Differentiating parotid tumors by quantitative signal intensity evaluation on MR imaging

Accepted Manuscript Differentiating parotid tumors by quantitative signal intensity evaluation on MR imaging

Eiji Matsusue, Yoshio Fujihara, Eiken Matsuda, Yusuke Tokuyasu, Shu Nakamoto, Kazuhiko Nakamura, Toshihide Ogawa PII: DOI: Reference:

S0899-7071(17)30119-5 doi: 10.1016/j.clinimag.2017.06.009 JCT 8268

To appear in: Received date: Revised date: Accepted date:

26 December 2016 19 June 2017 28 June 2017

Please cite this article as: Eiji Matsusue, Yoshio Fujihara, Eiken Matsuda, Yusuke Tokuyasu, Shu Nakamoto, Kazuhiko Nakamura, Toshihide Ogawa , Differentiating parotid tumors by quantitative signal intensity evaluation on MR imaging, (2016), doi: 10.1016/j.clinimag.2017.06.009

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Differentiating parotid tumors by quantitative signal intensity evaluation on MR imaging. Eiji Matsusue 1, Yoshio Fujihara 1, Eiken Matsuda 2, Yusuke Tokuyasu 3, Shu Nakamoto 3, Kazuhiko Nakamura 1, Toshihide Ogawa 4

PT

1: Department of Radiology, Tottori Prefectural Central Hospital, Tottori, Japan 2: Department of Otorhinolaryngology, Tottori Prefectural Central Hospital, Tottori, Japan

RI

3: Department of Pathology, Tottori Prefectural Central Hospital, Tottori, Japan 4: Division of Radiology, Department of Pathophysiological Therapeutic Science, Tottori

SC

University, Tottori, Japan

Corresponding author: Eiji Matsusue

MA

NU

Department of Radiology, Tottori Prefectural Central Hospital, 730 Ezu, Tottori, Tottori 680-0901, Japan E-mail：[email protected]

PT E

D

Key Words Magnetic resonance imaging, parotid tumor, diffusion weighted imaging, conventional MR imaging Highlights ・Differentiating parotid tumors by quantitative signal intensity evaluation

CE

was assessed on MRI. ・Evaluations using spinal cord as an internal reference on conventional

AC

MRI. ・ADC maps were useful for differentiating pleomorphic adenomas from any other tumors. ・T2WI and CE-T1WI were useful for discriminating pleomorphic adenomas from any other tumors. ・ Discrimination between Warthin tumors and malignant tumors was difficult using any MR sequence. Declaration of conflicting interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. 1

ACCEPTED MANUSCRIPT

Funding The author(s) received no financial support for the research, authorship, and/or publication of this article.

PT

Abstract

RI

Background: There have been no reports about quantitative evaluations of enhancing components of parotid tumors on conventional MR imaging.

SC

Purpose: To evaluate the signal intensity of the enhancing components of tumors, including pleomorphic adenomas (PAs), Warthin tumors (WTs) and malignant tumors (MTs), on various MR sequences and to assess the usefulness of quantitative

NU

evaluations for differentiation among the three groups of tumors. Material and Methods: A total of 39 histologically proven tumors, including 15 PAs, 17

MA

WTs and 7 MTs, were enrolled in this study. The tumor-to-spinal cord contrast ratio (TSc-CR) was calculated by dividing the signal intensity of the lesion by that of the spinal cord on MR sequences, including T1-weighted imaging (T1WI), T2-weighted imaging (T2WI) and postcontrast gadolinium-enhanced T1WI (CE-T1WI). The mean

D

apparent diffusion coefficient (ADC) value was selected in each tumor. Furthermore, the

PT E

differences in the TSc-CRs and the ADC values among the three groups of tumors were statistically evaluated. Cutoff values were determined for the prediction of tumor histology.

CE

Results: The TSc-CRs of PAs were significantly higher than those of WTs and MTs on T2WI and CE-T1WI. The sensitivities and specificities were 100% and 87.5%, respectively, at a cutoff value of 1.31 for the TSc-CR of T2WI; and 83.3% and 100%,

AC

respectively, at a cutoff value of 2.00 for the TSc-CR of CE-T1WI. For the ADC values, sensitivity and specificity for discriminating PAs from WTs or MTs were both 100% when the cutoff value of the ADC was set at 1.40×10－3mm2/s. Conclusion: ADC maps and the quantitative evaluations using the TSc-CRs on T2WI and CE-T1WI were useful for discriminating WTs or MTs from PAs. However, discrimination between WTs and MTs was difficult using any MR sequence.

Introduction Up to 80% of salivary gland neoplasms occur in the parotid gland. Pleomorphic 2

ACCEPTED MANUSCRIPT

adenoma (PA) is the most common parotid tumor and accounts for approximately half of tumors found in the parotid gland. Warthin tumors (WTs) are the second most common parotid tumor of the parotid gland after PA [1, 2] although WTs are declining in incidence since they are tobacco related. It is important to discriminate between these common benign tumors and malignant tumors preoperatively because this information strongly influences the surgical procedure. Fine needle aspiration cytology (FNAC) is

PT

widely used for the diagnosis of parotid gland lesions and can predict whether the lesion is benign or malignant with an accuracy of 81-98% [3]. However, FNAC is not always

RI

conclusive because of sampling difficulties and the great heterogeneity of parotid gland tumors [4, 5]. Therefore, preoperative imaging has an important role in surgical

SC

planning. Computed tomography (CT), magnetic resonance imaging (MRI) and ultrasonography (US) are commonly used to evaluate parotid gland lesions. On conventional MRI, tumor lesions, including solid parts, borders and cystic/necrotic

NU

regions, have been qualitatively evaluated [6-10]. In addition, dynamic contrast MRI and diffusion-weighted imaging (DWI) have been performed to quantitatively assess the

MA

enhancing components of tumors [11-19].

Quantitative evaluations of the tumor lesions are considered to be more objective compared to the visual evaluations of the lesions on conventional MRI. Besides, the quantitative values as well as threshold values, obtained from the quantitative analysis,

D

are also expected as indicators for the qualitative evaluations of parotid tumors on MRI.

PT E

However, there have been no reports about quantitative evaluations of enhancing components of tumors on conventional MRI. The purpose of this study was to evaluate the signal intensity of the enhancing

CE

components of parotid tumors, including PAs, WTs and malignant tumors (MTs), on various MR sequences and to assess the usefulness of quantitative evaluations for

AC

differentiation among the three groups of tumors.

Material and methods

Patients Between April 2010 and May 2016, MRI examinations were performed in 51 consecutive patients in our hospital with suspected parotid tumors. Of these patients, 36 patients who underwent gadolinium-enhanced MRI and diffusion-weighted MRI were enrolled in the present study, including WTs (17 tumors in 14 patients; 12 men and 2 women; mean age, 65 years; age range, 50-71 years), PAs (15 tumors in 15 patients; 4 3

ACCEPTED MANUSCRIPT

men and 11 women; mean age, 53 years; age range, 32-84 years) and MTs (7 tumors in 7 patients; 4 men and 3 women; mean age, 67 years; age range, 54-84 years). The remaining 15 patients were excluded because 12 patients underwent no enhanced MR imaging and three patients had lesions of other histological types (2 patients with basal cell adenoma and one patient with oncocytoma). FNAC was performed for the diagnosis of nine of eighteen WTs, ten of fifteen PAs and all seven MTs before surgery. Three of

PT

nine WTs were diagnosed as probable WTs, and six of nine WTs were diagnosed as definite WTs. Five of ten PAs were diagnosed as probable PAs, and five of ten PAs were

RI

diagnosed as definite PAs. Six of seven MTs were diagnosed as malignancy or carcinoma. One case of MTs was diagnosed as definite malignant lymphoma. Tumor types were

SC

proven histopathologically after surgical excision except one case with malignant lymphoma. In addition, all seven MTs were diagnosed as primary parotid tumors (Table 1). This retrospective study was approved by the institutional review board of our

NU

hospital, and the informed consent requirement was waived.

MA

MRI

MR examinations were performed within two months before surgery in all patients. MRI was performed using a 1.5T MRI system (Excite HD; GE Healthcare, Milwaukee,

D

Wisc., USA) with head and neck array coils. The T1- and T2-weighted imaging

PT E

parameters were as follows. For the axial pre- and post-contrast gadolinium-enhanced T1-weighted spin-echo sequence, the parameters were repetition time, 500msec; echo time, 10msec; matrix, 256×256; section thickness, 5mm; intersection gap, 1mm; field of

CE

view, 22cm; and two signals averaged. For the axial T2-weighted fast spin-echo sequence, the parameters were repetition time, 3500msec; echo time, 90msec; matrix, 256×256; section thickness, 5mm; intersection gap, 1mm; field of view, 22cm; and two

AC

signals averaged. The parameters for DWI were as follows. For the axial spin-echo single-shot echo-planar sequence, the parameters were repetition time, 4300msec; echo time, 80msec; matrix, 128×128; section thickness, 5-mm, intersection gap, 1-mm; field of view, 22cm; and two signals averaged.

Sensitizing diffusion gradients were applied

sequentially in the x, y, and z directions with b values of 0 and 800 sec/mm 2. Apparent diffusion coefficient (ADC) maps were also generated.

MRI evaluations and statistical analyses MR data were interpreted using a clinical viewer (F-report; FUJIFILM, Tokyo, 4

ACCEPTED MANUSCRIPT

Japan). Images were analyzed by one radiologist (E.M., 20 years of experience in head and neck MR imaging). For quantitative evaluation of MR images, the signal intensity of the parotid tumors was measured on the image with the maximum tumor size. For evaluation of ADC mapping, a region of interests (ROI) drawn freehand as large as possible, including the solidly enhancing component of the tumor, was manually placed, and the mean ADC values of each lesion were obtained. Localization of each ROI

PT

was confirmed using T1-weighted imaging (T1WI), T2-weighted imaging (T2WI) and postcontrast gadolinium-enhanced T1WI (CE-T1WI) with visual exclusion of large

RI

cystic or necrotic areas and large vessels. If there were multiple solidly enhancing components in the tumor, e.g. cystic tumor, a ROI was placed including the largest

SC

enhancing component of them. For evaluation of other MR sequences including T1WI, T2WI and CE-T1WI, a ROI drawn was manually placed, and the signal intensity values were obtained in the same manner. As for an internal reference, parotid gland includes

NU

fat tissue as its substantial component [20]. Also, adipose tissue in the healthy parotid gland increases with age [21]. Thus, parotid gland was not considered to be the suitable

MA

reference tissue. Meanwhile, the spinal cord is present in the field of view during a head and neck study. Therefore, this tissue can potentially be used as a reference. It also has the advantage of being rarely affected by malignancy. Hence, we used spinal cord in each case as an internal reference. An oval shaped ROI was drawn on the gray matter of

D

the spinal cord on each MR sequence as an internal reference. Representative images of

PT E

the selected ROIs are shown in Fig.2-4. On each MR sequence, the tumor-to-spinal cord contrast ratio (TSc-CR) was calculated by dividing the SI of the enhanced tumor lesion by that of the spinal cord.

CE

In order to evaluate intrarater reproducibility, the process of the ROI measurement was performed three times on three different days. The average ADC value of the three trials was used for the correlation analysis. Also, the average TSc-CR of the three trials

AC

was also used for the correlation analysis in each sequence. Differences in the TSc-CRs in each MR sequence and in the ADC values were statistically evaluated among the three groups of tumors (WTs, PAs and MTs) using the Kruskal-Wallis test followed by the Mann-Whitney U test. To determine the sensitivity and specificity of the images for the prediction of tumor histology, we also performed receiver operating characteristic (ROC) analysis. Cutoff values were determined using the Youden index. P ＜ .05 was considered indicative of a significant difference. Intrarater reproducibility was evaluated by an intraclass correlation coefficient based on the values of three trials in each MR sequence. An intraclass correlation coefficient value of >0.75 was considered “good to excellent”. 5

ACCEPTED MANUSCRIPT

All statistical analyses were performed with EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), which is a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria) [22].

PT

Results

RI

ROI size ranged, depending on lesion size, from 21-1474 (median, 179) mm2 on ADC map; 15-1218 (median, 143) mm2 on T1WI; 15-1218 (median, 147) mm2 on T2WI; and

SC

15-1218 (median, 137) mm2 on CE-T1WI, respectively. The ADC values and TSc-CRs in WTs, PAs and MTs are shown in Table 2. The intraclass correlation coefficient of the three trials was good to excellent on ADC values (0.99), TSc-CRs of T1WI (0.98),

NU

TSc-CRs of T2WI (0.95) and TSc-CRs of CE-T1WI (0.97), respectively. Scatterplots of the ADC values and of the TSc-CRs on all MR sequences are shown in Fig. 1.

MA

For the ADC mapping, the ADC values of WTs, PAs and MTs were in the range of 0.82-1.36 (median, 0.99), 1.40-2.49 (median, 1.95), and 0.45-1.36 (median, 1.02), respectively, and were significantly different among the 3

groups (P<.001,

Kruskal-Wallis tests). The ADC value of malignant lymphoma classified as MTs was

D

0.41×10-3 mm2/sec, which was the lowest value of all tumors. Significant differences

PT E

were observed between PAs and WTs (P＜.001, Mann-Whitney U tests) and between PAs and MTs (P＜.001, Mann-Whitney U tests). No significant difference was observed between WTs and MTs (P=.924, Mann-Whitney U tests).

CE

For the T1WI, the TSc-CRs of WTs, PAs and MTs were in the range of 0.91-1.56 (median, 1.21), 0.66-1.25 (median, 1.02), and 0.80-1.19 (median, 0.99), respectively, and the TSc-CRs of the tumors did not differ significantly different among WTs, PAs and

AC

MTs (P＞.05, Kruskal-Wallis tests). For the T2WI, the TSc-CRs of WTs, PAs and MTs were in the range of 0.87-1.47 (median, 1.13), 1.31-2.62 (median, 1.59), and 0.93-1.31 (median, 1.03), respectively, and were significantly different among the 3 groups (P<.001, Kruskal-Wallis tests). Significant differences were observed between PAs and WTs (P ＜ .001, Mann-Whitney U tests) and between PAs and MTs (P ＜ .001, Mann-Whitney U tests). No significant difference was observed between WTs and MTs (P=.757, Mann-Whitney U tests). For the CE-T1WI, the TSc-CRs of WTs, PAs and MTs were in the range of 1.42-1.99 (median, 1.67), 1.82-2.65 (median, 2.19), and 1.22-1.85 (median, 1.65), respectively, and were significantly different among the 3 groups (P<.001, Kruskal-Wallis tests). Significant differences were observed between PAs and 6

ACCEPTED MANUSCRIPT

WTs (P ＜ .001, Mann-Whitney U tests) and between PAs and MTs (P ＜ .001, Mann-Whitney U tests). No significant difference was observed between WTs and MTs (P=.541, Mann-Whitney U tests). ROC analyses showed that for the ADC values, the sensitivity and specificities for discriminating PAs from WTs or MTs were both 100 % when the cutoff value of the ADC was 1.40×10-3 mm2/s. For the TSc-CR of T2WI, the sensitivities and specificities for

PT

discriminating PAs from WTs or MTs were 100 % and 87.5 %, respectively, at a cutoff value of 1.31. For the TSc-CR of CE-T1WI, the sensitivities and specificities for

RI

discriminating PAs from WTs or MTs were 83.3 % and 100 %, respectively, at a cutoff

SC

value of 2.00. The cutoff values of all MR sequences are shown in Fig. 1.

Discussion

NU

On our quantitative analyses of the parotid tumors, including WTs, PAs and MTs, the TSc-CRs of PAs were significantly higher than those of WTs or MTs on T2WI and

MA

CE-T1WI. The sensitivities and specificities were 100% and 87.5%, respectively, at a cutoff value of 1.31 for the TSc-CR of T2WI; and 83.3% and 100%, respectively, at a cutoff value of 2.00 for the TSc-CR of CE-T1WI. Furthermore, the mean ADC values of PAs were significantly higher than those of WTs or MTs and the sensitivity and

D

specificities were both 100% when the cutoff value of the ADC was 1.40×10-3 mm2/s.

PT E

The other hand, no significant difference was observed between WTs and MTs on any MR sequences. Our quantitative evaluations of parotid tumors were considered to have the following two strengths; 1. Evaluations using spinal cord as an internal reference,

CE

and 2. Evaluations using solidly enhancing components as the solid parts of the tumors. For the qualitative evaluation of tumor lesions on conventional MRI, the signal intensities of the tumors are usually compared to those of normal parotid glands, whose

AC

signal intensities are not certain due to their physiological fat tissue [21, 22]. Therefore, we used spinal cord as a stable internal reference for the evaluations of the tumors. Besides, we evaluated signal changes of the solidly enhancing components of the tumor; Cystic mass tumors are frequently seen in PAs, WTs and MTs. Those kinds of tumors, consisting of solid and cystic components, seem to be difficult to evaluate qualitatively because the qualitative evaluation of the solid components might be prone to be influenced by the signal intensities of the cystic components adjacent to the solid components. Hence, our quantitative evaluation of the enhancing component of the tumor lesions, using ADC values, the TSc-CRs on T2WI or CE-T1WI, were considered to be more objective, compared to the visual evaluations of the tumor lesions. 7

ACCEPTED MANUSCRIPT

PAs and WTs account for the majority of parotid tumors. Clinically, differentiation of PAs from WTs is essential, because the surgery plan differs between them; a partial parotidectomy is suitable for PAs because PAs are associated with a 2-25% risk of malignant degeneration over time. In contrast, enucleation is sufficient for WTs because WTs usually do not recur [23, 24]. T2-hyperintensity, persistent enhancement after contrast administration and increased ADC values are well-known as specific MRI

PT

findings of PAs and their findings are histologically consistent with fibromyxoid stroma [7, 12 25]. The other hand, cellular epithelial and lymphoid components of WTs can be

RI

recognized as hypointensity on T2WI and conventional CE-T1WI [26]. In addition, hypointensity on conventional CE-T1WI is usually observed in WTs with early

SC

enhancement and a high washout rate, namely washout time-signal intensity curve (TIC) pattern on dynamic contrast-enhanced MRI [11, 13, 18]. Furthermore, WTs, which consist of epithelial and lymphoid stroma with microscopic cysts filled with

NU

proteinaceous fluid, have low ADC values [26]. In this study, the TSc-CRs of PAs were significantly higher than those of WTs on T2WI and CE-T1WI. However, there was

MA

slight overlap between WTs and PAs both on T2WI and CE-T1WI. Meanwhile, a significant proportion of PAs show intermediate to hypointensity on T2WI and also reduced ADC values, whose findings reflected hypercellularity with less-myxoid stroma and might cause the overlap between PAs and WTs. Still, ADC maps allow diagnoses of

D

tumors with intermediate to low intensity on T2WI or CE-T1WI, which may suggest

PT E

WTs, as PAs if the tumors have high ADC values. In this study, the mean ADC values of PAs were significantly higher than those of WTs. The cutoff value of 1.40×10-3 mm2/s was useful for discriminating between PAs and WTs on ADC maps. Besides, there was

CE

no overlap between MTs and PAs on ADC maps. T2 hypointensity of MTs, such as carcinoma and malignant lymphoma, has been previously linked to highly cellular tumors [8]. In addition, hypointensity on

AC

conventional CE-T1WI is usually observed in MTs on dynamic contrast-enhanced MRI with the following two patterns; 1. Washout TIC pattern as seen in WTs and 2. Moderate increase rate and moderate washout rate, namely plateau TIC pattern [11, 13, 18]. Besides, MTs exhibit hypercellularity, which reduces the extracellular matrix and the diffusion space of water protons in the extracellular and intracellular dimensions, with a resultant decrease in ADCs [27]. In this study, the TSc-CRs of MTs were significantly lower than those of PAs on T2WI and CE-T1WI. Meanwhile, malignant parotid tumor can be histologically classified as low, intermediate or high grade according to intracystic components, mitotic figures, neural invasion, necrosis and cellular anaplasia. On T2WI, high-grade malignancies show low to intermediate 8

ACCEPTED MANUSCRIPT

intensity, reflecting high-cellularity [28, 29, 30]. On the other hand, low-grade malignancies, such as mucoepidermoid carcinoma, acinic cell carcinoma and adenoid cystic carcinoma, show high signal intensity, reflecting cystic architectural pattern because of the existence of abundant mucin-secreting cells [19, 30, 31]. In this study, there was no overlap between MTs and PAs on T2WI. However, the possibility of the overlap between MTs showing high signal intensity, such as low-grade malignancies of

PT

MTs, and PAs showing low signal intensity, such as above mentioned hypercellularity PAs, should be considered for discrimination MTs and PAs on T2WI. Generally, MTs

RI

show lower ADC values compared to those of PAs. Yet, there have been no reports evaluating ADC values of MTs, classifying by their grades. In this study, the mean ADC

SC

values of MTs were significantly lower than those of PAs. Furthermore, the sensitivities and specificities were both 100% when the cutoff value of the ADC was 1.40×10-3 mm2/s. Hypointensity both on T2WI and CE-T1WI are usually observed in WTs as well as

NU

in MTs [8, 11, 18, 26]. No significant difference was observed between WTs and MTs on any MR sequences in this study. For discrimination between WTs and MTs based on

MA

ADC values, it has been reported that the average ADC value of WTs is significantly lower than that of MTs [26]. Overlap between WTs and MTs has also been observed [16], and several authors have concluded that ADC values cannot differentiate WTs from MTs [14, 19]. In this study, the median mean ADC value of WTs was 0.99×10-3 mm2/s

D

and the median mean ADC value of MTs was 1.02×10-3 mm2/s. Therefore, discriminating

PT E

WTs from MTs based on ADC values is likely to be difficult. Meanwhile, the average ADC value of malignant lymphoma classified as MTs in our cases was 0.41×10-3 mm2/sec, which was the lowest value of all tumors. Wang et al reported that the ADC

CE

value of malignant lymphoma (0.66 ±0.17×10-3 mm2/s) was significantly lower than those of carninomas (1.13±0.43×10-3 mm2/s) [27]. According to the previous studies, WTs have ADC values of 0.72 to 0.96×10-3 mm2/s [16，18, 19, 26], which were higher

AC

than those of malignant lymphoma. Therefore, tumor lesions with prominent low ADC values, approximately less than 0.7 ×10-3 mm2/s, might be considered malignant lymphoma as MTs rather than WTs. In this study, FNAC was performed for the diagnosis of nine of eighteen WTs, ten of fifteen PAs, all of seven MTs before surgery. All cases were correctly diagnosed whether the lesion was benign or malignant. All of the cases with WTs and PAs were able to be definitely or probably diagnosed with FNAC. Also, one case of MTs was definitely diagnosed as malignant lymphoma. FNAC is a relatively inexpensive and quick procedure that can be easily performed in the outpatient clinic although it is invasive and specimens may be insufficient for a conclusive diagnosis. On the other 9

ACCEPTED MANUSCRIPT

hand, MRI is superior in defining tumor characteristics and extension. Additionally, MRI can contribute anatomic information that may be useful for surgical planning although the drawback of MRI includes the higher cost and longer examination time. FNAC can be recommended as initial preoperative assessment of parotid gland tumors. However, it is not necessary to perform FNAC on all cases with parotid gland tumors because most of PAs were able to be differentiated from WTs or MTs by conventional

PT

MRI and ADC maps. Ideally, FNAC and MRI should be used in combination to reliably establish the diagnosis of parotid tumors.

RI

Our study has several limitations. First, we mainly included PAs and WTs because MTs were few and histologically heterogeneous; thus, it was not possible to identify

SC

typical MR findings for each malignant pathological entity. Second, tumor locations, growth patterns, margins and signal intensity of cystic/necrotic content were not evaluated because we focused on the assessment of enhancing components of tumors.

NU

Third, evaluations of tumors were not performed on dynamic contrast-enhancement MRI because our study was designed for evaluations of the head and neck region using

MA

a standard neck MRI protocol. Fourth, the qualitative evaluations using fat suppressed T2WI and CE-T1WI were not performed in this study. Fat suppression is quite useful method to detect tumor lesions qualitatively. However, on these fat suppressed imaging sequences, several parotid glands of our cases were difficult to perform correct

D

quantitative evaluations due to those inhomogeneous fat suppressions. Finally, the

PT E

single observer who selected the location of the ROI was not blinded to the results of other imaging studies and to the clinical data. Because of these limitations, the results of this study should be considered a preliminary pilot investigation. Further validation

CE

is needed with a larger number of cases. However, this is the first report on the quantitative evaluations of parotid tumors using the tumor-to-spinal cord contrast ratio on standard MR sequences to characterize parotid tumors. Furthermore, this simple

AC

evaluation method of parotid tumors using a cutoff value is expected to be as a useful indicator for discrimination PAs from the other tumors. In conclusion, ADC maps and the quantitative evaluations using the tumor-to-spinal cord contrast ratio on T2WI and CE-T1WI were useful for discriminating WTs or MTs from PAs. However, discrimination between WTs and MTs was difficult using any MR sequence.

10

ACCEPTED MANUSCRIPT

PT

Figure legends

RI

Fig. 1.

Scatterplots of the TSc-CRs on MR sequences and of the ADC values

SC

Note.- WT indicates Warthin tumor; PA, pleomorphic adenoma; MT, malignant tumor; PG, parotid gland; TSc-CR, tumor to spinal cord contrast ratio.

NU

Fig. 2.

Representative images of the selected region of interests (ROIs) in a 52-year-old woman

MA

with pleomorphic adenoma.

A. Axial ADC mapping. B. Axial T1-weighted image (T1WI). C. Axial T2-weighted image (T2WI). D. Axial contrast enhanced T1-weighted image (CE-T1WI). A ROI drawn as large as possible, including the solidly enhancing component of the (intermittent yellow lines in A-D). Localization of each ROI was

D

tumor, is placed

PT E

confirmed using other sequences with visual exclusion of cystic or necrotic areas (arrowed in A-D). Each on T1WI, T2WI and CE-T1WI, an oval shaped ROI is also drawn on the gray matter of the spinal cord

(intermittent yellow lines in B-D) as an

mm2/sec

for the ADC value, 0.99 for the TSc-CR

CE

internal reference. Values of

2.46×10-3

of T1WI, 2.65 for the TSc-CR of T2WI, and 2.19 for the TSc-CR of CE-T1WI are

Fig. 3.

AC

consistent with pleomorphic adenoma.

Representative images of the selected ROIs in a 66-year-old man with Warthin tumor. A. Axial ADC mapping. B. Axial T1WI. C. Axial T2WI. D. Axial CE-T1WI. A ROI is placed including the largest solidly enhancing component of the multiple enhancing ones in the cystic tumor (intermittent yellow lines in A-D). Each on T1WI, T2WI and CE-T1WI, an oval shaped ROI is also drawn on the gray matter of the spinal cord (intermittent yellow lines in B-D). Values of 1.15×10-3 mm2/sec for the ADC value, 1.42 for the TSc-CR of T1WI, 1.53 for the TSc-CR of T2WI, and 1.98 for the TSc-CR of CE-T1WI are considered both a 11

ACCEPTED MANUSCRIPT

Warthin tumor and a malignant tumor. Fig. 4. Representative images of the selected ROIs in a 60-year-old man with carcinoma ex pleomorphic adenoma. A. Axial ADC mapping. B. Axial T1WI. C. Axial T2WI. D. Axial CE-T1WI.

PT

A ROI drawn as large as possible, including the solidly enhancing component of the tumor, is placed (intermittent yellow lines in A-D) with visual exclusion of cystic or

RI

necrotic areas (arrowed in A-D). Each on T1WI, T2WI and CE-T1WI, an oval shaped ROI is also drawn on the gray matter of the spinal cord (intermittent yellow lines in

SC

B-D).

Values of 1.12×10-3 mm2/sec for the ADC value, 0.91 for the TSc-CR of T1WI, 1.18 for the TSc-CR of T2WI, and 1.69 for the TSc-CR of CE-T1WI were considered both a Warthin

AC

CE

PT E

D

MA

NU

tumor and a malignant tumor.

12

ACCEPTED MANUSCRIPT

References 1. Eveson JW, Cawson RA. Salivary gland tumours: a review of 2410 cases with particular reference to histological types, site, age and sex distribution. J Pathol 1985;146:51-58.

PT

2. Eveson JW, Auclair P, Gnepp DR, El-Naggar AK. Tumours of the salivary glands: introduction. In: Barnes L, Eveson JW, Reichart P. Sidransky D (eds) : Pathology and

RI

genetics head and neck tumors. Lyon : IARC Press, 2005:212-13.

SC

3. Seethala RR, LiVolsi VA, Baloch ZW. Relative accuracy of fine-needle aspiration and frozen section in the diagnosis of lesions of the parotid gland. Head and Neck

NU

2005;27:217-23.

4. Schindler S, Nayar R, Dutra J, Bedrossian CW. Diagnostic challenges in aspiration

MA

cytology of the salivary glands. Semin Diagn Pathol 2001;18:124-46. 5. Schmidt RL, Hall BJ, Wilson AR, Layfield LJ. A systematic review and meta-analysis of the diagnostic accuracy of fine-needle aspiration cytology for parotid gland lesions.

PT E

D

Am J Clin Pathol 2011;136:45-59.

6. Freling NJ, Molenaar WM, Vermey A, Mooyaart EL, Panders AK, Annyas AA, Thijn CJ. Malignant parotid tumors: clinical use of MR imaging and histologic correlation.

CE

Radiology 1992;185:691-6.

7. Ikeda K, Katoh T, Ha-Kawa SK, Iwai H, Yamashita T, Tanaka Y. The usefulness of

AC

MR in establishing the diagnosis of parotid pleomorphic adenoma. AJNR Am J Neuroradiol 1996;17:555-9. 8. Okahara M, Kiyosue H, Hori Y, Matsumoto A, Mori H, Yokoyama S. Parotid tumors: MR imaging with pathological correlation. Eur Radiol 2003; 13 (Suppl 4):L25-33. 9. Ishibashi M, Fujii S, Kawamoto K, Nishihara K, Matsusue E, Kodani K, Kaminou T, Ogawa T. Capsule of parotid gland tumor: evaluation by 3.0 T magnetic resonance imaging using surface coils. Acta Radiol 2010;51:1103-10.

13

ACCEPTED MANUSCRIPT

10. Christe A, Waldherr C, Hallett R, Zbaeren P, Thoeny H. MR imaging of parotid tumors: typical lesion characteristics in MR imaging improve discrimination between benign and malignant disease. AJNR Am J Neurodadiol 2011;32:1202-7. 11. Yabuuchi H, Fukuya T, Tajima T, Hachitanda Y, Tomita K, Koga M. Salivary gland tumors: diagnostic value of gadolinium-enhanced dynamic MR imaging with

PT

histopathologic correlation. Radiology 2003;226:345-54.

RI

12. Motoori K, Yamamoto S, Ueda T, Nakano K, Muto T, Nagai Y, Ikeda M, Funatsu H, Ito H. Inter- and intratumoral variability in magnetic resonance imaging of pleomorphic

SC

adenoma: an attempt to interpret the variable magnetic resonance findings. J Comput Assist Tomogr 2004;28:233-46.

NU

13. Aghaghazvini L, Salahshour F, Yazdani N, Sharifian H, Kooraki S, Pakravan M, Shakiba M. Dynamic contrast-enhanced MRI for differentiation of major salivary

MA

glands neoplasms, a 3-T MRI study. Dentomaxillofac Radiol 2015;44:20140166. doi: 10.1259/dmfr.20140166.

14. Yerli H, Agildere AM, Aydin E, Geyik E, Haberal N, Kaskati T, Oguz D, Ozluoglu LN.

D

Value of apparent diffusion coefficient calculation in the differential diagnosis of parotid

PT E

gland tumors. Acta Radiol 2007;48:980-7. 15 Eida S, Sumi M, Sakihama N, Takahashi H, Nakamura T. Apparent diffusion

CE

coefficient mapping of salivary gland tumors: prediction of the benignancy and malignancy. AJNR Am J Neuroradiol 2007;28:116-21.

AC

16. Habermann CR, Arndt C, Graessner J, Diestel L, Petersen KU, Reitmeier F, Ussmueller JO, Adam G, Jaehne M. Diffuison-weighted echo-planar MR imaging of primary parotid gland tumors: is a prediction of different histologic subtypes possible? AJNR Am J Neuroradiol 2009;30:591-6. 17. Celebi I1, Mahmutoglu AS, Ucgul A, Ulusay SM, Basak T, Basak M. Quantitative diffusion-weighted magnetic resonance imaging in the evaluation of parotid gland masses: a study with histopathological correlation. Clin Imaging 2013;37:232-8. 18. Yabuuchi H1, Matsuo Y, Kamitani T, Setoguchi T, Okafuji T, Soeda H, Sakai S, 14

ACCEPTED MANUSCRIPT

Hatakenaka M, Nakashima T, Oda Y, Honda H. Partotid gland tumors: can addition of diffusion-weighted MR imaging to dynamic contrast-enhanced MR imaging improve diagnostic accuracy in characterization? Radiology 2008;249:909-16. 19. Kikuchi M, Koyasu S, Shinohara S, Imai Y, Hino M, Naito Y. Preoperative diagnostic strategy for parotid gland tumors using diffusion-weighted MRI and technetium-99m

PT

pertechnetate scintigraphy: A Prospective Study. PLoS One 2016;11:e0148973.

RI

20. Scott J, Flower EA, Burns J. A quantitative study of histological changes in the

SC

human parotid gland occurring with adult age. J Oral Pathol 1987;16:505-10. 21. Sumi M, Takagi Y, Uetani M, et al. Diffusion-weighted echoplanar MR imaging of

NU

the salivary glands. AJR Am J Roentgenol 2002;178:959-65.

22. Kanda Y. Investigation of the freely-available easy-to-use software “EZR” (Easy R)

MA

for medical statistics. Bone marrow Transplant. 2013: 48: 452-8. 23. Rong X, Zhu Q, Ji H, Li J, Huang H. Differentiation of pleomorphic adenoma and Warthin’s tumor of the parotid gland: ultrasonographic features. Acta Radiol 2014; 55:

PT E

D

1203-9.

24. Yuan WH, Hsu HC, Chou YH, Hsueh HC, Tseng TK, Tiu CM. Gray-scale and color Doppler ultrasonographic features of pleomorphic adenoma and Warthin’s tumor in

CE

major salivary glands. Clin Imaging 2009; 33: 348-53. 25. Tsushima Y, Matsumoto M, Endo K, Aihara T, Nakajima T. Characteristic bright

AC

signal of parotid pleomorphic adenomas on T2-weighted images with pathological correlation. Clin Radiol 1994;49:485-9. 26. Ikeda M, Motoori K, Hanazawa T, Nagai Y, Yamamoto S, Ueda T, Funatsu H, Ito H. Warthin tumor of the parotid gland: diagnostic value of MR imaging with histopathologic correlation. AJNR Am J Neuroradiol 2004;25:1256-62. 27. Wang J, Takashima S, Takayama F, Kawakami S, Saito A, Matsushita T, Momose M, Ishiyama T. Head and neck lesions: characterization with diffusion-weighted echo-planar MR imaging. Radiology 2001;220:621-30. 15

ACCEPTED MANUSCRIPT

28, Freling NJ, Molenaar WM, Vermey A, Mooyaart EL, Panders AK, Annyas AA, Thijn CJ. Malignant parotid tumors: clinical use of MR imaging and histologic correlation. Radiology 1992;185:691-6. 29. Motoori K, Iida Y, Nagai Y, Yamamoto S, Ueda T, Funatsu H, Ito H, Yoshitaka O.

PT

MR imaging of salivary duct carcinoma. AJNR Am J Neuroradiol 2005;26:1201-6.

RI

30. Kashiwagi N, Dote K, Kawano K, Tomita Y, Murakami T, Nakanishi K, Araki Y, Mori K, Tomiyama N. MRI findings of mucoepidermoid carcinoma of the parotid gland:

SC

correlation with pathological features. Br J Radiol 2012;85:709-13.

31. Sigal R, Monnet O, de Baere T, Micheau C, Shapeero LG, Julieron M, Bosq J, Vanel

NU

D, Piekarski JD, Luboinski B, et al. Adenoid cystic carcinoma of the head and neck: evaluation with MR imaging and clinical-pathologic correlation in 27 patients.

AC

CE

PT E

D

MA

Radiology 1992;184:95-101.

16

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

AC

CE

PT E

Figure 1

17

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

18

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

19

AC

CE

PT E

D

MA

NU

SC

RI

PT

ACCEPTED MANUSCRIPT

20

ACCEPTED MANUSCRIPT

Table 1. Histopathological diagnoses of parotid gland tumors Final diagnosis No. of FNAC No. of (diagnosis) MRI alone Warthin tumor (WT) (n = 17)

9

8

PT

3 (Probably WT)

RI SC

Pleomorphic adenoma (PA) (n =

MA

NU

15)

5 (Definitely PA) 7

PT E

Squamous cell carcinoma Adenoid cystic

3 (2 Carcinoma, 1 Malignancy) 1 (Malignancy) 1

carcinoma

(Malignancy)

AC

CE

ex

5 (Probably PA)

D

Malignant tumor (MT) (n = 7) Carcinoma PA

6 (Definitely WT) 10 5

Small carcinoma

cell

1 (Carcinoma)

Malignant lymphoma

1 (Malignant lymphoma)

21

0

ACCEPTED MANUSCRIPT

Table 2. The ADC values and TSc-CRs in the three types of parotid tumors WT PA MT p value

range (median)

range (median)

range (median)

PT

(Kruskal-Wallis Test)

RI

0.82-1.36 (0.99) 1.40-2.49 (1.95) 0.45-1.35 (1.02) p ＜0.001

ADC value

SC

(×10-3mm2/sec)

MA

NU

TSc-CR

0.91-1.56 (1.21) 0.66-1.25 (1.02) 0.80-1.19 (0.99) p ＞0.05

T2WI

0.87-1.47 (1.13) 1.31-2.62 (1.59) 0.93-1.31 (1.03) p ＜0.001

CE-T1WI

1.42-1.99 (1.67) 1.82-2.65 (2.19) 1.22-1.85 (1.65) p ＜0.001

CE

PT E

D

T1WI

AC

Note.- WT indicates Warthin tumor; PA, pleomorphic adenoma; MT, malignant tumor; TSc-CR, tumor to spinal cord contrast ratio.

22