Research Topics: Pulmonary Risk Stratification in Predicting PPCs

1. Validation and Comparison of Existing Risk Scoring Tools

• ARISCAT (Assess Respiratory Risk in Surgical Patients in Catalonia) - the most widely recommended multifactorial score (age, preoperative SpO2, recent respiratory infection, surgical site, operative duration, hemoglobin, emergency surgery). Its external validity varies across non-European populations, opening research gaps.
• LAS VEGAS Score vs. ARISCAT in cardiac and thoracic surgery (Siddiqui et al., 2024, PMID 38989237)
• NSQIP Pneumonia Risk Tool and NSQIP Respiratory Failure Risk Tool - derived from >180 U.S. hospitals, validated across large multicenter datasets
• Gupta Respiratory Failure Calculator - procedure-specific regression model
• Research question: Which score performs best across non-Western, elderly, obese, or immunocompromised surgical populations?

2. Systematic Review and Meta-Analysis of Prediction Models

• A 2025 meta-analysis in the British Journal of Anaesthesia (Huang et al., PMID 40473567) found discriminative ability and clinical impact of PPC prediction models remains inconsistent - pointing to a major research opportunity.
• Evaluation of model calibration, discrimination (AUROC), and clinical net benefit across studies.
• The 2022 Lancet Digital Health study (PMID 35750401) externally validated multiple prognostic models for major abdominal surgery patients internationally.
• Research question: What is the methodological quality (PROBAST criteria) of existing PPC prediction models, and how many are ready for clinical deployment?

3. Machine Learning and AI-Based Risk Prediction Models

• Explainable ML (SHAP analysis) applied to thoracoscopic surgery patients to identify the most important modifiable predictors (Wang et al., 2025, PMID 40688297)
• Automated ML (AutoML) pipelines for non-small cell lung cancer (NSCLC) patients undergoing thoracoscopic surgery (Qiu et al., 2025, PMID 41004509)
• Intraoperative respiratory dynamic features (real-time ventilator data streams) as ML model inputs (Li et al., 2024, PMID 38637267)
• Deep neural network models in geriatric patients (multicenter prospective, PMID 36017511)
• ML models after neurosurgery - a relatively understudied surgical subgroup (Xu et al., 2025, PMID 39947880)
• Research question: Can real-time intraoperative AI models outperform preoperative scoring tools in predicting PPCs?

4. Pulmonary Function Testing (PFT) as a Stratification Tool

• A 2022 systematic review (Dankert et al., PMID 34939921) found PFTs add limited incremental value for noncardiothoracic surgery, but remain indispensable before lung resection.
• Key thresholds: FEV1 <2 L, MVV <50% predicted, PCO2 ≥45 mmHg, PO2 ≤50 mmHg (Harrison's Principles of Internal Medicine, 22e)
• Research question: What PFT thresholds or composite indices optimally stratify risk for specific surgical subgroups (thoracic, upper abdominal, spine)?

5. New Prediction Models for Lung Resection Candidates

• Novel models incorporating cardiopulmonary exercise testing (CPET) variables, VO2 max, and ventilatory equivalents (Svoboda et al., 2024, ERJ Open Research, PMID 39286055)
• Integration of predicted postoperative (ppo) lung function with functional capacity measures
• Research question: What combination of preoperative assessments (spirometry + CPET + imaging) provides the most accurate PPC risk estimate for thoracic oncology patients?

6. Operative and Procedure-Specific Risk Stratification

• High-risk procedures: aortic surgery, esophagectomy, thoracotomy, upper abdominal surgery - procedure-related risk outweighs patient-related risk (Murray & Nadel's Textbook of Respiratory Medicine)
• Minimally invasive (VATS/RATS) vs. open thoracotomy - differential PPC rates and risk model applicability
• Upper abdominal surgery (Garg et al., 2025, PMID 38526466) - spirometry and clinical predictors
• Research question: Should procedure-specific PPC risk models replace universal scores for high-risk surgeries?

7. Patient-Specific Risk Factor Research

• Obstructive Sleep Apnea (OSA): STOP-BANG screening and its predictive role for PPCs and postoperative desaturation
• Frailty and functional dependence in elderly patients - frailty indices vs. ASA class
• Obesity and BMI - controversial predictor; uncomplicated obesity may not substantially increase PPC risk (Fishman's Pulmonary Diseases)
• Malnutrition (serum albumin <3.5 g/dL) as an independent risk factor
• Preoperative anemia (Hb <10 g/dL) - included in ARISCAT
• Research question: How does multimorbidity clustering (e.g., COPD + OSA + low albumin) compound PPC risk beyond individual factor scoring?

8. Lung Ultrasound as a Perioperative Risk Stratification Tool

• Lung Ultrasound Score (LUS) for predicting PPCs - an emerging and debated topic (Vetrugno et al., 2026, European Journal of Anaesthesiology, PMID 41340139)
• Pre- and postoperative LUS for early detection of atelectasis, consolidation, and pleural effusion
• Research question: What is the additive value of preoperative LUS over clinical risk scores in stratifying PPC risk?

9. Risk Stratification in Specific Patient Populations

• Geriatric patients: Deep neural network models (PMID 36017511); ASA class and functional dependence are stronger predictors than comorbidities alone
• Lung cancer surgery patients: Explainable ML incorporating tumor characteristics and staging (Chen et al., 2025, PMID 40791887)
• Cardiac surgery patients: Nomograms post-CABG (Khanna et al., 2023, PMID 34689983)
• Neurosurgery patients: Multicenter ML model development (PMID 39947880)
• Urological surgery: ARISCAT validation in major urological procedures (PMID 40948164)
• Esophageal cancer surgery: Unique PPC profile given one-lung ventilation and thoracoabdominal access (Duff et al., 2022, PMID 35393612)
• Research question: Are universal PPC scoring tools valid in surgical subspecialties, or should population-specific models be developed and mandated?

10. Intraoperative Risk Modification and Its Interaction with Preoperative Stratification

• Lung-protective ventilation (LPV): Low tidal volume + PEEP titration - meta-analysis showing PEEP 5 cmH2O + plateau pressure ≤16 cmH2O reduces PPC risk (Miller's Anesthesia 10e)
• Driving pressure-guided ventilation - 2024 meta-analysis (PMID 38937217) supports it as a modifier of PPC risk
• Neuromuscular blockade reversal with sugammadex vs. neostigmine - meta-analysis showing superiority of sugammadex in PPC prevention (PMID 37027443)
• Neuraxial vs. general anesthesia - epidural and spinal techniques reduce PPCs in abdominal surgery
• Research question: Do preoperative risk scores retain their discriminative ability when patients receive standardized lung-protective intraoperative protocols?

11. Standardization of PPC Definitions and Outcome Measurement

• The StEP-COMPAC (Standardized Endpoints in Perioperative Medicine) and EPCO (European Perioperative Clinical Outcome) definitions as reference frameworks
• Heterogeneity in composite PPC endpoints vs. individual outcome reporting (pneumonia, atelectasis, ARDS, respiratory failure, PE, pneumothorax, reintubation)
• Research question: How do different PPC outcome definitions affect the apparent performance of risk stratification tools?

12. Preoperative Optimization Strategies Guided by Risk Stratification

• Inspiratory muscle training (IMT) in high-risk patients pre-surgery
• Prehabilitation programs combining exercise, nutrition, and smoking cessation
• Oral hygiene protocols and decolonization to reduce postoperative pneumonia
• Research question: Does individualized preoperative optimization based on PPC risk stratification reduce actual complication rates vs. standard care?

Summary Table of Key Research Areas

• Huang et al. (2025) - Prediction models for PPCs: systematic review & meta-analysis. Br J Anaesth [PMID 40473567]
• LePledge et al. / Lancet Digital Health (2022) - External validation of prognostic models for abdominal surgery [PMID 35750401]
• Li et al. (2024) - Intraoperative respiratory dynamics + explainable ML [PMID 38637267]
• Dankert et al. (2022) - PFTs for PPC prediction: systematic review [PMID 34939921]
• Gu et al. (2024) - Driving pressure-guided ventilation meta-analysis [PMID 38937217]

Preoperative Tests for Predicting PPCs (Excluding CPET)

1. Pulse Oximetry (SpO2)

2. Spirometry / Pulmonary Function Tests (PFTs)

• Lung resection (thoracic surgery): Spirometry and DLCO are mandatory and well-validated. Key thresholds before resection:
  • FEV1 <2 L or <80% predicted - elevated risk
  • MVV <50% predicted
  • ppoFEV1 <40% - high risk for resection
  • DLCO <40% predicted - contraindication to pneumonectomy
• Noncardiothoracic surgery: Spirometry adds limited incremental value beyond clinical history and physical exam. A 2022 systematic review (Dankert et al., PMID 34939921) concluded that evidence for PFTs in non-thoracic surgery remains weak, though subgroup analysis suggests benefit in upper abdominal surgery.

• All patients undergoing lung resection
• Suspected but undiagnosed COPD
• Known COPD/asthma where baseline function cannot be assessed clinically
• Patients with symptoms disproportionate to their known disease

• Asymptomatic patients
• Cardiac surgery without new/worsened cardiopulmonary symptoms ("Choosing Wisely" recommendation)
• General elective surgery without pulmonary comorbidities

3. Chest Radiograph (CXR)

• History of known cardiopulmonary disease
• Age ≥50 undergoing upper abdominal, thoracic, or aortic surgery
• Severe obesity (BMI >40 kg/m2) - to screen for undiagnosed CHF, pulmonary HTN, or cardiomegaly
• Clinical suspicion of infection or bullous emphysema
• Abnormal findings on physical exam

4. Arterial Blood Gas (ABG)

• PaO2 <60 mmHg on room air - increased PPC risk
• PaCO2 ≥45 mmHg - associated with increased perioperative risk, especially for thoracic surgery (Cummings Otolaryngology; Maingot's Abdominal Operations)
• Elevated serum HCO3- in patients with suspected OSA - suggests chronic CO2 retention and possible obesity hypoventilation syndrome (OHS), which substantially increases risk

• Lung resection: Provides useful baseline for patients with chronic CO2 retention.
• Noncardiothoracic surgery: Limited value - hypercapnia is usually clinically apparent and ABG rarely alters management (Murray & Nadel's). A 2025 study (IJCMAAS) found ABG was a poor predictor in elective abdominal surgery (AUC 0.611, non-significant).
• Suspected OSA: Elevated serum bicarbonate on routine labs is a reasonable surrogate; if elevated, consider full ABG.

• Thoracic surgery and lung resection planning
• New or worsened respiratory symptoms requiring workup
• Suspected OHS or severe COPD (to guide postoperative ventilation planning)

5. Serum Albumin

• Albumin <3.5 g/dL: OR 2.53 for pulmonary complications (Miller's Anesthesia 10e)
• Included as a risk factor in the ACP 2006 guidelines and Arozullah respiratory failure index
• Hypoalbuminemia reflects poor nutritional reserve, impaired respiratory muscle strength, and reduced immune defense against pneumonia

6. Blood Urea Nitrogen (BUN)

• Elevated BUN >7.5 mmol/L (>21 mg/dL): OR 4.81 for pulmonary complications (Miller's Anesthesia 10e)
• Included in some PPC risk models as a marker of end-organ dysfunction
• Additive value above clinical assessment is uncertain (Murray & Nadel's)

7. Hemoglobin / Complete Blood Count (CBC)

• Hemoglobin ≤10 g/dL is one of the 7 variables in the ARISCAT score (11 risk points; OR 3.0 for PPCs)
• Anemia impairs oxygen-carrying capacity and reduces physiologic reserve, predisposing to hypoxemia after surgery
• Preoperative WBC and neutrophil ratio are emerging ML model predictors in thoracoscopic surgery populations

8. Serum Bicarbonate (HCO3-)

9. Lung Ultrasound (LUS)

10. Ventilation-Perfusion (V/Q) Scan and Quantitative CT

• Quantitative CT volumetry can calculate predicted postoperative lung function (ppoFEV1, ppoDLCO) based on regional lung contribution
• V/Q scintigraphy (SPECT): Measures fractional contribution of each lung segment; used when CT is not available or when regional assessment is needed
• Both are superior to simple spirometric thresholds for estimating functional reserve after resection

Summary: When to Use Which Test