Role of Quantitative Diffusion-Weighted Imaging in Differentiating Benign and Malignant Orbital Masses

Abstract

Aim To determine the role of diffusion-weighted imaging (DWI) with apparent diffusion coefficient (ADC) values in differentiating benign and malignant orbital masses.

Materials and Methods After obtaining institutional ethical board approval and informed consent from all patients, an observational study was done for a period of 24 months in the radiology department of a tertiary care hospital in South India. Conventional magnetic resonance imaging and DWI using a 3T scanner was done for all patients with suspected orbital mass lesion. ADC value and clinicohistopathological correlation were studied for every patient. Chi-square test was used to compare the signal characteristics of DWI and ADC maps between benign and malignant lesions. A comparison of mean ADC values for benign and malignant masses was performed using Student’s t-test for independent samples. The cut-off value for ADC was obtained using the receiver operating characteristic (ROC) curve.

Results Of 44 patients with orbital lesions, 70% were benign and 30% were malignant. There was a significant difference in the mean ADC values of benign and malignant orbital masses. Using ROC curve analysis, an optimal ADC threshold of 1.26 × 10⁻³ mm²/s was calculated for the prediction of malignancy with 100% sensitivity, 80.65% specificity, and 86.36% accuracy (95% confidence interval: 0.872, 1.00, p < 0.0001). Two ADC thresholds were used to characterize the orbital masses with more than 90% confidence.

Conclusion Quantitative assessment of ADC is a useful noninvasive diagnostic tool for differentiating benign and malignant orbital masses. Malignant orbital lesions demonstrate significantly lower ADC values as compared with benign lesions.

Keywords

ADC value - benign - diffusion-weighted imaging - malignant - orbital mass

Introduction

The orbit is a small anatomical space containing various important structures.[1] Orbital pathologies assume critical importance as the eyeball forms the most important sense organ in humans. The conditions that affect vision have a great impact on several aspects of an individual’s life. Various pathologies affecting the orbit can be broadly categorized as congenital, vascular, infectious/inflammatory, and neoplastic (includes both benign and malignant lesions).[2] A good understanding of orbital anatomy is crucial for orbital imaging.[3] [4] A compartment-based approach toward the evaluation of orbital masses simplifies the diagnostic approach and helps to narrow down the differential diagnoses.[5] As there are numerous possible etiologies of orbital masses, arriving at the correct clinical diagnosis can be challenging. Orbital imaging is commonly performed for patients presenting with proptosis, visual impairment, diplopia, leukocoria, eye pain, and trauma. Magnetic resonance imaging (MRI) has superseded other modalities to become the most accurate imaging modality due to superior soft-tissue contrast, better multiplanar examinations, and lack of irradiation.[6] It provides crucial information regarding the ocular structures involved, the extent of involvement, and intracranial extension.[7]

The role of advanced imaging techniques such as diffusion-weighted imaging (DWI) in the initial staging and posttreatment follow-up of head and neck cancers is being increasingly recognized. DWI is being used to differentiate benign and malignant thyroid nodules, cervical lymph nodes, and parotid masses.[8] The application of DWI in the orbit is limited as compared with other head and neck cancers due to geometric distortion caused by the adjacent air, bone, and soft tissue interfaces. Newer imaging techniques such as single- and multishot echo-planar DWI are being developed for adequate evaluation of orbital masses and optic nerve pathologies. Parallel imaging techniques are being developed to overcome the limitations of magnetic susceptibility artifacts caused by echo-planar imaging (EPI).[1]

Qualitative assessment of signal intensity on diffusion-weighted images may not reveal significant differences between benign and malignant lesions due to the presence of the T2 shine-through effect in benign lesions. Hence, a quantitative assessment using apparent diffusion coefficient (ADC) maps is required for more accurate characterization of lesions since ADC is only affected by changes in diffusivity.[9] [10] [11]

Most of the studies in the literature reported heterogeneity of equipment and scanning techniques, study cohort of specific predominant lesions, comparison of particular orbital masses like pseudotumor versus lymphoma, and interobserver variability.[12] [13] [14] This study was undertaken to eliminate the biases related to the machine, scanning techniques, and interobserver variations. A more heterogeneous group of benign and malignant masses was included in this study for better generalization of results. As earlier studies have reported different ADC cut-off values to differentiate benign and malignant orbital masses, we proposed a two-threshold model for better characterization of orbital lesions.

The present study was aimed to determine the role of DWI with ADC values in differentiating benign and malignant orbital masses.

Materials and Methods

Patient Population

An observational study was performed in the radiology department of a tertiary care hospital in South India from April 2017 till April 2019. After institutional ethical board approval, written informed consent was obtained from all the patients. All patients with clinically suspected orbital mass lesions referred to the radiology department for orbital MRI were included. Patients operated for orbital mass lesions and referred for follow-up study, patients with general contraindications for MRI, and those not willing to undergo MRI were excluded. Clinicohistopathological correlations were studied for every patient.

Data acquisition: The patient demographic data were recorded in the case record form. Patients were instructed to remove all metallic belongings before the examination. Pediatric patients and noncooperative patients were sedated to minimize motion artifacts. All the patients were subjected to conventional MRI and DWI. MRI was performed using a 3T (Philips Ingenia) scanner, using a 16-channel head coil with the patient in the supine position.[15] [16] A scout sagittal T1-weighted image was obtained to confirm the position of the patient and to act as a localizer for subsequent image acquisitions.[17] Initially, 3D FLAIR brain screening was performed in all cases. All the patients underwent conventional orbital imaging, including T1-weighted and T2-weighted imaging with fat suppression. Multiplanar contrast-enhanced T1-weighted fat-suppressed image was obtained in axial, coronal, and sagittal planes after the administration of 0.1 mmol/kg dose of gadolinium-based intravenous contrast (Magnevist/Multihance/Omniscan). DWI was performed in the axial plane using a single-shot echo-planar imaging sequence with b-values of 0 and 1,000 second mm². The following imaging parameters were used: average TR/TE of 6,500/65 ms for DWI, field of view of 16 to 18 cm, slice thickness of 4 mm, interslice gap of 1 to 2 mm, and a matrix of 256 × 256. Additional sequences like gradient and balanced steady-state sequences were used if clinically indicated.

Data analysis: Images were sent to the workstation and the PACS server for analysis. The high-resolution T2-weighted images were assessed first to localize the lesion in each case. An ADC map was generated for every patient. The lesion was identified on the ADC map image using multiplanar cross-reference with the conventional images. The quantitative ADC value was calculated by drawing a freehand region of interest within the lesion on the ADC map. The peripheral aspects of the lesion and areas of necrosis seen on the ADC map were excluded to avoid the effect of partial volume averaging. The areas of calcification were not excluded while calculating the ADC value.

Statistical Analysis

The diagnostic parameters of DWI with ADC were obtained in terms of sensitivity, specificity, positive predictive value, and negative predictive value. Chi-square test was used to compare the signal characteristics of the DWI and ADC maps of benign and malignant lesions. The comparison of mean ADC values for benign and malignant masses was performed using Student’s t-test for independent samples. The cut-off value for ADC was obtained using the receiver operating characteristic (ROC) curves. All the analyses were performed using IBM SPSS Statistics 20 and a p-value of less than 0.05 was considered statistically significant.

Results

The baseline characteristics of patients with suspected orbital mass lesions included in our study for a period of 24 months are shown in [Table 1]. It was observed that patients belonged to all age groups ranging from 1 year to 73 years with a mean age of 32.93 years. We found a female preponderance in our study with 29 (66%) female patients and 15 (34%) male patients. There were 31 (70%) benign orbital lesions and 13 (30%) malignant lesions.

Table 1 Demographics and baseline characteristics

Characteristics	Categories	Descriptive statistics
Abbreviation: SD, standard deviation.
Age in years [mean ± SD]		32.93 ± 22.94
Gender [No. (%)]	Male	15 (34.09)
	Female	29 (65.91)
Tumor type [No. (%)]	Benign	31 (70.45)
	Malignant	13 (29.54)

The final diagnosis was established by histopathological examination in majority of the cases (n = 34) or based on the patient’s therapeutic response to antibiotics, anti-inflammatory drugs, and use of corticosteroids (n = 10). [Table 2] illustrates the correlation between the imaging and clinical/histopathology findings.

Table 2 Orbital MRI and clinical*/histopathological correlation (C/HPR)

C/HPR	Benign (n = 31) [No. (%)]	Malignant (n= 13) [No. (%)]	MRI
Abbreviation: MRI, magnetic resonance imaging. *Final diagnosis based on therapeutic response (n = 10).
Benign (n = 27)	26 (83.87)	1 (7.69)
Malignant (n = 17)	5 (16.12)	12 (92.30)

[Table 3] demonstrates the diffusion-weighted appearances of different orbital lesions. It shows that the distributions of DWI appearances of benign and malignant lesions were significantly different from each other (p = 0.002).

Table 3 Distribution of orbital tumors according to appearance on DWI

Classification of tumor	DWI [No. (%)]			p
	Isointense	Hypointense	Hyperintense
Abbreviation: DWI, diffusion-weighted imaging. aObtained using Chi-square test.
Benign (n = 31)	2 (6.45)	16 (51.61)	13 (41.93)	0.002a
Malignant (n = 13)	6 (46.15)	1 (7.69)	6 (46.15)

[Table 4] shows the distribution of all the benign and malignant orbital masses encountered in our study along with their mean ADC values. The mean ADC for the benign group was 1.47 ± 0.28 × 10⁻³ mm²/s and that for the malignant group was 0.94 ± 0.210⁻³ mm²/s. The difference between the mean ADC values for the two groups was found to be statistically significant (p < 0.00001).

Table 4 Descriptive statistics for ADC value according to the type of orbital tumor

Type of tumor (No.)	Minimum ADC ( 10−3 mm2/s)	Maximum ADC ( 10−3 mm2/s)	ADC (10−3 mm2/s) (mean ± SD)
Abbreviation: ADC, apparent diffusion coefficient.
Benign (31)	0.94	2.04	1.47 ± 0.28
Vascular (8)	1.37	1.71	1.55 ± 0.12
Inflammatory (7)	0.94	2.04	1.57 ± 0.36
Optic nerve sheath complex Neoplasm (7)	0.98	1.73	1.35 ± 0.25
Infection (4)	0.94	1.43	1.20 ± 0.24
Peripheral nerve sheath tumor (3)	1.33	1.42	1.38 ± 0.04
Lacrimal gland neoplasm (1)	1.78	1.78	1.78
Congenital (1)	1.92	1.92	1.92
Malignant (13)	0.63	1.25	0.94 ± 0.21
Metastases (6)	0.97	1.25	1.14 ± 0.11
Primary ocular malignancy (3)	0.71	0.83	0.77 ± 0.06
Lymphoproliferative (2)	0.63	0.78	0.71 ± 0.11
Others (2)	0.84	0.90	0.87 + 0.04

We also assessed the risk of malignancy with respect to the age of the patient, gender, and size of the lesion using the odds ratio as shown in [Table 5]. We found 23 lesions (52%) greater than 15 mm in size.

Table 5 Risk of orbital malignancy associated with different characteristics

Characteristics	Categories	Benign [No. (%)]	Malignant [No. (%)]	Odds ratio	Lower limit	Upper limit	p a
Note: Bold p-values indicate statistical significance. *Obtained using Chi-square test.
Age (y)	≤20 (n = 16)	13 (81.25)	3 (18.75)	1.000
	21–40 (n = 12)	11 (91.66)	1 (8.33)	0.436	0.014	4.381	0.435
	41–60 (n = 10)	4 (40.00)	6 (60.00)	5.841	1.020	42.476	0.031
	61–80 (n = 6)	3 (50.00)	3 (50.00)	3.959	0.476	36.541	0.413
Gender	Male (n = 15)	10 (66.67)	5 (33.33)	1.000
	Female (n = 29)	21 (72.41)	8 (27.59)	0.763	0.193	3.178	0.692
Lesion size	<10 mm (n = 16)	7 (43.75)	9 (56.25)	1.000
	11–15 mm (n = 5)	4 (80.00)	1 (20.00)	0.226	0.007	2.104	0.156
	>15 mm (n =23)	20 (86.95)	3 (13.05)	0.128	0.022	0.574	0.004

[Figs. 1 ] [2] [3] [4 ] show the imaging appearances of some of the benign and malignant orbital masses in our study.

Fig. 1 (A–E) Retinoblastoma is seen as a hypointense right intraocular mass lesion on T2WI (A) with blooming on SWI (B). There is an extension to the right optic nerve reaching up to the orbital apex. The lesion shows avid heterogeneous enhancement on contrast-enhanced T1WI (C). DWI (D) and ADC maps (E) show diffusion restriction. ADC, apparent diffusion coefficient; DWI, diffusion-weighted imaging; SWI, susceptibility-weighted imaging; T1WI, T1-weighted imaging.

Fig. 2 (A-D) Optic nerve glioma is seen as a fusiform enlargement of the intraorbital segment of left optic nerve. It appears isointense on T2WI (A) with avid enhancement on contrast-enhanced T1W image (B). DWI (C) and ADC maps (D) demonstrate signal intensity similar to normal gray matter suggestive of benign lesion. ADC, apparent diffusion coefficient; DWI, diffusion-weighted imaging; T1WI, T1-weighted imaging.

Fig. 3 (A-D) Lymphangioma is seen involving both intraconal and extraconal compartments of the right orbit with typical fluid–fluid levels on T2W (A) image. There is a mild patchy enhancement in the contrast-enhanced T1W image (B). DWI (C) and ADC maps (D) suggest the benign nature of the lesion. ADC, apparent diffusion coefficient; DWI, diffusion-weighted imaging; T1W, T1 weighted; T2W, T2-weighted imaging.

Fig. 4 (A-D) Lymphoma is seen as a mild T2 hypointense mass (A) involving the extraconal compartment of the left orbit with extension to the left ethmoid sinus. There is heterogeneous enhancement in the contrast-enhanced T1W image (B). DWI (C) and ADC maps (D) show restricted diffusion. ADC, apparent diffusion coefficient; DWI, diffusion-weighted imaging; T1W, T1 weighted.

ROC curve analysis was performed to calculate the optimal ADC threshold for the diagnosis of malignancy. The area under the curve was measured as 0.938 and the optimal ADC threshold was identified at the point of maximum accuracy. [Fig. 5] shows that 1.26 × 10⁻³ mm²/s was identified as the optimal cut-off value of ADC for characterizing orbital masses with a sensitivity of 100%, a specificity of 80.65%, and an accuracy of 86.36%.

Fig. 5 Receiver operating characteristic curve analysis of ADC values to calculate an optimal ADC threshold to distinguish between benign and malignant orbital lesions. ADC, apparent diffusion coefficient.

The sensitivity and specificity at different ADC threshold values are shown in [Table 6]. It was seen that an ADC threshold of less than 1.04 × 10⁻³ mm²/s was more than 90% specific for malignancy, and an ADC threshold of less than 1.22 × 10⁻³ mm²/s was more than 90% sensitive for malignancy.

Table 6 Descriptive statistics for different ADC threshold values (n = 44)

ADC threshold value ( 10−3 mm2/s)	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)	Accuracy (%)
Abbreviation: ADC, apparent diffusion coefficient; NPV, negative predictive value; PPV, positive predictive value. aOptimal cut-off obtained using Youden’s index.
1.04	69.20	90.00	75.00	87.50	84.09
1.22	92.30	83.80	70.59	96.30	86.36
1.26a	100.00	80.65	68.42	100.00	86.36

Discussion

DWI uses the diffusivity of moving water protons in different tissues to produce image contrast and thus provides useful information for the assessment of pathological tissues.[12] [18] Malignant lesions demonstrate restricted diffusion due to higher cellularity and higher nuclear–cytoplasmic ratio within the cells. There are few studies reported in the literature evaluating the role of quantitative diffusion-weighted images in differentiating benign and malignant orbital masses. Some of the earlier studies have suggested that DWI has a role in the prediction of malignancy in orbital masses, differentiation of orbital lymphoma from pseudotumors, evaluation of orbital cellulitis, identification of orbital abscess, and assessment of optic nerve lesions and infarct.[8] [17] [19] [20] [21] [22]

To the best of our knowledge, this is the first study evaluating the role of DWI and ADC values in orbital mass lesions among the Indian population. In our study, a total of 44 patients clinically suspected of having an orbital mass lesion underwent conventional MRI along with DWI with an aim of evaluating the role of DWI and ADC values in differentiating benign and malignant orbital lesions.

In our study, we found that majority of the patients had benign orbital masses (70%) and only one-third had malignant masses (30%). Our results were in agreement with the studies reported in the literature by various researchers. Shields et al reported 64% benign and 36% malignant lesions in a retrospective study.[23] Hemat also found 68% benign orbital lesions while 32% malignant lesions in his study.[17] However, Ro et al reported almost equal distribution of benign and malignant orbital lesions in one of their studies.[24]

Histopathological examination was done to attain the final diagnosis in 34 cases. The final diagnosis was established from the clinical course of the disease after medical management in the remaining 10 patients. Imaging and clinical/histopathology findings were in agreement in 84% benign cases and 92% malignant cases.

In our study, the majority of the benign lesions were vascular in origin consisting of lymphangioma (5), hemangioma (2), and orbital varix (1). Inflammatory masses included pseudotumor (3), granulomatosis with polyangiitis (2), and a solitary case of Erdheim–Chester disease and immunoglobulin G 4 (IgG4)-related disease each. Benign optic nerve sheath complex lesions composed of six cases of optic nerve glioma and a case of optic nerve sheath meningioma. Orbital infections included a case of endophthalmitis, panophthalmitis, orbital abscess, and orbital cellulitis with subretinal abscess each. Peripheral nerve sheath tumors included two cases of schwannoma and a case of neurofibroma. Other benign lesions were a solitary fibrous tumor of the lacrimal gland and a dermoid. Choroid metastases (6) were the most common malignant lesion in our study followed by retinoblastoma (3). Other malignant lesions included a malignant optic glioma, an aggressive optic nerve sheath meningioma, and two lymphomas. In a similar study published by Soliman et al (n = 50), the descriptive analysis of benign and malignant lesions was not mentioned.[25]

From the diffusion-weighted images, we found that 52% of benign lesions appeared hypointense, while 42% appeared hyperintense and 6% appeared isointense. Among the malignant lesions, diffusion hyperintensity was seen in 46% cases, 46% were isointense, and 8% were hypointense. The distributions of DWI appearances of benign and malignant lesions were found to be significantly different from each other (p = 0.002).

In our study, mean ADC values for benign and malignant orbital lesions were found to be 1.47 ± 0.28 × 10⁻³ and 0.94 ± 0.2 × 10⁻³ mm²/s, respectively, suggesting a significant difference between them (p < 0.00001). This was in agreement with the results of Razek et al, Sepahdari et al, Fatima et al, Şerifoğlu et al, and Hemat who also reported that malignant lesions had much lower ADC values as compared with benign lesions.[8] [11] [14] [17] [22] The malignant cases of optic nerve glioma and optic nerve sheath meningioma both displayed areas of diffusion restriction with lower ADC values as compared with the benign ones.

Also due to the small size of the choroid metastasis, some amount of partial volume averaging was present while measuring their ADC values. This is likely to result in falsely elevated ADC values for choroid metastasis and hence broader distribution of ADC values for malignant lesions. Thus, the mean ADC value for malignant lesions was also calculated after excluding choroid metastasis and was found to be 0.78 ± 0.09 × 10⁻³ mm/s. Sepahdari and Kapur et al also observed the effect of partial volume averaging on ADC values and concluded that ADC values for retinoblastoma had a strong inverse correlation with the size of the lesion.[12]

In our study, we observed a significant correlation of the age of the patient (p = 0.031) and the lesion size (p = 0.004) with respect to the risk of malignancy. However, this needs validation in a larger group due to the small sample size of our study.

In this study, ROC curve analysis identified 1.26 × 10⁻³ mm²/s as the optimal ADC cut-off value for characterizing orbital masses with a sensitivity of 100%, specificity of 80.65%, and accuracy of 86.36%. Hemat and Razek et al used ADC cut-off values of 0.93 × 10⁻³ and 1.15 × 10⁻³ mm/s to differentiate between malignant and benign lesions with more than 80 and 90% accuracy, respectively.[8] [17] Sepahdari et al reported an ADC cut-off value of 1.0 × 10⁻³ mm²/s with 84% specificity; however, a more stringent cut-off value of 0.9 × 10⁻³ mm/s increased the specificity to 97% at the cost of reduced sensitivity.[13] Fatima et al used a lower ADC threshold value of 0.84 × 10⁻³ mm/s to distinguish malignant lesions from benign lesions with more than 80% sensitivity and specificity.[22] Roshdy et al did not calculate an optimal ADC cut-off value but observed an overlap between the ADC values of benign and malignant lesions.[26] Soliman et al reported an ADC cut-off value of 0.9 × 10⁻³ mm/s with 76% sensitivity and 96% specificity.[25]

Of the 31 benign lesions in our study, four displayed ADC values lower than the threshold. Orbital abscess and orbital cellulitis with subretinal abscess demonstrated ADC values lower than the cut-off due to the presence of inflammatory cells. Other benign lesions to show falsely low ADC values were IgG4-related disease and optic nerve sheath meningioma. The lymphoplasmacytic infiltrates seen in IgG4-related disease and increased cellularity in meningioma could account for this. The mean ADC value of benign lesions after excluding the false-negative cases was 1.54 ± 0.210⁻³ mm²/s, which was higher than the overall mean ADC of benign lesions (1.47 ± 0.28 × 10⁻³ mm²/s). However, ADC values of all the malignant lesions were lower than the threshold.

Different ADC threshold values were used in previous studies and each one provided a different sensitivity, specificity, and accuracy. Therefore, a single ADC cut-off value cannot be considered sensitive and specific to distinguish between malignant and benign masses. Hence, we used a two-threshold model to characterize the orbital masses based on ADC values. Orbital masses were divided into three categories: likely malignant masses having >90% probability of being malignant and an ADC value of less than 1.04 × 10⁻³ mm²/s (18% in our study); likely benign masses having >90% probability of being benign and an ADC value of more than 1.22 × 10⁻³ mm²/s (71% in our study); and indeterminate masses with an ADC value between 1.04 and 1.22 × 10⁻³ mm/s (11% in our study). This two-threshold model for characterization of orbital mass lesions was also used by Sepahdari et al and Hemat.[14] [17]

The limitation of this study includes partial volume averaging and susceptibility artifacts which were observed in choroid metastasis with the potential to impact the measured ADC values. The role of improved EPI and non-EPI DWI techniques should be explored to overcome this challenge. Another limitation of our study was the small sample size. Hence, we recommend prospective studies with a larger sample size to validate our results.

Conclusion

We conclude that malignant orbital lesions have a significantly lower ADC value as compared with benign lesions. Two ADC thresholds were used to characterize most of the orbital masses with more than 90% confidence. Therefore, DWI with ADC is a quantitative, noninvasive tool adjunct to conventional MRI for the characterization of benign and malignant nature of the orbital masses.

Conflicts of Interest

There are no conflicts of interest.

Financial Support and Sponsorship

Nil.

• References

• 1 Purohit BS, Vargas MI, Ailianou A. et al. Orbital tumours and tumour-like lesions: exploring the armamentarium of multiparametric imaging. Insights Imaging 2016; 7 (01) 43-68
• 2 O’Neill M, Vega G, Cofnas P, Nagornaya N, Bhatia R. CT and MRI imaging spectrum of orbital masses: a pictorial essay and review of the literature. Cureus 2012; 4: 279
• 3 Mafee MF. Orbit: embryology, anatomy and pathology. In: Som PM, Curtin HD, eds. Head and Neck Imaging. 4th ed.. St Louis, MI: Mosby 2003; 529-654
• 4 Aviv RI, Casselman J. Orbital imaging: part 1. Normal anatomy. Clin Radiol 2005; 60 (03) 279-287
• 5 Goh PS, Gi MT, Charlton A. Tan C, Gangadhara Sundar JK, Amrith S. Review of orbital imaging. Eur J Radiol 2008; 66 (03) 387-395
• 6 Dunarintu S, Birsasteanu F, Onet D. et al. Radio-imaging diagnosis of the ocular and orbital tumours. J Exp Med Surg Res 2008; 15: 5-12
• 7 Tailor TD, Gupta D, Dalley RW, Keene CD, Anzai Y. Orbital neoplasms in adults: clinical, radiologic, and pathologic review. Radiographics 2013; 33 (06) 1739-1758
• 8 Razek AA, Elkhamary S, Mousa A. Differentiation between benign and malignant orbital tumors at 3-T diffusion MR-imaging. Neuroradiology 2011; 53 (07) 517-522
• 9 Wang J, Takashima S, Takayama F. et al. Head and neck lesions: characterization with diffusion-weighted echo-planar MR imaging. Radiology 2001; 220 (03) 621-630
• 10 Eida S, Sumi M, Sakihama N, Takahashi H, Nakamura T. Apparent diffusion coefficient mapping of salivary gland tumors: prediction of the benignancy and malignancy. AJNR Am J Neuroradiol 2007; 28 (01) 116-121
• 11 Şerifoğlu İ, Oz İİ, Damar M, Tokgöz Ö, Yazgan Ö, Erdem Z. Diffusion-weighted imaging in the head and neck region: usefulness of apparent diffusion coefficient values for characterization of lesions. Diagn Interv Radiol 2015; 21 (03) 208-214
• 12 Kapur R, Sepahdari AR, Mafee MF. et al. MR imaging of orbital inflammatory syndrome, orbital cellulitis, and orbital lymphoid lesions: the role of diffusion-weighted imaging. AJNR Am J Neuroradiol 2009; 30 (01) 64-70
• 13 Sepahdari AR, Kapur R, Aakalu VK, Villablanca JP, Mafee MF. Diffusion-weighted imaging of malignant ocular masses: initial results and directions for further study. AJNR Am J Neuroradiol 2012; 33 (02) 314-319
• 14 Sepahdari AR, Politi LS, Aakalu VK, Kim HJ, Razek AA. Diffusion-weighted imaging of orbital masses: multi-institutional data support a 2-ADC threshold model to categorize lesions as benign, malignant, or indeterminate. AJNR Am J Neuroradiol 2014; 35 (01) 170-175
• 15 Wang Y, Xiao LH. Orbital schwannomas: findings from magnetic resonance imaging in 62 cases. Eye (Lond) 2008; 22 (08) 1034-1039
• 16 Xian J, Zhang Z, Wang Z. et al. Evaluation of MR imaging findings differentiating cavernous haemangiomas from schwannomas in the orbit. Eur Radiol 2010; 20 (09) 2221-2228
• 17 Hemat EM. Characterization of orbital masses by diffusion-weighted magnetic resonance imaging (DWI) and apparent diffusion coefficient (ADC) value. Egypt J Radiol Nucl Med 2016; 48: 115-123
• 18 Politi LS, Forghani R, Godi C. et al. Ocular adnexal lymphoma: diffusion-weighted mr imaging for differential diagnosis and therapeutic monitoring. Radiology 2010; 256 (02) 565-574
• 19 Sepahdari AR, Aakalu VK, Setabutr P, Shiehmorteza M, Naheedy JH, Mafee MF. Indeterminate orbital masses: restricted diffusion at MR imaging with echo-planar diffusion-weighted imaging predicts malignancy. Radiology 2010; 256 (02) 554-564
• 20 Lope LA, Hutcheson KA, Khademian ZP. Magnetic resonance imaging in the analysis of pediatric orbital tumors: utility of diffusion-weighted imaging. J AAPOS 2010; 14 (03) 257-262
• 21 Ragheb AS, Abd El-Rahman HM, Khattab HA. Can DWI & ADC differentiate orbital lymphoma, non-specific orbital inflammation and orbital cellulitis?. Egypt J Radiol Nucl Med 2012; 43: 157-164
• 22 Fatima Z, Ichikawa T, Ishigame K. et al. Orbital masses: the usefulness of diffusion-weighted imaging in lesion categorization. Clin Neuroradiol 2014; 24 (02) 129-134
• 23 Shields JA, Shields CL, Scartozzi R. Survey of 1264 patients with orbital tumors and simulating lesions: the 2002 Montgomery Lecture, part 1. Ophthalmology 2004; 111 (05) 997-1008
• 24 Ro SR, Asbach P, Siebert E, Bertelmann E, Hamm B, Erb-Eigner K. Characterization of orbital masses by multiparametric MRI. Eur J Radiol 2016; 85 (02) 324-336
• 25 Soliman A, Aggag M, Abdelgawwad A, Aly W, Yossef A. Role of diffusion weighted MRI in evaluation of orbital lesions. Al-Azhar Int Med J 2020; 1: 331-336
• 26 Roshdy N, Shahin M, Kishk H. et al. MRI in diagnosis of orbital masses. Curr Eye Res 2010; 35 (11) 986-991

Publication History

Article published online:
23 May 2021