Clinical Commander

All dossiers
pulm.pe.core.v1

Pulmonary Embolism (acute + extended)

pulmonologyacuteadultacuteinpatientoutpatienttransition
Start encounter →Print handoutPRODUCTION

PE = canonical Bayesian rule-out engine. §5.5.2 wired: pretest bands (Wells/Geneva low/intermediate/high) set the prior; sequential LRs (PERC-negative LR− ≈ 0.18 PMID 18318689; age-adjusted D-dimer LR− conditional on Wells PMID 24643601/28549662; CTPA segmental+ LR+ ≥50; V/Q high-prob LR+ / normal LR−); decision thresholds T_test ≈ 1.8-2% (Pauker-Kassirer) and treat-while-imaging T_treat zone; YEARS/PEGeD as phase logic. Two conditional dependencies explicitly modeled and flagged as do-not-multiply-independently: (1) D-dimer LR | Wells band (severity_triggers d_dimer_negative_conditional_on_wells, perc_negative_low_pretest_rule_out; INITIAL_WORKUP phase logic); (2) troponin / RV-strain | time-since-onset for risk stratification (RISK_STRATIFICATION phase, Category D trigger — repeat troponin 6-12 h if early draw). Effect sizes wired with PMIDs (≥8): PEITHO death/decompensation 2.6% vs 5.6% OR 0.44 + bleeding 6.3% vs 1.2% + stroke 2.4% vs 0.2% (24716681); PEITHO long-term ~33% functional limitation, CTEPH ~2-3% (28335835); AMPLIFY recurrence 2.3% vs 2.7% RR 0.84, bleed 0.6% vs 1.8% RR 0.31 (23808982); AMPLIFY-EXT recurrence 1.7% vs 8.8% (23216615); EINSTEIN-PE recurrence 2.1% vs 1.8% HR 1.12, bleed HR 0.49 (22449293); RE-COVER VTE HR 1.10 bleed HR 0.82 (19966341); Caravaggio recurrence 5.6% vs 7.9% HR 0.63 (32223112); Hokusai-VTE Cancer composite 12.8% vs 13.5% HR 0.97 (29231094); CLOT recurrence 9% vs 17% HR 0.48 (12853587); sPESI 30-d mortality ~1.0% vs ~10.9% (20696966). Cross-dossier routing via workups[].branches_to + sibling_differentiation[].sibling_engine_id using real engine_ids: cardio.dvt.core.v1, cardio.acute-hf.core.v1, cardio.aortic-dissection.core.v1, pulm.cap.core.v1, id.sepsis.core.v1, pulm.pulmonary_htn_group2_to_5.v1 (CTEPH route). Look-alike pivots authored: PE vs ACS/HF (troponin+ECG+CT/echo RV-vs-LV), PE vs aortic dissection (tPA-catastrophic warning), PE vs pneumonia/sepsis (D-dimer degraded by infection), PE vs anxiety. Special populations: pregnancy (LMWH only; V/Q over CTPA to limit breast dose), cancer (apixaban/edoxaban non-luminal-GI vs LMWH/dalteparin luminal-GI — Caravaggio/Hokusai-VTE-Cancer/SELECT-D/CLOT), renal (Cockcroft-Gault CrCl NOT CKD-EPI eGFR is the binding DOAC metric — distinction noted in calculators/contraindication_rules), obesity (apixaban/rivaroxaban OK per ISTH 2021 SSC; avoid dabigatran/edoxaban), triple-positive APS (warfarin not DOAC — TRAPS PMID 33128325), CTEPH → pulm.pulmonary_htn_group2_to_5.v1. Regimen axis: anticoagulation by Category A-E with loading (apixaban 10 BID×7d; rivaroxaban 15 BID×21d) / maintenance / parenteral lead-in (edoxaban, dabigatran) / duration logic (provoked vs unprovoked vs cancer; HERDOO2/DASH stop-decision in monitoring) / reversal agents (idarucizumab, andexanet, 4F-PCC) / monitoring / extended reduced-dose deprescribing (AMPLIFY-EXT). Matches pulm.copd.core.v1 regimen depth. SCHEMA-GAP NOTES: (1) _types.ts has no first-class field for Bayesian LR tables / decision thresholds — encoded in severity_triggers, phase purpose/advance_when, calculator guideline_basis, regimen rationale, and the .depth.md LR table; (2) no field for conditional-dependency graph — modeled narratively in severity_triggers + INITIAL_WORKUP/RISK_STRATIFICATION phase logic; (3) RequiredCalculator.drives enum lacks "diagnostic_gate" — Wells/PERC reuse risk_stratification/screening; (4) no pretest-prevalence field — captured in calculator guideline_basis + .depth.md. DEPTH-PASS-2 2026-05-17 (shard-07-cardio-chronic CL-4, golden-template mirror of cardio.htn.core.v1): (1) co-located src/lib/dossiers/pulm.pe.core.v1._design-brief.md + ._research-bundle.md authored per §5.5 items 1+2 (8-section bundle: guideline-freshness vs floor, evidence-anchor table with 14 verified PMIDs + effect sizes + 95% CI + retrieval 2026-05-17, dose-effect anchors, RxCUI log, MCP-fallback log, pretest priors, T_test≈2%/T_treat≈5% Pauker-Kassirer/Klok derivation, cross-dossier routing); design_brief: field repointed from legacy package path to the co-located brief. (2) Bayesian seed prisma/seed/ros-and-ddx/pulm.pe.core.v1.{differentials,ros,finding-lrs}.ts authored mirroring htn Prisma shapes exactly: 12 differentials (acute-dyspnea/pleuritic-pain partition + PE severity strata, cohort-anchored base_prevalence w/ PMID), 14 ROS, 33 LR rows = 16 LR+/17 LR− (Wells bands, PERC, age-adjusted & YEARS D-dimer, CTPA filling-defect LR+ ≥50, RV strain echo/CT, troponin conditional on time, sPESI bands), 3 conditionalDependencies (D-dimer|Wells; troponin|time-since-onset; D-dimer-age-adjusted|age), T_test≈2%/T_treat≈5% in header. (3) RxCUI bugs fixed vs DrugEffectProfile + RxNav-validated intervention registry: apixaban 1364445→1364430 (×3), dabigatran 1037045→1037042, tenecteplase 1804735→259280 (RxNav ingredient; profile-registry 259081 SCD:REJECT in known 96-fail baseline, not used), dalteparin 69749→67109, andexanet alfa 1927855→2045114, idarucizumab 1605373→1716191, 4F-PCC 1359574→1670383; all NEEDS_RXNAV_VALIDATION placeholders removed; rivaroxaban/edoxaban/enoxaparin/heparin/warfarin/alteplase/norepinephrine/dobutamine confirmed OK by npm run research:rxnav:validate. (4) Dose-effect: every regimen drug rationale enriched with time-to-effect + recurrent-VTE/mortality/major-bleed HR with 95% CI from live-verified DOAC/thrombolysis trials (AMPLIFY RR 0.84/0.31, EINSTEIN-PE HR 1.12/0.49, Hokusai-VTE HR 0.89/0.81, RE-COVER HR 1.10/0.82, Caravaggio HR 0.63/0.82, Hokusai-VTE Cancer HR 0.97, CLOT HR 0.48, PEITHO OR 0.44, TRAPS HR 6.9, AMPLIFY-EXT 1.7% vs 8.8%, all 95% CI). (5) Content refresh: evidence.last_reconciled 2026-05-17; primary guideline confirmed 2026 AHA/ACC/ACCP/ACEP/CHEST PE (PMID 41712677, DOI 10.1161/CIR.0000000000001415, Circulation 153(12):e977-e1051, Categories A-E) live-verified current vs floor (2019 ESC PMID 31504429 + CHEST 2021 + ASH 2020); age-adjusted & YEARS D-dimer, sPESI/Bova, DOAC-first-line, CDT/thrombectomy for intermediate-high (PEERLESS win-ratio 5.01, 95% CI 3.68–6.97), extended reduced-dose anticoagulation all retained and re-anchored to verified PMIDs. SCHEMA-GAP NOTES (carried): (1) _types.ts has no first-class field for Bayesian LR tables / decision thresholds — encoded in severity_triggers, phase purpose/advance_when, calculator guideline_basis, regimen rationale, and the seed finding-lrs.ts header; (2) no field for conditional-dependency graph — modeled in finding-lrs.ts conditionalDependencies[] + severity_triggers + INITIAL_WORKUP/RISK_STRATIFICATION phase logic; (3) RequiredCalculator.drives enum lacks "diagnostic_gate" — Wells/PERC reuse risk_stratification/screening; (4) no pretest-prevalence field — captured in differentials base_prevalence + bundle §6. RESIDUAL non-blocking items (out of scope for this depth shard): Wells/Geneva/PERC/sPESI/Hestia/Bova not all in clinical-tools-registry; no engine-specific test file under tests/ (covered by dossier-contract.test.ts); package path under cardio.pe.acute/ (legacy) — manifest/package pointers untouched per shard scope; profile-registry 96-fail baseline (incl. tenecteplase profile CUI) NOT touched per scope (profiles owned by another lane). DEPTH-PASS-3 2026-05-26 (lane-E): +NMA +USPSTF +Cochrane +ICER stubs +decision thresholds, side-car at pulm.pe.core.v1._depth-pass-3.md.

Entry points (5)

  • symptom
    Acute dyspnea, pleuritic chest pain, syncope, or hemoptysis (ESC 2019 §5; AHA/ACC 2026)
    acute_dyspnea_or_pleuritic_pain
  • vital_abnormality
    Unexplained tachycardia + hypoxia (ESC 2019 §5)
    tachycardia_hypoxia
  • lab_abnormality
    Elevated D-dimer with PE-compatible pretest probability (ESC 2019 §6.3) — interpret conditional on Wells band
    d_dimer_elevated
  • imaging
    Echo or CT showing RV strain / dilation — risk not diagnosis (ESC 2019 §7.3)
    rv_strain
  • problem_list
    Concurrent DVT raising index of suspicion (ESC 2019 §6.4) → see cardio.dvt.core.v1
    concurrent_dvt

Required inputs (22)

  • agerequired
    demographic • used at CONTEXT
    Age-adjusted D-dimer threshold (>50 yr: age × 10 ng/mL FEU) shifts the test LR− favorably (ADJUST-PE Righini JAMA 2014 PMID 24643601)
  • wells_geneva_featuresrequired
    symptom • used at CONTEXT
    Pretest probability band (Wells / revised Geneva) — sets the prior for Bayesian update; D-dimer LR is conditional on this band
  • sbprequired
    vital • used at CONTEXT
    Sustained SBP <90 → high-risk Category E (AHA/ACC 2026; ESC 2019 Table 4)
  • hrrequired
    vital • used at CONTEXT
    Wells HR >100 (Wells 2000); sPESI HR ≥110 (Jiménez Arch Intern Med 2010 PMID 20696966)
  • spo2required
    vital • used at CONTEXT
    sPESI SpO2 <90 (Jiménez 2010 PMID 20696966); ESC 2019 severity stratification
  • rr
    vital • used at CONTEXT
    Tachypnea + hemodynamic compromise (ESC 2019)
  • d_dimer
    lab • used at INITIAL_WORKUP
    Rule-out test in low/intermediate pretest probability; LR− depends on Wells band (conditional dependency — do not multiply independently). YEARS van der Hulle Lancet 2017 PMID 28549662
  • troponin
    lab • used at RISK_STRATIFICATION
    Risk (not diagnosis): troponin elevation → intermediate-high (Category D). Interpreted conditional on time-since-onset — early draw can be falsely negative (ESC 2019 Table 4; AHA/ACC 2026)
  • bnp
    lab • used at RISK_STRATIFICATION
    RV strain biomarker — intermediate-high risk (ESC 2019 Table 5)
  • creatininerequired
    lab • used at CONTEXT
    CrCl (Cockcroft-Gault, NOT eGFR) for DOAC selection; eGFR/CrCl divergence matters at extremes of weight/age (CHEST 2021; ASH 2020)
  • hemoglobinrequired
    lab • used at CONTEXT
    Bleeding risk + transfusion threshold for thrombolysis (AHA/ACC 2026)
  • plateletsrequired
    lab • used at CONTEXT
    Thrombocytopenia precludes thrombolysis (<50k) and full-dose LMWH; screen HIT if heparin-exposed (AHA/ACC 2026; CHEST 2021)
  • ctpa
    imaging • used at INITIAL_WORKUP
    Definitive PE diagnosis — segmental+ filling defect LR+ ≥ 50; also clot burden + RV:LV ratio (ESC 2019 Class I; AHA/ACC 2026)
  • vq_scan
    imaging • used at INITIAL_WORKUP
    V/Q when CTPA contraindicated (renal, contrast allergy, pregnancy) — high-probability scan LR+ high; normal scan LR− very low (ESC 2019 §6)
  • echo
    imaging • used at RISK_STRATIFICATION
    RV dysfunction + McConnell sign — risk stratification + reperfusion decision, not primary diagnosis (ESC 2019 §7.3; AHA/ACC 2026)
  • malignancy
    history • used at CONTEXT
    Cancer-associated VTE — DOAC vs LMWH by tumor site; extended therapy (Caravaggio NEJM 2020 PMID 32223112; Hokusai-VTE Cancer PMID 29231094; SELECT-D PMID 29746227)
  • pregnancy
    history • used at CONTEXT
    LMWH (NOT DOAC/VKA); CTPA vs V/Q radiation/breast-dose trade-off (ESC 2019 §10; ASH 2020)
  • antiphospholipid
    history • used at CONTEXT
    Triple-positive APS → warfarin, NOT DOAC (TRAPS Pengo; rivaroxaban arm stopped for excess arterial events; JTH 2020 PMID 33128325)
  • weight
    demographic • used at CONTEXT
    Obesity (BMI ≥40 / >120 kg): apixaban & rivaroxaban acceptable per ISTH 2021 SSC; dabigatran/edoxaban less data — informs agent
  • prior_vte_or_provokedrequired
    history • used at CONTEXT
    Provoked (transient) vs unprovoked vs cancer → duration logic (CHEST 2021; ESC 2019 §7.5)
  • recent_bleeding_or_high_riskrequired
    history • used at CONTEXT
    Active bleeding, recent surgery, recent ICH — contraindicates anticoag/thrombolysis (AHA/ACC 2026 Table; ESC 2019 Table 10)
  • current_anticoagulant
    medication • used at CONTEXT
    Breakthrough PE on therapeutic anticoagulation → reassess adherence/APS/cancer + escalate (CHEST 2021; ESC 2019)

12-phase flow (12)

  1. 1FRAME
    Confirm acute PE scope (NOT chronic CTEPH — route persistent post-PE PH to pulm.pulmonary_htn_group2_to_5.v1) and adult population (ESC 2019 §1; AHA/ACC 2026)
    advance: Acute presentation confirmed
  2. 2ENTRY
    Trigger from acute symptoms, abnormal vitals, or imaging finding (ESC 2019 §5)
    inputs: age
    advance: Entry symptom/vitals captured
  3. 3CONTEXT
    Compute pretest probability (Wells / revised Geneva band) — this sets the Bayesian prior. Capture cancer, pregnancy, APS, weight, prior VTE, current anticoagulant, bleeding risk, CrCl, platelets, Hgb (AHA/ACC 2026; ESC 2019)
    inputs: wells_geneva_features, sbp, hr, spo2, creatinine, hemoglobin, platelets, malignancy, pregnancy, antiphospholipid, weight, prior_vte_or_provoked, recent_bleeding_or_high_risk, current_anticoagulant
    actions: calc.wells_pe
    advance: Pretest band assigned + risk inputs ready
  4. 4RED_FLAGS
    Hemodynamic instability Category E — empiric UFH NOW (treat-while-imaging: posterior at T_treat zone) + thrombolysis/thrombectomy team activation; arrest → ECPR (AHA/ACC 2026 Class I; ESC 2019 Table 4). Also: PERC screen in low pretest probability (Kline JTH 2008 PMID 18318689)
    inputs: sbp, spo2
    actions: calc.perc
    advance: Stability achieved or PERT/ECMO activated
  5. 5INITIAL_WORKUP
    Bayesian gate: PERC-negative + low pretest → STOP (posterior < T_test ≈ 1.8-2%). Else age-adjusted D-dimer (LR− conditional on Wells band) → if positive or Wells-high, CTPA (segmental+ LR+ ≥50) or V/Q if contrast contraindicated. Encode YEARS / PEGeD as branching logic. ECG (S1Q3T3, RBBB) supportive only (ESC 2019 §6; ADJUST-PE PMID 24643601; YEARS PMID 28549662)
    inputs: d_dimer, ctpa, vq_scan
    actions: pe_full, calc.perc
    advance: PE confirmed (posterior > T_treat) or ruled out (posterior < T_test)
  6. 6BRANCHING_WORKUP
    Confirmed PE: echo for RV dysfunction (ESC 2019 §7.3); lower-extremity Doppler if DVT not documented (→ cardio.dvt.core.v1); thrombophilia / APS testing in selected unprovoked (off anticoag, ≥3 mo). If PE excluded: pivot to look-alikes — ACS (troponin + ECG + CT pivot → see cardio.acute-hf.core.v1 for RV-vs-LV failure), aortic dissection (→ cardio.aortic-dissection.core.v1), pneumonia/sepsis (→ pulm.cap.core.v1, id.sepsis.core.v1), anxiety/hyperventilation
    inputs: echo
    actions: pe_full
    advance: RV function + DVT presence assessed OR alternative dx routed
  7. 7DIFFERENTIAL
    PE vs ACS vs aortic dissection vs pneumothorax vs pneumonia vs decompensated HF vs anxiety; anatomic categories saddle vs lobar vs segmental vs subsegmental (ESC 2019 §5; AHA/ACC 2026)
    advance: Confirmed PE with location + thrombus burden, or alternate dx
  8. 8RISK_STRATIFICATION
    AHA/ACC 2026 Clinical Categories A-E (A asymptomatic→home; B symptomatic low-score→early discharge; C hospitalize; D incipient cardiopulmonary failure; E persistent hypotension). sPESI (Jiménez 2010 PMID 20696966) / Hestia (Zondag JTH 2013 PMID 23336721) for outpatient; PESI (Aujesky 2005 PMID 16020800) for 30-d mortality; Bova for normotensive intermediate; troponin/RV-strain interpreted conditional on time-since-onset
    inputs: sbp, hr, spo2, troponin, bnp, echo
    actions: calc.spesi
    advance: Category assigned + outpatient candidacy decided
  9. 9TREATMENT
    DOAC over VKA Class I (AHA/ACC 2026; AMPLIFY PMID 23808982). LMWH preferred parenteral / cancer (Caravaggio PMID 32223112). Category E → alteplase 100 mg/2 h (ESC 2019 Class I) or thrombectomy per PERT (PEERLESS PMID 39470698). Cancer: DOAC (non-luminal-GI) or LMWH (Hokusai-VTE Cancer PMID 29231094). Pregnancy → LMWH (ASH 2020). Triple-positive APS → warfarin (TRAPS PMID 33128325). Outpatient for Category A-B sPESI/Hestia 0 (HoT-PE PMID 31120118)
    inputs: creatinine, prior_vte_or_provoked, malignancy, pregnancy, antiphospholipid, weight
    advance: Anticoagulant + duration + escalation pathway documented
  10. 10DISPOSITION
    Outpatient HoT-PE/Hestia for Category A-B low-risk; ward Category C; ICU Category D/E; PERT for Category D/E (AHA/ACC 2026; HoT-PE Barco EHJ 2020 PMID 31120118)
    advance: Disposition set
  11. 11MONITORING
    Anticoagulant adherence + bleeding surveillance; 3-month re-evaluation for extended therapy (CHEST 2021); CTEPH screen at 3-6 months if persistent dyspnea — V/Q + echo + RHC (ESC 2019 §9; PEITHO long-term Konstantinides JACC 2017 PMID 28335835 — ~33% persistent functional limitation, CTEPH ~2-3%)
    advance: Monitoring plan with bleeding precautions documented
  12. 12FOLLOWUP
    Reassess provoked vs unprovoked vs cancer; extended anticoagulation Class I for unprovoked (AHA/ACC 2026); reduced-dose DOAC (AMPLIFY-EXT Agnelli NEJM 2013 PMID 23216615; EINSTEIN-CHOICE); HERDOO2 (women) / DASH scores to guide stopping unprovoked; PEmb-QoL; occupational return
    advance: Duration decision + return precautions documented