surfaced that family prominently; do not let it leak into the combination.

---

10. Bottom line

• Claim 1 is exposed to a strong § 103 attack, and possibly a § 102 attack, on the combination of US 2007/0259424 A1 (admitted art) + US 2007/0059718 A1 + US 2008/0138809 A1, optionally with Huang et al. 2004 / U.S. 5,837,115 / WO 2004/113877 for the channel-spanning, size-graded, staggered array and with the ordinary weir/tapered-constriction filter art for the funnel geometry. The patent's own BACKGROUND supplies the motivation.
• Claim 2 is exposed to a strong § 103 attack on the same combination plus US 2007/0059781 A1 ("transport cells into and out of the … capture modules") and the reverse-flow-retrieval teaching of US 2014/0087456 A1 — subject to resolving that reference's § 102(e) date, which turns on whether its reverse-flow disclosure is supported by the 2009-04-24 provisional (61/172,250). If it is not, substitute general backwash art (WO 2009/087375 / US 2009/0188864 A1) and expect a harder fight on the "turn upside down" step.
• The decisive open factual question is the § 102(e) date of the NUS reference. The decisive legal question is how broadly "forward funnel" is construed — if broadly, the case is easy; if narrowly (paired tapered guides with rounded corners defining inter-funnel reverse passages), much of the strength shifts to the dependent claims and to the design-choice arguments in § 8.
• Caveats carried forward: the printed dependent-claim set and numbering remain unverified; the patent's own "Patent Citations" table was not retrievable in this session, so additional references cited on its face may exist and should be pulled from the file wrapper before finalizing any chart; and WO 2010/118427 A1, despite appearing on the patent's citation page, is not prior art against the '134 because it published 2010-10-14, after the 2010-04-15 priority date.

Sources:

• US 9,372,134 B2 record and specification — https://patents.google.com/patent/US9372134/en ; https://patents.justia.com/patent/9372134
• US 2007/0259424 A1 (cited in the '134 specification); US 8,986,966 B2 (Toner et al., listing it) — https://patentimages.storage.googleapis.com/16/35/c9/7dfc24f8f49fe5/US8986966.pdf
• US 2007/0059718 A1 — https://www.patents-review.com/a/20070059718-systems-methods-enrichment-analytes.html
• US 2007/0059781 A1 — https://patentimages.storage.googleapis.com/f1/0b/1a/ba2d074bce29d0/US20070059781A1.pdf
• US 2008/0138809 A1 — https://www.freepatentsonline.com/y2008/0138809.html
• US 2007/0099207 A1 — http://patentimages.storage.googleapis.com/5c/da/70/b5230a4525488b/US20070099207A1.pdf
• US 2014/0087456 A1 (NUS) — https://patents.google.com/patent/US20140087456 ; https://patents.google.com/patent/WO2012128717A1 ; https://www.freepatentsonline.com/y2014/0087456.html
• WO 2010/118427 A1 (date-flagged, not prior art) — https://patents.google.com/patent/WO2010118427A1
• Same-family (not prior art): US 2013/0071304 A1 — https://patents.justia.com/patent/20130071304 ; WO 2011/129649 A2 — https://patents.google.com/patent/WO2011129649A2/en ; KR101254679B1 — https://worldwide.espacenet.com/publicationDetails/description?CC=KR&NR=101254679B1