| Gene | Category | Predicted log2FC | Direction | Gap Detected |
|---|
Analyze receptor occupancy and pathway modifications from concurrent medications.
Parameter uncertainty propagated through the full mechanistic model to quantify output variability.
Functional Monte Carlo, n = 512, plus global Sobol sensitivity (base N = 256;
2,304 evaluations). Seven hormone-layer parameters perturbed uniform ±25% around model
nominal and propagated to cycle day 21. Source: ASME V&V 40 Credibility
Dossier v1.0, §5 (outputs/uq_montecarlo_v1.0.json,
outputs/uq_sobol_v1.0.json).
| Output (cycle day 21) | Mean | 95% CI | CV |
|---|---|---|---|
| E2 peak (pg/mL) | 102.8 | [72.7, 136.4] | 16.9% |
| P4 peak (ng/mL) | 9.33 | [6.26, 13.29] | 19.1% |
| Decidualization | 0.779 | [0.703, 0.832] | 3.9% |
| Receptivity | 0.555 | [0.262, 0.688] | 19.4% |
Most uncertain output: receptivity (CV 19.4%), driven by luteal progesterone
— production k_P4_luteal (Sobol ST 0.59) and clearance
d_P4 (ST 0.49). Decidualization is well-buffered (CV 3.9%).
Endogenous baseline at cycle day 21 across the three hormonally distinct life stages
(OTA-26-004 §5). Source: outputs/grant_roa_triad_v1.0.json.
| Life stage | Peak E2 (pg/mL) | Decidualization | Receptivity |
|---|---|---|---|
| Reproductive | 100.6 | 0.780 | 0.597 |
| Perimenopause | 82.7 | 0.856 | 0.359 |
| Postmenopause | 40.0 | 0.008 | 0.000 |
Postmenopausal endometrium shows near-zero receptivity (atrophic, low-estrogen milieu),
consistent with physiology. New this build: life-stage parameterization (§5) and
toxicity go/no-go scalars (§4.5); see outputs/grant_capability_matrix_v1.0.json.
Competitive receptor binding propagated to tissue outputs (OTA-26-004 Deliverable 1).
Values shown baseline → on-treatment at the indicated cycle day. Source:
outputs/grant_drug_pharmacology_v1.0.json.
| Drug (class), context | Receptivity | Decidualization | Mechanism |
|---|---|---|---|
| Tamoxifen (SERM), reproductive | 0.60→0.46 | 0.78→0.71 | ERα antagonism at high E2 |
| Tamoxifen (SERM), postmenopause | — | prolif 0.031→0.042 ↑ | paradoxical agonism (low E2) |
| Mifepristone (SPRM antagonist) | 0.60→0.08 | 0.78→0.67 | PR antagonism collapses receptivity |
| Leuprolide (GnRH agonist) | 0.60→0.00 | 0.78→0.07 | flare then ~90% E2 suppression (73→7 pg/mL) |
| MPA (progestin), proliferative phase | — | 0.076→0.086 ↑ | PR agonism drives decidualization |
| Letrozole (aromatase inhibitor) | 0.60→0.57 | 0.78→0.77 | E2 synthesis suppression (73→64 pg/mL) |
Tamoxifen's opposite effect at high vs low estrogen (antagonist reproductive, paradoxical agonist postmenopause) emerges from the competitive-binding mechanism, not a hardcoded rule. Research-use, hypothesis-generating; not for clinical decisions.
Global variance-based total-order sensitivity (ST) for the clinical readout (receptivity) at cycle day 21. Saltelli sampling, Jansen estimator, base N = 256 (2,304 evaluations); bootstrap 95% CI. Source: V&V 40 Dossier v1.0 (outputs/uq_sobol_v1.0.json).
| Parameter | Total-order sensitivity (receptivity) | ST | 95% CI |
|---|---|---|---|
k_P4_luteal | 0.59 | ±0.12 | |
d_P4 | 0.49 | ±0.12 | |
d_E2 | 0.05 | ±0.01 | |
k_E2_luteal | 0.01 | ±0.00 | |
k_P4_basal | 0.00 | ±0.00 | |
k_E2_follicular | 0.00 | ±0.00 | |
t_ovulation | 0.00 | ±0.00 |
Total-order indices (ST) capture each parameter's full effect including interactions. Receptivity is dominated by luteal progesterone kinetics; estradiol-peak timing (d_E2, t_ovulation) instead governs the E2 output. Full per-output indices are in the Credibility Dossier v1.0.
Spatial zone decomposition of endometrial tissue compartments and pathway scores.
Simulate follicular growth, predict oocyte yield, assess OHSS risk, and compare stimulation protocols using POSEIDON stratification.
Probabilistic posterior over coarse cycle phases from a single serum-style E2/P4 observation (literature-shaped Gaussian likelihoods). Quantifies phase support and ambiguity (entropy) — not histologic ground truth.
Scans a literature-shaped simulated hormone trajectory for six adverse states (hypoestrogenic, endometrial atrophy, hyperstimulation, premature luteinization, progesterone resistance, breakthrough-bleeding risk). Research-grade heuristics — not validated for clinical decision-making.
The two pre-specified OTA-26-004 Aim-2 acceptance metrics over a longitudinal scenario: (1) sustained proliferation > 0.7 for ≥ 4 consecutive weeks; (2) hyperplasia-prone attractor entry rate (proliferation > 0.7 & WNT > 0.5 & apoptosis < 0.3). Reduced longitudinal surrogate; research-grade hypothesis-generation only.
The same assessment across three hormonally distinct life stages (reproductive, perimenopause, postmenopause), using representative mid-luteal hormones and receptor-baseline scaling. Uses the condition selected above. Receptor scales are research priors (uncalibrated).
One-compartment oral pharmacokinetics (first-order absorption/elimination, dose superposition) with an Emax PD bridge to E2/P4. PK priors from published literature. Research-grade.
34 named clinical regimens with E2/P4 modifier factors, grouped by indication (endometriosis, adenomyosis, fibroids, hyperplasia, cancer, PCOS, RIF, luteal support, thin endometrium, IVF/FET).
Research-grade proxy pairing an efficacy axis (proliferation suppression for proliferative disease, or decidualization/receptivity gain for support indications) against residual hyperplasia-prone risk, from applying a regimen's E2/P4 factors. Uses the condition selected above.
Section 3060(a) of the 21st Century Cures Act (Pub. L. 114-255, enacted December 13, 2016) amended Section 520 of the Federal Food, Drug, and Cosmetic Act (21 U.S.C. 360j) to add subsection (o), which excludes certain clinical decision support software from the definition of a "device" under Section 201(h) of the FD&C Act. Software meeting all four criteria of Section 520(o)(1)(E) is not a medical device and requires no FDA clearance, approval, or registration.
FDA guidance: "Changes to Existing Medical Software Policies Resulting from Section 3060 of the 21st Century Cures Act" (January 2026, superseding September 2022 version). Available at: FDA.gov Guidance Documents.
| # | Statutory Criterion | EndoTwin-W (prior internal assessment) | Met? |
|---|---|---|---|
| 1 | Not intended to acquire, process, or analyze a medical image, a signal from an in vitro diagnostic device, or a pattern or signal from a signal acquisition system. | EndoTwin-W accepts numerical laboratory values (glycodelin-A and CA-125 concentrations) manually entered from a standard blood draw report. It does not process medical images, raw IVD instrument signals, or signal acquisition patterns. | Yes |
| 2 | Intended for the purpose of displaying, analyzing, or printing medical information about a patient or other medical information. | EndoTwin-W analyzes and displays patient-specific medical information: Progesterone Resistance Score, pathway disruption profiles, predicted gene expression changes, biomarker assessment, and window-of-implantation timing predictions. | Yes |
| 3 | Intended for the purpose of supporting or providing recommendations to a health care professional about prevention, diagnosis, or treatment of a disease or condition. | EndoTwin-W provides recommendations to OB/GYN physicians regarding endometrial receptivity assessment, progesterone resistance classification, and condition-specific pathway disruptions as a research aid for studying endometrial receptivity. | Yes |
| 4 | Intended for the purpose of enabling such health care professional to independently review the basis for such recommendations so that it is not the intent that the professional rely primarily on any of such recommendations to make a clinical decision regarding an individual patient. | EndoTwin-W displays full mechanistic transparency: individual pathway disruption scores, per-gene predicted expression changes with fold-change magnitudes, underlying biomarker values, radar visualization of pathway contributions, and gap detection analysis. The clinician can independently evaluate the basis for every recommendation and is expected to exercise independent clinical judgment. | Yes |
References: 21st Century Cures Act, Pub. L. 114-255, Section 3060(a) (2016); 21 U.S.C. 360j(o)(1)(E); FDA Guidance, "Changes to Existing Medical Software Policies Resulting from Section 3060 of the 21st Century Cures Act" (January 2026).