Discussion
We evaluated the clinical utility of DWI for the pre-operative
localization of PTAs. While the PTAs were faintly identified with
intense distortion on SS-EPI, they were fairly identified with moderate
distortion on RS-EPI. The inter-rater reliability was substantial or
perfect, and the validity of DW-MRI for the localization of PTAs was
demonstrated. In addition, the PTAs could be discriminated from cervical
lymph nodes and thyroid glands by measuring the ADC values, and thus
RS-EPI is expected to become the optimal diagnostic modality for the
localization of PTAs.
With the introduction of 3T MRI, improved diagnostic performance for the
detection of PTAs has been reported in relation to the higher spatial
resolution and contrast-to-noise ratio.(1) PTAs are usually T2
hyperintense, and fat suppression is effective to achieve an accurate
depiction.(1) However, fat suppression is not easy to use in the neck
region, especially in the mediastinum, because of the magnetic field
inhomogeneities due to the proximity of lungs and upper airways. It thus
remains challenging to obtain good images of PTAs with MRI, and there is
currently no consensus on the optimal MRI protocol for the evaluation of
PTAs.
The diagnostic value of 4D MRI and of the IDEAL sequence for the
detection of PTAs has been reported.(19, 20) Although the use of these
imaging techniques improved the diagnostic sensitivity for the detection
of PTAs, false-positive diagnoses are occasionally attributed to
cervical lymph nodes or adjacent/ectopic thyroid tissues. To distinguish
PTAs from lymph nodes and thyroid tissues, Yildiz et al. used
conventional DWI provided by a 1.5T system, and they reported higher ADC
values in parathyroid lesions compared to lymph nodes and thyroid
tissues; however, the statistical significance of these differences was
not reported.(9)
Although SS-EPI is a well-established conventional method for the
acquisition of DWI data with short scan times, it suffers from geometric
distortion, signal dropout, and image blurring. Alternatively, RS-EPI is
a multi-shot sequence that reduces susceptibility artifact and blurring
arising from T2* decay, and thus lower distortion can be achieved.(21)
The clinical utility of RS-EPI has been well documented in some regions,
including the head and neck.(13) In our present case series, the
acquisition parameters were quite different between RS-EPI and SS-EPI,
because they were independently set to optimize their image quality
within clinically acceptable acquisition times (2–3 min). Since a
sufficient signal-to-noise ratio (SNR) was not expected for SS-EPI, the
number of excitations were increased to four. A sufficient SNR was
expected for RS-EPI, so we increased the number of readout segments to
five, to reduce the distortion. The results demonstrated that the PTAs
were well identified with less distortion by RS-EPI compared to SS-EPI,
and the conclusions of the two independent readers showed a good
correlation.
RS-EPI is thought to be a useful and reliable imaging technique for the
evaluation of PTAs. However, we observed herein that even with RS-EPI,
some PTAs were poorly identified with intense distortion, which might
mask the improvement in the lesion identification score. The distortion
might come from (1) the relatively large field of view (FOV) needed to
obtain whole neck images, (2) susceptibility artifact that remained even
with RS-EPI, and/or (3) motion artifact from respiration. In addition,
it is not still easy to obtain good-quality DW images at the upper
mediastinum due to the complex stricture facing the air, and the
establishment of a feasible imaging protocol of the cervical region to
the upper mediastinum using DW-MRI is expected.
In this study, we used ADC values for discriminating the PTAs and other
cervical organs, and the PTAs were well distinguished from the thyroid
glands and cervical lymph nodes. Cervical lymph nodes exhibit low ADC
values,(22) and benign neck pathologies including adenomas are reported
to show relative high ADC values,(23, 24) which is consistent with our
present observations. Higher ADC values were observed on RS-EPI compared
to SS-EPI in our case series. Differences in ADC values between RS-EPI
and SS-EPI are still controversial. Although higher ADC values on RS-EPI
have been reported in the human mammary gland and in a phantom
study,(21, 25) Koyasu et al. reported no significant differences in the
ADCs of salivary gland lesions.(13) In our present cases, the accurate
evaluation of lesions due to less distortion on 3T RS-EPI may have
contributed to the precise selection of the area, while more distorted
lesions on 1.5T SS-EPI might have been contaminated by surrounding fat
tissue and consequently resulted in a decreased ADC, which we observed
particularly in the thyroid glands. Adding DWI using RS-EPI to common
sequences such as STIR and T2WI could thus be beneficial for the
pre-operative evaluation of PTAs.
This is the first study to investigate the clinical utility of RS-EPI in
the pre-operative localization of PTAs, to our knowledge. One of the
limitations of our study was its retrospective design. In addition, the
acquisition protocols for SS-EPI and RS-EPI were independently set to
maximize their own features (rather than a comparison of the two
protocols), which resulted in different acquisition schemes. The sample
size was small, and there would be a selection bias because selected
patients with surgically confirmed pathologies were included. Further
prospective studies with larger sample sizes are warranted to clarify
the diagnostic utility, indications, and limitations of RS-EPI for the
localization of PTAs.