MTX-PNEUMO: Transparent Methotrexate-Associated Pneumonitis Risk Stratification in Rheumatic and Autoimmune Disease

Authors: Dr. Erick Zamora-Tehozol, DNAI, RheumaAI
ORCID: 0000-0002-7888-3961

Abstract

Methotrexate remains the anchor conventional synthetic DMARD for rheumatoid arthritis and other autoimmune disease, yet methotrexate-associated pneumonitis remains one of its most feared noninfectious toxicities because the syndrome is uncommon, diagnostically messy, and potentially severe. The bedside problem is rarely ignorance that lung toxicity exists. The harder clinical task is deciding when new respiratory symptoms during methotrexate exposure should trigger urgent drug-toxicity concern instead of routine watchful waiting. We present MTX-PNEUMO, an executable Python skill for transparent methotrexate-pneumonitis risk stratification before or during treatment. The model integrates age, time since methotrexate initiation, weekly dose, pre-existing interstitial lung disease or fibrosis, abnormal baseline chest imaging, prior DMARD lung toxicity, diabetes, hypoalbuminemia, chronic kidney disease, dyspnea, dry cough, fever, hypoxemia, eosinophilia, diffuse interstitial or ground-glass imaging pattern, and whether infection has been excluded. Outputs include visible component scores, categorical risk classes, recommended actions, alerts, and explicit limitations. In demonstration scenarios, a stable low-dose patient without pulmonary symptoms is LOW risk, an older patient early after methotrexate initiation with cough and fever is HIGH risk, and a patient with baseline ILD plus acute dyspnea, hypoxemia, and diffuse infiltrates soon after start is VERY HIGH risk. MTX-PNEUMO is intended as an auditable triage aid rather than a diagnostic gold standard and does not replace imaging interpretation, infectious workup, or specialist pulmonary judgment.

Keywords: methotrexate, pneumonitis, rheumatoid arthritis, interstitial lung disease, pulmonary toxicity, clinical decision support, DeSci, RheumaAI

1. Clinical problem

Methotrexate has transformed autoimmune care because it is effective, affordable, and familiar. But pulmonary toxicity remains a major source of anxiety because methotrexate pneumonitis can present abruptly with cough, fever, dyspnea, hypoxemia, and diffuse interstitial or ground-glass radiographic change. The event is uncommon, yet missing it can cause serious harm. Over-calling it also causes harm because infection and rheumatoid arthritis-associated interstitial lung disease are common competing explanations.

This creates a real decision problem: which patients deserve immediate methotrexate interruption and escalated evaluation, and which patients mainly need structured monitoring? A transparent score can help make that reasoning visible.

2. Methodology

2.1 Design principles

The score follows five practical principles:

1. Pulmonary reserve matters. Pre-existing ILD/fibrosis and abnormal baseline imaging lower safety margin.
2. Timing matters. Acute or subacute symptoms in the first months after methotrexate initiation fit the classic presentation.
3. Clinical pattern matters. Dyspnea, dry cough, fever, hypoxemia, eosinophilia, and diffuse interstitial imaging jointly raise concern.
4. Host vulnerability matters. Older age, diabetes, hypoalbuminemia, CKD, and prior DMARD lung toxicity worsen risk context.
5. Uncertainty must stay explicit. Infection is a major competing diagnosis and must be actively excluded.

2.2 Model structure

The implementation computes four visible components:

• Host vulnerability — age, pre-existing ILD/fibrosis, abnormal baseline imaging, prior DMARD lung toxicity, diabetes, hypoalbuminemia, and CKD
• Exposure pattern — methotrexate exposure, weekly dose, and early timing after treatment initiation
• Presentation severity — dyspnea, dry cough, fever, hypoxemia, eosinophilia, and diffuse interstitial or ground-glass pattern
• Diagnostic uncertainty — whether infection remains unexcluded

Interaction terms intensify concern when baseline ILD coexists with new diffuse infiltrates, when early-onset symptoms include both fever and dyspnea, and when hypoalbuminemia coexists with diabetes. A modest down-adjustment occurs when the patient is asymptomatic or when infection has already been reasonably excluded.

2.3 Output logic

The skill returns:

• Total score
• Risk class: LOW, INTERMEDIATE, HIGH, VERY HIGH
• Recommended actions
• Safety alerts
• Explicit limitations

3. Executable skill

3.1 Implementation

The implementation is standalone Python using only the standard library and is stored locally at:

skills/mtx-pneumo/mtx_pneumo.py

3.2 Demo output summary

RA patient on low-dose methotrexate without pulmonary symptoms -> LOW
Older patient early after methotrexate start with cough and fever -> HIGH
Baseline ILD with acute dyspnea, hypoxemia, and diffuse infiltrates soon after start -> VERY HIGH

Representative very-high-risk output:

total_score: 124.0
risk_class: VERY HIGH
alert: Hypoxemia/desaturation is a severity signal and should not be handled as routine outpatient cough triage.

4. Why this solves a real problem

Methotrexate lung toxicity is often discussed either too vaguely or too absolutely. In real practice, clinicians face competing diagnoses, partial data, and short timelines. MTX-PNEUMO does not pretend to solve the entire diagnostic problem. It solves a narrower but important one: making the structure of concern explicit enough that teams can escalate faster when the pattern is dangerous and document uncertainty honestly when it is not.

5. Limitations

1. This is an evidence-informed heuristic tool, not a validated probability calculator for methotrexate pneumonitis.
2. Methotrexate pneumonitis is uncommon and lacks a single definitive diagnostic test; competing infection and RA-ILD remain major confounders.
3. Weights are clinically derived from reviews and observational risk patterns rather than prospective multivariable calibration.
4. The tool is not appropriate as a stand-alone decision engine for ICU transfer, bronchoscopy, or glucocorticoid initiation.
5. Definitive management requires clinician review of imaging, oxygenation, alternative causes, and overall rheumatoid disease context.

References

1. Fragoulis GE, Nikiphorou E, Larsen J, Korsten P, Conway R. Methotrexate-Associated Pneumonitis and Rheumatoid Arthritis-Interstitial Lung Disease: Current Concepts for the Diagnosis and Treatment. Front Med (Lausanne). 2019;6:238. DOI: 10.3389/fmed.2019.00238
2. Conway R, Low C, Coughlan RJ, O'Donnell MJ, Carey JJ. Methotrexate and Lung Disease in Rheumatoid Arthritis: A Meta-Analysis of Randomized Controlled Trials. Arthritis Rheumatol. 2014;66(4):803-812. DOI: 10.1002/art.38322
3. Barrera P, Laan RF, van Riel PL, Dekhuijzen PN, Boerbooms AM, van de Putte LB. Methotrexate-related pulmonary complications in rheumatoid arthritis. Ann Rheum Dis. 1994;53(7):434-439. DOI: 10.1136/ard.53.7.434
4. Dawson JK, Graham DR, Desmond J, Fewins HE, Lynch MP. Investigation of the chronic pulmonary effects of low-dose oral methotrexate in patients with rheumatoid arthritis: a prospective study incorporating HRCT scanning and pulmonary function tests. Rheumatology (Oxford). 2002;41(3):262-267. DOI: 10.1093/rheumatology/41.3.262
5. Fraenkel L, Bathon JM, England BR, et al. 2021 American College of Rheumatology Guideline for the Treatment of Rheumatoid Arthritis. Arthritis Rheumatol. 2021;73(7):1108-1123. DOI: 10.1002/art.41752

Executable skill_md

#!/usr/bin/env python3
"""
MTX-PNEUMO — Methotrexate-Associated Pneumonitis Risk Stratification

Transparent clinical skill for estimating concern for methotrexate-associated
pneumonitis before or during methotrexate exposure in rheumatic and autoimmune
disease.

Authors: Dr. Erick Zamora-Tehozol (ORCID:0000-0002-7888-3961), DNAI, RheumaAI
License: MIT
"""

from dataclasses import dataclass, asdict
from typing import Dict, Any, List, Optional
import json


@dataclass
class MtxPneumoInput:
    age: int
    diagnosis: str
    planned_or_current_methotrexate: bool = True
    months_since_start: Optional[float] = None
    preexisting_ild_or_fibrosis: bool = False
    abnormal_baseline_chest_imaging: bool = False
    prior_dmard_lung_toxicity: bool = False
    diabetes: bool = False
    hypoalbuminemia: bool = False
    chronic_kidney_disease: bool = False
    current_dose_mg_per_week: float = 0.0
    dyspnea: bool = False
    dry_cough: bool = False
    fever: bool = False
    resting_hypoxemia_or_desaturation: bool = False
    eosinophilia: bool = False
    diffuse_interstitial_or_ground_glass_pattern: bool = False
    infection_not_yet_excluded: bool = True


def host_vulnerability(inp: MtxPneumoInput) -> float:
    score = 0.0
    if inp.age >= 70:
        score += 2.0
    elif inp.age >= 60:
        score += 1.2
    if inp.preexisting_ild_or_fibrosis:
        score += 2.6
    if inp.abnormal_baseline_chest_imaging:
        score += 1.4
    if inp.prior_dmard_lung_toxicity:
        score += 1.8
    if inp.diabetes:
        score += 1.0
    if inp.hypoalbuminemia:
        score += 1.2
    if inp.chronic_kidney_disease:
        score += 1.0
    return score


def exposure_pattern(inp: MtxPneumoInput) -> float:
    score = 0.0
    if inp.planned_or_current_methotrexate:
        score += 1.2
    if inp.current_dose_mg_per_week >= 20:
        score += 0.8
    elif inp.current_dose_mg_per_week >= 15:
        score += 0.4
    if inp.months_since_start is not None:
        if inp.months_since_start <= 6:
            score += 1.8
        elif inp.months_since_start <= 12:
            score += 0.8
    return score


def presentation_severity(inp: MtxPneumoInput) -> float:
    score = 0.0
    if inp.dyspnea:
        score += 1.6
    if inp.dry_cough:
        score += 1.2
    if inp.fever:
        score += 1.1
    if inp.resting_hypoxemia_or_desaturation:
        score += 2.0
    if inp.eosinophilia:
        score += 0.9
    if inp.diffuse_interstitial_or_ground_glass_pattern:
        score += 1.8
    return score


def uncertainty_penalty(inp: MtxPneumoInput) -> float:
    return 1.2 if inp.infection_not_yet_excluded else 0.0


def total_score(inp: MtxPneumoInput) -> float:
    score = (
        host_vulnerability(inp)
        + exposure_pattern(inp)
        + presentation_severity(inp)
        + uncertainty_penalty(inp)
    )
    if inp.preexisting_ild_or_fibrosis and inp.diffuse_interstitial_or_ground_glass_pattern:
        score += 1.2
    if inp.months_since_start is not None and inp.months_since_start <= 6 and inp.fever and inp.dyspnea:
        score += 1.0
    if inp.hypoalbuminemia and inp.diabetes:
        score += 0.8
    if not inp.dyspnea and not inp.dry_cough and not inp.fever:
        score -= 1.0
    if not inp.infection_not_yet_excluded:
        score -= 0.6
    return round(max(score, 0.0) * 5.0, 1)


def classify(score: float) -> str:
    if score >= 60:
        return "VERY HIGH"
    if score >= 35:
        return "HIGH"
    if score >= 18:
        return "INTERMEDIATE"
    return "LOW"


def recommendations(inp: MtxPneumoInput, score: float) -> List[str]:
    out: List[str] = []
    if score < 18:
        out.append("Current methotrexate pneumonitis concern is low, but baseline lung history and symptom review should still be documented.")
    elif score < 35:
        out.append("Meaningful pulmonary caution: review baseline imaging/pulmonary history and intensify symptom surveillance after methotrexate exposure.")
    elif score < 60:
        out.append("High concern for methotrexate-associated pneumonitis: urgent reassessment, chest imaging, pulse oximetry, and methotrexate interruption should be considered while alternatives are evaluated.")
    else:
        out.append("Very high concern for methotrexate-associated pneumonitis: treat this as a potentially serious drug toxicity until proven otherwise.")
        out.append("Immediate drug cessation, expedited imaging, infection workup, and specialist-level pulmonary assessment are appropriate.")
    if inp.infection_not_yet_excluded:
        out.append("Infection remains a competing diagnosis and should be actively excluded rather than assumed away.")
    else:
        out.append("Even when infection seems unlikely, methotrexate pneumonitis remains a diagnosis of exclusion supported by the overall clinical pattern.")
    return out


def alerts(inp: MtxPneumoInput, score: float) -> List[str]:
    out: List[str] = []
    if inp.preexisting_ild_or_fibrosis:
        out.append("Pre-existing ILD/fibrosis reduces reserve and can make methotrexate-related lung toxicity harder to distinguish from background disease.")
    if inp.months_since_start is not None and inp.months_since_start <= 6:
        out.append("Acute/subacute symptoms in the first months after methotrexate initiation fit the classic timing of methotrexate pneumonitis.")
    if inp.resting_hypoxemia_or_desaturation:
        out.append("Hypoxemia/desaturation is a severity signal and should not be handled as routine outpatient cough triage.")
    if score >= 35:
        out.append("This tool supports transparent triage and monitoring; it does not replace imaging interpretation, bronchoalveolar evaluation, or specialist judgment.")
    return out


def run_mtx_pneumo(inp: MtxPneumoInput) -> Dict[str, Any]:
    score = total_score(inp)
    return {
        "input_summary": asdict(inp),
        "host_vulnerability": round(host_vulnerability(inp), 2),
        "exposure_pattern": round(exposure_pattern(inp), 2),
        "presentation_severity": round(presentation_severity(inp), 2),
        "diagnostic_uncertainty": round(uncertainty_penalty(inp), 2),
        "total_score": score,
        "risk_class": classify(score),
        "recommended_actions": recommendations(inp, score),
        "alerts": alerts(inp, score),
        "limitations": [
            "Evidence-informed heuristic model, not a validated probability calculator for methotrexate pneumonitis.",
            "Methotrexate pneumonitis is uncommon and lacks a single definitive diagnostic test; competing infection and RA-ILD remain major confounders.",
            "Weights are clinically derived from reviews and observational risk patterns rather than prospective multivariable calibration.",
            "The tool is not appropriate as a stand-alone decision engine for ICU transfer, bronchoscopy, or glucocorticoid initiation.",
            "Definitive management requires clinician review of imaging, oxygenation, alternative causes, and overall rheumatoid disease context."
        ]
    }


if __name__ == '__main__':
    demos = [
        (
            'RA patient on low-dose methotrexate without pulmonary symptoms',
            MtxPneumoInput(age=46, diagnosis='Rheumatoid arthritis', current_dose_mg_per_week=12.5, months_since_start=18, infection_not_yet_excluded=False),
        ),
        (
            'Older patient early after methotrexate start with cough and fever',
            MtxPneumoInput(age=63, diagnosis='Rheumatoid arthritis', current_dose_mg_per_week=15, months_since_start=4, dry_cough=True, fever=True),
        ),
        (
            'Baseline ILD with acute dyspnea, hypoxemia, and diffuse infiltrates soon after start',
            MtxPneumoInput(age=72, diagnosis='Rheumatoid arthritis', current_dose_mg_per_week=20, months_since_start=2, preexisting_ild_or_fibrosis=True, abnormal_baseline_chest_imaging=True, diabetes=True, hypoalbuminemia=True, dyspnea=True, dry_cough=True, fever=True, resting_hypoxemia_or_desaturation=True, eosinophilia=True, diffuse_interstitial_or_ground_glass_pattern=True),
        ),
    ]

    print('=' * 78)
    print('MTX-PNEUMO — Methotrexate-Associated Pneumonitis Risk Stratification')
    print('Authors: Dr. Erick Zamora-Tehozol, DNAI, RheumaAI')
    print('=' * 78)
    for label, demo in demos:
        result = run_mtx_pneumo(demo)
        print(f'\n--- {label} ---')
        print(json.dumps(result, indent=2))

Skill documentation

MTX-PNEUMO

Methotrexate-associated pneumonitis risk stratification in rheumatic and autoimmune disease

What it does

MTX-PNEUMO is a transparent clinical skill that estimates concern for methotrexate-associated pneumonitis before methotrexate starts or while evaluating new respiratory symptoms during exposure.

Inputs

• Age and diagnosis
• Planned or current methotrexate exposure
• Time since methotrexate initiation and weekly dose
• Pre-existing ILD/fibrosis or abnormal baseline chest imaging
• Prior DMARD lung toxicity
• Diabetes, hypoalbuminemia, and chronic kidney disease
• Dyspnea, dry cough, fever, hypoxemia/desaturation
• Eosinophilia
• Diffuse interstitial or ground-glass imaging pattern
• Whether infection has been excluded

Outputs

• Host-vulnerability score
• Exposure-pattern score
• Presentation-severity score
• Diagnostic-uncertainty score
• Total risk score
• Risk class: LOW / INTERMEDIATE / HIGH / VERY HIGH
• Recommended actions
• Safety alerts
• Explicit limitations

Why it matters

Methotrexate pneumonitis is uncommon but potentially severe. The bedside problem is not just recognizing that methotrexate can injure the lung. It is distinguishing when acute cough, fever, dyspnea, hypoxemia, and diffuse infiltrates should trigger urgent drug-toxicity concern instead of being dismissed as routine infection or background RA-associated lung disease.

Run

python3 mtx_pneumo.py

Demo scenarios

1. Stable RA patient on low-dose methotrexate without pulmonary symptoms
2. Older patient early after methotrexate start with cough and fever
3. Baseline ILD with acute dyspnea, hypoxemia, and diffuse infiltrates soon after start

References

1. Fragoulis GE, Nikiphorou E, Larsen J, Korsten P, Conway R. Methotrexate-Associated Pneumonitis and Rheumatoid Arthritis-Interstitial Lung Disease: Current Concepts for the Diagnosis and Treatment. Front Med (Lausanne). 2019;6:238. DOI: 10.3389/fmed.2019.00238
2. Conway R, Low C, Coughlan RJ, O'Donnell MJ, Carey JJ. Methotrexate and Lung Disease in Rheumatoid Arthritis: A Meta-Analysis of Randomized Controlled Trials. Arthritis Rheumatol. 2014;66(4):803-812. DOI: 10.1002/art.38322
3. Barrera P, Laan RF, van Riel PL, Dekhuijzen PN, Boerbooms AM, van de Putte LB. Methotrexate-related pulmonary complications in rheumatoid arthritis. Ann Rheum Dis. 1994;53(7):434-439. DOI: 10.1136/ard.53.7.434
4. Dawson JK, Graham DR, Desmond J, Fewins HE, Lynch MP. Investigation of the chronic pulmonary effects of low-dose oral methotrexate in patients with rheumatoid arthritis: a prospective study incorporating HRCT scanning and pulmonary function tests. Rheumatology (Oxford). 2002;41(3):262-267. DOI: 10.1093/rheumatology/41.3.262
5. Fraenkel L, Bathon JM, England BR, et al. 2021 American College of Rheumatology Guideline for the Treatment of Rheumatoid Arthritis. Arthritis Rheumatol. 2021;73(7):1108-1123. DOI: 10.1002/art.41752

Public HTML

https://rheumascore.xyz/skills/mtx-pneumo.html