The most important fact about the litigation over Cigna’s PxDx claims-review system is also the easiest to lose: nobody has won the case.
Patients allege that Cigna used PxDx to reject certain medical claims without the individualized physician review required by their benefit plans and California law. Cigna denies wrongdoing, disputes that PxDx is an “algorithm,” and says the process matches procedure and diagnosis codes after care has already been delivered.
In March 2025, U.S. District Judge Dale A. Drozd allowed fiduciary-duty claims by three plaintiffs and a related California unfair-competition theory to move forward. But he dismissed the plaintiffs’ direct claim for unpaid benefits because the complaint did not identify specific plan terms entitling them to those benefits. He also dismissed the PxDx-based claims of three other named plaintiffs after considering Cigna’s evidence that their claims had not gone through PxDx.
That mixed ruling did not establish that PxDx unlawfully denied a single claim. It determined which legal theories and which plaintiffs could continue into discovery.
What PxDx does is itself disputed
The plaintiffs’ third amended complaint describes PxDx as an automated system that identifies mismatches between a billed procedure and diagnosis, then enables Cigna medical directors to approve denials in batches without reviewing individual medical records. The complaint alleges that this practice displaced the judgment promised in benefit-plan documents and required by California’s licensed-physician review rules.
Those are allegations, not findings. Cigna’s answer denies that PxDx is an algorithm. The company does admit using PxDx “to determine whether claims for certain services are medically necessary” under applicable plan terms or coverage policies, as well as to facilitate payment of high-volume, low-cost services.
In a separate public explanation, Cigna calls PxDx simple code-matching software used after treatment. It says the process applies to about 50 relatively low-cost tests and procedures, that most reviewed claims are paid automatically, and that denied claims can be resubmitted with a corrected diagnosis code or appealed. Cigna also says PxDx is not artificial intelligence or machine learning.
The parties therefore disagree not only about legality, but about the correct vocabulary. “Automated,” “algorithmic,” and “AI” are not interchangeable. The public filings support describing PxDx as software-assisted claims review. Whether that process unlawfully substituted automation for clinical judgment remains contested.
The court rejected part of the plaintiffs’ factual theory
The named plaintiff in the case title, Suzanne Kisting-Leung, alleged that Cigna denied payment for an ultrasound through PxDx. Cigna submitted a declaration from a medical officer stating that PxDx denials generate separate letters and that no such letters were associated with the claims of Kisting-Leung or two other plaintiffs, Cristina Thornhill and Amanda Bredlow.
The court found no genuine dispute, on the evidence then presented, that those three plaintiffs’ claims were not subjected to PxDx review. Their fiduciary-duty and California claims were dismissed for lack of standing. That finding materially narrows any accurate account of the case: the court did not accept the proposition that every challenged Cigna denial — including the lead plaintiff’s — came from PxDx.
Three plaintiffs remained on the PxDx theories: Samantha Dababneh, Randall Rentsch and Abdulhussein Abbas. The court held that they had plausibly alleged a fiduciary breach by claiming Cigna entrusted medical-necessity determinations to PxDx despite plan language requiring a medical director to make those decisions. At this stage, plausibility is enough; proof comes later.
What survived — and what did not
The surviving case is not a straightforward demand that Cigna pay disputed medical bills. The court dismissed the ERISA claim seeking benefits under Section 502(a)(1)(B), with permission to amend, because the plaintiffs relied on coverage policies and general references to covered services rather than the governing plan provisions. The plaintiffs later elected not to amend that claim.
The ERISA fiduciary-duty claim under Section 502(a)(3) survived for Dababneh, Rentsch and Abbas. Their theory is procedural and equitable: that Cigna violated plan terms by delegating medical-necessity determinations to PxDx and should be ordered to change the process or provide other equitable relief.
The court also allowed a California Unfair Competition Law claim to proceed for the remaining plaintiffs. The judge concluded that they had adequately alleged a violation based on California requirements for licensed-physician involvement and accurate reasons for denials, and that the theory was not preempted at the pleading stage.
What the ruling means: surviving a motion to dismiss means a claim may be litigated. It does not mean the allegations are true, that a class will be certified, or that Cigna violated the law.
Cigna’s defense is broader than “we do not use AI”
In court, Cigna argued that several plaintiffs lacked standing, that one plaintiff’s claims were untimely, that the complaint failed to identify breached plan terms, that some plaintiffs had not exhausted internal appeals, and that the California claim was inadequately pleaded and preempted by federal law.
The judge accepted some of those arguments and rejected others. Cigna’s May 2025 answer then denied the remaining liability allegations and asserted dozens of defenses, including failure to exhaust plan remedies, statutes of limitation, plan exclusions and the contention that the case cannot properly proceed as a class action.
Cigna’s public statement adds a practical defense: PxDx is post-service review, not prior authorization; it says patients are not denied treatment and most in-network patients should not be billed when a PxDx-reviewed service is unpaid. The patient-side answer is that a denial of payment can still leave a person or provider fighting over a bill and that contractual review duties apply even after care occurs. The litigation has not resolved that tension.
The next consequential question is class certification
After discovery, the remaining plaintiffs moved for class certification on October 30, 2026. The docket entry establishes that the request was filed, not that it will succeed. The public docket available at publication did not yet include a merits ruling on that request, and Cigna’s opposition was not yet on file.
Class certification will require the plaintiffs to show, among other things, that common questions can be resolved across the proposed class and that the named plaintiffs can adequately represent it. Cigna has already signaled that it will contest those points. Variations among plan terms, procedure codes, denial reasons, appeals and patient financial responsibility could become central.
The evidentiary question beneath the legal fight is concrete: what did Cigna’s medical directors actually review before approving PxDx-associated denials, and did that review satisfy each plan’s promises and governing law? Discovery may produce the answer. The March 2025 ruling does not.
Primary sources
- Third amended complaint, filed June 14, 2024 (plaintiffs’ allegations).
- Cigna’s motion to dismiss, filed August 23, 2024 (defense arguments and supporting evidence).
- Order granting in part and denying in part Cigna’s motion to dismiss, filed March 31, 2025.
- Cigna’s answer to the third amended complaint, filed May 2, 2025.
- Cigna’s public PxDx explanation.
- Federal docket mirror for subsequent scheduling and filing history. The October 30, 2026 class-certification filing is also reflected in the court’s public docket feed.
Disclosure: This article is based on public records and Cigna’s published response. Patient Intelligence News did not rely on unnamed sources. No party reviewed the article before publication.