A preliminary diagnostic accuracy study of quantitative MRI biomarkers for differentiating parotid tumor types

Introduction

It is important to differentiate between types of parotid tumors since this can influence patient management. While fine needle aspiration is accurate in identifying malignancy in parotid gland lesions, diagnostic imaging also plays a role in assessment (1). Ultrasound is a fast, non-invasive, topical imaging modality that lacks ionizing radiation, but it does not possess sufficient sensitivity or specificity in this setting and may miss deeper lesions (2). Computed tomography (CT) is also fast and may detect deeper parotid lesions but radiates the patient’s head and neck and lacks the contrast resolution to delineate infiltrating masses. Therefore, assessment of salivary gland tumors is best performed with magnetic resonance imaging (MRI) (3). While imaging is mainly used to characterize the location and extent of parotid tumors, certain features can be suggestive of tumor type. For example, on conventional MRI, it has been observed that low T2 signal intensity and ill-defined margins of a parotid tumor are suggestive of malignancy (4). Conversely, T2 hyperintensity is suggestive of benign parotid tumor histology, though this singular characteristic is not sufficient to be a lone discriminator. However, tumor classification based on qualitative radiological imaging is subjective and often limited. The purpose of this study is to quantitatively evaluate the T2 signal intensity and surface regularity as a surrogate for margin status of parotid tumors on MRI. We present the following article in accordance with the STARD reporting checklist (available at https://gs.amegroups.com/article/view/10.21037/gs-22-88/rc).

Methods

Patients and scans

The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by the institutional review board of the University of Chicago (No. IRB14-0749) and individual consent for this retrospective analysis was waived. Adult patients with pathology-proven parotid tumors depicted on MRI available at a single institution were included. The MRI scans were performed at 1.5 or 3 T and included axial fat-suppressed T2-weighted (TR: 3,839, TE: 80, NEX: 2) with 3 mm slice thickness using head and neck neurovascular coils.

MRI analysis

Tumor regions of interest (ROIs) were drawn on axial T2-weighted images in MATLAB using in-house developed software under the supervision of a board certified neuroradiologist. The outlines were then slightly modified by dilation, followed by gray level thresholding in order to more accurately conform to the outline of the contrast-enhanced region (Figure 1). The resulting ROIs were entered into an image processing algorithm and further constructed to form a 3D image of the tumor in order determine surface irregularity. Tumor surface regularity was calculated for each imaging set by creating a dimensionless ratio where TV is the segmented tumor volume (in cubic centimeters) and where TS is the surface area (in square centimeters) of the tumor being analyzed:

Surfaceregularity=6πTV(TS)3[1]

Figure 1 Quantitative T2 heat maps in milliseconds show lower signal in the adenoid cystic carcinoma (A) than in the pleomorphic adenoma (B).

Statistical analysis

MATLAB’s Linear Discriminant Analysis algorithm was used to separate classes using the features described previously in order to populate the receiver operating characteristic (ROC) curves. P values were calculated between the produced ROC curve and the guessing line using the method described by Hanley and McNeil (5). A P value of <0.05 is considered significant.

Results

A total of 35 tumors, including 21 benign and 14 malignant neoplasms (Table 1) were included in this analysis. There were a total of 19 females and 16 males with an average age of 60. Among patients with benign tumors, there were 11 females and 10 males with an average age of 54 years. Among patients with malignant tumors, there were 8 females and 6 males with an average age of 66. For differentiating the benign versus malignant parotid tumors, T2 signal and surface regularity combined yielded an area under the curve of 0.62 (P value =0.2) through the ROC analysis (Figure 2). However, for the pleomorphic adenomas versus other types of parotid tumors, using both T2 signal and surface regularity yielded an area under the curve of 0.81 (P value: 0.007) through the ROC analysis (Figure 3), which meets exceeds the threshold of 0.80 for an acceptable diagnostic test (6).

Figure 2 ROC curve for T2 signal and surface regularity for benign versus malignant parotid tumors. ROC, receiver operating characteristic; SR, surface regularity; AUC, area under the curve.

Figure 3 ROC curve for T2 signal and surface regularity for pleomorphic adenoma versus other parotid tumors. ROC, receiver operating characteristic; SR, surface regularity; AUC, area under the curve.

Discussion

In the evaluation of parotid gland tumors with MRI, poorly defined margins, T2-signal hypointensity, and invasion of surrounding structures are suggestive of malignancy but not definitive (4,7,8). Malignant tumors demonstrate ill-defined borders in 59% of cases, which is almost three times more than benign tumors. In addition to poorly defined margins, malignant tumors commonly have a diffuse or multifocal pattern of growth and spread into the subcutaneous tissues and masticator space. Lymphadenopathy is also more common in malignant tumors (4).

Signal intensity is another useful feature in the differentiation of parotid tumors. High signal intensity on T2 weighted images (T2WI) is a well-known characteristic of benign tumors, including pleomorphic adenomas, and likely represents serous and mucinous components. Conversely, low signal intensity on T2WI suggests an aggressive lesion with high cellularity (4,8-12). However, low signal intensity can also be found in benign Warthin tumors, which is related to cellular components and cysts containing proteinaceous fluid and multiple cell-types. However, the signal is significantly lower than that of malignant tumors (13).

This study shows that a combination of surface regularity and T2 signal can differentiate pleomorphic adenomas from other types of parotid tumors. This is important given that pleomorphic adenomas are the most common salivary gland tumors and the surgical management of these tumors shifted away from enucleation towards superficial or total parotidectomy (14). Thus, the identification of pleomorphic adenomas on preoperative imaging can be helpful in surgical planning and follow-on care.

Although the combination of T2 signal and surface regularity was unable to differentiate benign and malignant tumors in general, there was a trend that was perhaps limited due to the small sample size. Thus, further evaluation of this technique is warranted in a larger cohort. The absence of post contrast imaging further limits evaluation. The presence of contrast has been shown to affect margin evaluation with ill-defined borders being more conspicuous after contrast administration (4).

Conclusions

A combination of T2 signal and surface regularity can reliably differentiate pleomorphic adenomas from other types of parotid tumors and can potentially be used as a predictive imaging biomarker.

Acknowledgments

We thank Annie Xiao for compiling the DICOM images used in this study and performing a preliminary assessment.

Funding: None.

Footnote

Reporting Checklist: The authors have completed the STARD reporting checklist. Available at https://gs.amegroups.com/article/view/10.21037/gs-22-88/rc

Data Sharing Statement: Available at https://gs.amegroups.com/article/view/10.21037/gs-22-88/dss

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://gs.amegroups.com/article/view/10.21037/gs-22-88/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. The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by the institutional review board of the University of Chicago (No. IRB14-0749) and individual consent for this retrospective analysis was waived.

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. Feinstein AJ, Alonso J, Yang SE, et al. Diagnostic Accuracy of Fine-Needle Aspiration for Parotid and Submandibular Gland Lesions. Otolaryngol Head Neck Surg 2016;155:431-6. [Crossref] [PubMed]
2. Wu S, Liu G, Chen R, et al. Role of ultrasound in the assessment of benignity and malignancy of parotid masses. Dentomaxillofac Radiol 2012;41:131-5. [Crossref] [PubMed]
3. 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]
4. Piludu F, Marzi S, Ravanelli M, et al. MRI-Based Radiomics to Differentiate between Benign and Malignant Parotid Tumors With External Validation. Front Oncol 2021;11:656918. [Crossref] [PubMed]
5. Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:29-36. [Crossref] [PubMed]
6. Nahm FS. Receiver operating characteristic curve: overview and practical use for clinicians. Korean J Anesthesiol 2022;75:25-36. [Crossref] [PubMed]
7. Xu Z, Chen M, Zheng S, et al. Differential diagnosis of parotid gland tumours: Application of SWI combined with DWI and DCE-MRI. Eur J Radiol 2022;146:110094. [Crossref] [PubMed]
8. Coudert H, Mirafzal S, Dissard A, et al. Multiparametric magnetic resonance imaging of parotid tumors: A systematic review. Diagn Interv Imaging 2021;102:121-30. [Crossref] [PubMed]
9. Rosai J, Ackerman S. Surgical Pathology. 8th ed. St. Louis: Mosby, 1986:815-55.
10. Harnsberger HR. Handbook of Head and Neck Imaging. 2nd ed. St. Louis: Mosby, 1990:60-73.
11. Ikeda K, Katoh T, Ha-Kawa SK, et al. The usefulness of MR in establishing the diagnosis of parotid pleomorphic adenoma. AJNR Am J Neuroradiol 1996;17:555-9.
12. Som PM, Biller HF. High-grade malignancies of the parotid gland: identification with MR imaging. Radiology 1989;173:823-6. [Crossref] [PubMed]
13. Ikeda M, Motoori K, Hanazawa T, et al. Warthin tumor of the parotid gland: diagnostic value of MR imaging with histopathologic correlation. AJNR Am J Neuroradiol 2004;25:1256-62.
14. Zbären P, Vander Poorten V, Witt RL, et al. Pleomorphic adenoma of the parotid: formal parotidectomy or limited surgery? Am J Surg 2013;205:109-18. [Crossref] [PubMed]