Introduction
Parotid gland tumors pose a diagnostic challenge, with benign lesions vastly outnumbering malignant ones. Diffusion-weighted imaging (DWI) and apparent diffusion coefficient (ADC) histogram analysis offer a noninvasive approach to characterize these tumors preoperatively. This article summarizes findings from a retrospective study that evaluated the diagnostic performance of ADC histogram parameters and their interobserver reliability in differentiating benign from malignant parotid gland tumors using 1.5 T MRI.
Study Design and Population
The investigation was a retrospective, single-center observational study approved by the institutional review board. Among 103 identified cases with parotid lesions, 90 patients contributing 100 parotid lesions met inclusion criteria after histopathological confirmation. Lesions were diagnosed via fine-needle aspiration biopsy or surgical excision. The mean patient age was mid-adulthood, with both genders represented. Histopathology served as the reference standard for distinguishing benign from malignant lesions.
MRI Protocol and Image Analysis
All examinations used a 1.5 Tesla MRI system with standard neck coils. The protocol included T1-weighted, T2-weighted, post-contrast T1-weighted imaging, and DWI with b values of 0, 500, and 1000 s/mm2. An ADC map was generated for quantitative analysis. Two experienced neuroradiologists, blinded to pathology, conducted separate qualitative assessments and then performed quantitative ADC analysis by drawing ROIs around the tumor’s solid component on the ADC map. ADC-derived histogram features analyzed included min, max, mean, and statistical descriptors such as skewness and kurtosis.
Interobserver agreement was evaluated, and ROC analysis was employed to determine optimal cutoffs for differentiating benign and malignant tumors. The study examined conventional MRI features in parallel to diffusion metrics to understand their combined diagnostic potential.
Key Imaging Findings
Benign parotid lesions predominantly included pleomorphic adenomas and Warthin tumors, while malignant lesions encompassed acinic cell carcinoma, mucoepidermoid carcinoma, adenoid cystic carcinoma, and others. Notably, malignant tumors more often exhibited restricted diffusion, presenting high signal on high b-value images and low ADC values, a pattern with strong statistical significance compared with benign lesions.
ADC metrics revealed that malignant tumors had significantly lower ADC values across multiple measures. For observer 1, benign lesions had a median ADCmean around 1.4 × 10^-3 mm2/s, while malignant lesions showed a median near 0.9 × 10^-3 mm2/s. Histogram analysis further highlighted differences: malignant lesions had lower mean ADC, lower min and max ADC values, and different distribution characteristics such as skewness indicating asymmetric diffusion. Interobserver agreement for ADC values was excellent (ICC up to 0.96), underscoring the robustness of diffusion metrics in clinical practice.
Diagnostic Performance of ADC Parameters
The study identified a key cut-off for identifying malignancy: an ADC mean value < 1.15–1.18 × 10^-3 mm2/s yielded robust discrimination. Specifically, using a single ADC value, the area under the ROC curve (AUC) reached about 0.84–0.86 with sensitivity in the mid-70s and specificity in the mid-80s, translating to strong overall accuracy. Mean ADC emerged as the most reliable histogram parameter, outperforming minimum and maximum ADC values in aggregate diagnostic performance. While maximum ADC offered higher sensitivity in some analyses, its lower specificity reduced overall accuracy.
Comparisons with prior literature showed general agreement: malignant parotid tumors tend to have lower ADC values and more restricted diffusion than benign lesions. Some studies reported higher accuracies using alternative methods or combining imaging modalities, but the present findings reinforce the value of routine ADC histogram analysis as a practical preoperative tool.
Clinical Implications and Limitations
ADC histogram analysis provides a noninvasive, reproducible means to aid in preoperative decision-making for parotid tumors. High interobserver reliability supports its integration into routine radiologic assessment. However, limitations include the study’s retrospective design, single-center setting, and 1.5 T scanner parameters, which may influence ADC values. Variability in ROI placement and histological heterogeneity among tumors also impact diffusion metrics. Future prospective, multicenter studies using standardized acquisition and analysis protocols are warranted to validate cutoff values and refine diagnostic algorithms.
Conclusion
Diffusion-weighted MRI–derived ADC histogram analysis demonstrates meaningful diagnostic performance in distinguishing benign from malignant parotid gland tumors, with mean ADC as the most reliable parameter. When combined with conventional MRI features, ADC metrics can enhance preoperative characterization, improve sampling strategy, and support targeted treatment planning.
