A Nomogram to Predict Pneumothorax Requiring Chest Tube Placement following Percutaneous CT-guided Lung Biopsy

Masha Bondarenko, Jianxiang Zhang, Ulysis Baal, Brian Lam, Gunvant Chaudhari, Yoo Jin Lee, Jamie Schroeder, Maya Vella, Brian M. Haas, Thienkhai Vu, Jonathan Liu, Kimberly Kallianos, Shravan Sridhar, Brett M. Elicker, Jae Ho SohnmedRxiv 2024 · DOI

Abstract

Background: Pneumothorax requiring chest tube placement is a significant complication of CT-guided transthoracic lung biopsy, leading to increased hospital stays, clinical risks, and healthcare costs. Identifying high-risk patients pre-procedurally can improve patient selection and post-procedural care.

Purpose: To develop and externally validate a risk nomogram for pneumothorax requiring chest tube placement, utilizing a quantitative emphysema algorithm.

Materials & Methods: This two-center retrospective study (1994–2023) analyzed 2,512 patients. Emphysema was quantified using a 3x3x3 kernel algorithm and Hounsfield thresholding. A risk calculator was developed via logistic regression and forward variable selection, then validated using bootstrapping and Harrell’s C-index.

Results: Of the 2,512 patients, 157 (6.7%) required chest tubes. The final model identified three key predictors: age > 60 (OR 1.80), non-prone positioning (OR 2.48), and severe emphysema (OR 1.99). The nomogram achieved a C-index of 0.664 internally and 0.715 externally.

Conclusion: A simple three-factor nomogram effectively predicts the risk of severe pneumothorax, offering a practical tool for clinical risk stratification.

Introduction

CT-guided transthoracic lung biopsy is a cornerstone of modern diagnostic medicine, yet it carries a well-documented risk of complications. The most frequent is pneumothorax, occurring in approximately 20-30% of cases. While many instances are self-limiting, between 2% and 17% of biopsies result in a pneumothorax severe enough to require chest tube placement. This intervention necessitates hospitalization, increases the risk of hospital-acquired infections, and adds a substantial financial burden to the healthcare system.

Previous attempts to predict this risk have often relied on subjective radiologist assessments of emphysema or complex "black box" machine learning models that lack transparency and ease of use in a clinical setting. This study sought to create a simple, objective, and externally validated nomogram to identify high-risk patients before the procedure begins.

Methodology

Data Collection and Quantitative Analysis

The researchers conducted a retrospective study across two distinct institutions: University of California, San Francisco (UCSF), which served as the training set (n=1,899), and Zuckerberg San Francisco General Hospital (ZSFG), which served as the external validation set (n=613). This dual-center approach ensured the model was tested against different patient demographics and biopsy protocols.

To eliminate the variability of subjective assessments, the team employed a quantitative emphysema algorithm. This tool analyzes CT scans using 3x3x3 voxel kernels and Hounsfield Unit (HU) thresholding to score lung tissue destruction on a 3-point scale: None, Mild/Moderate, or Severe.

Model Development

The predictive model was built using logistic regression. To maximize clinical usability, the researchers performed "aggressive pruning" via forward variable selection based on the Akaike Information Criterion (AIC). This process stripped away dozens of potential variables to isolate the most impactful predictors, resulting in a streamlined tool that clinicians can use quickly without complex calculations.

Key Results

Incidence and Risk Factors

Across the total cohort of 2,512 patients, the incidence of pneumothorax requiring chest tube placement was 6.7%. The statistical analysis identified three primary independent risk factors:

  1. Non-prone Patient Position: Patients in supine or lateral positions faced the highest risk (Odds Ratio [OR] 2.48).
  2. Severe Emphysema: Quantitatively measured severe tissue destruction nearly doubled the risk (OR 1.99).
  3. Age Over 60: Older patients were significantly more susceptible to this complication (OR 1.80).

Model Performance and Validation

The resulting nomogram demonstrated strong performance characteristics:

  • Discrimination: The model achieved a Harrell’s C-index of 0.664 in the internal cohort and an even stronger 0.715 in the external validation cohort.
  • Calibration: The mean absolute error in calibration was only 0.5%, indicating that the predicted risks were highly consistent with observed outcomes.
  • Risk Stratification: The tool categorizes patients into Low (<4.5%), Moderate (4.5–7.9%), and High (>8%) risk groups. Notably, a patient presenting with all three risk factors has a 14.7% probability of requiring a chest tube.

Discussion

The Importance of Positioning

One of the most actionable insights from this study is the impact of patient positioning. The significantly higher risk associated with non-prone positioning suggests that, whenever clinically and technically feasible, placing the patient in a prone position (on their stomach) may serve as a simple, no-cost intervention to mitigate the risk of severe pneumothorax.

Objective vs. Subjective Assessment

By utilizing a quantitative algorithm for emphysema rather than relying on qualitative radiologist descriptions, the model bypasses the inter-observer variability that often plagues clinical risk scores. This objectivity contributes to the model's robust performance in external validation, suggesting it can be reliably applied across different hospital systems with varying equipment and protocols.

Clinical Utility

The simplicity of the three-factor model—age, position, and emphysema severity—allows for the creation of a transparent graphical nomogram and a web-based risk tool. This enables clinicians to have informed consent discussions with patients, consider alternative biopsy routes (such as bronchoscopy) for high-risk individuals, or prepare for more intensive post-procedural monitoring.

Limitations and Future Directions

As a retrospective study, the findings are limited by the historical nature of the data, where clinical thresholds for chest tube placement may have varied between individual practitioners. Furthermore, while the emphysema algorithm is objective, it represents a simplified version of lung analysis that may not account for the specific needle path through the parenchyma.

Future research should focus on prospective trials to determine if the implementation of this nomogram actually reduces complication rates or hospital stay durations. Additionally, integrating more granular data regarding the specific biopsy trajectory could further refine the model's predictive power.

Summary updated

Abstract

ABSTRACT Background Pneumothorax requiring chest tube after CT-guided transthoracic lung biopsy is one of the common complications, and the required hospital stay after chest tube placement represents an added clinical risk to patients and cost to the healthcare system. Identifying high-risk patients can prompt alternative biopsy modes and/or better preparation for more focused post-procedural care. Purpose To develop and externally validate a risk nomogram for pneumothorax requiring chest tube placement following CT-guided lung biopsy, leveraging quantitative emphysema algorithm. Materials &amp; Methods This two-center retrospective study included patients who underwent CT-guided lung biopsy from between 1994 and 2023. Data from one hospital was set aside for validation (n=613). Emphysema severity was quantified and categorized to 3-point scale using a previously published algorithm based on 3×3×3 kernels and Hounsfield thresholding, and a risk calculator was developed using forward variable selection and logistic regression. The model was validated using bootstrapping and Harrell’s C-index. Results 2,512 patients (mean age, 64.47 years +/-13.38 [standard deviation]; 1250 men) were evaluated, of whom 157 (6.7%) experienced pneumothorax complications requiring chest tube placement. After forward variable selection to reduce the covariates to maximize clinical usability, the risk score was developed using age over 60 (OR 1.80 [1.15-2.93]), non-prone patient position (OR 2.48 [1.63-3.75]), and severe emphysema (OR 1.99 [1.35-2.94]). The nomogram showed mean absolute error of 0.5% in calibration and Harrell’s C-index of 0.664 in discrimination in the internal cohort. Conclusion The developed nomogram predicts age over 60, non-prone position during biopsy, and severe emphysema to be most predictive of pneumothorax requiring chest tube placement following CT-guided lung biopsy.
BibTeX
@misc{bondarenko2024nomogram,
  author = {Masha Bondarenko and Jianxiang Zhang and Ulysis Baal and Brian Lam and Gunvant Chaudhari and Yoo Jin Lee and Jamie Schroeder and Maya Vella and Brian M. Haas and Thienkhai Vu and Jonathan Liu and Kimberly Kallianos and Shravan Sridhar and Brett M. Elicker and Jae Ho Sohn},
  title = {{A Nomogram to Predict Pneumothorax Requiring Chest Tube Placement following Percutaneous CT-guided Lung Biopsy}},
  year = {2024},
  doi = {10.1101/2024.02.01.24302030},
  howpublished = {medRxiv},
}
Figure