Original Research
Neuroradiology/Head and Neck Imaging
February 28, 2017

Differentiation of Benign and Malignant Head and Neck Lesions With Diffusion Tensor Imaging and DWI

Abstract

OBJECTIVE. The purpose of this study was to determine whether diffusion tensor imaging (DTI) can be used to differentiate between benign and malignant head and neck lesions.
MATERIALS AND METHODS. This retrospective study included patients with head and neck lesions who underwent clinical MRI at 1.5 or 3 T with DWI or DTI parameters. ROI analysis was performed, with lesion-to-medulla apparent diffusion coefficient (ADC) ratios generated.
RESULTS. Sixty-five patients with head and neck lesions were included (71 benign, 40 malignant). Twenty-one patients had multiple lesions. Statistically significant differences (p < 0.001) were seen in the mean ADC values ± SD of malignant and benign lesions (0.55 × 10−3 ± 0.14 × 10−3 mm2/s vs 0.89 × 10−3 ± 0.29 × 10−3 mm2/s, respectively) and in the mean ADC ratios of malignant and benign lesions (0.88 ± 0.21 vs 1.40 ± 0.44, respectively) with DTI parameters. DTI and DWI parameters produced similar mean ADC ratio values for malignant (0.88 ± 0.21 and 0.92 ± 0.54, respectively) and benign lesions (1.40 ± 0.44 and 1.79 ± 0.52, respectively). ADC ratio thresholds for predicting malignancy for DTI (ADC ratio ≤ 1) and DWI (ADC ratio ≤ 0.94) were also similar.
CONCLUSION. DTI is a useful predictor of malignancy for head and neck lesions, with ADC values of malignant lesions significantly lower than those of benign lesions. DTI ADC values were lower than DWI ADC values for all head and neck lesions in our study group, often below reported malignant DWI threshold values. Normalization of ADC values to an internal control resulted in similar ADC ratios on DWI and DTI.
DWI can serve as a useful adjunct to routine MRI pulse sequences in the evaluation of head and neck lesions. Multiple prior studies investigating the role of DWI in head and neck imaging have found significant differences in apparent diffusion coefficient (ADC) values of benign and malignant lesions, with malignant lesions tending to have lower ADC values [15]. Prior work has shown the utility of DWI for discriminating between head and neck squamous cell carcinoma (HNSCC) and lymphoma, staging HNSCC, differentiating treatment changes from tumor recurrence, and predicting early response to therapy in HNSCC [69]. However, knowledge gaps remain, particularly with regard to the widespread implementation of DWI for routine head and neck clinical use. Published threshold values for predicting benignity and malignancy of head and neck lesions are predicated on ADC values obtained with routine DWI parameters. No published studies to our knowledge have investigated the application of diffusion tensor imaging (DTI) parameters for the evaluation of head and neck lesions. Commonly employed in brain imaging, these DTI parameters include higher b values and higher numbers of diffusion probing gradients.
In this study, we examined the utility of ADC values obtained with DTI parameters for the evaluation of head and neck lesions. We compared the mean ADC values and a normalized lesion-to-medulla ADC ratio of benign and malignant head and neck lesions obtained with both DWI and DTI parameters, offering useful clinical threshold values for predicting benignity.

Materials and Methods

This retrospective review of more than 9 years of data (January 2005 through August 2014) from head and neck imaging archives at the University of Utah was approved by the institutional review board and complied with HIPAA guidelines. We included patients imaged at the University of Utah or referred with MR images from outside facilities, including routine clinical head and neck MRI using DWI or DTI parameters (or both). For purposes of this study, we defined DWI parameters as b values of 0 and 1000 s/mm2 and three orthogonal directions of diffusion probing gradients, as commonly reported in multiple prior studies. DTI parameters are defined as b values of 0 and 2000 s/mm2 with 20 directions of diffusion probing gradients. Diffusion imaging was included as part of clinical MRI protocols for imaging of the brain, temporal bones, orbits, paranasal sinuses, skull base, and soft tissues of the neck. All imaging was performed at either 1.5- or 3-T field strength, including Siemens Healthcare (Avanto, Aera, Harmony, TrioTim, Espree, and Verio), GE Healthcare (Genesis Signa, Signa Excite, and Signa HDxt), and Philips Healthcare (Achieva, Ingenia, and Intera) scanners. Diffusion parameters, magnetic field strengths, matrix sizes, and scanner models were documented after systematic review of the DICOM metadata. DWI and DTI sequences were performed in the axial plane using a standard 8-channel head and neck coil. Matrix sizes varied between scanners, typically 128 × 128 or 192 × 192 with a 230 × 230 mm FOV. Diffusion sequences were performed in addition to routine head and neck pulse sequences, which were available for review by the scoring radiologists.
We included patients who were at least 18 years old with histopathologically or clinicoradiologically proven head and neck lesions and MRI with DWI, DTI, or both. Histopathology was documented through query of the electronic medical record. Clinicoradiologic diagnoses were determined by combination of provided clinical history (including previously established diagnoses from outside institutions before referral to our institution), review of the electronic medical record, and accepted pathognomic imaging findings. For example, in some cases the imaging diagnosis of fibrous dysplasia was made on the basis of polyostotic expansile bone changes with characteristic ground-glass and cystic attenuation on CT (Fig. 1), and the diagnosis of paraganglioma was made on the basis of characteristic location within the carotid space and a classic salt-and-pepper signal intensity on T1-weighted imaging and T2-weighted imaging. We excluded patients with prior chemo-therapy, prior radiation therapy, studies limited by artifacts (per consensus of reviewers), unknown disease or abnormality, and lesions smaller than 1 cm. Using these criteria, we identified 65 patients (36 male, 29 female) with 111 target lesions, including 71 benign and 40 malignant. Twenty-one patients had multiple lesions.
Fig. 1A —42-year-old woman with fibrous dysplasia.
A, Axial contrast-enhanced T1-weighted MR image with fat saturation shows heterogeneously enhancing mass involving clivus and sphenoid bone.
Fig. 1B —42-year-old woman with fibrous dysplasia.
B, Axial CT scan shows polyostotic involvement of central skull base and occipital bone with characteristic areas of osseous expansion, ground-glass attenuation, and cystic change typical of fibrous dysplasia.
Fig. 1C —42-year-old woman with fibrous dysplasia.
C, Central skull base involvement is relatively dark on diffusion-tensor imaging trace (C) but has relatively high signal intensity on apparent diffusion coefficient (ADC) map (D). Evaluation of 3.3-cm2 freehand ROI (dotted line, D) revealed mean ADC value of 1.63 × 10−3 ± 0.28 × 10−3 mm2/s. Ratio of ADC values of freehand ROI and medulla ROI (not shown) was 1.65. These findings suggest benign lesion.
Fig. 1D —42-year-old woman with fibrous dysplasia.
D, Central skull base involvement is relatively dark on diffusion-tensor imaging trace (C) but has relatively high signal intensity on apparent diffusion coefficient (ADC) map (D). Evaluation of 3.3-cm2 freehand ROI (dotted line, D) revealed mean ADC value of 1.63 × 10−3 ± 0.28 × 10−3 mm2/s. Ratio of ADC values of freehand ROI and medulla ROI (not shown) was 1.65. These findings suggest benign lesion.
Freehand ROI evaluation was performed by consensus of two board- and subspecialty-certified neuroradiologists with 20 and 2 years of experience (Figs. 1 and 2). All ROI evaluation was performed at an iSite PACS workstation (software version 2006, Koninklijke Philips Healthcare). ROIs spared the peripheral 2 mm of lesions, excluding cystic and necrotic portions, as described by Srinivasan et al. [1]. Additional ROI evaluation of the medulla was performed at the level of the foramina of Luschka. To account for variation in ADC values among scanners, magnetic field strengths (which can lower ADC values as they increase), and matrix sizes, a normalized ADC ratio was calculated by dividing the mean ADC value of the lesion by the mean ADC value of the medulla.
Fig. 2A —48-year-old man with sinonasal neuroendocrine carcinoma.
A, Axial contrast-enhanced T1-weighted image with fat saturation shows heterogeneously enhancing transspatial mass centered in left maxillary sinus with transosseous invasion into retromaxillary fat pad and infrazygomatic masticator space, as well as into nasal cavity.
Fig. 2B —48-year-old man with sinonasal neuroendocrine carcinoma.
B, Mass appears bright on diffusion-tensor imaging trace image (B) and dark on corresponding apparent diffusion coefficient (ADC) map (C), consistent with reduced diffusivity in high cellularity lesion.
Fig. 2C —48-year-old man with sinonasal neuroendocrine carcinoma.
C, Mass appears bright on diffusion-tensor imaging trace image (B) and dark on corresponding apparent diffusion coefficient (ADC) map (C), consistent with reduced diffusivity in high cellularity lesion.
Fig. 2D —48-year-old man with sinonasal neuroendocrine carcinoma.
D, Evaluation of freehand ROIs (dotted lines) on ADC map shows mass (top ROI: 10.3 cm2, 0.39 × 10−3 ± 0.07 × 10−3 mm2/s) has significantly lower signal intensity than medulla used as internal control (bottom ROI: 0.7 cm2, 0.66 × 10−3 ± 0.13 × 10−3 mm2/s), resulting in ADC ratio of 0.59. This finding is suspicious for malignant lesion.
Mean ADC values and ADC ratios between benign and malignant groups were compared using an unpaired t test with unequal variance. A p value less than 0.05 was considered statistically significant. Additional ROC and cut-point analysis was performed using commercial statistical analysis software (Stata, StataCorp) to determine sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV).

Results

The most common benign lesions encountered in our study were paraganglioma (n = 12), meningioma (n = 10), oncocytoma (n = 9), neurofibroma (n = 8), schwannoma (n = 7), fibrous dysplasia (n = 5), and cavernous hemangioma of the orbit (n = 4). The most common malignant lesions included systemic metastases (n = 10; renal cell carcinoma, carcinoid, breast carcinoma, and prostate carcinoma), lymphoma (n = 6), myeloma or plasmacytoma (n = 6), melanoma (n = 3), adenoid cystic carcinoma (n = 3), squamous cell carcinoma (n = 2), sinonasal undifferentiated carcinoma (n = 2), and sinonasal neuroendocrine carcinoma (n = 2). Histopathologic diagnosis was made in 79 lesions; clinicoradiologic diagnosis was made in 32 lesions.
In our study, 46 patients had MRI that included DTI and 30 patients had MRI that included DWI. Significant differences were found between mean ADC values of benign and malignant lesions using DTI parameters (0.89 × 10−3 ± 0.29 × 10−3 mm2/s vs 0.55 × 10−3 ± 0.14 × 10−3 mm2/s, respectively) and DWI parameters (1.43 × 10−3 ± 0.39 × 10−3 mm2/s vs 0.77 × 10−3 ± 0.53 × 10−3 mm2/s, respectively) (Table 1). Significant differences were also found between mean ADC ratios of benign and malignant lesions using DTI parameters (1.40 ± 0.44 vs 0.88 ± 0.21, respectively) and DWI parameters (1.79 ± 0.52 vs 0.92 ± 0.54, respectively) (Table 1).
TABLE 1: Apparent Diffusion Coefficient (ADC) Values and Ratios of Benign Versus Malignant Lesions Using Diffusion-Tensor Imaging (DTI) and DWI Parameters
ParameterMean ADC Value ± SD (× 10−3 mm2/s)Mean ADC Ratio ± SD
DTI  
 Benign0.89 ± 0.29 (0.80-0.99)1.40 ± 0.44 (1.26-1.54)
 Malignant0.55 ± 0.14 (0.50-0.60)0.88 ± 0.21 (0.81-0.95)
pa< 0.001< 0.001
DWI  
 Benign1.43 ± 0.39 (1.29-1.56)1.79 ± 0.52 (1.61-1.97)
 Malignant0.77 ± 0.53 (0.41-1.13)0.92 ± 0.54 (0.56-1.28)
pa0.00290.0004

Note—Values in parentheses are 95% CI. A p value less than 0.05 was considered statistically significant.

a
Calculated using t test.
To assess the accuracy of mean ADC value and ADC ratio for predicting benignity, we performed ROC analysis. For mean ADC value, ROC analysis yielded AUCs of 0.89 with DTI parameters and 0.87 with DWI parameters (Fig. 3). For mean ADC ratio, ROC analysis yielded AUCs of 0.88 using DTI parameters and 0.99 using DWI parameters (Fig. 3). For predicting malignancy, cut-point analysis found optimal mean ADC threshold values of ≤ 0.701 × 10−3 mm2/s (sensitivity, 90.9%; specificity, 73.7%; PPV, 75.0%; NPV, 90.3%) and ≤ 0.832 × 10−3 mm2/s (sensitivity, 90.0%; specificity, 85.7%; PPV, 64.3%; NPV, 96.8%) for DTI and DWI parameters, respectively. For ADC ratios, cut-point analysis found optimal mean threshold values of ≤ 1 (sensitivity, 79.3%; specificity, 89.5%; PPV, 85.2%; NPV, 85.0%) and ≤ 0.94 (sensitivity, 90.0%; specificity, 94.3%; PPV, 81.8%; NPV, 97.1%) for DTI and DWI parameters, respectively.
Fig. 3A —ROC curve analysis of different methods for predicting benignity of head and neck lesions.
A, Mean diffusion-tensor imaging (DTI) apparent diffusion coefficient (ADC) value. AUC = 0.89.
Fig. 3B —ROC curve analysis of different methods for predicting benignity of head and neck lesions.
B, Mean DWI ADC value. AUC = 0.87.
Fig. 3C —ROC curve analysis of different methods for predicting benignity of head and neck lesions.
C, DTI ADC ratio. AUC = 0.88.
Fig. 3D —ROC curve analysis of different methods for predicting benignity of head and neck lesions.
D, DWI ADC ratio. AUC = 0.99.

Discussion

In this study, we found a statistically significant difference in mean ADC values between benign and malignant lesions using both DWI and DTI parameters. For DWI parameters, this result corroborates those published by other authors [1, 36]. We also investigated the application of DTI parameters for the assessment of head and neck lesions, which has not been reported in the literature to our knowledge. Our results show that DTI ADC values for malignant lesions tend to be lower than those for benign lesions, mirroring the trend reported with DWI ADC values. However, our results show that ADC values obtained with DTI parameters are universally lower than ADC values obtained with DWI parameters, often lower than previously reported malignant threshold values [1, 3, 5, 8]. We also describe a method of normalizing mean ADC values using the medulla as an internal control to account for differences in scanners, field strengths, and matrix sizes encountered in clinical practice.
Variability in reported mean ADC values and discriminatory threshold values has limited the routine application of diffusion head and neck imaging in clinical practice. In an early study, Wang et al. [3] examined ADC values of 97 head and neck lesions obtained with DWI parameters at 1.5-T field strength. Their study found statistically significant differences in ADC values among lymphomas, carcinomas, benign solid tumors, and benign cystic lesions [3]. In our study, lymphomas had the lowest mean ADC value (0.66 × 10−3 ± 0.17 × 10−3 mm2/s), which was significantly less than the mean ADC value of carcinomas (1.13 × 10−3 ± 0.43 × 10−3 mm2/s) (p < 0.001). The mean ADC value of carcinomas was significantly less than that of benign solid tumors (1.56 × 10−3 ± 0.51 × 10−3 mm2/s) (p = 0.002), which in turn was significantly less than the mean ADC value of benign cystic lesions (2.05 × 10−3 ± 0.62 × 10−3 mm2/s) (p = 0.035). After performing ROC analysis of their dataset, Wang et al. suggested a threshold value less than 1.22 × 10−3 mm2/s was appropriate for predicting malignancy of a generic head and neck lesion (accuracy, 86%; sensitivity, 84%; specificity, 91%).
Subsequent work by Srinivasan and colleagues [1] found significant differences in mean ADC values of benign and malignant head and neck lesions imaged at 3-T field strength. In their study, 33 patients with 17 benign and 16 malignant lesions were imaged using DWI parameters, yielding a statistically significant difference in mean ADC value between benign (1.505 × 10−3 ± 0.487 × 10−3 mm2/s) and malignant lesions (1.071 × 10−3 ± 0.293 × 10−3 mm2/s) (p = 0.004). As in other studies, Srinivasan et al. found some overlap between mean ADC values of benign and malignant lesions, but it was outside of the 95% CI in their study group. On the basis of these data, they suggested ADC values less than 1.3 × 10−3 mm2/s predicted malignancy.
Prior studies have also reported the ability of DWI to predict benignity at specific sub-sites within the head and neck, including the orbit, skull base, thyroid, and lymph nodes. For orbital tumors, Abdel Razek and colleagues [10] reported a statistically significant difference in mean ADC values between benign (1.57 × 10−3 ± 0.33 × 10−3 mm2/s) and malignant tumors (0.84 × 10−3 ± 0.34 × 10−3 mm2/s) (p = 0.001) in a retrospective study of 47 patients imaged at 3-T field strength. They suggested a threshold value of 1.15 × 10−3 mm2/s for differentiating between benign and malignant lesions of the orbit (accuracy, 93%; sensitivity, 95%; specificity, 91%).
In in a separate retrospective study of 45 patients with lesions in the skull base, Abdel Razek et al. [11] found a statistically significant difference in mean ADC values between benign (1.63 × 10−3 ± 0.29 × 10−3 mm2/s) and malignant lesions (1.002 × 10−3 ± 0.21 × 10−3 mm2/s) (p = 0.001). On the basis of those results, they suggested a threshold value of 1.3 × 10−3 mm2/s for differentiating between benign and malignant skull base lesions (AUC, 0.932; accuracy, 94%; sensitivity, 94%; specificity, 93%; PPV, 93%; NPV, 94%).
With regard to thyroid nodules, Bozgeyik and colleagues [12] examined images obtained at 1.5 T with b values of 100, 200, and 300 s/mm2 and found a statistically significant difference between mean ADC values of benign (3.06 × 10−3 ± 0.71 × 10−3, 1.80 × 10−3 ± 0.60 × 10−3, and 1.15 × 10−3 ± 0.43 × 10−3 mm2/s, respectively) and malignant nodules (0.96 × 10−3 ± 0.65 × 10−3, 0.56 × 10−3 ± 0.43 × 10−3, and 0.30 × 10−3 ± 0.20 × 10−3 mm2/s, respectively) (p < 0.05). The authors concluded that 300 s/mm2 was the optimal b value and suggested a threshold ADC value of 0.62 × 10−3 mm2/s for differentiating between benign and malignant thyroid nodules (AUC, 0.977; sensitivity, 90%; specificity, 100%) [12].
For cervical lymph nodes, the use of DWI remains controversial and confusing, both in discriminatory threshold values and in its general utility. Early work by Sumi et al. [4] reported that benign reactive lymph nodes have a significantly lower mean ADC value (0.302 × 10−3 ± 0.062 × 10−3 mm2/s) than do metastatic cervical lymph nodes (0.410 × 10−3 ± 0.105 × 10−3 mm2/s) (p < 0.01). Their study found that nodal lymphoma had an even lower mean ADC value (0.223 × 10−3 ± 0.056 × 10−3 mm2/s; p < 0.05) relative to benign lymphadenopathy. In a study evaluating cervical metastatic disease in HNSCC, de Bondt and colleagues [13] reported an optimal mean ADC value threshold of less than 1.0 × 10−3 mm2/s (sensitivity, 92.3%; specificity, 83.9%) for detecting metastatic cervical nodes and concluded that diagnostic criteria that include mean ADC value outperform those that are based solely on size, margin, and signal intensity of a lymph node (AUC = 0.98 vs AUC = 0.91, respectively). In a separate study of the utility of DWI in assessment of HNSCC metastatic disease, Vandecaveye and colleagues [7] found a statistically significant difference (p < 0.0001) between mean ADC values of benign cervical lymph nodes (1.19 × 10−3 ± 0.22 × 10−3 mm2/s) and HNSCC nodal metastases (0.85 × 10−3 ± 0.27 × 10−3 mm2/s) and suggested a threshold value of 0.94 × 10−3 mm2/s for differentiating between benign nodes and metastatic disease (accuracy, 91%; sensitivity, 84%; specificity, 94%). A study by Lim et al. [14] evaluated DWI findings in 26 consecutive patients with head and neck malignancy including 91 benign and 25 metastatic nodes; in this small study, no significant difference in mean ADC values was found between benign and metastatic nodes.
Another limitation in the widespread application of DWI has been the variability and reproducibility of ADC values between equipment and institutions [15, 16]. Kolff-Gart and colleagues [15] found that variation in ADC values is smallest when imaging is performed on the same equipment using the same MR pulse sequences and that mean ADC measurements of the spinal cord were the most precise and reproducible. Follow-up imaging consistently performed on the same scanners with the same MR pulse sequences would be the ideal situation for interpreting neuroradiologists, but reality dictates that variability in equipment and pulse sequences is common (if not the rule) both within an individual department and between institutions.
In an attempt to minimize the variability that equipment and scan parameters, including magnetic field strength and selected b values, have on mean ADC values, we used the medulla as an internal control by which a relative reduction in diffusivity could be standardized. We selected the medulla as our reference standard because it exhibits many of the same attributes that Kolff-Gart et al. [15] cited for their choice of the spinal cord as an internal control. The medulla is virtually always visualized within the FOV of any head and neck imaging study (including orbits and paranasal sinus MRI, which may exclude the spinal cord as a possible internal control). It is rarely affected by head and neck malignancy, is less commonly affected by chronic microvascular disease than other regions of brain parenchyma, and offers the easy reproducibility of a standard location for placing a control ROI (we employed the medulla at the level of the foramina of Luschka). Additionally, our study used data acquired on numerous scanner manufacturers and models, which parallels typical clinical practice. Lastly, our data show that a normalized ADC ratio is a useful predictor of benignity of a head and neck lesion that accounts for differences in mean ADC values resulting from differences in scanner make and model, magnetic field strength, and variable acquisition matrix size.
Our study had some limitations. First, it was retrospective; future prospective studies will be required to ascertain the true clinical utility of DTI in evaluating head and neck tumors. Second, although efforts were made to limit clinical bias, this study was not blinded, and reviewers had access to all routine pulse sequences. However, the ability to visualize lesions on diffusion and conventional MR sequences mirrors clinical practice. Third, some selection bias was present, including a disproportionate number of suprahyoid neck lesions in our study group and relative paucity of HNSCC (the most common head and neck malignancy), which is not commonly imaged with routine MRI at our institution. Fourth, we included only pretreatment lesions, so utility of DTI for differentiating residual or recurrent disease from posttreatment changes is unknown. Fifth, our study did not specifically look at nodal metastatic disease but rather at primary mucosal or parenchymal lesions and extranodal or osseous metastatic deposits. Lastly, we included data obtained from multiple MRI scanner manufacturers and models, yielding variability of mean ADC values, but this scenario also parallels clinical imaging trends and was accounted for by the use of a normalized ADC ratio.
In conclusion, both DWI and DTI may be useful in characterizing benignity of head and neck lesions. Given that mean ADC values derived from DTI parameters are characteristically lower than ADC values derived from DWI parameters (often lower than reported malignant threshold values), the two should not be used interchangeably. A potential workaround is the normalization of ADC values to an internal control (e.g., the medulla), thus creating an ADC ratio that has the potential to reduce false-positive results that could arise when using published mean ADC threshold values obtained with DWI parameters on studies employing DTI parameters. Our data indicate that an ADC ratio threshold value of approximately 1 is a useful discriminator for both DWI and DTI techniques in that lesions with an ADC ratio less than 1 were typically malignant. Lastly, we stress that diffusion pulse sequences are a complementary adjunct to routine, conventional MR pulse sequences but should not be viewed as a standalone imaging technique for predicting the benignity of head and neck lesions.

References

1.
Srinivasan A, Dvorak R, Perni K, et al. Differentiation of benign and malignant pathology in the head and neck using 3T apparent diffusion coefficient values: early experience. AJNR 2008; 29:40–44
2.
Srinivasan A, Mohan S, Mukherji SK. Biologic imaging of head and neck cancer: the present and the future. AJNR 2012; 33:586–594
3.
Wang J, Takashima S, Takayama F, et al. Head and neck lesions: characterization with diffusion-weighted echo-planar MR imaging. Radiology 2001; 220:621–630
4.
Sumi M, Sakihama N, Sumi T, et al. Discrimination of metastatic cervical lymph nodes with diffusion-weighted MR imaging in patients with head and neck cancer. AJNR 2003; 24:1627–1634
5.
Abdel Razek AA, Gaballa G, Elhawarey G, et al. Characterization of pediatric head and neck masses with diffusion-weighted MR imaging. Eur Radiol 2009; 19:201–208
6.
Maeda M, Kato H, Sakuma H, et al. Usefulness of the apparent diffusion coefficient in line scan diffusion-weighted imaging for distinguishing between squamous cell carcinomas and malignant lymphomas of the head and neck. AJNR 2005; 26:1186–1192
7.
Vandecaveye V, De Keyzer F, Vander Poorten V, et al. Head and neck squamous cell carcinoma: value of diffusion-weighted MR imaging for nodal staging. Radiology 2009; 251:134–146
8.
Abdel Razek AA, Kandeel AY, Soliman N, et al. Role of diffusion-weighted echo-planar MR imaging in differentiation of residual or recurrent head and neck tumors and posttreatment changes. AJNR 2007; 28:1146–1152
9.
Kim S, Loevner L, Quon H, et al. Diffusion-weighted magnetic resonance imaging for predicting and detecting early response to chemoradiation therapy of squamous cell carcinomas of the head and neck. Clin Cancer Res 2009; 15:986–994
10.
Razek AA, Elkhamary S, Mousa A. Differentiation between benign and malignant orbital tumors at 3-T diffusion MR-imaging. Neuroradiology 2011; 53:517–522
11.
Abdel Razek A, Mossad A, Ghonim M. Role of diffusion-weighted MR imaging in assessing malignant versus benign skull-base lesions. Radiol Med 2011; 116:125–132
12.
Bozgeyik Z, Coskun S, Dagli AF, et al. Diffusion-weighted MR imaging of thyroid nodules. Neuroradiology 2009; 51:193–198
13.
de Bondt RB, Hoeberigs MC, Nelemans PJ, et al. Diagnostic accuracy and additional value of diffusion-weighted imaging for discrimination of malignant cervical lymph nodes in head and neck squamous cell carcinoma. Neuroradiology 2009; 51:183–192
14.
Lim HK, Lee JH, Baek HJ, et al. Is diffusion-weighted MRI useful for differentiation of small non-necrotic cervical lymph nodes in patients with head and neck malignancies? Korean J Radiol 2014; 15:810–816
15.
Kolff-Gart AS, Pouwels PJ, Noij DP, et al. Diffusion-weighted imaging of the head and neck in healthy subjects: reproducibility of ADC values in different MRI systems and repeat sessions. AJNR 2015; 36:384–390
16.
Hoang JK, Choudhury KR, Chang J, et al. Diffusion-weighted imaging for head and neck squamous cell carcinoma: quantifying repeatability to understand early treatment-induced change. AJR 2014; 203:1104–1108

Information & Authors

Information

Published In

American Journal of Roentgenology
Pages: 1110 - 1115
PubMed: 28245145

History

Submitted: March 20, 2016
Accepted: October 24, 2016
Version of record online: February 28, 2017

Keywords

  1. apparent diffusion coefficient
  2. diffusion-tensor imaging
  3. DWI
  4. head and neck lesions

Authors

Affiliations

Nicholas A. Koontz
Department of Radiology, University of Utah Health Sciences Center, Salt Lake City, UT.
Department of Radiology and Imaging Sciences, Indiana University School of Medicine, 550 N University Blvd, Rm 0663, Indianapolis, IN 46202.
Richard H. Wiggins III

Notes

Address correspondence to N. A. Koontz ([email protected]).

Metrics & Citations

Metrics

Citations

Export Citations

To download the citation to this article, select your reference manager software.

Articles citing this article

View Options

View options

PDF

View PDF

PDF Download

Download PDF

Media

Figures

Other

Tables

Share

Share

Copy the content Link

Share on social media