Introduction
Differentiating benign from malignant parotid gland tumors remains a clinical challenge. Advances in diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) histogram analysis offer a noninvasive tool to aid preoperative characterization. This retrospective single-center study assesses the diagnostic performance and interobserver reliability of conventional MRI features and ADC histogram metrics in distinguishing benign from malignant parotid gland masses, using histopathology as the gold standard.
Study design and population
The investigation reviewed neck MRI examinations performed between January 2019 and December 2023 for suspected parotid lesions. After applying inclusion and exclusion criteria, 90 patients with 100 parotid lesions were analyzed. Histopathology—obtained via fine-needle aspiration biopsy or surgical excision—served as the reference standard. Benign tumors included pleomorphic adenoma and Warthin tumor, while malignant lesions encompassed a spectrum of histologies, with acinic cell carcinoma and mucoepidermoid carcinoma among the most common.
MRI protocol and image analysis
All images were acquired on a 1.5 T MRI system with a dedicated neck coil. The protocol included T1- and T2-weighted sequences, post-contrast T1-weighted imaging, and DWI with multiple b-values (0, 500, 1000 s/mm²). ADC maps were computed from DWI data. Two experienced neuroradiologists, blinded to histopathology, independently analyzed qualitative features (lesion location, size, margins, signal characteristics, and diffusion appearance) and performed quantitative ADC measurements.
For ADC histogram analysis, ROIs were drawn on the tumor’s solid portion on a representative axial ADC map, avoiding necrotic or cystic areas. ADC metrics derived from the histogram included minimum (min), maximum (max), mean, standard deviation, skewness, and kurtosis. The study also explored the diagnostic performance of single-parameter ADC values (ADCmean, ADCmin, ADCmax) and histogram-derived metrics in distinguishing benign from malignant tumors.
Key findings: ADC values and histogram metrics
Benign parotid lesions exhibited higher ADC values than malignant ones. Specifically, malignant tumors showed significantly lower ADCmean, ADCmin, and ADCmax values (P < 0.001). The mean ADC was particularly informative, with malignant lesions displaying a median around 1.05–1.15 × 10^-3 mm²/s compared with approximately 1.4–1.5 × 10^-3 mm²/s in benign tumors. Interobserver agreement for ADC measurements was excellent (ICC around 0.96), underscoring the robustness of ADC histogram analysis across readers.
The diagnostic performance of ADC alone showed an area under the ROC curve (AUC) around 0.80–0.85 for identifying malignancy. A common cut-off near 1.15 × 10^-3 mm²/s yielded sensitivity in the mid-70s and specificity in the mid-80s for observer information, with overall accuracy near 75–81%. Histogram parameters provided complementary insight, though mean ADC generally emerged as the most reliable single discriminator among the tested metrics.
Clinical implications
ADC histogram analysis augments conventional MRI by offering quantitative data that reflect tumor cellularity and diffusion characteristics. The observed trend—lower ADC in malignant lesions and higher dispersion or skewness in certain histogram parameters—aligns with known diffusion patterns in salivary gland cancers. Incorporating ADC histogram analysis into preoperative planning can enhance radiologic confidence, inform biopsy decisions, and potentially influence surgical strategy, particularly when distinguishing pleomorphic adenomas from malignant transformations or higher-grade cancers.
Comparisons with existing literature
Findings corroborate prior work showing lower ADC values in malignant parotid tumors and high interobserver reliability for diffusion metrics. Some studies reported superior performance with whole-lesion histogram analyses or diffusion kurtosis imaging, while others highlighted the strong diagnostic value of mean ADC. Across diverse cohorts, ADC-based assessment consistently supports its role as a noninvasive discriminator in parotid gland oncology.
Limitations and future directions
As a retrospective, single-center study, results may reflect specific scanner parameters or patient populations. Future multicenter prospective trials with standardized imaging protocols and integration of radiomics/texture analysis could refine cut-off values and improve diagnostic accuracy. Combining ADC histogram features with other MRI biomarkers and clinical data may further enhance differentiation between benign and malignant parotid gland tumors.
Conclusion
ADC histogram analysis derived from routine neck MRI provides valuable, reproducible information for differentiating benign from malignant parotid gland tumors. With an excellent interobserver agreement and meaningful diagnostic performance, ADC metrics—especially mean ADC—should be considered as part of standard preoperative evaluation to support histopathology-guided management decisions.
