shard-3 neuro-sym Phase E autonomous authoring (2026-05-16): chronic adult trigeminal neuralgia (classical / secondary / idiopathic) at INTEGRATED with full §5.5 contract depth. Diagnosis per ICHD-3 (29368949) + ICOP 1st ed (32103673); classification + therapy per EAN 2019 Bendtsen guideline (30860637) and AAN-EFNS 2008 (Cruccu 18721143) + AAN/EFNS 2008 practice parameter (Gronseth 18716236). MRI brain (± MRA, FIESTA/CISS trigeminal protocol) MANDATORY in ALL — exclude secondary cause + assess neurovascular compression; trigeminal reflex testing surrogate where MRI unavailable/equivocal. Quantitative content (≥10 effect sizes): carbamazepine NNT ~1.7–2.6 (PMID 18721143/30860637), oxcarbazepine equally effective + better tolerated, lamotrigine add-on superior to placebo P=0.011 (Zakrzewska 9415510), BTX-A significant VAS/frequency reduction (Wu 22492424; Zhang 25263254), MVD 70% pain-free at 10 yr + recurrence <1%/yr (Barker NEJM 1996 8598865), SRS ~70% relief with ~1-mo latency + higher recurrence, percutaneous ~90% initial relief with sensory-loss/recurrence trade-offs, 49% concomitant continuous pain (TN type 2 — Maarbjerg 25231219), 89% Na-channel-blocker response (Maarbjerg 25231219), ~15% secondary TN. RxCUIs curl-verified bidirectionally via RxNav REST 2026-05-16: carbamazepine 2002, oxcarbazepine 32624, lamotrigine 28439, baclofen 1292, gabapentin 25480, pregabalin 187832, phenytoin 8183, fosphenytoin 72236, lidocaine 6387, onabotulinumtoxinA 860189, sodium valproate 9919, eslicarbazepine 1482502. Surgery (MVD / gamma-knife SRS / percutaneous rhizotomy) + MRI = non_pharm. PMID correction vs task prompt: the AAN-EFNS 2008 management guideline is PMID 18721143 (Cruccu, Eur J Neurol) — NOT 18796700. 18716236 IS valid and distinct (Gronseth, Neurology 2008 practice parameter). EAN 2019 (30860637) is the primary always-latest floor; NRDP 2024 (38816415) added as the most recent synthesis. Special populations (≥4): MS-associated secondary TN (route neuro.ms-flare.core.v1; combination AEDs; percutaneous/SRS over MVD), pregnancy/childbearing potential (AVOID sodium valproate; carbamazepine teratogenicity; lamotrigine clearance changes; specialist co-management), elderly (carbamazepine hyponatraemia/falls/cardiac/cognitive — oxcarbazepine or slow titration; MVD vs percutaneous candidacy), renal/hepatic dose adjustment (gabapentin/pregabalin via calc.ckd_epi_2021), acute crisis unable-to-eat/drink (admission + IV lidocaine/fosphenytoin/phenytoin). Psychosocial/safety: TN "suicide disease" — high depression/anxiety/suicidality (NRDP 2024 38816415); PHQ-9/GAD-7 + suicide-risk screening wired; nutrition/hydration compromise from trigger avoidance is a severity trigger. 6 severity_triggers (acute crisis → admission+IV; medically refractory/intolerant → early neurosurgery; secondary red-flags → urgent MRI; suicidality/severe depression; severe carbamazepine ADR SJS-TEN/hyponatraemia; MS-associated secondary TN). 2 setting playbooks (outpatient diagnosis+medical-titration+surgical-referral; ed/inpatient acute crisis unable-to-eat/IV therapy + transition to neurosurgery). 1 regimen axis tn_pharmacotherapy (3 steps). 4 sibling-differentiation rows (neuro.ms-flare.core.v1, neuro.cluster-headache.core.v1, symptom.acute_vision_loss.v1, neuro.migraine.core.v1). Schema-blocked (forbidden by shard file scope; surfaced as tickets in research bundle): no TN-specific calculator in clinical-tools-registry.ts (no ICHD-3/ICOP TN classifier, no paroxysm-severity scorer, no secondary-TN likelihood-ratio scorer, no Barker MVD-recurrence predictor). Pain scales encoded narratively + as severity_triggers — no fabricated calc ids. calc.phq9/calc.gad7 used for psych/suicidality comorbidity; calc.ckd_epi_2021 for renal AED dosing. No router/panel entry — matches in-shard neuro.bell-palsy.v1 (new INTEGRATED neuro dossiers have no panel this shard). Depth-pass-2 (2026-05-18, CL-3): added §5.5.2 Bayesian differential layer — prisma/seed/ros-and-ddx/neuro.trigeminal-neuralgia.v1.{ros,differentials,finding-lrs}.ts (12 ROS items, 11 differentials with sourced pre-test priors, 22 finding x diagnosis LR rows; gbs.core exemplar import/type style; LR bounds 0.01<=lr_positive<=100, 0.01<=lr_negative<=1). §5.5.2 high-discrimination targets met: MRI NVC-with-morphological-change LR+ 18 (sens 0.85 / (1-spec 0.95)=17, EAN 2019 PMID 30860637) and MRI demyelinating-plaque/structural-lesion LR+ 40; normal-protocolised-MRI LR- 0.05 (=(1-sens 0.95)/spec 0.90 ~= 1/20, effective rule-out >=15x, AAN/EFNS practice parameter PMID 18716236). Three documented conditional-dependency notes: [CD-1] MRI-NVC LR conditional on classical-vs-idiopathic NOT secondary (contact-alone is non-discriminatory); [CD-2] trigeminal-reflex-testing LR conditional on secondary-TN suspicion and correlated with the objective-sensory-deficit finding (substitute, not independent addition — AAN/EFNS PMID 18716236); [CD-3] MRI-normal and trigeminal-reflex-normal rows correlated (intact-pathway) — MRI primary, reflex confirmatory where MRI equivocal. Depth-pass-2 §5.5.1 quantified inline provenance (>=10 distinct effect sizes, units + trial/year/PMID): carbamazepine vs placebo NNT ~1.7-2.6 for pain relief with ~70% initial response (AAN-EFNS 2008 Level A PMID 18721143; EAN 2019 PMID 30860637); carbamazepine number-needed-to-harm ~3.4 minor / ~24 major (AAN/EFNS practice parameter PMID 18716236); oxcarbazepine equally effective for pain, better tolerated, fewer DDIs, less enzyme induction (AAN-EFNS 2008 Level B PMID 18721143; EAN 2019 PMID 30860637); lamotrigine 400 mg/day add-on vs placebo composite efficacy P=0.011, 11/13 better on lamotrigine (Zakrzewska Pain 1997 RCT PMID 9415510); botulinum toxin type A 75 U vs saline significant VAS + attack-frequency reduction (Wu Cephalalgia 2012 RCT n=42 PMID 22492424); two-dose BTX-A both doses superior to placebo (Zhang Cephalalgia 2014 RCT n=84 classical TN PMID 25263254); microvascular decompression 70% excellent (pain-free, no medication) at 10 yr + recurrence <1%/yr after yr 10 (Barker/Jannetta NEJM 1996 n=1185 PMID 8598865); MVD operative mortality ~0.2%, ipsilateral hearing loss ~1% (Barker NEJM 1996 PMID 8598865; EAN 2019 PMID 30860637); stereotactic radiosurgery (gamma knife) ~70% initial relief with ~1-month median latency to effect + higher recurrence than MVD + sensory loss in a minority (EAN 2019 PMID 30860637); percutaneous procedures (RF thermocoagulation / glycerol / balloon) ~90% initial relief with higher recurrence + sensory-loss / anaesthesia-dolorosa / corneal-anaesthesia trade-offs (EAN 2019 PMID 30860637; AAN/EFNS PMID 18716236); concomitant continuous pain (TN type 2) in 49% of TN and 89% Na-channel-blocker response in a 158-patient prospective cohort (Maarbjerg Headache 2014 PMID 25231219); ~15% of all TN is secondary (EAN 2019 PMID 30860637; NRDP 2024 PMID 38816415). Depth-pass-2 resolving cross-dossier engine_id routes (>=2; targets verified to exist as dossier files in src/lib/dossiers/): neuro.ms-flare.core.v1 (MS-associated secondary TN — demyelinating plaque on MRI; treat the MS), neuro.cluster-headache.core.v1 (prominent ipsilateral cranial autonomic features → TAC / SUNCT-SUNA ladder), symptom.acute_vision_loss.v1 (age >50 + jaw-claudication GCA mimic — never delay steroids), neuro.migraine.core.v1 (longer throbbing photophobic facial-pain pivot). No unresolved cross-ref engine_ids; no dental/ENT engine file exists in-shard so odontogenic/TMD/sinusitis look-alikes are encoded as differentials with referral_significance narrative (dental/ENT/orofacial-pain referral) rather than an engine route. Depth-pass-2 special-population branches encoded as data: pregnancy / childbearing potential — AVOID sodium valproate (teratogenic, neurodevelopmental risk; rxcui 9919 carries NOT_childbearing_potential trigger), carbamazepine teratogenicity (neural-tube) + hyponatraemia, lamotrigine clearance increases in pregnancy → level monitoring, minimise drugs + MFM-neurology co-management (EAN 2019 PMID 30860637); elderly — carbamazepine SIADH/hyponatraemia, falls, AV-block / cardiac conduction, cognitive/sedation → lower start dose, oxcarbazepine or slow titration, weigh posterior-fossa MVD risk vs percutaneous rhizotomy / SRS (AAN-EFNS 2008 PMID 18721143; Barker NEJM 1996 PMID 8598865); renal/hepatic — gabapentin/pregabalin renally dose-adjusted via calc.ckd_epi_2021, carbamazepine/oxcarbazepine/eslicarbazepine/valproate hepatic + Na monitoring; HLA-B*15:02 carbamazepine SJS/TEN screening BEFORE use in Han Chinese / SE-Asian / Thai ancestry (EAN 2019 PMID 30860637; encoded in severity_triggers + contraindication_rules). Depth-pass-2 deprescribing / MVD-referral trigger encoded as data: medically_refractory_or_intolerant severity_trigger fires neurosurgical referral now (do not delay) + MVD-vs-SRS-vs-percutaneous counselling (EAN 2019 PMID 30860637; Barker NEJM 1996 PMID 8598865); step_down_when on the tn_pharmacotherapy axis = sustained remission → taper to lowest effective dose + attempt monitored drug-holiday (deprescribing pathway). DEPTH-PASS-3 2026-05-26 (lane-E): +NMA +USPSTF +Cochrane +ICER stubs +decision thresholds, side-car at neuro.trigeminal-neuralgia.v1._depth-pass-3.md.