Why we are not a medical device, yet.

Whenever the underlying science is real and the demos in the room look impressive, there is pressure inside an early company to let the marketing get out in front of the regulatory file. A lot of consumer health technology has done it. Most of those companies have paid for it eventually, either through a warning letter, a recall, or a quiet erosion of credibility with the clinicians they were trying to reach. We are not going to do that. The next few pages are a plain account of where Chronic Trace sits on the FDA pathway today, the claims we can and cannot make from a labeling perspective, and what the word "yet" in the title is supposed to carry.

What Software-as-a-Medical-Device means

The U.S. Food and Drug Administration regulates medical devices, which under the relevant statute means anything intended for use in diagnosing, curing, treating, or preventing disease.^[1] Software falls under this when its intended use meets one of those criteria. The FDA's term of art for this category is Software-as-a-Medical-Device, abbreviated SaMD.^[2] It is not a new category. It is the same Center for Devices and Radiological Health regulating the same kind of clinical claim, with the added complexity that the device is code.

Software that claims to diagnose a concussion is a medical device. Software that claims to assess concussion likelihood as decision support for a qualified clinician is also a medical device. Software that flags wellness signals warranting evaluation by a qualified clinician is sometimes a medical device and sometimes not, depending on how the intended use is written. The line is set by intent, not by what the software can do. Two pieces of code that perform the same computation can sit on different sides of the line depending on what the company says the software is for.

Chronic Trace today is positioned as wellness screening. The platform produces measurements and indicators. It does not diagnose, does not treat, does not prescribe. In clinical output we avoid the word "concussion" entirely. The composite report uses language closer to "wellness signals warranting evaluation by a qualified healthcare professional" than anything that resembles a diagnostic label. That is the appropriate regulatory posture for a platform whose validation data is still being collected.

The 510(k) pathway

Most novel SaMD products clear the FDA through one of three pathways: 510(k) clearance,^[3] de novo authorization,^[4] or premarket approval (PMA).^[5] The pathway depends on the device classification and on whether a substantially equivalent predicate device already exists.

Our planned pathway is 510(k) clearance for the concussion- likelihood and TBI severity indications. Two FDA-cleared predicates exist in the rough class of mobile concussion- assessment tools, which is the regulatory anchor that makes 510(k) viable rather than de novo. The 510(k) process requires analytical validation (does the device measure what it claims to measure, with characterized precision and accuracy across the intended use environment) and clinical validation (does the measurement correlate with the clinical outcome it claims to inform, against reference instruments and in the intended use population).

The Suicide Risk Index in the platform's index family is a different situation. There is no equivalent FDA-cleared predicate. That submission targets the de novo authorization pathway, which is for novel devices without predicates and is more involved.

Both pathways start with the same step: a pre-submission meeting with the FDA Center for Devices and Radiological Health. The Q-Sub, as it is called, lets the company present the device, the intended use, the planned analytical and clinical validation, the predicate or de novo argument, and the proposed labeling and indications for use. The FDA responds with non-binding feedback on what the eventual submission needs to contain to be reviewable.^[6] Our planned Q-Sub for the concussion and TBI severity indications is scheduled for the third quarter of 2026.

Predetermined Change Control Plans

One of the harder problems in regulating SaMD is the inherent tension between a static regulatory submission and software that learns and updates. A traditional medical device is cleared once and shipped unchanged. AI-enabled SaMD, by contrast, is built to improve. The FDA's solution, finalized in December 2024, is the Predetermined Change Control Plan framework.^[7]

PCCP is a section of the regulatory submission in which the sponsor describes, in advance, the types of changes the software is expected to undergo over its lifecycle. If the FDA accepts the PCCP, the sponsor can make those pre-authorized changes without filing a new 510(k) for each one. The PCCP framework is specifically designed for adaptive AI medical devices and is the right framework for a platform like ours, which is going to be tuned on validation data, refined on cohort outcomes, and updated as the patent-protected scoring methodology compounds across indications.

The FDA maintains a public list of AI/ML-enabled medical devices authorized for market; that list grew by record-level annual cohorts through 2024 and continues to widen.^[8] The regulatory door for adaptive SaMD has not been wider. That is one of the reasons the timing of the platform makes sense now in a way it would not have made sense five years ago.

What we can claim today

Until 510(k) clearance lands, the platform's public claims sit inside a narrow set. We can describe the architecture: a commodity-device capture that measures involuntary biomarker channels and fuses them through a proprietary methodology into a calibrated longitudinally-comparable index. We can describe the science: every individual biomarker channel the platform measures has decades of published clinical literature behind it. We can describe the intellectual property: nine U.S. provisional patent applications filed between January and May 2026 across the scoring foundation, the optical modalities, the stroke and psychiatric extensions, the compute architecture, and the neurodegeneration, oncologic, and infectious disease indications.

What we cannot claim today: any specific clinical sensitivity or specificity against a reference instrument; any diagnostic capability; any claim that the platform identifies, prevents, or treats a specific medical condition; any comparison to a cleared device that implies clinical equivalence. These claims wait for the validation data.

I know that frustrates readers who would like to evaluate the platform's clinical utility before the pivotal cohort closes. The blunter answer is that anyone who hands you clinical sensitivity numbers without a powered prospective cohort to back them is, in clinical research terms, making them up. We are not interested in joining that group.

What sits between here and clearance

The work between today and a 510(k) submission is not mysterious. It is a specific sequence of deliverables with defined inputs and defined outputs.

1. Q-Sub completion with FDA CDRH for the concussion-likelihood and TBI severity indications. Target: third quarter 2026.
2. Pivotal validation cohort enrollment begins. Target: fourth quarter 2026. The cohort design includes the paired biomarker-and-outcome data the analytical and clinical validation argument requires.
3. Department of Defense and Veterans Affairs pilot deployment for TBI screening. Target: second quarter 2027.
4. 510(k) submission for the Concussion Likelihood Index and TBI Severity. Target: third quarter 2027.
5. Initial commercial deployment in sports and military channels behind the cleared indications. Target: fourth quarter 2027.

Behind these milestones is the engineering work that does not show up on a regulatory roadmap but determines whether the roadmap lands. The platform's capture pipeline is migrating to native first-party operating-system computer-vision APIs on its primary deployment hardware, which materially strengthens the analytical-validation story by replacing a third-party model dependency with a vendor-supported first-party API on a defined device matrix. That migration is independent engineering work in flight this sprint and the next.

Why the careful framing matters

Three reasons.

First, the clinical claim carries the weight.A company that overstates clinically tends to end up in one of two places: either the overstatement gets caught and the regulatory and reputational consequences cascade, or it does not get caught and a clinician somewhere makes a real decision about a real patient using a number that does not have the validation it implies. Both outcomes are worse than moving slowly.

Second, the regulatory record is permanent.Every public claim we make this year is going to be read by an FDA reviewer next year, checking whether the company's framing of the device lines up with the indications for use the submission is asking to be cleared for. The cleaner that read is, the smoother the submission moves. The more friction we create today, the more corrective work we accumulate later.

Third, the people we want to reach can tell.The clinical advisor evaluating the platform, the institutional investor doing technical diligence, the sports medicine director considering a pilot, all of them have seen the alternative. They have seen companies that overclaim and watched what happened next. Careful framing reads to them as a credibility signal, not a hedge. Treating regulatory posture as something the marketing inherits from, rather than something the marketing departs from, is among the most durable trust signals a research-stage SaMD company can send.

The word "yet"

The title of this essay ends with the word "yet" for a reason. It is the difference between a company that has a real pathway, knows the work between it and clearance, and is doing that work in a defined sequence, versus a company that hopes the regulatory environment will eventually accommodate it. We are the first kind.

The platform is not a medical device today. The biomarker science under the platform is decades old. The way we are combining those biomarkers and the device they ride on is new. The clinical validation that turns the combination into a cleared device is in flight. The honest answer to "when?" is: when the cohort reads out, when the Q-Sub feedback is worked, when the submission goes in, when the review clears. Each of those is a specific event with a date attached, not a marketing line.

Until those events have happened, we are research-stage Software-as-a-Medical-Device in development. That phrase appears on every page that touches a prospective user, every email to a prospective investor, and every conversation with a clinician evaluating the platform. It appears because it is accurate, and because the company we want to be in five years is the company that said it before anyone forced us to.