🫀 cardio-echo-suite MedGemma Reasoning

Clinical reasoning powered by MedGemma 4B (Google Health AI).

Enter AI findings from an echo or ECG study, optionally paste a prior clinical note, and MedGemma will generate:

Differential diagnosis (ranked by probability)
Guideline recommendations (with evidence class + citation)
Uncertainty explanation (why the model is unsure)
Prior comparison (if prior EF available)

This Space is the reasoning backend for cardio-echo-suite.

Study Type

AI-Predicted EF (%)

Heart Rate (bpm)

Patient Age

Patient Sex

Max New Tokens

256 2048

AI Findings (JSON)

Prior Clinical Note (optional)

Reasoning Output (JSON)

API Usage

This Space exposes a Gradio API. cardio-echo-suite calls it via:

from gradio_client import Client
client = Client("Dalfino/cardio-medgemma")
result = client.predict(
    ef=32, hr=92, study_type="echo",
    findings_json='[{"name":"EF","value":"32%","confidence":0.9}]',
    prior_note="LVEF 45% prior.",
    patient_age=67, patient_sex="M",
    max_new_tokens=1024,
    api_name="/reason"
)

Or set REASONING_URL=https://Dalfino-cardio-medgemma.hf.space in cardio-echo-suite's .env.local.