Tumor peripheral bronchial signature predicts spread through air spaces in resected invasive mucinous adenocarcinoma of the lung

Abstract

Background: While a higher incidence of spread through air spaces (STAS) has been reported in invasive mucinous adenocarcinoma (IMA) of lung, most studies have only focused on non-mucinous adenocarcinoma (ADC). In this study, tumor peripheral bronchial signature (TPBS) was defined as indicators to describe bronchial morphological changes and bronchial distribution characteristics around the tumor in the preoperative computed tomography (CT) images. The value of TPBS was examined in predicting STAS in patients with resected IMA.

Methods: In the training cohort, the least absolute shrinkage and selection operator (LASSO) method was adopted to identify TPBS that was most strongly associated with STAS, and a formula of TPBS score was constructed. Subsequently, variations in clinical characteristics, radiological features, and TPBS scores were analyzed and compared using both univariate and multivariate statistical methods. Receiver operation characteristic (ROC) curve, decision curve, and calibration curve were employed to evaluate the model's efficacy.

Results: The best predictors included clinical T classification, spiculated margin, combined pneumonia, consolidation-tumor ratio (CTR) and TPBS score. In addition, three different STAS prediction models were developed and the corresponding area under the curve (AUC) values were 0.904, 0.877 and 0.838, respectively. Calibration curve illustrated that the predicted probability value generated by the hybrid model aligned well with the actual STAS status. Moreover, decision curve analysis (DCA) suggested that the hybrid model provided superior clinical utility and application value compared to both the basic model and the TPBS model.

Conclusions: TPBS score was identified as an independent predicting factor for STAS status. The nomogram, utilizing radiological features and TPBS score, demonstrated a high level of diagnostic accuracy and efficiency in predicting the status of STAS.

Keywords: Spread through air spaces (STAS); invasive mucinous adenocarcinoma (IMA); prediction; tumor peripheral bronchial signature (TPBS).

Figure 1

The flow chart for patient selection. Hospital 1, The First Affiliated Hospital of Ningbo University; Hospital 2, Ningbo Medical Center Lihuili Hospital. CT, computed tomography; IMA, invasive mucinous adenocarcinoma; STAS, spread through air spaces.

Figure 2

TPBS of a 65-year-old male patient with IMA and a positive STAS status was shown in the CT image. (A-C) The CT image shows a solid nodule in the right middle lobe of the lung, with bronchus truncation sign (orange arrows), cavity/cystic airspace sign (red arrows), and air bronchogram sign (blue arrows). (D) The closest distance from tumor to the bronchiole was measured. (E) The number of bronchioles within the safety surgical margin was measured. The black circle represented the safety surgical margin (2 cm around the tumor). (F) The tumor distance from the mediastinum (black line I) and lung width (black line II) was measured. CT, computed tomography; IMA, invasive mucinous adenocarcinoma; STAS, spread through air spaces; TPBS, tumor peripheral bronchial signature.

Figure 3

Selection of tumor peripheral bronchial signature using LASSO model. (A) LASSO coefficient profiles of the tumor peripheral bronchial signature of STAS. A coefficient profile plot was produced against the log (lambda) sequence. (B) Selection of tuning parameter lambda in the LASSO regression using tenfold cross-validation via minimum criteria. Selection for overall survival with λ=0.0295. LASSO, least absolute shrinkage and selection operator; STAS, spread through air spaces.

Figure 4

The ROC curve analysis of basic, TPBS and hybrid models in two cohorts. (A) The ROC curve of basic model in the training cohort. (B) The ROC curve of TPBS model in the training cohort. (C) The ROC curve of hybrid model in the training cohort. (D) The ROC curve of basic model in the external validation cohort. (E) The ROC curve of TPBS model in the external validation cohort. (F) The ROC curve of hybrid model in the external validation cohort. Cut-off value in the AUC curve was described by (1−specificity, sensitivity). AUC was described by average value (95% CI). AUC, area under the curve; CI, confidence interval; ROC, receiver operation characteristic; TPBS, tumor peripheral bronchial signature.

Figure 5

The nomogram for the preoperative prediction of the STAS status based on radiological features in resected IMA. CTR, consolidation-tumor ratio; IMA, invasive mucinous adenocarcinomas; STAS, spread through air spaces; TPBS, tumor peripheral bronchial signature.

Figure 6

The calibration curves of the hybrid model in two cohorts. The training cohort (A) and (B) the external validation cohort.

Figure 7

The decision curves of basic, TPBS and hybrid models in two cohorts. The training cohort (A) and (B) the external validation cohort. TPBS, tumor peripheral bronchial signature.

Figure 8

Kaplan-Meier curves of 5-year DFS for TPBS model divided by the high- and low-risk groups. (A) Kaplan-Meier curve of 5-year DFS for TPBS model in the training cohort. (B) Kaplan-Meier curve of 5-year DFS for TPBS model in the external validation cohort. (C) Kaplan-Meier curve of 5-year DFS for TPBS model further stratified according to STAS status in the training cohort. (D) Kaplan-Meier curve of 5-year DFS for TPBS model further stratified according to STAS status in the external validation cohort. DFS, disease-free survival; TPBS, tumor peripheral bronchial signature.