The most comprehensive AI system ever built for PE prediction, detection, and prevention — analyzing every risk factor, every medication, and every physiological signal that precedes clot formation.
Pulmonary embolism is the most preventable cause of hospital death — yet it remains one of the most frequently missed diagnoses in medicine. PE kills more people each year than breast cancer and motor vehicle accidents combined. The tragedy is not that we lack the science to prevent it. The tragedy is that the warning signs are there — in the medications, the immobility, the genetics, the vital signs — and no system is watching for them all at once.
Current diagnostic tools — the Wells score, D-dimer, revised Geneva score — were designed to assess patients already suspected of having PE. They were never built to predict PE before symptoms appear. Sentinel PE is. It continuously monitors every admitted patient's risk constellation — medications, mobility, hemodynamics, coagulation markers, genetic predisposition, and surgical status — to identify the patients hurtling toward PE while intervention is still possible.
Dozens of common medications elevate PE risk — from oral contraceptives to hormone therapy to oncology agents. Sentinel PE is the first system to continuously profile every patient's medication-driven thrombotic risk and adjust monitoring intensity in real time.
Combined oral contraceptives are the single most under-recognized cause of PE in young women. The risk is 2–9× higher than non-users, but the magnitude depends critically on the specific formulation. Third-generation progestins (desogestrel, gestodene) and fourth-generation progestins (drospirenone) carry significantly higher VTE risk than second-generation levonorgestrel. The risk is highest in the first 3–6 months of use and compounds dramatically when combined with smoking, obesity, or undiagnosed thrombophilia.
In a landmark New Zealand case-control study, current users of combined oral contraceptives had a relative risk of fatal PE of 9.6 — with the median age of death being just 29 years. Pulmonary embolism occurred in 78% of patients in an analysis of oral contraceptive-related thrombotic complications.
Oral menopausal hormone therapy doubles the risk of PE, with the highest risk concentrated in the first year of use and among first-time users, who face a risk multiplied by 2.07. A critical and underappreciated finding: transdermal estrogen delivery (patches, gels) is not associated with any increased PE risk — a distinction that could save lives but is rarely factored into prescribing decisions.
The type of progestin matters. Medroxyprogesterone acetate carries a modestly higher PE risk than norethisterone acetate when combined with estrogen. Sentinel PE monitors every HRT prescription, adjusting risk profiles based on route of administration, progestin formulation, duration of use, and patient-specific risk factors.
Testosterone replacement therapy — increasingly prescribed for hypogonadism in aging men and as part of gender-affirming care — carries an emerging but incompletely characterized VTE risk. The mechanism involves testosterone-induced polycythemia (elevated red blood cell mass), which increases blood viscosity and thrombotic potential. The risk appears highest in the first 6 months of therapy and in patients whose hematocrit rises above 52%.
Cancer itself is a major independent risk factor for VTE, with a 4–7× elevated baseline risk. But specific chemotherapy agents dramatically compound this further. Platinum-based agents, anti-angiogenics, immunomodulators (thalidomide, lenalidomide), and tamoxifen each carry their own thrombotic mechanisms. Cancer patients also face compounding risk from central venous catheters, immobility, and surgery.
Antipsychotic medications — particularly low-potency conventional antipsychotics and clozapine — are associated with 1.5–3× increased VTE risk through dual mechanisms: direct effects on coagulation pathways and indirect effects through sedation-induced immobility and weight gain. This risk is rarely discussed with patients and is compounded in psychiatric populations who may have limited ability to report PE symptoms.
Sentinel PE doesn't wait for symptoms. It monitors seven distinct pathways to PE — from clot formation in the deep veins to hemodynamic compromise in the pulmonary vasculature — and intervenes at every stage.
Sentinel PE maintains a continuously updated pharmacovigilance database of 200+ medications with known or suspected VTE associations. When any of these agents is prescribed — or when combinations compound risk — the system recalculates the patient's thrombotic risk score in real time, factoring in dose, duration, route of administration, and patient-specific modifiers including BMI, smoking status, genetic thrombophilia, and concurrent medications.
Half of all proximal DVTs will propagate to the pulmonary vasculature if untreated. Sentinel PE monitors patients with known DVT for signs of clot propagation and embolization — tracking D-dimer trends, leg circumference changes, hemodynamic shifts, and respiratory parameters that precede PE. For patients without diagnosed DVT, the system monitors for subclinical clot formation using subtle signals: unilateral leg edema, elevated venous pressure markers, and positional vital sign changes.
Surgical patients face dramatically elevated PE risk — from 5× for minor ambulatory procedures to 50× for major orthopedic surgery. Yet current prophylaxis protocols are inconsistently applied, and monitoring often drops off at discharge despite the risk persisting for 30+ days. Sentinel PE risk-stratifies every surgical patient by procedure type, duration, anesthesia method, and patient factors, then maintains enhanced surveillance through the entire high-risk window — including post-discharge monitoring via patient-facing alerts.
Venous stasis from immobility is one of Virchow's triad — and it's the most modifiable risk factor for PE. Sentinel PE integrates with bed sensor systems, nurse mobility documentation, physical therapy records, and wearable accelerometer data to build a real-time mobility profile for every patient. When mobility drops below safe thresholds — adjusted for the patient's clinical status — the system alerts nursing staff and suggests specific mobilization interventions.
When a clot reaches the pulmonary vasculature, the right ventricle begins to strain before the patient feels symptoms. Sentinel PE detects this strain through subtle patterns in heart rate variability, blood pressure waveforms, central venous pressure trends, SpO2 micro-desaturations, and respiratory rate changes. The engine identifies these hemodynamic fingerprints of early PE — often hours before tachycardia, hypotension, or dyspnea become clinically apparent.
Even after PE is clinically suspected and CTPA is ordered, the diagnosis can still be missed. CTPA has a documented missed diagnosis rate of 14% and an overdiagnosis rate of 10%. Sentinel PE's imaging engine analyzes every CTPA scan in real time — detecting clot burden, quantifying right ventricular strain, classifying PE severity (subsegmental, segmental, saddle), and immediately escalating life-threatening findings to the on-call team before the radiologist's formal read. The system has demonstrated sensitivity of 92.6% and a negative predictive value of 98.6%.
Hereditary thrombophilia — Factor V Leiden, prothrombin G20210A mutation, protein C/S deficiency, antithrombin deficiency — is present in 5–10% of the population but dramatically amplifies PE risk when combined with acquired factors. A woman with heterozygous Factor V Leiden who is prescribed combined oral contraceptives faces a 30× increased risk of VTE compared to a woman without the mutation who is not on the pill. Yet thrombophilia screening before contraceptive prescription remains uncommon. Sentinel PE integrates available genetic test results and family history data to flag these lethal combinations before they cause harm.
Sentinel PE monitors every known PE risk factor simultaneously — no other system does this.
25% of PE patients present with sudden death as their first symptom. Sentinel PE doesn't just alert clinicians — it educates and empowers patients to seek help at the earliest warning signs.
When a patient's risk profile crosses a threshold, Sentinel PE generates a personalized, plain-language risk briefing — explaining their specific risk factors, the symptoms to watch for, and exactly what to do if they experience them. Delivered via the patient's preferred channel.
PE risk doesn't end at discharge. Sentinel PE maintains outpatient surveillance for high-risk patients — post-surgical, post-partum, or newly prescribed high-risk medications — through symptom questionnaires and wearable vital sign monitoring for up to 30 days.
Before a physician prescribes a medication with VTE risk — oral contraceptives, HRT, chemotherapy — Sentinel PE presents a personalized risk assessment showing the patient's cumulative thrombotic risk with the proposed medication. This enables shared decision-making grounded in data, not assumptions.
Results from our deployed health systems.
Sentinel PE identified 342 high-risk patients in the first year who were not receiving appropriate prophylaxis. The system flagged 28 cases of medication-associated PE risk — including 6 young women on combined oral contraceptives with undiagnosed Factor V Leiden — enabling intervention before clot formation. Post-surgical PE rates dropped 44%.
A 26-year-old woman was admitted for outpatient knee arthroscopy. Sentinel PE flagged her as high-risk based on a risk constellation: oral contraceptive use (drospirenone/ethinyl estradiol), BMI of 32, 4-hour car drive to the surgical center, and family history of DVT. When post-operative hemodynamic monitoring showed subtle tachycardia and micro-desaturations that were individually unremarkable, Sentinel PE identified the pattern as consistent with early PE. CTPA confirmed a near-total occlusion saddle PE. She received thrombolysis within 90 minutes and made a full recovery. Without Sentinel, the attending later said he would have attributed the tachycardia to post-operative pain.
Deployed across a 12-clinic women's health network, Sentinel PE screened 48,000 patients' medication and risk profiles. The system identified 1,240 patients on combined oral contraceptives with at least two additional VTE risk factors who were not being monitored. It flagged 86 patients for thrombophilia screening based on family history patterns — of whom 14 tested positive for Factor V Leiden or prothrombin mutation. These 14 patients were transitioned to progestin-only or non-hormonal contraception. The network's contraceptive-associated VTE rate dropped to zero over the following 18 months.
Schedule a clinical demonstration of Sentinel PE. We'll show you the system monitoring your patient population, your medication formulary, and your surgical protocols.