Exploratory Analysis of Quantitative CT Metrics for Predicting Tumor Aggressiveness and Nodal Metastasis in Head and...

Preoperative assessment of Head and Neck Squamous Cell Carcinoma (HNSCC) aggressiveness is often hindered by the sampling errors of incisional biopsies. While Contrast-Enhanced Computed Tomography (CECT) is the standard for staging, its potential to serve as a non-invasive complementary radiological tool of the entire tumor volume remains underutilized.

To evaluate the predictive performance of preoperative CECT-derived tumor volume, densitometric values, and morphological features as predictors of histopathological grade and lymph node metastasis (pN) in HNSCC. The primary outcome was predicting lymph node metastasis (pN+), and the secondary outcome was predicting histopathological grade.

This retrospective observational study analyzed 42 patients with SCC of the oral cavity, larynx, or maxilla. Quantitative (3D volume, Hounsfield Units [HU], HU Delta) and qualitative (margins, lobulations, necrosis) CT parameters were correlated with definitive histopathology. Diagnostic performance was assessed using Receiver Operating Characteristic (ROC) curve analysis and Spearman's rank correlation.

High-grade tumors (G2/G3) demonstrated significantly larger median volumes (18.1 vs. 2.9 cm³, p = 0.006), lower contrast density (55 vs. 68 HU, p = 0.010), and reduced vascular wash-in (23 vs. 30 HU Delta, p = 0.008) compared to G1 lesions. ROC analysis identified a volume threshold of ≥9.43 cm³ for high-grade disease (AUC = 0.865; sensitivity 67.6%, specificity 100%). For regional metastasis (pN+), tumor volume was the only significant predictor (25.4 vs. 6.2 cm³, p = 0.036), with an optimal cut-off of ≥6.76 cm³ (AUC = 0.769; sensitivity 100%). Strong negative correlations were observed between contrast enhancement and internal necrosis (r = -0.812, p < 0.001).

Preoperative CECT parameters show promise as non-invasive imaging surrogates of HNSCC aggressiveness. A paradoxical reduction in contrast enhancement characterizes high-grade biology, reflecting disorganized neo-angiogenesis and internal hypoxia. Integrating 3D volumetric analysis and morphological markers shows potential as a complementary exploratory tool that, pending future prospective validation, may support risk stratification and surgical planning alongside traditional histopathological assessment.