Multimodal magnetic resonance imaging for the diagnosis of parotid gland malignancies: systematic review and meta-analysis

Introduction

The pathological types of parotid gland tumors (PGTs) are complex, and the clinical manifestations lack specificity (1). Some benign tumors are prone to relapse and malignancy, and some low-grade malignant tumors may have similar clinical characteristics to benign tumors. Some low-grade malignancies may have clinical features similar to those of benign tumors. Imaging modalities, such as MRI and computed tomography, can be used to identify the location and size of PGT. Fine needle biopsy is the primary method for identifying tumor types, but its sensitivity in identifying malignant PGT is low (70–80%) (2,3). Therefore, accurate preoperative diagnosis is of great significance to the formulation of a surgical plan (4).

Magnetic resonance imaging (MRI) has good tissue contrast resolution and can clearly display both parotid gland tumor itself and its surrounding tissues (5). At present, MRI has become It is an ideal diagnostic method for parotid gland tumor, including evaluation of tumor scope, local invasion, peripheral nerve diffusion, and distant lymph node metastasis. Apparent diffusion coefficient (ADC) values from diffusion-weighted imaging (DWI) are expected to identify parotid tumors. However, there is still a lot of controversy in this area and the results of different studies are inconsistent. There are large differences in the results, with a sensitivity ranging from 36% to 100% and a specificity ranging from 25% to 97% (6-8). Therefore, it is necessary to use meta method to analyze the significance of MRI multimodal diagnosis in parotid gland tumors. We present the following article based on the PRISMA reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-22-695/rc).

Methods

Search strategy

An independent search was conducted by two researchers for all studies comparing the recent outcomes of parotid gland tumors and multimodal MRI published up to 1 October, 2021. The databases searched included the Cochrane Library, PubMed, Web of Science, Embase and Chinese Biomedical Literature (CBM). Additionally, the references of identified articles were manually screened for potentially suitable studies. The search terms included “parotid gland”, “tumor”, “cancer”, “multimodal magnetic resonance imaging”, “MRI”, “MR”, et al. Diagnosis of parotid gland tumors by multimodal MRI, and guide the clinical treatment of parotid gland tumors. The flow chart is shown in Figure 1.

Figure 1 Flow chart of the literature screening.

Inclusion criteria

(I) Included articles had true-positive, false-negative, false-positive, and true-negative data for sensitivity and specificity indicators or indicators for which sensitivity and specificity could be calculated. (II) The age range of enrolled patients was 18-65 years old. (III) Articles published in English were accepted. (IV) Patients without other tumors.

Exclusion criteria

(I) Used animal subjects. (II) Were case reports, review articles, commentaries or meeting abstracts. (III) Literature that could not provide valid data for analysis.

Data extraction

All data extraction was completed independently by two evaluators. True positive, false negative, false positive, and true negative values of the ADC were evaluated for the unintentional study. To minimize selection bias, we selected the first reader or method from studies using multiple readers or methods. The first author, publication year, country, tumor/control, method and gold standard were also extracted.

Study quality assessment

Two reviewers independently used the Diagnostic Accuracy Research Quality Assessment Tool (QUADAS-2) to assess the risk of bias for each study. If necessary, the third party will make the final decision. Quality assessment is performed using Stata15 software (The Cochrane Collaboration, London, UK).

Statistical analysis

Diagnostic sensitivity and specificity were calculated using the formulas: sensitivity = true positive/(true positive + false negative); and specificity = true negative/(true negative + false positive). The area under the receiver operating characteristic (ROC) curve was used to evaluate the feasibility of MRI in the diagnosis accuracy of parotid gland tumors. Publication bias was evaluated using Deeks asymmetry funnel plot test, and a P value <0.05 was considered to indicate publication bias. Two-tailed P values of <0.05 were considered statistically significant. All statistical analysis was performed using Stata version 15.0.

Results

Basic characteristics of included studies

In this study, the databases of PubMed, the Cochrane Library, Web of Science, Embase, were searched. Repeated publications and studies were excluded by reading of titles and abstracts. A total of 1,879 relevant articles were retrieved and 5 articles published between 2016 and 2019, involving a total of 505 patients with parotid lesions (including 145 tumor patients and 360 control) (9-13). Different reports of the same clinical study and articles inconsistent with the content of this study were excluded, and references to relevant articles were investigated to prevent literature omission (Table 1).

Quality evaluation of the included studies

The 5 included studies (9-13), were of relatively good overall quality (based on QUADAS-2 criteria) in Figure 2.

Figure 2 Risk of bias and applicability concerns summary about each domain for each included study. -, low risk of bias; ?, unclear risk of bias; +, high risk of bias.

Pooled diagnostic sensitivity and specificity

The MRI-DWI combined sensitivity and specificity were 0.54 (0.22–0.83) and 0.93 (0.79–0.98), respectively, as shown in Figure 3. The MRI-DCE combined sensitivity and specificity were 0.81 (0.48–0.95) and 0.95 (0.92–0.97), respectively, as shown in Figure 4.

Figure 3 Forest plot: sensitivity and specificity of MRI-DWI. MRI-DWI, MRI with diffusion-weighted imaging; MRI, magnetic resonance imaging.

Figure 4 Forest plot: sensitivity and specificity of MRI-DCE. MRI-DCE, MRI with dynamic contrast-enhanced; MRI, magnetic resonance imaging.

Pooled receiver operating characteristic curves

The pooled area under the curve (AUC) of the MRI-DWI was 0.89 (95% CI: 0.86–0.91) as shown in Figure 5 and the pooled area under the curve (AUC) of the MRI-DCE was 0.96 (95% CI: 0.94–0.97) as shown in Figure 6.

Figure 5 The SROC curve of MRI-DWI diagnosis characteristics. SROC, summary receiver operating characteristic; AUC, area under the curve; MRI-DWI, MRI with diffusion-weighted imaging; MRI, magnetic resonance imaging.

Figure 6 The SROC curve of MRI-DCE diagnosis characteristics. SROC, summary receiver operating characteristic; AUC, area under the curve; MRI-DCE, MRI with dynamic contrast-enhanced; MRI, magnetic resonance imaging.

Publication analysis

The Deeks’ Funnel plot shows that there is no publication bias, and Figure 7 shows that there is no publication bias in diagnostic studies.

Figure 7 Deeks’ funnel plot asymmetry test. ESS: effect sizes.

Discussion

Salivary gland neoplasms are relatively rare, accounting for about 5% of all benign and malignant neoplasms, except for those of the skin (13). Among the different anatomical parts of salivary glands, parotid gland tumors have the highest incidence, with various types and complicated pathologic types. In terms of the frequency and location of parotid gland tumors, 80% of salivary gland tumors occur in the parotid gland, 80% are benign, and 80% are located in the superficial lobe. Surgical treatment is the preferred treatment for parotid gland tumors (14). Preoperative diagnosis of benign and malignant parotid gland tumors is critical to the selection of surgical methods. However, most parotid malignancies have clinical features similar to those of benign tumors, especially in the early stages, where they tend to grow slowly, replace rather than invade adjacent structures, and some malignancies may be mobile (15). Additionally, as a result of tumor cell diffusion and the risk of injury to structures such as the facial nerve, open biopsy is contraindicated in principle for parotid gland tumors. Although fine needle aspiration biopsy (FNA) is less invasive and relatively simple to perform, it carries the risk of tumor cell implantation, and the accuracy of FNA for small or deep tumors is controversial (16).

Therefore, multimodal MRI plays an important role in the preoperative differential diagnosis of parotid gland tumors. Clinicians should include specific diagnostic indicators, and high-accuracy and practical prediction models for benign and malignant diagnosis of parotid gland tumors (17).

At present, there is still a lack of unified guidelines for the assessment of benign and malignant parotid tumors before surgery. Most clinicians can diagnose them solely on personal clinical experience (18). This is more subjective and uncertain in empirical medicine, which may cause misdiagnosis and affect the treatment and prognosis of patients. Therefore, the mathematical prediction model of benign and malignant parotid tumors is being paid increasing attention by clinicians. It involves evidence-based medicine based on empirical medicine and is a new development and new form of clinical research. This prediction model can objectively and accurately judge the nature of parotid tumors (19-21).

The parotid gland is rich in saliva and adipose tissue, and is surrounded by a dense parotid masseter fascia, which is in sharp contrast to the surrounding muscle and bone tissue, enabling multimodal MRI to determine the location of the parotid gland tumor, morphology, boundary, and its relationship to adjacent important blood vessels. Due to its ability to produce MRI scans quickly, with high tissue and spatial resolution, multimodal preoperative MRI has gradually become the preferred means of assessing parotid gland tumor patients. There have been many studies on the analysis of benign and malignant parotid tumors by using multimodal MRI, but we did not locate any previous complete systematic evaluation of benign and malignant parotid tumors combined with multimodal MRI (22).

The systematic evaluation and meta-analysis had some limitations: (I) the inclusion of grey literature was not sufficient; (II) the random effects model was used for meta-analysis of some experimental results, without sensitivity analysis, thus reducing the reliability of the conclusions. There was statistical heterogeneity among the studies, which may be mainly attributed to the varying severity of the disease. In view of the clinical homogeneity of the included study, the random effects model was used for analysis. As the intervention measures and observation indicators of the articles included in this evaluation were different, all of them were low-quality studies, and all of them failed to describe the blind method and allocation hiding method in detail, sensitivity analysis was not performed; (III) the quality of the included literature was not high, a multi-center study had not been implemented, the efficacy indicators were not unified, and the follow-up data were not available, among other detracting features, which will have had a certain impact on the authenticity of the research conclusions.

Acknowledgments

Funding: The study was supported by Hainan Natural Science Foundation, Project No. 822RC824.

Footnote

Reporting Checklist: The authors have completed the PRISMA reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-22-695/rc

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-22-695/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Cai X, Zhu Q, Zeng Y, et al. Manganese Oxide Nanoparticles As MRI Contrast Agents In Tumor Multimodal Imaging And Therapy. Int J Nanomedicine 2019;14:8321-44. [Crossref] [PubMed]
2. Gudmundsson JK, Ajan A, Abtahi J. The accuracy of fine-needle aspiration cytology for diagnosis of parotid gland masses: a clinicopathological study of 114 patients. J Appl Oral Sci. 2016;24:561-7. [Crossref] [PubMed]
3. Schmidt RL, Hall BJ, Wilson AR, et al. A systematic review and meta-analysis of the diagnostic accuracy of fine-needle aspiration cytology for parotid gland lesions. Am J Clin Pathol 2011;136:45-59. [Crossref] [PubMed]
4. Biegger P, Ladd ME, Komljenovic D. Multifunctional Magnetic Resonance Imaging Probes. Recent Results Cancer Res 2020;216:189-226. [Crossref] [PubMed]
5. Orhan Soylemez UP, Atalay B. Differentiation of Benign and Malignant Parotid Gland Tumors with MRI and Diffusion Weighted Imaging. Medeni Med J 2021;36:138-45. [Crossref] [PubMed]
6. Dai YL, King AD. State of the art MRI in head and neck cancer. Clin Radiol 2018;73:45-59. [Crossref] [PubMed]
7. Abraham J. Imaging for head and neck cancer. Surg Oncol Clin N Am 2015;24:455-71. [Crossref] [PubMed]
8. Abdel Razek AAK, Mukherji SK. State-of-the-Art Imaging of Salivary Gland Tumors. Neuroimaging Clin N Am 2018;28:303-17. [Crossref] [PubMed]
9. Zhu L, Wang J, Shi H, et al. Multimodality fMRI with perfusion, diffusion-weighted MRI and 1 H-MRS in the diagnosis of lympho-associated benign and malignant lesions of the parotid gland. J Magn Reson Imaging 2019;49:423-32. [Crossref] [PubMed]
10. Tao X, Yang G, Wang P, et al. The value of combining conventional, diffusion-weighted and dynamic contrast-enhanced MR imaging for the diagnosis of parotid gland tumours. Dentomaxillofac Radiol 2017;46:20160434. [Crossref] [PubMed]
11. Yuan Y, Tang W, Tao X. Parotid gland lesions: separate and combined diagnostic value of conventional MRI, diffusion-weighted imaging and dynamic contrast-enhanced MRI. Br J Radiol 2016;89:20150912. [Crossref] [PubMed]
12. Elmokadem AH, Abdel Khalek AM, Abdel Wahab RM, et al. Diagnostic Accuracy of Multiparametric Magnetic Resonance Imaging for Differentiation Between Parotid Neoplasms. Can Assoc Radiol J 2019;70:264-72. [Crossref] [PubMed]
13. Zheng N, Li R, Liu W, et al. The diagnostic value of combining conventional, diffusion-weighted imaging and dynamic contrast-enhanced MRI for salivary gland tumors. Br J Radiol 2018;91:20170707. [Crossref] [PubMed]
14. Saku K, Tahara N, Takaseya T, et al. Multimodal imaging of cardiac-calcified amorphous tumor. J Nucl Cardiol 2020;27:682-5. [Crossref] [PubMed]
15. Choi YH, Yu JW, Kang MY, et al. Combination of multiparametric magnetic resonance imaging and transrectal ultrasound-guided prostate biopsies is not enough for identifying patients eligible for hemiablative focal therapy for prostate cancer. World J Urol 2019;37:2129-35. [Crossref] [PubMed]
16. Munshi A, Ganesh T, Gupta RK, et al. Perfusion magnetic resonance imaging in contouring of glioblastoma patients: Preliminary experience from a single institution. J Cancer Res Ther 2020;16:1488-94. [PubMed]
17. Kinnaird A, Sharma V, Chuang R, et al. Risk of Prostate Cancer after a Negative Magnetic Resonance Imaging Guided Biopsy. J Urol 2020;204:1180-6. [Crossref] [PubMed]
18. Huang X, Yang J, Li J, et al. Comparison of magnetic resonance imaging and 18-fludeoxyglucose positron emission tomography/computed tomography in the diagnostic accuracy of staging in patients with cholangiocarcinoma: A meta-analysis. Medicine (Baltimore) 2020;99:e20932. [Crossref] [PubMed]
19. Martino M, Fodor D, Fresilli D, et al. Narrative review of multiparametric ultrasound in parotid gland evaluation. Gland Surg 2020;9:2295-311. [Crossref] [PubMed]
20. Lin Y, Zhang K, Zhang R, et al. Magnetic nanoparticles applied in targeted therapy and magnetic resonance imaging: crucial preparation parameters, indispensable pre-treatments, updated research advancements and future perspectives. J Mater Chem B 2020;8:5973-91. [Crossref] [PubMed]
21. Karaman CZ, Tanyeri A, Özgür R, et al. Parotid gland tumors: comparison of conventional and diffusion-weighted MRI findings with histopathological results. Dentomaxillofac Radiol 2021;50:20200391. [Crossref] [PubMed]
22. Ding J, Wang W, Peng W, et al. MRI and CT imaging characteristics of myoepithelioma of the parotid gland. Acta Radiol 2016;57:837-43. [Crossref] [PubMed]

(English Language Editor: J. Jones)