The Test That Decides Treatment: Why Companion Diagnostics Is the Most Invisible — and Most Valuable — IVD Category
Every targeted oncology drug approved in 2026 came with a mandatory test. CDx is the regulatory link between a molecular result and a treatment decision, not a product.
Every targeted oncology therapy approved in the first half of 2026 arrived with something attached: a mandatory test. Not a recommendation. Not a guideline preference. A regulatory requirement. You cannot prescribe the drug without running the assay, and the assay must be the one the regulator cleared. Not a lookalike. Not the version your laboratory happened to have on its menu. The test does not accompany the treatment. The test is the treatment decision.
This is the companion diagnostic, CDx, and it is the most invisible category in in-vitro diagnostics. Routine chemistry fills tenders. Immunoassay fills menus. Molecular platforms fill catalogues. CDx fills none of those. It sits in a regulatory file, referenced by a drug label, and it decides who receives therapy. In a sector that measures value in test volume, CDx measures value in treatment decisions. The unit of value is not the report. It is the patient who received the right drug because the right result was produced, and the patient who was spared the wrong drug because the test said so.
What a companion diagnostic actually is
A companion diagnostic is not a test that happens to be useful. It is a test whose result is required, by regulatory label, to prescribe a specific therapy. The link is formal. The drug’s approval is contingent on the diagnostic’s approval, and vice versa. Two separate dossiers. Two separate review processes. One linked decision.
In the United States, the FDA evaluates the drug and the diagnostic in parallel, but the diagnostic’s clearance specifies it as the authorised test for identifying patients eligible for the therapy. In Europe, the IVDR framework applies its own evidence requirements. Two separate regulatory tracks. The clinical logic is straightforward: the test identifies the biomarker, the biomarker predicts response, the drug works in responders. The regulator’s job is to verify that the chain holds.
The companion diagnostic does not reward the test. It rewards the link between a molecular result and a therapeutic decision, validated, audited, and enshrined in a drug label.
This is why CDx is structurally different from every other IVD category. A tumour marker can be ordered or not. A thyroid panel can run on any validated platform. A CDx assay, once linked to a drug, is the gatekeeper for that drug. Replacing it is not a menu decision. It is a regulatory event.
The pattern: every checkpoint inhibitor drags a test
Five CDx approvals landed in the first half of 2026. They span ovarian cancer, advanced melanoma, prostate, and bladder. The biomarkers are different (PD-L1 expression, PTEN loss, MRD detection), but the structural pattern is the same. Each targeted therapy and checkpoint inhibitor that depends on a molecular predictor arrives with a diagnostic obligation attached.
The portfolio of PD-L1 pharmDx assays, the tests that determine eligibility for checkpoint inhibitor therapy, now covers ovarian cancer in addition to the lung, gastric, and cervical indications already approved. One of three established IHC CDx vendors accumulated clearances across multiple indications this year. The reference laboratory that accumulates approvals obtained FDA clearance for a CDx in advanced melanoma and launched a PTEN IHC companion test for prostate cancer on the dominant IHC platform.
Read that not as a product list but as a structural fact: every checkpoint inhibitor is tied to a test. The manufacturer who develops the drug does not always develop the test. But the test must exist, be cleared, and be run. The CDx is not an accessory to precision oncology. It is the infrastructure.
MRD and CDx: from monitoring to selection
The most consequential shift in the last twelve months is not a new biomarker. It is the migration of minimal residual disease detection (MRD) from a monitoring tool to a treatment selector.
MRD testing detects molecular evidence of residual disease after treatment, typically through circulating tumour DNA in a blood sample. Until recently, its role was prognostic: detect MRD, and the patient is at higher risk of recurrence. The clinician monitors. The test informs surveillance.
In 2026, that role is changing. The FDA updated its companion diagnostic list to include a blood-based MRD test. An MRD assay seeking CDx positioning was presented to Japan’s PMDA as a companion diagnostic in bladder cancer. The acquisition that consolidated the MRD stack, a reference genomics company buying a specialised MRD developer for USD 1.5 billion, was not a monitoring play. It was a positioning play: the company that owns the MRD assay and the data platform is the company positioned to link MRD status to therapy selection.
MRD is migrating from the surveillance report to the treatment decision. When it arrives, the test that told you the patient was at risk will become the test that tells you whether to escalate, de-escalate, or switch therapy entirely.
This convergence, MRD as CDx, is the structural development to watch. The laboratories and manufacturers that built MRD as a monitoring service are now competing for a different prize: the regulatory link between a residual disease result and a treatment choice.
Decentralisation: from central lab to hospital
The companion diagnostic has been, for most of its history, a centralised test. The assay runs in a reference laboratory. The sample ships. Turnaround is measured in days. That model works when the CDx is used for initial treatment selection: there is time, and the central lab’s standardisation justifies the logistics.
It does not work when the CDx becomes a tool for ongoing treatment decisions (escalation, de-escalation, switching) in patients already on therapy. That requires the test to run closer to the patient.
The oncology analysis platform’s multi-year agreement with a pharmaceutical company for decentralised CDx deployment is a signal. The intent is to move CDx capability from central reference labs to hospital laboratories, not by shipping samples faster but by placing validated assays and analysis pipelines where the treatment decision is made. The decentralisation of CDx is not a logistics optimisation. It is a clinical model change: from “send the sample, wait for the result” to “run the test where the patient is treated.”
The LATAM gap: regulatory, not technological
In Latin America, the companion diagnostic conversation is not about technology. The assays exist. The platforms are installed. The oncology analysis platform operates in hospitals across the region. The dominant IHC platform has a substantial installed base. The question is not whether the test can be run. It is whether the regulatory framework recognises it as a companion diagnostic, and whether that recognition enables or obstructs access to the linked therapy.
COFEPRIS observes FDA and EMA decisions. When the FDA clears a CDx, the Mexican regulator evaluates whether to follow. The lag is not measured in months. It is measured in the absence of a dedicated pathway: COFEPRIS does not have a structured CDx framework that links diagnostic clearance to a drug label. The therapy may be available. The test may be available. The formal link between them, the regulatory bridge that makes the test a gatekeeper for the drug, is not.
In LATAM, the companion diagnostic gap is regulatory, not technological. The instruments are there. The reagents are there. The regulatory bridge between result and treatment is not.
This matters because the laboratories that understand CDx (which assays carry regulatory weight, which biomarkers are linked to which therapies, what it takes to validate a CDx locally) are the laboratories that will capture the specialised oncology testing that follows precision medicine into the region. The laboratories that treat CDx as another line on the molecular menu will discover that the value was never in running the assay. It was in the link.
What to do now
If you direct a laboratory in LATAM and this still reads as an abstraction, consider three concrete actions.
First, map your installed base against the CDx landscape: which platforms you have, which IHC and molecular assays you run, and which map to therapies with active or pending CDx clearances. You may be closer to the category than you think, but missing the regulatory step that converts capability into clinical authority.
Second, track the MRD-to-CDx convergence. The test that monitors recurrence today will select therapy tomorrow. If your laboratory offers MRD testing, understand whether the assay you run is the one the regulator will recognise, or whether a different platform is positioning itself as the CDx. The distance between a monitoring test and a companion diagnostic is one regulatory filing.
Third, watch the EU’s Project COMBINE. The framework for co-developing drug, device, and IVD in a single clinical trial is the model that will define how CDx is built in the next decade. It is the template that regulators without a CDx pathway, including those in LATAM, will eventually adapt.
The companion diagnostic is the test that decides treatment. The laboratories that understand that, not as a technical detail but as a shift in where diagnostic value lives, will not be reading the market. They will be shaping it.