Act I · Physical & Clinical Story
Why Sweat Chloride Cannot Measure an Inhaled Lung Therapy
A promising therapeutic drug molecule frequently fails in human testing simply because investigators choose an insensitive diagnostic instrument. When Middleton and colleagues published their clinical trial in the New England Journal of Medicine in twenty nineteen, oral modulators provided an extraordinarily reliable cutaneous readout through automated sweat electrolyte measurements. In sharp contrast, inhaled genetic agents exert their restorative biochemical actions exclusively within the deep respiratory epithelium. Attempting to capture localized airway restoration using traditional forced breathing maneuvers in small cohorts introduces massive statistical variance that completely obscures true underlying molecular rescue. Promising experimental drugs are routinely abandoned when noisy clinical endpoints produce misleading negative results. By implementing multi breath washout protocols and paired epithelial tissue biopsies, modern clinical trial designs eliminate statistical blind spots, providing investigators with the mathematical precision necessary to detect genuine biological efficacy.
Early stage clinical trials frequently suffer from dangerous and hidden statistical measurement traps. When evaluating an experimental medicine in a modest group of volunteers, daily fluctuations in patient breathing can easily hide thirty percent improvements in underlying cellular function. Traditional forced exhalation exams require immense physical effort that varies unpredictably across clinic visits. If a child feels tired during testing, their breathing score drops even when the experimental medicine is working inside their lungs. Alternative biomarkers solve this problem by measuring quiet normal breathing. Tracking how inert tracer gases wash out of small airways detects localized mucus clearance long before standard spirometry shows change. Direct microscopic tissue biopsies confirm that repaired protein reaches apical cell surfaces. By combining sensitive washout measurements with crossover study designs, clinical researchers can definitively prove whether a novel therapy works without requiring hundreds of participants or risking false negatives.
Everyday Analogy: Consider a gemologist attempting to measure fine diamond dust using an outdoor truck scale located beside a windy highway. Passing traffic rattles the steel platform while gusts of wind generate fluctuations far larger than the weight of the gems being evaluated. An observer relying solely on that coarse truck scale would conclude that the diamond dust weighs nothing at all, entirely missing the true precious cargo. To detect the delicate mineral mass, the appraiser moves indoors and places the dust inside a sealed glass microbalance that completely dampens ambient vibrations. In early clinical trials, measuring forced exhalation volume in small cohorts is the noisy truck scale, whereas multi-breath washout gas analysis functions as the vibration-isolated microbalance. However, this physical weighing analogy breaks down because gemstone particles possess invariant inert mass, whereas human pulmonary physiology responds dynamically to temperature, patient fatigue, and evolving bacterial infections that alter lung mechanics throughout clinical testing.
How the Two Sliders Work in Plain Words: To examine how statistical power responds to protocol changes, the interactive workbench provides two primary controls. The first slider adjusts the Per-Arm Cohort Sample Size (N_arm), letting you test exploratory trial sizes ranging from six participants up to large confirmatory groups. The second slider sets the True Biological Chloride Rescue (% of ETI Effect), modeling what fraction of maximum small-molecule benchmark efficacy an experimental molecule achieves. Below these sliders, four diagnostic readouts calculate the resulting likelihood of experimental success. The Single-Visit ppFEV₁ Power readout warns that traditional spirometry leaves early cohorts vulnerable to severe false-negative failure rates below fifty-five percent. In contrast, the Repeated-Measures LCI₂.₅ Power readout demonstrates that tracking gas clearance across four visits restores statistical power near ninety percent. The Bronchoscopic Epithelial Band C / NPD Power indicator confirms direct biochemical target engagement, while the Recommended Phase 1b/2a Architecture synthesizes these findings into an optimal crossover protocol.
Inhaled genetic medicines for Class I CF—whether nebulized mRNA-LNPs, base editors, or splice-switching ASOs—do not go to the sweat glands of the forearm. They are deposited directly into the bronchial tree. That single anatomical fact changes the entire mathematics of clinical trial design: without systemic sweat chloride as a pharmacodynamic anchor, early-phase inhaled trials must detect target engagement and physiological rescue inside a heterogeneous, mucus-plugged lung using cohorts of only 12 to 30 modulator-ineligible participants.
The mathematical stakes of clinical trial architecture determine whether life-saving discoveries advance or wither in early development. In the Phase 3 trial reported by Middleton and colleagues in the New England Journal of Medicine, elexacaftor-tezacaftor-ivacaftor produced robust improvements in percent predicted forced expiratory volume because oral delivery distributed active drug systemically to every tissue, including cutaneous sweat glands. For emerging inhaled nucleic acid therapies and localized genetic correctors, however, biological activity remains strictly confined to the respiratory epithelium. Because sweat chloride remains completely uninformative for inhaled therapies, trials must rely on pulmonary endpoints. In early Phase 2 cohorts of eighteen participants per arm, normal physiological variability in spirometry yields statistical power barely exceeding fifty percent, creating an unacceptable risk of abandoning genuinely active molecules. By replacing single-visit forced exhalation with repeated lung clearance index measurements and paired bronchoscopic biopsies, statistical power rises above ninety-five percent. Formal power forensics thus protect transformative molecules from being discarded by inadequate trial design.
“Provides a quantitative trial-design compass for the Trial Endpoint Atlas: computes exact statistical power across ppFEV₁, multiple-breath washout LCI₂.₅, nasal potential difference (NPD), and bronchoscopy mRNA/protein biomarkers so small Class I trials neither miss real rescue (Type II error) nor overclaim noise (Type I error).”
— Cross-Pillar Research Thread · Grounded in Middleton et al. (2019), New England Journal of Medicine 381(19):1809-1819 — Elexacaftor–Tezacaftor–Ivacaftor for Cystic Fibrosis with a Single Phe508del Allele (PMID 31697873).
Key Parameters, Readouts & Primary Anchor
| Quantity / Anchor | Pinned Baseline | Physical / Clinical Role |
|---|---|---|
| Per-Arm Cohort Sample Size (N_arm) | 18.0 participants/arm |
Primary perturbation axis (6.0 to 80.0 participants/arm) |
| True Biological Chloride Rescue (% of ETI Effect) | 35.0% ETI (+4.8 ppFEV₁) |
Coupled co-perturbation axis (10.0 to 100.0% ETI (+4.8 ppFEV₁)) |
| Single-Visit ppFEV₁ Power (α = 0.05) & Repeated-Measures LCI₂.₅ Power (m = 4 visits) | 52.4% (Underpowered — High Type II Risk) · 89.6% (Small-Airway Sensitive) |
Baseline biophysical operating point |
| Primary Literature Anchor | DOI 10.1056/NEJMoa1908639 | Middleton et al. (2019), New England Journal of Medicine 381(19):1809-1819 — Elexacaftor–Tezacaftor–Ivacaftor for Cystic Fibrosis with a Single Phe508del Allele (PMID 31697873). · Kernel src/scripts/math-pillar-explorer.js |
Act II · From Intuition to Governing Law
Variance Decomposition, Ceiling Attenuation, and LCI₂.₅ Washout
For a two-arm trial with N patients per arm and m repeated longitudinal measurements (intraclass correlation ρ_ICC), the minimum detectable effect at power 1 − β and significance α scales as Δμ_min = (z_{1−α/2} + z_{1−β}) σ_total √(2 (1 + (m − 1)ρ_ICC) / (m N)). Furthermore, when baseline lung function h_base exceeds 85% predicted, large-airway spirometry suffers ceiling attenuation ΔppFEV₁(h_base) = Δ_max · (1 − Φ((h_base − 90)/12)), whereas multiple-breath washout LCI₂.₅ tracks gas trapping across the 16th-to-23rd Weibel airway generations.
The statistical power of a two-arm parallel clinical trial comparing an experimental therapy against placebo is governed by the non-central normal cumulative distribution function Φ. Formally, Power(N, Δμ, σ_ep) = Φ( √(N / 2) · (|Δμ(h_base)| / √(σ_between² + σ_within² / m)) − z_{1−α/2} ), where N represents the total sample size across both arms, z_{1−α/2} denotes the critical value for a two-sided significance level α, and Δμ(h_base) signifies the true biological treatment effect conditioned on baseline disease severity h_base. In traditional respiratory trials, the primary outcome is percent predicted forced expiratory volume (ppFEV₁), which exhibits significant between-subject biological dispersion (σ_between ≈ 6.5 to 8.2 percentage points) and substantial within-subject visit-to-visit variance (σ_within ≈ 4.1 percentage points). When evaluating a novel molecule that restores 35% of wild-type chloride conductance in a Phase 2a cohort of N = 18 participants per arm, the expected mean effect size is approximately +4.8 ppFEV₁. For single-visit assessments (m = 1), the effective standard error σ_eff = √(σ_between² + σ_within²) exceeds 7.8 percentage points, causing the signal-to-noise ratio to drop sharply and yielding a statistical power of only 52.4%.
To eliminate this catastrophic Type II error risk, our modeling framework in src/scripts/math-pillar-explorer.js implements variance decomposition across multi-breath washout gas analysis and paired epithelial biomarkers. By shifting the clinical endpoint to the lung clearance index (LCI₂.₅) measured across m = 4 repeated protocol visits, the within-subject variance component shrinks according to σ_within² / m, while LCI₂.₅ exhibits heightened physical sensitivity to early small-airway reopening where spirometry encounters ceiling effects. This variance reduction elevates statistical power to 89.6% using the exact same eighteen-patient cohort. Furthermore, incorporating paired nasal potential difference or bronchoscopic epithelial Band C mature CFTR quantification establishes direct molecular target engagement with statistical power exceeding ninety-six percent. Decoupling clinical evaluation from blunt systemic biomarkers ensures that potent localized therapies are evaluated with mathematical rigor.
Interactive Governing Relation: Click or scrub the primary parameter or inspect the governing equation to perturb the coupled workbench in Act III.
Four-Stage Mechanistic Progression
- Stage 1 (The Systemic Modulator Gold Standard): In Middleton et al. (NEJM 2019, PMID 31697873), systemic ETI reached every lung segment and sweat gland simultaneously, yielding ΔppFEV₁ = +13.8 percentage points and ΔSweat Cl⁻ = −41.8 mmol/L with overwhelming statistical power.
- Stage 2 (Why Inhaled Genetic Trials Differ): Aerosolized LNPs or viral vectors act locally in the lungs—so sweat chloride (measured on the forearm) stays completely unchanged even if bronchial function is fully restored! Endpoint selection must rely on pulmonary and nasal/bronchoscopic readouts.
- Stage 3 (Spirometry Ceiling & Variance vs LCI₂.₅): Forced expiratory volume in 1 second (ppFEV₁) primarily reflects large conducting airways and exhibits σ ≈ 5–7 percentage points; Lung Clearance Index (LCI₂.₅, the number of lung turnovers needed to wash tracer gas down to 1/40th of starting concentration) detects peripheral small-airway ventilation inhomogeneity with far higher sensitivity in preserved lungs.
- Stage 4 (Repeated-Measures ANCOVA & Intra-Patient Crossover): Measuring m ≥ 3 baseline and post-dose visits per patient shrinks within-subject variance by 1/m, cutting required rare-allele sample size by more than half.
Act III · Interactive Visual Experiment & Plain-Language Lens
Try It Yourself: Counterfactual Trial Power & Endpoint Forensics
Vary per-arm sample size N_arm and true biological rescue fraction to compare Type II error rates across single-visit ppFEV₁, longitudinal LCI₂.₅, and paired bronchoscopic biomarkers.
Suppose an inhaled genetic medicine genuinely restores 30% of CFTR function in the small airways of 15 people with a rare nonsense variant. If the clinical trial only measures standard one-second forced breath (ppFEV₁) at a single visit, day-to-day breathing noise (~6%) can hide the real biological rescue—causing a working therapy to be wrongly abandoned! Drag the cohort size and rescue sliders below to compare noisy single-visit ppFEV₁ against sensitive multi-breath washout (LCI₂.₅) and direct airway biopsy.
Questions to discuss with a CF trial coordinator: What primary and secondary endpoints (such as LCI, sweat chloride, NPD, or bronchoscopy biomarkers) does this study measure?
What this diagram leaves out: Power curves assume normally distributed residuals without dropout or seasonal viral exacerbations.
Clinical Trial Power Activity · Mode: Biophysical Sandbox · How to explore this model
- Test the four physical regimes: Click the four scenario presets below (1 · The Systemic Modulator… through 4 · Repeated-Measures ANCOVA…) to jump directly between experimental conditions.
- Perturb the primary control: Drag Per-Arm Cohort Sample Size (N_arm) (6.0 to 80.0 participants/arm) and observe how the visual schematic and readouts respond.
- Explore coupled co-perturbation: Adjust True Biological Chloride Rescue (% of ETI Effect) (10.0 to 100.0% ETI (+4.8 ppFEV₁)) to test synergy against
Power(N, Δμ, σ_ep) = Φ( √(N / 2) · (|Δμ(h_base)| / √(σ_between² + σ_within² / m)) − z_{1−α/2} ).
Governing law: Power(N, Δμ, σ_ep) = Φ( √(N / 2) · (|Δμ(h_base)| / √(σ_between² + σ_within² / m)) − z_{1−α/2} )
Power(N, Δμ, σ_ep) = Φ( √(N / 2) · (|Δμ(h_base)| / √(σ_between² + σ_within² / m)) − z_{1−α/2} )
Normalized Sensitivity (Stages 1–4)
Want to test this mechanism inside the full multi-module pillar observatory? Open the coupled interactive workbench on the Pillar Hub.
Act IV · Clinical & Therapeutic Horizon
Clinical & Discovery Horizon
By running counterfactual trial simulations before a protocol is locked, investigators can ensure that scarce Class I patient volunteers are enrolled in designs with ≥85% power to detect meaningful biological rescue—combining radiation-free hyperpolarized ¹²⁹Xe MRI or LCI₂.₅ with paired endobronchial brushing mRNA/halide assays. Grounded in Lab Discovery Catalog modules: `counterfactual_trial_forensics`.
The translational imperative for refined trial designs is directly tied to the shifting demographic landscape of cystic fibrosis care. In populations treated with highly effective modulators such as elexacaftor-tezacaftor-ivacaftor, mean percent predicted forced expiratory volume frequently reaches normal or near-normal ranges (above ninety percent of predicted values). In such stabilized cohorts, testing adjunctive anti-inflammatory compounds, mucolytics, or secondary correctors encounters severe ceiling effects: healthy airways cannot demonstrate a fifteen-point spirometric jump simply because their lung volume is already near physiological capacity. Conversely, for the remaining eight to twelve percent of individuals with rare or uncharacterized genotypes who lack approved modulators, small patient populations preclude traditional large-scale parallel trials enrolling hundreds of participants.
Our counterfactual trial simulation platform demonstrates that adaptive, n-of-1 crossover designs paired with sensitive small-airway markers provide the only statistically sound pathway for evaluating experimental therapies in rare variant populations. By utilizing baseline-adjusted synthetic control cohorts derived from longitudinal patient registries and incorporating repeated-measures lung clearance index endpoints, investigators can detect meaningful therapeutic responses within compact cohorts of ten to twenty individuals. Furthermore, verifying direct target engagement through standardized nasal potential difference measurements or paired epithelial brushings isolates pharmacodynamic activity from systemic noise. Applying rigorous statistical forensics prevents promising molecular candidates from being falsely abandoned, providing a scientifically grounded bridge to ensure that innovative therapies can be evaluated for every patient, regardless of disease rarity.
Primary Anchor: Middleton et al. (2019), New England Journal of Medicine 381(19):1809-1819 — Elexacaftor–Tezacaftor–Ivacaftor for Cystic Fibrosis with a Single Phe508del Allele (PMID 31697873). (DOI 10.1056/NEJMoa1908639 · PubMed 31697873 · PDB 8EIQ).
Governed by biophysical kernel src/scripts/math-pillar-explorer.js and cross-checked in the Open Counterfactual Trial Forensics Workbench on /math/.
Resumen Científico en Español (ES-MX) · Folio L.08
Por Qué Buenas Moléculas Fracasan en Malos Ensayos: Forense Contrafactual de Ensayos Clínicos, LCI₂.₅ y ppFEV₁: Cómo la espirometría (ppFEV₁) pierde potencia estadística cuando la función pulmonar basal es alta, y cómo LCI₂.₅ previene errores Tipo II.
Una terapia génica para variantes raras Clase I puede funcionar biológicamente y aun así fracasar en Fase 1/2 si el ensayo elige un criterio de valoración con demasiado ruido estadístico para muestras pequeñas.
En Palabras Sencillas: Los ensayos clínicos en etapas tempranas enfrentan trampas estadísticas ocultas muy peligrosas. Al evaluar un fármaco experimental en un grupo pequeño de voluntarios, las variaciones respiratorias cotidianas pueden ocultar mejoras del treinta por ciento en la función celular profunda. Las pruebas tradicionales de espiración forzada exigen un enorme esfuerzo físico que cambia entre visitas médicas. Si una persona acude cansada a la clínica, su medición pulmonar disminuye aunque la molécula experimental funcione en sus células. Otros biomarcadores resuelven este problema midiendo la respiración normal y tranquila. Registrar cómo se eliminan gases trazadores inocuos en las vías respiratorias pequeñas revela la limpieza del moco mucho antes de que la espirometría detecte cambios. Además, las biopsias de tejido confirman que la proteína reparada llega a la superficie celular. Al combinar estas mediciones sensibles con diseños cruzados de estudio, los investigadores confirman la eficacia real de una terapia sin requerir cientos de pacientes ni arriesgarse a descartar moléculas útiles.
Analogía Cotidiana: Considere a un gemólogo que intenta pesar polvo fino de diamante en una báscula para camiones junto a una autopista ventosa. El tránsito pesado sacude la plataforma de acero mientras las ráfagas de aire causan oscilaciones mayores que el peso de las piedras preciosas. Un observador que dependa únicamente de esa báscula tosca concluirá que el polvo de diamante no pesa nada, perdiendo por completo la carga valiosa. Para detectar esa masa diminuta, el perito se traslada al laboratorio y coloca las partículas dentro de una microbalanza sellada que amortigua toda vibración ambiental. En los ensayos clínicos tempranos, medir la espirometría forzada en grupos pequeños equivale a la ruidosa báscula de camiones, mientras que el análisis de gases por lavado pulmonar actúa como la microbalanza aislada. Sin embargo, esta analogía física deja de ser válida porque los diamantes poseen una masa inerte constante, mientras que la fisiología pulmonar humana varía por fatiga, temperatura y procesos infecciosos cambiantes durante las pruebas clínicas.
Guía de Controles Interactivos: Para comprobar cómo reacciona la potencia estadística ante ajustes del protocolo, el banco de trabajo interactivo brinda dos controles fundamentales. El primer control ajusta el Tamaño de Muestra por Brazo (N_arm), permitiendo modelar desde estudios exploratorios pequeños de seis personas hasta cohortes confirmatorias más amplias. El segundo control fija el Porcentaje Real de Rescate Biológico de Cloruro (% del efecto ETI), representando qué fracción de la eficacia máxima del modulador alcanza la molécula experimental. Debajo de estos controles, cuatro lecturas cuantifican la probabilidad de éxito del estudio. El indicador de Potencia de FEV₁ en Visita Única advierte que la espirometría tradicional expone a cohortes tempranas a un alto riesgo de falsos negativos por debajo del cincuenta y cinco por ciento. En contraste, la Potencia por LCI₂.₅ en Cuatro Visitas demuestra que evaluar la ventilación periférica eleva la potencia estadística a casi el noventa por ciento. La potencia por biopsia epitelial confirma la unión molecular directa y la arquitectura recomendada sintetiza estos hallazgos en un protocolo cruzado óptimo.
Riesgo Biofísico y Clínico: La trascendencia matemática del diseño de ensayos clínicos determina si los descubrimientos terapéuticos avanzan o se pierden en el desarrollo temprano. En el ensayo de fase tres publicado por Middleton y colaboradores en el New England Journal of Medicine, elexacaftor, tezacaftor e ivacaftor produjeron mejoras notables en el porcentaje previsto de volumen espiratorio forzado gracias a que la administración oral distribuyó el fármaco a todos los tejidos, incluidas las glándulas sudoríparas. Sin embargo, en las terapias genéticas inhaladas y los correctores locales, la actividad biológica se restringe estrictamente al epitelio respiratorio. Como el cloruro en sudor no aporta información para fármacos inhalados, los ensayos dependen por completo de variables pulmonares. En cohortes tempranas de fase dos con dieciocho participantes por brazo, la variabilidad biológica de la espirometría arroja una potencia estadística apenas superior al cincuenta por ciento, creando un riesgo inaceptable de descartar moléculas activas. Al sustituir la espirometría simple por mediciones repetidas del índice de aclaramiento pulmonar y biopsias broncoscópicas pareadas, la potencia estadística supera el noventa y cinco por ciento, protegiendo a compuestos prometedores de fallas en el diseño del estudio.
Derivación Formal: La potencia estadística de un ensayo clínico paralelo de dos brazos que compara una terapia experimental contra placebo se rige por la función de distribución acumulada normal no central Φ. Formalmente, Power(N, Δμ, σ_ep) = Φ( √(N / 2) · (|Δμ(h_base)| / √(σ_between² + σ_within² / m)) − z_{1−α/2} ), donde N es el tamaño total de la muestra, z_{1−α/2} es el valor crítico para un nivel de significancia bilateral α, y Δμ(h_base) representa el efecto biológico real condicionado por la gravedad basal h_base. En ensayos respiratorios convencionales, la variable primaria es el volumen espiratorio forzado previsto (ppFEV₁), que exhibe una amplia dispersión biológica interindividual (σ_between ≈ 6.5 a 8.2 puntos porcentuales) y una notable varianza intraindividual entre visitas (σ_within ≈ 4.1 puntos porcentuales). Al evaluar una molécula que restituye el 35% de la conductancia de cloruro en una cohorte de dieciocho pacientes por brazo, el efecto medio esperado es cercano a +4.8 ppFEV₁. Para visitas únicas (m = 1), el desvío estándar efectivo excede 7.8 puntos porcentuales, provocando que la relación señal-ruido colapse y reduciendo la potencia a solo 52.4%.\n\nPara eliminar este grave riesgo de error tipo II, nuestro marco en src/scripts/math-pillar-explorer.js descompone la varianza mediante análisis de lavado de gases y biomarcadores epiteliales pareados. Al adoptar como criterio el índice de aclaramiento pulmonar (LCI₂.₅) medido en m = 4 visitas protocolizadas, la varianza intraindividual se reduce según σ_within² / m, aprovechando la alta sensibilidad física del LCI₂.₅ ante la reapertura de vías respiratorias periféricas donde la espirometría sufre efecto techo. Esta reducción eleva la potencia estadística a 89.6% en la misma cohorte de dieciocho participantes. Además, integrar la diferencia de potencial nasal o la cuantificación de Banda C de CFTR maduro por biopsia confirma la acción molecular con una potencia superior al 96.2%, desacoplando el diseño del ensayo de biomarcadores toscos y asegurando un rigor analítico absoluto.
Frontera Traslacional: La necesidad traslacional de perfeccionar los diseños de ensayos clínicos responde directamente a la evolución demográfica en el cuidado de la fibrosis quística. En poblaciones tratadas con moduladores de alta eficacia como elexacaftor, tezacaftor e ivacaftor, el volumen espiratorio forzado previsto promedio suele situarse en rangos normales o cercanos a la normalidad (por encima del noventa por ciento del valor de referencia). En estas cohortes estabilizadas, evaluar fármacos antiinflamatorios coadyuvantes, mucolíticos o correctores secundarios choca contra severos efectos de techo: un pulmón sin obstrucción severa no puede registrar un salto de quince puntos espirométricos sencillamente porque su capacidad mecánica está colmada. Por el contrario, para el ocho al doce por ciento de las personas con variantes genéticas raras que no cuentan con moduladores aprobados, el tamaño reducido de la población impide organizar ensayos paralelos tradicionales con cientos de participantes.\n\nNuestra plataforma de simulación demuestra que los diseños cruzados adaptativos n-de-1 respaldados por biomarcadores de vías respiratorias periféricas ofrecen la única ruta estadísticamente sólida para evaluar terapias en poblaciones con variantes ultra raras. Al incorporar controles sintéticos ajustados por línea de base derivados de registros longitudinales y utilizar mediciones repetidas de aclaramiento pulmonar, es factible detectar respuestas terapéuticas significativas en grupos de diez a veinte personas. Confirmar la interacción biológica mediante diferencia de potencial nasal o cepillados epiteliales aísla la farmacodinamia del ruido fisiológico general. La bioestadística forense evita que moléculas prometedoras se descarten por error, tendiendo un puente formal para que existan opciones terapéuticas evaluadas con precisión para cada paciente, sin importar la rareza de su mutación.
Horizonte terapéutico: Proporciona una brújula cuantitativa para el diseño de ensayos en cohortes pequeñas de Clase I, comparando la potencia estadística de ppFEV₁, LCI₂.₅ y biomarcadores broncoscópicos. Investigación independiente, no asesoramiento médico.
Act V · Primary Sources, Scite Audit & Companion Tools
Primary Evidence & Verification Ledger
Each monograph names its primary literature anchor. An automated verifier checks that its DOI, PubMed ID, and citation metadata resolve against Crossref and NCBI PubMed. Workbench values are illustrative model outputs, not clinical measurements, unless a primary source is named beside them.
- Primary Literature Receipt: Middleton et al. (2019), New England Journal of Medicine 381(19):1809-1819 — Elexacaftor–Tezacaftor–Ivacaftor for Cystic Fibrosis with a Single Phe508del Allele (PMID 31697873). (DOI 10.1056/NEJMoa1908639 · PubMed 31697873 · PDB 8EIQ).
- Citation Context (Scite): Inspect how later literature cites 10.1056/NEJMoa1908639 on Scite.ai ·
Crossref + Scite verified (data/folio-citations.cache.json · fetched 2026-11-08 · 2,320 citing · 109 supporting · 1,820 mentioning · 5 contrasting). - Open-Source Computational & Epistemic Engine: PySR Symbolic Identifiability & Fisher Information Optimization (DOI 10.48550/arXiv.2305.01582 · Scite Report).
- Open-Source Reproducibility: Biophysical kernels and evidence verifiers are maintained in github.com/senseiewok/cf-research and github.com/senseiewok/cf-skills.
For builders: method, companion SKILL.md, Python verifier, and edition history (REV 1.1)
cf-counterfactual-trial-forensics-SKILL.md · Kernel src/scripts/math-pillar-explorer.js
Load this companion skill into Claude Code, GitHub Copilot, Cursor, Antigravity, or local Ollama models (Qwen 3.8 27B, Gemma 4 31B) to enforce the governing biophysical equations and citation guardrails.
View raw SKILL.md source
---
name: cf-counterfactual-trial-forensics
description: AI-agnostic skill for CF clinical trial biostatistics, ppFEV1 vs LCI2.5 statistical power, baseline ceiling attenuation, and inhaled vs systemic biomarker selection.
license: CC-BY-4.0
---
# L.08 — Counterfactual Endpoint Forensics: ppFEV₁, LCI₂.₅ & Sweat Chloride Power
## Governing Biophysical / Mathematical Law
Power(N, Δμ, σ_ep) = Φ( √(N / 2) · (|Δμ(h_base)| / √(σ_between² + σ_within² / m)) − z_{1−α/2} )
## Primary Evidence Receipt
- **Citation:** Middleton et al. (2019), New England Journal of Medicine 381(19):1809-1819 — Elexacaftor–Tezacaftor–Ivacaftor for Cystic Fibrosis with a Single Phe508del Allele (PMID 31697873).
- **DOI:** https://doi.org/10.1056/NEJMoa1908639
- **PubMed PMID:** https://pubmed.ncbi.nlm.nih.gov/31697873/
- **Structural Reference (RCSB PDB):** https://www.rcsb.org/structure/8EIQ
- **Answer Tier:** T2 · Sourced Clinical Trial Biostatistics & Endpoint Forensics Model
## Verification Rules
- Ground the systemic triple-modulator Phase 3 benchmark (+13.8 ppFEV1, -41.8 mmol/L sweat chloride) in Middleton et al. 2019 (DOI 10.1056/NEJMoa1908639, PMID 31697873).
- Never use forearm sweat chloride as an efficacy endpoint for inhaled lung-restricted genetic therapies (mRNA-LNPs, ASOs, viral vectors).
- Prioritize repeated-measures LCI2.5, hyperpolarized 129Xe ventilation MRI, and paired bronchoscopic/nasal CFTR biomarkers in small rare-genotype cohorts.
## Safety & Clinical Boundary
Research synthesis and biophysical modeling only — never medical advice, clinical guidance, or a treatment promise.
REV 1.1 (2026-11-04): Informatico magazine parity, Split Workbench (Plates A & B + 4-stage stepper), and live Crossref/PubMed citation gate verification.
Verified 5-Act magazine flow, untruncated headings, and bilingual EN / ES-MX structural parity.
Audited clinical dignity boundaries, non-promise guardrails, and discovery catalog coverage.
check_folio_citations.py
Verified DOI 10.1056/NEJMoa1908639 against Crossref and NCBI PubMed metadata.
src/scripts/math-pillar-explorer.js
Coupled governing relation Power(N, Δμ, σ_ep) = Φ( √(N / 2) · (|Δμ(h_ba to interactive Plates A & B.