CT-based radiomics shows variable accuracy for predicting occult lymph node metastasis in early-stage lung adenocarcinoma

This systematic review and meta-analysis evaluated the diagnostic accuracy of CT-based radiomics for predicting occult lymph node metastasis in patients with early-stage lung adenocarcinoma. The analysis included 10 studies comprising 6,349 patients, though specific study settings and comparator methods were not reported. The primary outcome was diagnostic accuracy for predicting occult lymph node metastasis.

In internal validation cohorts, pooled sensitivity was 0.85 and pooled specificity was 0.78, with an area under the summary receiver operating characteristic curve of 0.89. However, performance decreased in external validation cohorts, where pooled sensitivity was 0.72 and pooled specificity was 0.75, with an area under the curve of 0.80. These results indicate moderate to good diagnostic performance that appears less robust when applied to independent datasets.

Subgroup analyses of external validation cohorts suggested possible variation in diagnostic performance according to sample size, CT protocol, and segmentation method. Safety and tolerability data were not reported in the included studies. Key limitations include heterogeneity in methodology across studies and the absence of prospective validation in clinical practice settings.

While CT-based radiomics shows potential for non-invasive prediction of occult lymph node metastasis in early-stage lung adenocarcinoma, the observed performance variation between internal and external validation highlights methodological challenges. Further multicenter prospective studies with harmonized imaging protocols and radiomics pipelines are needed to confirm clinical applicability before this approach can be recommended for routine practice.