Introduction
Distinguishing benign from malignant parotid gland tumors remains a clinical challenge. This retrospective study from a single tertiary care center investigates how diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) histogram analysis on neck MRI can aid in differentiating parotid lesions. By evaluating both conventional MRI features and ADC histogram parameters, the research aimed to identify reliable imaging markers that correlate with histopathology.
Study design and population
The analysis reviewed neck MRI studies performed between January 2019 and December 2023 for suspected parotid lesions. After applying inclusion criteria and histopathological confirmation, 100 parotid lesions from 90 patients (51 males, 39 females; age 20–83) were analyzed. Histology—via FNAB or surgical excision—served as the gold standard. Exclusions included poor image quality, incomplete imaging, lack of histopathology, or prior treatment. The cohort encompassed a mix of benign tumors, including pleomorphic adenomas (39%) and Warthin tumors (20%), and a substantial malignant group (41%), with acinic cell carcinoma, mucoepidermoid carcinoma, and adenoid cystic carcinoma among the subtypes.
Imaging protocol and analysis
All studies used a 1.5 Tesla MRI system with a dedicated head/neck coil. Sequences included T1-, T2-weighted images, and post-contrast T1-weighted imaging, along with diffusion-weighted imaging at b values of 0, 500, and 1000 s/mm² to generate ADC maps. Two experienced neuroradiologists, blinded to pathology, performed qualitative and quantitative assessments, then calculated mean ADC and histogram-derived metrics (min, max, mean, skewness, kurtosis) by manually outlining solid tumor portions on ADC maps.
ADC values for lesions were compared against histopathology. Interobserver reliability was evaluated, and ROC analysis determined optimal cutoffs to distinguish malignant from benign lesions. The study also tracked associations between imaging features (e.g., lesion margins, diffusion behavior, enhancement patterns) and tumor type.
Key diffusion findings
Benign parotid lesions demonstrated higher ADC values than malignant ones. Across observers, malignant lesions showed a marked decrease in ADC measurements, with mean and minimum/maximum ADCs significantly lower in cancer. The ADC histogram analysis provided additional discriminative power: mean ADC, minimum ADC, and maximum ADC all differed significantly between groups, with malignant tumors showing lower central tendency and altered distribution shapes.
ADC thresholds and diagnostic performance
Using ROC analysis, the following cutoffs emerged (Observer 1 as reference):
– ADC (single value) < 1.15 × 10⁻³ mm²/s yielded AUC 0.847, sensitivity 73.2%, specificity 86.4%.
– Mean ADC from histogram analysis < 1.175 × 10⁻³ mm²/s yielded AUC 0.793, sensitivity 75.6%, specificity 67.8%.
– Minimum ADC showed fair performance (AUC 0.720), Maximum ADC achieved higher sensitivity (80.5%) but lower specificity (59.3%), with AUC 0.712.
Overall, mean ADC was the most robust histogram parameter for differentiating malignant from benign parotid tumors in this dataset, with excellent interobserver agreement (ICC near 0.96 for ADC values).
Clinical implications
ADC histogram analysis complements conventional MRI features by providing quantitative, reproducible metrics that reflect cellular density and microstructural differences between tumor types. The study observed that malignant lesions more often displayed ill-defined margins, higher diffusion restriction, lymph node involvement, skin or bone invasion, and other aggressive features, all contributing to a higher likelihood of cancer. The high interobserver reliability underscores the potential of ADC-based metrics to standardize reporting and support preoperative decision-making.
Context within the literature
Comparable studies have reported consistent trends: malignant parotid tumors tend to have lower ADC values and distinct histogram characteristics. While some investigations note higher diagnostic performance for whole-tumor histogram measures or combined diffusion and morphologic assessments, the present findings align with a broader literature suggesting diffusion metrics are valuable in salivary gland tumor characterization. Differences in imaging parameters, patient cohorts, and analytic methods likely explain variations in exact cutoffs and AUCs across studies.
Conclusion
In this retrospective cohort, ADC and ADC histogram parameters demonstrated meaningful capacity to differentiate benign from malignant parotid gland tumors, with mean ADC emerging as a reliable discriminator and excellent interobserver agreement. When integrated with conventional MRI features and clinical assessment, ADC histogram analysis can enhance preoperative characterization, guide biopsy decisions, and inform surgical planning.
