The number that decides a rare-disease forecast is usually the one nobody has measured: the patients who have the disease but not the diagnosis. In most rare conditions the diagnosed population is a fraction of the true one, so a forecast anchored to who is coded today understates the market that actually exists. Transthyretin amyloidosis of the heart is the clearest case, and it shows why the undiagnosed pool is not a caveat on the number. It is the number.
The diagnosed are the minority
Transthyretin cardiac amyloidosis affects an estimated 100,000 to 150,000 people in the United States, and most of them do not know it. The disease hides inside heart failure with preserved ejection fraction, where a patient-flow view counts more than 500,000 people over 70 who likely carry undiagnosed wild-type disease, the form that accounts for roughly 80 percent of cardiac ATTR. Against that pool, only 70,000 to 100,000 patients are diagnosed and treated today. The treated market is the visible tip; the disease is the iceberg.
Why the patients hide
ATTR stays hidden because its signature looks like ordinary heart failure and the test that confirms it was not, until recently, part of the work-up. A breathless elderly patient with a thick-walled heart is managed as common heart failure, and the amyloid underneath is never named. The lesson generalises across rare disease: the rate-limiting step is not efficacy or access, it is whether the patient is ever identified, and that sits upstream of everything a brand plan usually optimises.
Case-finding is an engineering problem, not a fixed rate
The undiagnosed pool is not a constant. It converts as fast as the health system is built to find it. Cardiac ATTR can now be confirmed non-invasively by technetium-pyrophosphate scintigraphy, which replaced biopsy, so diagnostic-scan capacity rather than disease biology sets the pace. The United Kingdom is the clearest natural experiment: a national scintigraphy pathway introduced in 2021 took annual diagnoses from under 500 to several thousand across roughly 30 centres. The Gulf shows the mirror image, where fewer than eight scintigraphy centres across six countries convert the same underlying burden far more slowly, regardless of price or formulary status.
You forecast the case-finding rate, not the prevalence
A defensible ATTR forecast runs on the diagnostic funnel, not the prevalence estimate. You start from the undiagnosed pool, apply a scan-suspicion rate, a scan rate, and a confirmation yield, and you run the sensitivity on scintigraphy penetration, on centres added per year and referrals per centre, rather than on price. Every incremental scanner, not every rebate, converts a findable patient into a treated one. The undiagnosed pool is not a footnote to the addressable market. It is the addressable market, and sizing it is the work.
The shape holds across rare disease
The market a team can see is bounded by the diagnostic pathway; the market it could have is bounded only by how hard the system looks. Treat the undiagnosed pool as the opportunity, size the case-finding rate as the variable, and the forecast reflects the disease rather than the coding. Ignore it and the number describes today's diagnosed minority as if it were the whole. This is one of the shifts behind the rare-disease sizing discipline: the right question is not how many patients are diagnosed, but how many the system can be built to find.
AXLRx sizes rare-disease markets from the undiagnosed pool down, with the case-finding rate made explicit and stress-tested by geography. See the ATTR amyloidosis briefs, or commission a model scoped to your asset.