DCT
2:25-cv-00459
Veracyte Inc v. Sonic Healthcare USA Inc
I. Executive Summary and Procedural Information
- Parties & Counsel:
- Plaintiff: Veracyte, Inc. (Delaware)
- Defendant: Sonic Healthcare USA, Inc. (Delaware) and Clinical Pathology Laboratories, Inc. (Texas)
- Plaintiff’s Counsel: The Dacus Firm, P.C.; Wilson Sonsini Goodrich & Rosati, P.C.
- Case Identification: 2:25-cv-00459, E.D. Tex., 05/01/2025
- Venue Allegations: Plaintiff alleges venue is proper because Defendants have regular and established places of business within the Eastern District of Texas, where they have allegedly committed acts of patent infringement by making, using, or selling the accused testing services.
- Core Dispute: Plaintiff alleges that Defendants’ ThyroSeq line of molecular tests for thyroid nodules infringes three patents related to methods for analyzing and classifying biological tissue samples using gene expression data and trained algorithms.
- Technical Context: The lawsuit concerns the field of molecular diagnostics, specifically tests designed to classify thyroid nodules as benign or malignant to help patients with indeterminate cytopathology results avoid unnecessary diagnostic surgeries.
- Key Procedural History: The complaint does not reference prior litigation or post-grant proceedings between the parties. It does, however, detail a competitive product development history, alleging that Defendants’ ThyroSeq test evolved over several versions to incorporate features, such as the use of RNA expression analysis and trained algorithms, that were first introduced by Plaintiff’s Afirma tests.
Case Timeline
| Date | Event |
|---|---|
| 2009-05-07 | Priority Date for U.S. Patent No. 12,110,554 |
| 2009-12-09 | Priority Date for U.S. Patent Nos. 10,114,924 & 10,672,504 |
| 2011-01-01 | Plaintiff's Afirma GEC test commercially launched |
| 2013-09-01 | Defendant's ThyroSeq v1 test announced |
| 2014-06-12 | Defendant's ThyroSeq v2 test announced |
| 2018-01-01 | Defendant's ThyroSeq v3 GC test launched (approx.) |
| 2018-10-30 | U.S. Patent No. 10,114,924 Issued |
| 2020-06-02 | U.S. Patent No. 10,672,504 Issued |
| 2024-10-08 | U.S. Patent No. 12,110,554 Issued |
| 2025-05-01 | Complaint Filing Date |
II. Technology and Patent(s)-in-Suit Analysis
U.S. Patent No. 10,114,924 - "Methods for Processing or Analyzing Sample of Thyroid Tissue"
The Invention Explained
- Problem Addressed: The patent’s background section describes deficiencies in traditional cancer diagnosis methods, particularly for thyroid nodules, which can be subjective, difficult to reproduce, and yield an "indeterminate" result, leading to a large number of unnecessary thyroid removal surgeries (Compl. ¶55; ’924 Patent, col. 1:33-50, col. 1:62-2:4).
- The Patented Solution: The invention is a multi-step method for resolving these indeterminate cases. It involves first subjecting a portion of a thyroid tissue sample to standard cytological testing; if that result is indeterminate, a second portion of the same sample is then assayed for gene expression product levels. This quantitative data is input into a programmed computer, which uses a pre-trained algorithm to classify the sample as positive or negative for thyroid cancer with high accuracy, thereby providing an unambiguous diagnosis (’924 Patent, Abstract; col. 2:5-25).
- Technical Importance: This method provided a novel, quantitative solution to the clinical problem of indeterminate thyroid cytology, aiming to reduce the high rate of unnecessary diagnostic surgeries performed on patients whose nodules were ultimately benign (Compl. ¶¶4, 56).
Key Claims at a Glance
- The complaint asserts at least independent claim 1 (Compl. ¶61).
- The essential elements of independent claim 1 include:
- (a) obtaining a sample of thyroid tissue from a subject, where the sample comprises gene expression products;
- (b) subjecting a first portion of the sample to cytological testing that indicates the sample is indeterminate;
- (c) upon this indeterminate finding, assaying gene expression products from a second portion of the sample to yield a data set;
- (d) inputting this data into a programmed computer with an algorithm trained with a plurality of training samples to generate a classification (positive or negative) with at least 90% accuracy; and
- (e) electronically outputting a report identifying the classification.
- The complaint does not explicitly reserve the right to assert dependent claims for this patent.
U.S. Patent No. 10,672,504 - "Algorithms for Disease Diagnostics"
The Invention Explained
- Problem Addressed: The patent, like the ’924 Patent, addresses deficiencies in traditional methods for detecting thyroid cancer, which can lack accuracy and objectivity (Compl. ¶79; ’504 Patent, col. 1:38-48).
- The Patented Solution: The invention provides a method where nucleic acid molecules from a thyroid tissue sample are sequenced to measure the expression levels of genes from a specific, recited group (e.g., ALK, CALCA, DICER1, etc.). The resulting data is processed in a computer by an algorithm that was trained on known benign and non-benign samples, which then generates a classification of the test sample as positive or negative for cancer with at least 85% accuracy (’504 Patent, Abstract, col. 111:53-112:5).
- Technical Importance: The claimed method provides a specific technical improvement in cancer detection by using sequencing to analyze a defined set of genetic markers and applying a trained algorithm to improve diagnostic accuracy (Compl. ¶79).
Key Claims at a Glance
- The complaint asserts at least independent claim 1 (Compl. ¶84).
- The essential elements of independent claim 1 include:
- (a) sequencing nucleic acid molecules from a thyroid tissue sample to yield data on gene expression levels corresponding to a plurality of genes from a specified Markush group;
- (b) using a trained algorithm in a computer to process the data and generate a classification (positive or negative) at an accuracy of at least 85%, where the algorithm is trained with a plurality of different known samples; and
- (c) electronically outputting a report identifying the classification.
- The complaint does not explicitly reserve the right to assert dependent claims for this patent.
U.S. Patent No. 12,110,554 - "Methods for Classification of Tissue Samples as Positive or Negative for Cancer"
- Patent Identification: U.S. Patent No. 12,110,554, "Methods for Classification of Tissue Samples as Positive or Negative for Cancer," issued October 8, 2024 (’554 Patent).
- Technology Synopsis: The patent describes methods for diagnosing cancer in biological samples where a prior diagnostic screening was ambiguous or indeterminate (Compl. ¶102; ’554 Patent, col. 1:12-16). The solution involves assaying nucleic acids from the sample to identify both (i) the level of gene expression products and (ii) one or more genetic mutations, and then using a trained algorithm to process both sets of data to generate a positive or negative classification for cancer (Compl. ¶¶102, 107).
- Asserted Claims: Independent claim 1 is asserted (Compl. ¶106).
- Accused Features: The complaint alleges that the ThyroSeq v3 GC test infringes by analyzing both DNA and RNA from indeterminate thyroid samples to identify gene expression levels and genetic mutations, and then using a trained algorithm to generate a cancer classification (Compl. ¶¶112-114).
III. The Accused Instrumentality
Product Identification
- The accused instrumentality is the ThyroSeq v3 Genomic Classifier ("ThyroSeq v3 GC") testing service offered by Defendants (Compl. ¶¶37, 60).
Functionality and Market Context
- The ThyroSeq v3 GC test analyzes fine-needle aspiration (FNA) samples from thyroid nodules, particularly those with indeterminate cytology results (Compl. ¶¶65, 67). It utilizes next-generation sequencing to measure both DNA and mRNA to detect genetic alterations, including mutations, gene fusions, and abnormal gene expression alterations (Compl. ¶¶37, 66). The resulting data is processed by a proprietary trained algorithm, the "Genomic Classifier," which classifies the nodule as positive or negative for cancer and generates an electronic report (Compl. ¶¶37, 69, 70). The complaint alleges that ThyroSeq v3 GC is a direct competitor to Plaintiff's Afirma tests and that they are recognized as the two major commercially available molecular tests in this market (Compl. ¶¶7, 40).
IV. Analysis of Infringement Allegations
’924 Patent Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| (a) obtaining said sample of thyroid tissue of said subject... wherein said sample of thyroid tissue comprises gene expression products | Defendants obtain a fine-needle aspiration tissue sample of a thyroid nodule for ThyroSeq testing, from which gene expression products (mRNA) are isolated. | ¶¶65, 66 | col. 21:1-8 |
| (b) subjecting a first portion of said sample of thyroid tissue to cytological testing that indicates that said first portion of said sample of thyroid tissue is indeterminate | The accused ThyroSeq test is ordered with Thyroid FNA Analysis, and all indeterminate FNA results are "reflexed to" ThyroSeq for further testing. | ¶67 | col. 107:16-20 |
| (c) upon identifying said first portion of said sample of thyroid tissue as indeterminate, assaying by sequencing... said gene expression products from a second portion... to yield a data set | Defendants use targeted amplification-based next-generation sequencing technology to detect abnormal gene expression alterations in the indeterminate samples. | ¶68 | col. 107:21-31 |
| (d) in a programmed computer, inputting said data... to a trained algorithm to generate a classification... at an accuracy of at least 90%... | The ThyroSeq algorithm was developed using a "training tissue set" to distinguish cancer from benign nodules with a reported accuracy of 92.1% in one set and 90.9% in a validation set. | ¶69 | col. 107:32-43 |
| (e) electronically outputting a report that identifies said classification of said sample of thyroid tissue as positive or negative for said thyroid cancer | Defendants produce an electronic report that classifies the sample as positive or negative. The complaint provides a visual example of a "THYROSEQ® V3 GC RESULTS SUMMARY" report identifying a sample as "NEGATIVE." (Compl. ¶70, p. 22). | ¶70 | col. 107:44-48 |
’504 Patent Infringement Allegations
| Claim Element (from Independent Claim 1) | Alleged Infringing Functionality | Complaint Citation | Patent Citation |
|---|---|---|---|
| (a) sequencing nucleic acid molecules from said sample... to yield data comprising one or more levels of gene expression products... which... correspond to a plurality of genes comprising two or more genes selected from the group consisting of: ALK, CALCA, DICER1... | The ThyroSeq v3 test uses sequencing technology on DNA and mRNA to analyze 112 genes, which allegedly includes all 11 genes listed in the Markush group of claim 1. | ¶¶88, 89, 90 | col. 111:53-62 |
| (b) using a trained algorithm in a computer to process said data... to generate a classification... at an accuracy of at least 85%... wherein said trained algorithm is trained with a plurality of training samples... | The ThyroSeq algorithm was developed using a training tissue set to distinguish cancer from benign nodules and reportedly achieves an accuracy of 90.9% in a validation set. | ¶91 | col. 111:63-112:1 |
| (c) electronically outputting a report that identifies said classification of said sample of thyroid tissue as positive or negative for said thyroid cancer | Defendants produce an electronic report classifying the sample. The complaint includes a visual of a "THYROSEQ® V3 GC RESULTS SUMMARY" report classifying a sample as "POSITIVE" with a "High (~90%)" probability of cancer. (Compl. ¶92, p. 29). | ¶92 | col. 112:2-5 |
Identified Points of Contention
- Scope Questions: A central question for the '924 Patent infringement analysis may be whether a single actor performs or controls all steps of the claimed method. The complaint alleges that indeterminate results from cytological testing are "reflexed to ThyroSeq®" (Compl. ¶67), which raises the question of whether Defendants "subject" the sample to cytological testing as required by claim 1(b), or merely receive a sample after that step has been performed by a third party, creating a potential divided infringement issue.
- Technical Questions: For the '504 Patent, a potential point of contention may be whether the accused ThyroSeq test's analysis of 112 genes to detect "abnormal gene expression alterations" (Compl. ¶88) constitutes measuring "levels of gene expression products" for the specific Markush group of genes, as required by the claim. The analysis will depend on the technical evidence showing precisely what the ThyroSeq test measures and how that maps to the claimed gene list.
V. Key Claim Terms for Construction
The Term: "subjecting a first portion of said sample of thyroid tissue to cytological testing" (’924 Patent, Claim 1)
- Context and Importance: The construction of this term is critical to determining direct infringement. Practitioners may focus on this term because if it requires the accused infringer to personally perform the cytological testing, Defendants may argue they do not directly infringe, as they receive samples that have already been deemed "indeterminate" by third-party clinicians or labs. This could shift the focus to theories of divided or induced infringement.
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The claim preamble recites a "method for processing or analyzing a sample," which may support an interpretation where the entire sequence of events is part of the claimed method, regardless of which entity performs the initial cytological screen. The patent's purpose is to provide a solution for samples already known to be indeterminate, suggesting the "subjecting" step could be a condition precedent rather than an active step required of the infringer.
- Evidence for a Narrower Interpretation: The use of the active verb "subjecting" suggests that the entity performing the method must itself perform this step. The specification describes a clinical workflow where a sample is obtained and then subjected to various analyses, which could support a reading that a single party or a party in control performs the sequence of claimed actions (’924 Patent, col. 21:1-25; Fig. 18A).
The Term: "trained algorithm" (’924 and ’504 Patents, Claim 1)
- Context and Importance: The definition of this term is central to whether the accused ThyroSeq "Genomic Classifier" falls within the claim scope. The dispute may focus on whether "trained" requires a specific type of machine learning process, as described in the specification, or if it can broadly cover any algorithm whose classification parameters (e.g., cutoffs) were developed using a set of known samples.
- Intrinsic Evidence for Interpretation:
- Evidence for a Broader Interpretation: The term itself is broad. The specification discloses that the "present invention provides algorithms and methods for classification of samples" and is not limited to any single type (’504 Patent, col. 111:21-31). This could support a construction that encompasses any algorithm developed or refined using a "training set" of data, as the complaint alleges was done for ThyroSeq (Compl. ¶¶69, 91).
- Evidence for a Narrower Interpretation: The detailed description in the patents gives specific examples of "trained algorithms," such as "support vector machines, naive Bayesian algorithm, random forest algorithms," and other machine learning techniques (’504 Patent, col. 111:23-28). A defendant may argue that these specific disclosures limit the term to algorithms that involve an iterative learning process, potentially excluding simpler rule-based classifiers whose parameters were merely derived from a static analysis of a training data set.
VI. Other Allegations
- Indirect Infringement: The complaint alleges inducement of infringement. This is based on allegations that Defendants provide instructions, marketing, and promotional materials encouraging and facilitating clinicians to order the ThyroSeq test for indeterminate samples, thereby causing them to participate in the performance of the claimed multi-step methods (Compl. ¶¶71, 93).
- Willful Infringement: The complaint alleges willful infringement based on Defendants’ alleged knowledge of the Asserted Patents by or shortly after their issue dates. This allegation is supported by claims that the parties are direct and major competitors in a niche market, that Defendants have closely monitored Plaintiff’s Afirma products, featured papers about Afirma on their own website, and copied patented features into successive versions of ThyroSeq (Compl. ¶¶49-51, 72, 94).
VII. Analyst’s Conclusion: Key Questions for the Case
- A core issue will be one of single-actor infringement: can Plaintiff establish that Defendants directly infringe the multi-step method claims, particularly the initial cytological testing step recited in the '924 Patent? The resolution will likely depend on claim construction and evidence of whether Defendants control or direct the entire claimed process, or whether the case will primarily turn on theories of divided or induced infringement.
- A key evidentiary question will be one of technical scope: does the accused ThyroSeq "Genomic Classifier" constitute a "trained algorithm" as contemplated by the patents? The case may depend on detailed technical evidence comparing the development and operation of Defendants' algorithm against the patent specifications' descriptions of machine learning and other algorithmic training methods.
- A third central question will be one of knowledge and intent: what evidence can be shown to establish that Defendants had pre-suit knowledge of the patents and willfully infringed? The outcome of this question will be highly dependent on discovery into Defendants' competitive monitoring, product development processes, and awareness of Plaintiff's patent portfolio.