Clinical and demographic data improve diagnostic accuracy of dynamic contrast-enhanced and diffusion-weighted MRI in differential diagnostics of parotid gland tumors

Deep learning-assisted diagnosis of benign and malignant parotid tumors based on ultrasound: a retrospective study

BACKGROUND

To develop a deep learning(DL) model utilizing ultrasound images, and evaluate its efficacy in distinguishing between benign and malignant parotid tumors (PTs), as well as its practicality in assisting clinicians with accurate diagnosis.

METHODS

A total of 2211 ultrasound images of 980 pathologically confirmed PTs (Training set: n = 721; Validation set: n = 82; Internal-test set: n = 89; External-test set: n = 88) from 907 patients were retrospectively included in this study. The optimal model was selected and the diagnostic performance evaluation is conducted by utilizing the area under curve (AUC) of the receiver-operating characteristic(ROC) based on five different DL networks constructed at varying depths. Furthermore, a comparison of different seniority radiologists was made in the presence of the optimal auxiliary diagnosis model. Additionally, the diagnostic confusion matrix of the optimal model was calculated, and an analysis and summary of misjudged cases' characteristics were conducted.

RESULTS

The Resnet18 demonstrated superior diagnostic performance, with an AUC value of 0.947, accuracy of 88.5%, sensitivity of 78.2%, and specificity of 92.7% in internal-test set, and with an AUC value of 0.925, accuracy of 89.8%, sensitivity of 83.3%, and specificity of 90.6% in external-test set. The PTs were subjectively assessed twice by six radiologists, both with and without the assisted of the model. With the assisted of the model, both junior and senior radiologists demonstrated enhanced diagnostic performance. In the internal-test set, there was an increase in AUC values by 0.062 and 0.082 for junior radiologists respectively, while senior radiologists experienced an improvement of 0.066 and 0.106 in their respective AUC values.

CONCLUSIONS

The DL model based on ultrasound images demonstrates exceptional capability in distinguishing between benign and malignant PTs, thereby assisting radiologists of varying expertise levels to achieve heightened diagnostic performance, and serve as a noninvasive imaging adjunct diagnostic method for clinical purposes.

Amide proton transfer-weighted magnetic resonance imaging for the differentiation of parotid gland tumors

Purpose

To assess the usefulness of amide proton transfer-weighted (APTw) imaging in the differentiation of parotid gland tumors.

Materials and methods

Patients with parotid gland tumors who underwent APTw imaging were retrospectively enrolled and divided into groups according to pathology. Two radiologists evaluated the APTw image quality independently, and APTw images with quality score ≥3 were enrolled. The maximum and average values of APTw imaging for tumor lesions (APTmax and APTmean) were measured. The differences in APTmax and APTmean were compared between malignant tumors (MTs) and benign tumors (BTs), as well as between MTs and pleomorphic adenomas (PAs) and between MTs and Warthin tumors (WTs). Independent-samples t-test, Kruskal-Wallis H test, and receiver operating characteristic (ROC) curve analyses were used for statistical analysis.

Results

Seventy-three patients were included for image quality evaluation. In this study, 32/73 and 29/73 parotid tumors were scored as 4 and 3, respectively. After excluding lesions with quality score ≤2 (12/73), the APTmean and APTmax of MTs were 4.15% ± 1.33% and 7.43% ± 1.61%, higher than those of BTs 2.74% ± 1.04% and 5.25% ± 1.54%, respectively (p < 0.05). The areas under the ROC curve (AUCs) of the APTmean and APTmax for differentiation between MTs and BTs were 0.819 and 0.821, respectively. MTs indicated significantly higher APTmean and APTmax values than those of PAs (p < 0.05) and WTs (p < 0.05). The AUCs of the APTmean and APTmax for differentiation between MTs and PAs were 0.830 and 0.815 and between MTs and WTs were 0.847 and 0.920, respectively.

Conclusion

Most APTw images for parotid tumors had acceptable image quality for APTw value evaluation. Both APTmax and APTmean can be used to differentiate MTs from BTs and to differentiate MTs from subtype parotid gland tumors.

Feasibility of Intravoxel Incoherent Motion (IVIM) and Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) in Differentiation of Benign Parotid Gland Tumors

Aim: The aim of this prospective study is to identify quantitative intravoxel incoherent motion and dynamic contrast-enhanced magnetic resonance imaging parameters of the most frequent benign parotid tumors, compare their utility and diagnostic accuracy. Methods: The study group consisted of 52 patients with 64 histopathologically confirmed parotid focal lesions. Parametric maps representing apparent diffusion coefficient (ADC), pure diffusion coefficient (D), pseudo-diffusion coefficient (D*), perfusion fraction (FP) and transfer constant (Ktrans), reflux constant (Kep), extra-vascular extra-cellular volume fraction (Ve), and initial area under curve in 60 s (iAUC) have been obtained from multiparametric MRI. Results: Statistically significant (p < 0.001) inter-group differences were found between pleomorphic adenomas (PA) and Warthin tumors (WT) in all tested parameters but iAUC. Receiver operating characteristic curves were constructed to determine the optimal cut-off levels of the most significant parameters allowing differentiation between WT and PA. The Area Under the Curve (AUC) values and thresholds were for ADC: 0.931 and 1.05, D: 0.896 and 0.9, Kep: 0.964 and 1.1 and Ve: 0.939 and 0.299, respectively. Lesions presenting with a combination of ADC, D, and Ve values superior to the cut-off and Kep values inferior to the cut-off are classified as pleomorphic adenomas. Lesions presenting with combination of ADC, D, and Ve values inferior to the cut-off and Kep values superior to the cut-off are classified as Warthin tumors. Conclusions: DWI, IVIM and quantitative analysis of DCE-MRI derived parameters demonstrated distinctive features of PAs and WT and as such they seem feasible in differentiation of benign parotid gland tumors.

Histogram analysis of dynamic contrast-enhanced magnetic resonance imaging in the differential diagnosis of parotid tumors

Objective

Dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) histograms were used to investigate whether their parameters can distinguish between benign and malignant parotid gland tumors and further differentiate tumor subgroups.

Materials and methods

A total of 117 patients (32 malignant and 85 benign) who had undergone DCE-MRI for pretreatment evaluation were retrospectively included. Histogram parameters including mean, median, entropy, skewness, kurtosis and 10th, 90th percentiles were calculated from time to peak (TTP) (s), wash in rate (WIR) (l/s), wash out rate (WOR) (l/s), and maximum relative enhancement (MRE) (%) mono-exponential models. The Mann–Whitney U test was used to compare the differences between the benign and malignant groups. The diagnostic value of each significant parameter was determined on Receiver operating characteristic (ROC) analysis. Multivariate stepwise logistic regression analysis was used to identify the independent predictors of the different tumor groups.

Results

For both the benign and malignant groups and the comparisons among the subgroups, the parameters of TTP and MRE showed better performance among the various parameters. WOR can be used as an indicator to distinguish Warthin’s tumors from other tumors. Warthin’s tumors showed significantly lower values on 10th MRE and significantly higher values on skewness TTP and 10th WOR, and the combination of 10th MRE, skewness TTP and 10th WOR showed optimal diagnostic performance (AUC, 0.971) and provided 93.12% sensitivity and 96.70% specificity. After Warthin’s tumors were removed from among the benign tumors, malignant parotid tumors showed significantly lower values on the 10th TTP (AUC, 0.847; sensitivity 90.62%; specificity 69.09%; P < 0.05) and higher values on skewness MRE (AUC, 0.777; sensitivity 71.87%; specificity 76.36%; P < 0.05).

Conclusion

DCE-MRI histogram parameters, especially TTP and MRE parameters, show promise as effective indicators for identifying and classifying parotid tumors. Entropy TTP and kurtosis MRE were found to be independent differentiating variables for malignant parotid gland tumors. The 10th WOR can be used as an indicator to distinguish Warthin’s tumors from other tumors.

A nomogram based on clinical information, conventional ultrasound and radiomics improves prediction of malignant parotid gland lesions

Although conventional ultrasound (CUS) allows for clear detection of parotid gland lesions (PGLs), it fails to accurately provide benign-malignant differentiation due to overlapping morphological features. Radiomics is capable of processing large-quantity volume of data hidden in CUS image undiscovered by naked eyes. The aim was to explore the potential of CUS-based radiomics score (Rad-score) in distinguishing benign (BPGLs) and malignant PGLs (MPGLs). A consecutive of 281 PGLs (197 in training set and 84 in test set) with definite pathological confirmation was retrospectively enrolled. 1465 radiomics features were extracted from CUS images and Rad-score was constructed by using Least Absolute Shrinkage and Selection Operator (LASSO) algorithm. Different nomogram models, including clinic-radiomics (Clin + Rad-score), CUS-clinic (CUS + Clin) and combined CUS-clinic-radiomics (CUS + Clin + Rad-score), were built using logistic regression. The diagnostic performance of different models were calculated and compared by area under receiver operating curve (AUC) and corresponding sensitivity and specificity. Finally, 26 radiomics features were independent signatures for predicting MPGLs, with MPGLs having higher Rad-scores in both cohorts (both P < 0.05). In the test population, CUS + Clin + Rad-score obtained an excellent diagnostic result, with significantly higher AUC value (AUC = 0.91) when compared to that of CUS + Clin (AUC = 0.84) and Clin + Rad-score (AUC = 0.74), respectively (both P < 0.05). In addition, the sensitivity of this combined model was higher than that of single Rad-score model (100.00% vs. 71.43%, P = 0.031) without compromising the specificity value (82.86% vs. 88.57%, P = 0.334). The calibration curve and decision curve analysis also indicated the clinical effectiveness of the proposed combined nomogram. The combined CUS-clinic-radiomics model may help improve the discrimination of BPGLs from MPGLs.

Quantitative Analysis and Pathological Basis of Signal Intensity on T2-Weighted MR Images in Benign and Malignant Parotid Tumors

Objective

To investigate the value of the signal intensity on T2-weighted magnetic resonance (MR) imaging using quantitative analysis in the differentiation of parotid tumors.

Materials and Methods

MR data of 80 pleomorphic adenomas (PAs), 68 Warthin tumors (WTs), and 34 malignant tumors (MTs) confirmed by surgery and histology were retrospectively analyzed. The signal intensities of tumor, normal parotid gland, spinal cord, and buccal subcutaneous fat were measured, and the signal intensity ratios (SIRs) between the tumor and the three references were calculated. Receiver operating characteristic curve was used to determine the optimal threshold and diagnostic efficiency of SIR for differentiating PAs, WTs, and MTs.

Results

The area under the curve (AUC) of tumor to parotid gland SIR (SIR_P), tumor to spinal cord SIR (SIR_C), and tumor to buccal subcutaneous fat SIR (SIR_F) for differentiating PAs and WTs was 0.922, 0.918, and 0.934, respectively. The sensitivity and specificity at an optimal SIR threshold were 86.3% and 91.2%, 80.0% and 97.1%, and 85.0% and 94.1%, respectively. The AUC of SIR_P, SIR_C, and SIR_F for distinguishing PAs from MTs was 0.793, 0.802, and 0.774, respectively. The sensitivity and specificity at an optimal SIR threshold was 86.3% and 61.8%, 80.0% and 73.5%, and 82.5% and 73.5%, respectively. The AUC of SIR_P, SIR_C, and SIR_F for distinguishing WTs from MTs was 0.716, 0.709, and 0.759, respectively. The sensitivity and specificity at an optimal SIR threshold were 61.8% and 82.4%, 55.9% and 82.4%, and 64.7% and 86.8%, respectively.

Conclusion

SIR_P, SIR_C, and SIR_F on T2-weighted MR images had high diagnostic efficiency for differentiating between PAs and WTs, while SIR_P and SIR_C for differentiating between PAs and MTs, and SIR_F for differentiating between WTs and MTs had relatively high diagnostic efficiency.