Companion Diagnostics: Key Trends Shaping Negotiations

Introduction

Companion diagnostics (“CDx”) are medical devices which provide information that is essential for the safe and effective use of a corresponding drug or biological product. Contracting around CDx has grown substantially more complex over the last several years, driven by advances in diagnostic technology and increasingly interdependent commercial relationships. 

What follows addresses three areas where that complexity is manifesting most acutely in negotiations, where legacy contract language is proving inadequate to the task, and what deal teams should consider to address these points. In particular, data rights and liability allocation have become central sources of transactional friction, and agreements that treat platform ambiguity as a threshold problem tend to underestimate what follows from it.

Technological Convergence and Platform Ambiguity

Complexity of the underlying technology incorporated in CDx is increasing. That complexity, such as the shift from single-analyte assays to multi-omic platforms, has introduced a structural problem that most CDx agreements were not designed to handle. For instance, a contract structured around a single biomarker does not translate cleanly when imposed on an arrangement involving a next-generation sequencing panel that simultaneously interrogates hundreds of variants, some relevant to the therapeutic indication and others potentially valuable for entirely different purposes.

The practical consequence of such structure/technology misalignment is ambiguity at the platform level. When a diagnostic partner’s assay generates data beyond what is needed for the approved CDx claim of the product manufacturer, the agreement often does not specify who owns that surplus data, who can commercialize it, or whether its generation triggers any exclusivity obligations. These are not hypothetical gaps. They are appearing with increasing frequency in disputes and renegotiations as platforms scale.

To address this issue, deal teams should consider defining a “CDx Field” with reference not just to the therapeutic indication covered by the approved CDx claim, but to the specific analytes and data outputs the assay is permitted to interrogate. Broader platform rights should be explicitly defined and allocated between the parties, with compensation terms reflecting the scope of access granted to each side.

Data Rights

This is the negotiation you must have. Because of this increased complexity, data rights are where most CDx negotiations are actually stalling. And, the reason is not hard to identify. Diagnostic platforms generate patient-level molecular data that has value well beyond confirming a treatment indication for its partners. That data can inform drug development programs, population health analytics, and future diagnostic product development. 

Pharma partners often want broad rights to that data for research purposes; diagnostic partners are increasingly reluctant to grant broad data rights, citing competitive concerns and data privacy obligations. That said, pharma sponsors have equally legitimate needs tied to regulatory requirements and long-term drug development strategy, including compliance with the Health Insurance Portability and Accountability Act and, where relevant, the General Data Protection Regulation and to support post-approval pharmacovigilance.

This asymmetry in the parties’ goals can be genuinely difficult to resolve. Both characterizations deserve scrutiny: pharma sponsors require sufficient data access to meet regulatory obligations, while diagnostic partners have legitimate grounds to resist terms that effectively transfer their core commercial asset without appropriate compensation. The negotiation tends to go better when it starts by disaggregating the data: regulatory submissions data, research use data, and commercialization data are all different things, are viewed with different significance by the parties, and should therefore be addressed in separate provisions with separate terms.

One dynamic worth noting from both sides: diagnostic partners who retain meaningful data rights may build a stronger long-term position than the royalty rate they are entitled to receive from their pharma product partner on product sales would suggest, while pharma sponsors who secure broad data rights may underestimate the downstream value of that access relative to other negotiated terms. Royalty streams in these deals are finite and tied to a single therapeutic’s commercial trajectory, while data rights can compound in value over time and across programs. Both parties should model the long-term value of data access before treating it as a secondary negotiating point.

Indemnification and Limitations of Liability

This is a negotiation that can hit on sensitivities. If data rights is where CDx negotiations stall, indemnification and limitations of liability is where they occasionally break down entirely. The commercial stakes are high enough, and the underlying risk genuinely difficult enough to allocate, that these provisions are increasingly the last items to close and the most likely to require senior business involvement to resolve.

The core tension is straightforward to describe and considerably harder to resolve. A diagnostic test that incorrectly identifies a patient as biomarker-positive may result in that patient receiving a therapy that is ineffective or, even worse, harmful. A test that incorrectly identifies a patient as biomarker-negative may result in that patient being denied a therapy that would have benefited them. Both failure modes carry potential liability, and the question of who bears it among the diagnostic developer, the pharma sponsor, the ordering physician, and the laboratory performing the test does not have a clean answer under current law.

Diagnostic partners have historically sought liability caps tied to fees paid under the agreement (which is a concept that is common among services agreements more generally), while pharma sponsors have sought exposure commensurate with the downstream consequences of a testing error. Both positions reflect rational assessments of asymmetric risk. However, the gap between the diagnostic company’s fee-based revenue and the pharma sponsor’s potential downstream exposure can be substantial, reflecting the structural difficulty of calibrating liability to risk in arrangements where the parties’ commercial outcomes are deeply interdependent.

Several specific issues are generating the most friction in current negotiations.

• Product liability allocation. When a patient suffers harm clearly or arguably attributable to an incorrect CDx result, both the diagnostic developer and the pharma sponsor are potential defendants. Agreements that do not address this scenario explicitly leave both parties exposed to joint litigation with misaligned interests at precisely the moment when coordinated defense would serve both parties better.

  It can be prudent to allocate primary defense obligations based on where the alleged failure occurred, distinguishing assay design from labeling from clinical use, with cross-indemnification obligations structured accordingly. Reaching that level of specificity can sometimes require input from regulatory counsel familiar with the FDA’s framework for in vitro diagnostics (“IVDs”).

• Consequential damages waivers and their limits. Consequential damages waivers are nearly universal in CDx agreements but present challenges for both parties: pharma sponsors find that their most significant potential losses may fall outside the waiver, while diagnostic partners face pressure to accept carve-outs that can, taken to the extreme, render their liability caps functionally meaningless. 

  The difficulty is that the most significant losses a pharma sponsor might suffer from a diagnostic failure, including lost revenues attributable to label restrictions, the cost of post-market study obligations, and expenses associated with product withdrawal, are precisely the categories a standard consequential damages waiver excludes. Pharma sponsors are increasingly seeking carve-outs from those waivers for these losses, while diagnostic partners continue to resist. The outcomes in how this tension is resolved in practice vary considerably depending on, e.g., the relative negotiating leverage of the parties and the degree to which each side has modeled its actual exposure.

• Indemnification for regulatory submissions. CDx co-development requires the diagnostic partner to support the pharma sponsor’s regulatory submissions, including the Pre-market Approval application or the applicable 510(k) pathway, with technical data and documentation relating to the CDx. 

  If that data later proves inaccurate or incomplete, the allocation of resulting regulatory exposure and the obligation to indemnify for the cost of responding to FDA inquiries or satisfying post-market requirements is frequently unaddressed in the original agreement. It is a gap that surfaces reliably, and almost always at the worst possible time for those who have not addressed it in the agreement.

• Artificial intelligence and distributed causation. As artificial intelligence (“AI”) and machine learning algorithms become integrated into CDx platforms, the liability analysis becomes materially more complex. When an AI-driven diagnostic tool produces an incorrect result, the failure may be attributable to the training data, the algorithm design, the clinical implementation, or some combination of all three. 

  Existing indemnification frameworks may not be designed to accommodate that kind of distributed causation, and the parties are increasingly negotiating algorithm-specific indemnification provisions that attempt to trace liability to the point of failure. The FDA’s evolving framework for AI-based Software as a Medical Device adds a further layer of complexity: as that framework develops, agreements that locked in a specific allocation of regulatory responsibility may find that the underlying regulatory obligations have shifted in ways that make the original indemnification structure difficult to apply and no longer fit-for-purpose.

These provisions require genuine negotiation grounded in a realistic assessment of where the risks actually lie, not the insertion of market-standard language that neither party has examined with appropriate care—and which, as noted along, may not be applicable in light of the every-changing CDx landscape. Agreements that do not appropriately and carefully address the issues outlined above are frequently the ones that generate the most expensive disputes later, typically in circumstances that neither party anticipated and that the agreement’s drafters would not recognize as the scenario they thought they were addressing.

Conclusion

The issues addressed here are not independent. For instance, a CDx agreement that handles platform scope loosely will almost certainly create data rights problems downstream and indemnification provisions that do not reflect a realistic assessment of where diagnostic failures actually occur. The drafting challenge is to address each area with specificity while keeping the overall structure coherent. That requires negotiators who understand both the commercial objectives and the operational realities of CDx development. The agreements that hold up are the ones where the parties did that work at the outset, not the ones that deferred the hard conversations to a future amendment or, worse, to litigation.