A Manifesto

Most of medicine operates on a model in which the physician sets the destination and the patient's role is to trust the expertise that leads there. When a tumor has to be removed or a fracture has to be set, this is exactly the right model, and it works because the goal itself is rarely in question.

Gynecology works differently, for reasons that have nothing to do with the quality of its physicians. A large share of gynecological decisions belong to a category the medical literature calls preference-sensitive: whether to operate on endometriosis now or manage it medically, whether to preserve the uterus, whether and when to freeze eggs, which hormonal treatment to accept and for how long. For these questions there is no single medically correct answer, because the best option depends on variables that live outside the clinic, such as the children a woman plans or does not plan, the demands of her work, the pain she is prepared to carry, and the risks she is willing to accept in exchange for relief.

The practical consequence is that the basic unit of good gynecological care is a conversation between two people who each hold information the other needs. The physician brings clinical expertise and, ideally, a precise understanding of what is happening in this particular body rather than in a population average. The patient brings the life the decision will land in, which no one else can supply. When either side enters that conversation underinformed, the decision degrades quietly, and the woman lives with the consequences for years or decades.

Our starting conviction is that every woman deserves to enter that conversation fully informed, and that her physician deserves the same. In practice, both are currently working with far less information than the stakes would justify.

The shortfall is a data problem with a long history, and it can be quantified.

Until 1993, women were not required to be included in clinical trials funded by the National Institutes of Health, which means a substantial part of the evidence base of modern medicine was built on male bodies and extrapolated to female ones. Conditions that affect only women entered the research agenda late and remain thinly funded to this day: by one widely cited estimate, female-specific conditions outside of oncology receive around one percent of global healthcare research funding.

Endometriosis, the condition where we begin, illustrates the pattern well. It affects roughly one woman in ten of reproductive age, close to two hundred million people worldwide, and the average interval between first symptoms and diagnosis is seven to ten years, a period typically spent between appointments and inconclusive results, accumulating explanations that never quite fit. Relative to its prevalence and its burden, endometriosis receives some of the lowest research investment per patient in medicine.

It is worth being precise about what kind of information is actually missing, because the internet holds an enormous amount of material about women's health. What is scarce is evidence of clinical grade, by which we mean data that is collected with explicit consent, structured to a consistent standard, labeled by specialists, followed over time, and tied to real outcomes. Very little of the available material meets that bar, and the shortfall compounds on itself: a model cannot learn from data that was never collected, and care cannot be benchmarked against a ground truth that does not exist. Research requires clinical-grade data, clinical-grade data requires infrastructure purpose-built to produce it, and in gynecology that infrastructure is largely missing.

Building it is slow and unglamorous work, done one properly consented and properly documented exam at a time. We think it is the most useful work available in this field, and it is the work we have chosen.

Matricis.ai's ambition is to be a foundational brick of clinical-grade evidence in women's health: an institution that produces the data and conducts the research, then returns the knowledge to the two people who need it most, the woman and her physician, via Geneviève's Institute.

We begin with imaging, and the choice follows from the economics of information. A pelvic MRI is one of the densest diagnostic objects in medicine, and when the exam is acquired to a dedicated gynecological protocol, a trained reader can learn a great deal about a woman's body from a single study.

A good pelvic MRI is the clearest windows we have into a woman's reproductive anatomy, and a specialist reading it well can see far more than most people imagine. It shows the uterus, the ovaries, and the surrounding tissue in enough detail to tell many conditions apart from one another, which matters because so many of them produce the same everyday symptoms of pain, heavy bleeding, or trouble conceiving.

It can identify endometriosis, including the deep and hidden forms that a standard ultrasound often misses, and show where in the body it has taken hold. It can find fibroids and describe their size and position, which is frequently what determines whether they need treatment at all. It can distinguish a harmless ovarian cyst from one that warrants a closer look, and it can pick up adenomyosis, a common and under-recognized cause of painful, heavy periods. It offers clues about fertility, from the health of the ovaries to conditions that can quietly interfere with conception. And it can flag the early warning signs of something more serious, the features that separate a benign mass from one that needs urgent attention, which is part of why careful imaging can matter so much.

Two things make this kind of exam especially valuable. It is objective, so a finding can be measured precisely rather than described in impressions, and it is comparable over time, so the same woman's scan a year later can show whether a condition is stable, improving, or advancing. Much of what a woman and her physician need in order to understand what is happening and to choose a path together is already contained in a single exam, waiting to be read properly.

Reading all of it well is genuinely difficult. Pelvic MRI is one of the most demanding studies in radiology, the expertise it requires is rare, and the way the exam is performed and reported varies widely from one place to the next. The findings that matter most to a woman's decision tend to be the subtlest ones, a small area of deep disease or an early change in the muscle of the uterus, and these are exactly the findings most likely to be missed when the exam is rushed or read by someone without the specific training. The imaging is very often already there. What is missing is the reading, with the time and the expertise it deserves, and we believe this reading has to be the heart of the program.

The imaging is the core, and we intend to connect it to everything around it through Geneviève Institute, because an image is most useful when it sits alongside the rest of the picture. A finding is read in the context of a woman's symptoms and history rather than on its own, and when the results call for specialized care, we introduce her to the physician best suited to her situation, so that the report becomes the beginning of the conversation described at the start of this document. Over time the same exam becomes a baseline she can be measured against, which turns a single snapshot into a record of how her condition actually behaves.

Each exam also does quiet work beyond the individual. With the woman's explicit consent, and handled to clinical standards of protection, it is structured and labeled by specialists, then added to the dataset, where it strengthens the evidence base for the women who come after her. Diagnosis funds research and research sharpens diagnosis, and this loop is the engine of the institution.

Precise diagnosis in routine care is one of the missing ingredients in the search for better treatments, and its absence is one reason progress in women's health has been so slow. The connection runs through three problems that all trace back to the same root.

The first is finding patients at all. A treatment can only be studied on women who have been correctly identified as having the condition, and when a disease is under-diagnosed, trials cannot recruit fast enough. Recruitment that drags on for years makes a study slower, more expensive, and riskier to run, and that cost quietly discourages companies from pursuing the disease in the first place. Under-diagnosis does not only delay individual women; it removes the incentive to develop the very treatments they need.

The second is knowing who is being treated. Endometriosis behaves less like one disease than like several patterns of disease that likely respond to different treatments. When the diagnosis at the door is imprecise, a trial fills with women whose conditions are subtly different, the real effect of a promising drug is diluted across a mixed group, and a treatment that would have helped a specific subset can appear to fail everyone. Imaging read by a consistent expert standard makes it possible to tell those groups apart, to enroll the right women in the right studies, and to see which therapy works for whom.

The third is the hardest, and the most overlooked. A clinical trial can only measure success as well as its instruments allow, and for endometriosis those instruments barely exist. There is no established, objective way to measure the burden of disease, no reliable measurement of lesion size that a trial can track before and after treatment. So studies fall back on what they can capture, mostly subjective pain scores, which are noisy, easily confounded, and slow to move. A drug can be shrinking disease in the body without that improvement showing up cleanly in the numbers a trial is allowed to report. Objective imaging, measured precisely and compared over time, offers the field a real endpoint: a way to see whether a treatment is actually changing the disease rather than only changing how a woman reports feeling on a given day. Better measurement is the difference between a trial that can prove a treatment works and one that cannot.

Better diagnosis in ordinary practice makes better research possible, and better research is how the field finally earns treatments that are targeted rather than approximate. Diagnosis funds research and research sharpens diagnosis, and each in turn makes better therapies possible.