Diagnostic performance evaluation of integration dynamic contrast-enhanced magnetic resonance imaging with susceptibility-weighted imaging in differentiating of parotid gland masses

Background: The histological spectrum of parotid gland lesions is remarkably diverse, necessitating a broad differential diagnosis. This study aims to evaluate the diagnostic efficacy of combining dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) with susceptibility-weighted imaging (SWI) in distinguishing masses in the parotid gland.

Methods: A retrospective study was conducted on 104 consecutive recruitment patients with parotid gland masses diagnosed pathologically or clinically (67 males and 37 females; median age 51.5 years (IQR:26.8 years; range:18-85 years)), who underwent DCE-MRI and SWI between January 2018 and December 2023.Of these, 106 benign and 13 malignant lesions were histopathologically or clinically confirmed. A multivariate logistic regression model was used to integrate the time-signal intensity curve (TIC), degree of intratumoral susceptibility signal intensity (ITSS), and the largest diameter of veins (Dv-max) in the lesions, analyze these three parameters between benign and malignant groups, and construct a combined diagnostic model. Some tests and receiver operating characteristic curve analysis were performed for statistical analyses.

Results: There were significant differences in Dv-max, ITSS and types of TICs between benign and malignant groups. Establishing a median Dv-max of 1.330 mm as the cut-off value, the diagnostic performance was optimized (area under the curve [AUC], 0.6851; sensitivity, 61.56%; specificity, 83.02%). Regarding “ITSS leveling 0 or 1” as benign lesions and “ITSS leveling 2 or 3” as malignant lesions, the diagnostic performance was optimized (AUC, 0.6861; sensitivity, 53.85%; specificity, 82.08%). Categorizing “curve type Ⅰ or Ⅱ” as benign lesions and “curve type Ⅲ” as malignant lesions, the diagnostic performance was optimized (AUC, 0.6970; sensitivity, 76.92%; specificity, 70.75%). Integration of Dv-max, ITSS, and the type of TIC yielded superior diagnostic performance relative to pattern-only assessment, with statistically significant improvements in key metrics: AUC = 0.8581, sensitivity = 92.31%, and specificity = 68.87% (P < 0.05).

Conclusion: While the integration of DCE-MRI and SWI holds the potential to assist in differentiating between benign and malignant parotid gland masses, its diagnostic performance remains below but approaches the threshold typically required for routine clinical utility. These findings suggest that further optimization of imaging protocols, inclusion of additional quantitative parameters, or validation in larger, multi-center cohorts may be necessary to enhance the clinical value of this approach. Future studies should also focus on refining cut-off values for parameters to improve sensitivity and specificity.