The following is a summary of “Mediastinal staging lymph node probability map in non-small cell lung cancer,” published in the March 2025 issue of Respiratory Research by Bordas-Martinez et al.
Accurate staging of mediastinal lymph nodes (LNs) is critical for guiding the management of non-small cell lung carcinoma. PET/CT and EBUS-TBNA are standard diagnostic modalities, and predictive algorithms have been developed for each technique individually. However, an optimized approach that integrates these diagnostic modalities along with clinical data could enhance the accuracy of malignancy prediction. This study aimed to develop and validate a predictive algorithm that combines PET/CT findings, EBUS features, and patient clinical characteristics to estimate the probability of malignancy in mediastinal LNs. A retrospective analysis was performed on consecutive patients with NSCLC staged using PET/CT and EBUS-TBNA.
LNs were categorized by nodal level (N1, N2, N3) and anatomical region (AR), including subcarinal, paratracheal, and hilar nodes. Standardized Uptake Value maximum (SUVmax) was determined for each sampled LN, while EBUS-derived parameters included short-axis diameter, morphology, border characteristics, echogenicity, and presence of a vascular hilum. A logistic regression model incorporating age, DSA, SUVmax, and AR was developed to estimate malignancy probability. The study included 116 patients with a mean age of 66, of whom 93% were male. A total of 358 LNs were analyzed, with malignancy confirmed in 51% of adenocarcinoma cases, 35% of squamous cell carcinoma cases, and 14% of unspecified NSCLC cases. The predictive model demonstrated strong diagnostic performance, achieving an area under the receiver operating characteristic (ROC) curve of 0.89, indicating high discriminative ability.
Additionally, a user-friendly application was developed to facilitate the implementation of the algorithm in clinical practice. These findings suggest that integrating clinical parameters with PET/CT and EBUS features can generate a malignancy probability map before LN sampling, potentially improving diagnostic efficiency and patient management. However, further prospective studies and external validation are necessary to confirm the model’s clinical applicability and robustness.
Source: respiratory-research.biomedcentral.com/articles/10.1186/s12931-025-03121-z
The post Integrated Model for Predicting Mediastinal Lymph Node Malignancy first appeared on Physician's Weekly.
