Spaces:
Running on Zero
Running on Zero
File size: 11,205 Bytes
9ec16e6 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 | """
MedVision Edge — Clinical Protocol Engine
Deterministic function calling: WHO IMCI protocols, drug dosage, referral urgency.
No hallucination possible — all outputs come from verified JSON lookup tables.
"""
import json
import os
from pathlib import Path
PROTOCOLS_DIR = Path(__file__).parent.parent / "protocols"
def _load_protocol(name):
path = PROTOCOLS_DIR / f"{name}.json"
with open(path) as f:
return json.load(f)
# Pre-load all protocols
_PROTOCOLS = {}
for name in ["pneumonia", "consolidation", "cardiomegaly", "effusion", "edema", "dosage", "referral"]:
try:
_PROTOCOLS[name] = _load_protocol(name)
except FileNotFoundError:
pass
# Mapping from finding names to protocol keys
_FINDING_TO_KEY = {
"Pneumonia": "pneumonia",
"Consolidation": "consolidation",
"Cardiomegaly": "cardiomegaly",
"Effusion": "effusion",
"Pleural Effusion": "effusion",
"Edema": "edema",
"Pulmonary Edema": "edema",
}
def get_clinical_protocol(finding: str, severity: str = "mild") -> dict:
"""
Retrieve WHO IMCI clinical protocol based on X-ray finding.
Args:
finding: Detected pathology (e.g., "Pneumonia", "Cardiomegaly")
severity: "mild", "moderate", or "severe"
Returns:
Dict with classification, treatment, referral criteria
"""
key = _FINDING_TO_KEY.get(finding, finding.lower())
protocol = _PROTOCOLS.get(key)
if not protocol:
return {"error": f"No protocol found for '{finding}'"}
severity = severity.lower()
if severity not in ("mild", "moderate", "severe"):
severity = "mild"
result = {
"condition": protocol.get("condition", finding),
"severity": severity,
"classification": protocol.get("classification", {}).get(severity, {}),
"treatment": protocol.get("treatment", {}).get(severity, {}),
"referral_criteria": protocol.get("referral", {}),
"source": protocol.get("source", "WHO IMCI 2024"),
}
return result
def calculate_drug_dosage(drug: str, weight_kg: float, age_years: int = None) -> dict:
"""
Calculate medication dosage based on weight and age.
Args:
drug: Drug name (e.g., "amoxicillin", "furosemide")
weight_kg: Patient weight in kg
age_years: Patient age in years (optional)
Returns:
Dict with dose, frequency, route, warnings
"""
dosage_db = _PROTOCOLS.get("dosage", {}).get("drugs", {})
drug_key = drug.lower().replace(" ", "_")
drug_info = dosage_db.get(drug_key)
if not drug_info:
return {"error": f"Drug '{drug}' not found in formulary"}
is_pediatric = age_years is not None and age_years < 18
if is_pediatric and "pediatric" in drug_info.get("dosing", {}):
ped = drug_info["dosing"]["pediatric"]
dose_per_kg = ped.get("dose_mg_per_kg", 0)
calculated_dose = dose_per_kg * weight_kg
max_dose = ped.get("max_dose_mg")
if max_dose and calculated_dose > max_dose:
calculated_dose = max_dose
# Find closest weight band
weight_band = None
for band in ped.get("weight_table", []):
wt = band.get("weight_kg", "")
if isinstance(wt, str):
if wt.startswith(">=") and weight_kg >= float(wt[2:]):
weight_band = band
elif "-" in wt:
lo, hi = wt.split("-")
if float(lo) <= weight_kg <= float(hi):
weight_band = band
result = {
"drug": drug_info.get("class", drug),
"patient": f"{weight_kg}kg, {age_years}yr",
"calculated_dose_mg": round(calculated_dose),
"frequency": ped.get("frequency", ""),
"route": ped.get("route", "oral"),
"duration_days": ped.get("duration_days", "as directed"),
"formulations": drug_info.get("formulations", []),
}
if weight_band:
result["recommended_dose"] = weight_band.get("dose", "")
if ped.get("caution"):
result["caution"] = ped["caution"]
return result
else:
adult = drug_info.get("dosing", {}).get("adult", {})
if isinstance(adult, dict) and "dose" in adult:
return {
"drug": drug_info.get("class", drug),
"patient": f"{weight_kg}kg, adult",
"dose": adult.get("dose", ""),
"frequency": adult.get("frequency", ""),
"route": adult.get("route", "oral"),
"formulations": drug_info.get("formulations", []),
}
else:
return {
"drug": drug_info.get("class", drug),
"patient": f"{weight_kg}kg, adult",
"dosing": adult,
"formulations": drug_info.get("formulations", []),
}
def determine_referral_urgency(findings: list, confidence_scores: dict = None) -> dict:
"""
Determine if patient needs referral and urgency level.
Args:
findings: List of detected pathologies (e.g., ["Pneumonia", "Effusion"])
confidence_scores: Optional dict of {pathology: confidence} (not used for rules)
Returns:
Dict with urgency level, criteria met, pre-transfer actions
"""
referral_db = _PROTOCOLS.get("referral", {})
referral_by_finding = referral_db.get("referral_by_finding", {})
urgency_levels = referral_db.get("urgency_levels", {})
if not findings:
return {
"urgency": "follow_up",
"color": "green",
"recommendation": "No acute findings. Routine follow-up recommended.",
}
# Determine highest urgency across all findings
max_urgency = "follow_up"
urgency_order = ["follow_up", "routine", "urgent", "emergency"]
matched_criteria = []
for finding in findings:
key = _FINDING_TO_KEY.get(finding, finding.lower())
rules = referral_by_finding.get(key, {})
# Multiple findings = higher urgency
if len(findings) >= 3:
level = "urgent"
elif len(findings) >= 2:
level = "routine"
else:
level = "follow_up"
# Check finding-specific referral rules
if rules.get("refer_emergency"):
matched_criteria.append(f"{finding}: {rules['refer_emergency']}")
if rules.get("refer_urgent"):
matched_criteria.append(f"{finding}: {rules['refer_urgent']}")
if urgency_order.index(level) < urgency_order.index("urgent"):
level = "routine" # At least routine if has urgent criteria
if urgency_order.index(level) > urgency_order.index(max_urgency):
max_urgency = level
# For single severe finding or multiple findings, escalate
severe_findings = {"Cardiomegaly", "Edema", "Pulmonary Edema"}
if any(f in severe_findings for f in findings) and len(findings) >= 2:
max_urgency = "urgent"
urgency_info = urgency_levels.get(max_urgency, {})
return {
"urgency": max_urgency,
"color": urgency_info.get("color", "green"),
"timeframe": urgency_info.get("timeframe", "Routine follow-up"),
"findings_detected": findings,
"referral_criteria": matched_criteria[:5], # Top 5
"pre_transfer_actions": urgency_info.get("pre_transfer_actions", []),
"recommendation": f"{'URGENT: ' if max_urgency in ('urgent', 'emergency') else ''}Refer for specialist evaluation. {len(findings)} patholog{'y' if len(findings)==1 else 'ies'} detected.",
}
def generate_clinical_summary(findings: dict, patient_age: int = None, patient_weight: float = None, language: str = "en") -> str:
"""
Generate a complete clinical summary with protocols for all detected findings.
Args:
findings: Dict of {pathology: True/False}
patient_age: Age in years (optional)
patient_weight: Weight in kg (optional)
language: Output language code (for header only)
Returns:
Formatted clinical summary string
"""
detected = [p for p, present in findings.items() if present]
if not detected:
return "**No significant pathologies detected.**\n\nRecommendation: Routine follow-up. No immediate intervention required.\n\n⚠️ *This is an AI screening tool. Findings should be confirmed by a qualified radiologist.*"
lines = []
lines.append(f"## Clinical Summary — {len(detected)} Finding{'s' if len(detected)>1 else ''} Detected\n")
for finding in detected:
protocol = get_clinical_protocol(finding, severity="mild")
if "error" not in protocol:
classification = protocol.get("classification", {})
treatment = protocol.get("treatment", {})
lines.append(f"### {finding}")
if classification.get("label"):
lines.append(f"**Classification:** {classification['label']}")
if isinstance(treatment, dict):
setting = treatment.get("setting", "")
if setting:
lines.append(f"**Setting:** {setting}")
if treatment.get("first_line"):
fl = treatment["first_line"]
if isinstance(fl, dict):
dose = fl.get('dose', '')
if not dose and fl.get('dose_mg_per_kg'):
dose = f"{fl['dose_mg_per_kg']} mg/kg"
lines.append(f"**First-line:** {fl.get('drug', '')} {dose} {fl.get('frequency', '')}")
elif isinstance(fl, list):
for med in fl:
lines.append(f"- {med.get('drug', '')} {med.get('dose', '')} {med.get('frequency', '')}")
lines.append("")
# Referral
referral = determine_referral_urgency(detected)
urgency_emoji = {"emergency": "🔴", "urgent": "🟠", "routine": "🟡", "follow_up": "🟢"}.get(referral["urgency"], "⚪")
lines.append(f"### Referral Assessment")
lines.append(f"{urgency_emoji} **{referral['urgency'].upper()}** — {referral.get('timeframe', '')}")
lines.append(f"{referral.get('recommendation', '')}")
lines.append("")
# Dosage if weight provided
if patient_weight and detected:
lines.append("### Suggested Dosing")
primary_drug = "amoxicillin" if "Pneumonia" in detected or "Consolidation" in detected else "furosemide"
dosage = calculate_drug_dosage(primary_drug, patient_weight, patient_age)
if "error" not in dosage:
if dosage.get('calculated_dose_mg'):
dose_str = f"{dosage['calculated_dose_mg']} mg"
elif dosage.get('dose'):
dose_str = dosage['dose']
else:
dose_str = "N/A"
lines.append(f"**{primary_drug.title()}**: {dose_str} {dosage.get('frequency', '')}")
lines.append("")
lines.append("---")
lines.append("⚠️ *AI screening tool. All findings must be confirmed by a qualified radiologist. Do not use as sole basis for clinical decisions.*")
return "\n".join(lines)
|