Synopsis
In
this study, we evaluated the advantages of using high b-value high-resolution diffusion-weighted
MR imaging (DWI) to differentiate between malignant and benign thyroid nodules.
This prospective study included 28 consecutive patients with thyroid nodules
(14 malignant nodules in 10 patients and 24 benign nodules in 20 patients).
Three b-values including b-values of 0, 800 and 2000 sec/mm2 and a
Readout Segmentation Of Long Variable Echo-trains (RESOLVE) imaging technique
were used in the high b-value high-resolution DWI. The results showed that the
high b-value high-resolution DWI can further increase the diagnostic accuracy
for thyroid nodules with the best sensitivity, specificity and area under
receiver operating characteristic curve (AUC) than low b-value or single high b-value.PURPOSE
High
b-value diffusion-weighted MR imaging (b-value > 1000
sec/mm
2) has proven to be powerful in diagnosing prostate cancers
and breast cancers, etc.
1,2. However, the use of high b-value DWI to
differentiate between malignant and benign thyroid nodules has never been
investigated before. The purpose of our study was to evaluate the advantages of
multi b-value (b = 0, 800, 2000 sec/mm
2) with a commercially available
high-resolution DWI imaging technique to differentiate thyroid nodules.
METHODS
The prospective study included 28 consecutive patients with thyroid
nodules (14 malignant nodules in 10 patients and 24 benign nodules in 20
patients). All MRI examinations were performed using a 3T MRI scanner (MAGNETOM
Skyra; Siemens Healthcare, Erlangen, Germany) and a bilateral 8-channel
phased-array surface coil. Images were acquired in axial orientation and in the
following order: T2-weighted fast spin echo sequence with fat suppression,
T1-weighted fast spin echo sequence with fat suppression, high-resolution DWI
with frequency-selective fat saturation. The high b-value high-resolution DWI images
were collected using the RESOLVE technique
1 with the following
parameters: iPAT factor = 2 (GRAPPA), repetition time / echo time = 7700 / 67
msec, field of view = 175 mm, average = 1, matrix = 120 × 120, slice thickness
= 3.0 mm, gap = 0.3 mm. Diffusion gradients were applied sequentially with
b-values of 0, 800 and 2000 sec/mm
2, and ADC maps were automatically
generated online based on the three b-values. The scan time of the
high-resolution DWI was 5 minutes and 45 seconds. The ADC value, signal intensity of each nodule (SI
nodule) and signal intensity of normal thyroid tissue (SI
NTT)
were measured on ADC maps and DWI images independently by 2 experienced radiologists. All
images including the T1- and T2-weighted images and nodular signals in the DWI
maps were qualitatively reviewed first, and region of interest (ROI) for each
nodule was drawn manually. Then the mean ADC value of each nodule was
calculated based on the radiologists’ measurements. The diagnostic accuracy of
ADC values was evaluated in terms of sensitivity, specificity and area under
receiver operating characteristic (ROC) curve (AUC). All the nodules were proven
by postoperative pathology.
RESULTS
Benign thyroid nodules included nodular goiter (n = 19), adenoma (n = 3)
and thyroiditis (n = 2). Malignant thyroid nodules included papillary carcinoma
(n = 12) and follicular adenocarcinoma (n = 2). On DWI images with b-value = 0 sec/mm
2,
the malignant thyroid nodules presented a lower ratio of SI
nodule/SI
NTT
than that of benign nodules (1.37 ± 0.64 vs. 2.14 ± 0.91, P = 0.009). If ≤ 1.56 is adopted as
the cutoff value, the sensitivity, specificity and AUC were 78.75%, 70.83%
and 0.75. On DWI images with b-value = 800 sec/mm
2, however, the
malignant thyroid nodules showed a higher ratio of SI
nodule/SI
NTT
than that of benign nodules (2.76 ± 1.23 vs. 1.37 ± 0.66, P < 0.001). If > 1.60 is selected as the cutoff value, the sensitivity, specificity
and AUC were 85.71%, 75.00% and 0.84. On DWI images with b-value =
2000 sec/mm
2, the malignant thyroid nodules showed a much higher ratio
of SI
nodule/SI
NTT than that of benign nodules (3.35 ± 1.98
vs. 0.89±0.47, P < 0.001, Fig. 1
and 2). If > 1.55 is applied as
the cutoff value, the sensitivity, specificity and AUC were 85.71%, 87.50%
and 0.90. Furthermore, on the ADC maps obtained by combining the 3 b-values, the
malignant thyroid nodules presented a lower mean ADC value than that of benign
nodules (0.89 ± 0.14 × 10
-3 mm
2/sec vs. 1.55 ± 0.23 × 10
-3
mm
2/sec, P < 0.001,
Fig. 3). If ≤ 1.105 × 10
-3
mm
2/sec is adopted as the cutoff value, the sensitivity, specificity
and AUC were 100.00% (95% CI, 76.80%-100.00%), 100.00% (95% CI,
85.80%-100.00%) and 1.00.
DISCUSSION
High-resolution and high
b-value DWI imaging pose a significant challenge for gradient hardware.
Increasing the matrix and b-value will lead to longer TE and lower SNR. RESOLVE
is a multi-shot EPI sequence with low susceptibility based image distortion
which allows the reduction of TE for each individually-acquired segment
and can thus obtain high-quality DWI images with both high resolution and high
b-value.
3 From the figures shown in this study, we can see that
distortion was reduced and SNR was increased by using RESOLVE in comparison
with single-shot EPI.
CONCLUSION
Compared with the conventional b-value DWI (b-value ≤ 1000
sec/mm
2), high b-value high-resolution DWI using RESOLVE can increase
the diagnostic accuracy for thyroid nodules. And the ADC value obtained by
combining DWI images with three b-values (0, 800, 2000 sec/mm
2) shows
the best performance.
Acknowledgements
No acknowledgement found.References
1. Wisner D. J.,
Rogers N., Deshpande V. S., et al. High-resolution diffusion-weighted imaging for
the separation of benign from malignant BI-RADS 4/5 lesions found on breast MRI
at 3T. Journal of Magnetic Resonance Imaging. 2014;40(3):674-681.
2. Liney G. P.,
Holloway L., Al Harthi T. M., et al. Quantitative evaluation of
diffusion-weighted imaging techniques for the purposes of radiotherapy planning
in the prostate. The British journal of radiology. 2015;88(1049):20150034.
3. Porter, D. A.,
Heidemann, R. M. High Resolution Diffusion Weighted Imaging Using Readout
Segmented Echo Planar Imaging, Parallel Imaging and Two Dimensional Navigator
based Reacquisition. Magnetic Resonance in Medicine. 2009;62(2):468-475.