Keywords
Dynamic contrast magnetic resonance imaging - pituitary microadenoma - signal time
curve
Introduction
Pituitary adenomas are diverse group of tumors arising from the pituitary gland that
are estimated to be present in 14.4%–22.5% of the general population.[1] They have been traditionally classified, based on size into macroadenoma (size >
1 cm) and microadenoma (size < 1 cm).
Magnetic resonance imaging (MRI) remains the mainstay of pituitary imaging and has
largely supplanted computed tomography (CT) for detection and localization of microadenomas.
Spin echo (SE) MR sequences were the first to be utilized in detection of pituitary
adenoma; however, the sensitivity of these sequences remained low, especially in cases
of microadenoma.[2],[3] Subsequent studies showed that microadenoma detection improved on contrast studies
and was dependent on time of acquisition after contrast administration. To augment
the contrast difference between normal gland and microadenoma, early dynamic scan
was proposed.[4],[5],[6] Dynamic pituitary MRI during contrast scanning has resulted in increased diagnostic
yield for these lesions and has presently become the criterion standard. A significant
problem with dynamic MRI is the increased false positivity.[2],[3],[7],[8] At present, none of the MRI sequence is found to be unequivocally optimal for their
detection, and thus, diagnosis depends on combination of images obtained before, during
and after the contrast injection.[9]
This study was intended to combine the merits of precontrast T1 SE sequence and dynamic
contrast enhanced magnetic resonance imaging of the pituitary gland in evaluation
of microadenoma, through the signal time curve (STC) analysis. More specifically,
we aimed to assess whether the evaluation of precontrast T1 signal intensity ratio
(SIR) of the suspicious lesion observed on DCE MRI can improve the diagnostic performance
for localizing microadenoma.
Materials and Methods
Patients
This study was approved by institutional review board (IRB) and an informed consent
was obtained from each patient enrolled in the study. We retrospectively reviewed
MRI images of consecutive 23 patients referred for dynamic postcontrast MRI of pituitary
gland. For our study, these patients were divided into two groups.
Group 1 (n = 15; male/female ratio, 4:11; mean age, 36 years) consisted of 15 patients who were
labeled as pituitary microadenoma in clinical records based on compelling clinical
and biochemical evidence compatible with adenomas. In all these patients, MRI had
reported a focus of differential enhancement sized 3–10 mm in the pituitary gland
on dynamic contrast study suggestive of microadenoma. Of these, 11 were labeled as
prolactinomas, who had galactorrhea, infertility, or amenorrhea with serum prolactin
>50 ng/mL. Four of these patients presented with acromegaly with growth hormone level
exceeding 20 ng/mL.
Group 2 (n = 8; male/female ratio, 1:7; mean age, 34 years) consisted of eight patients who
were not labeled as microadenomas based on hormonal evaluation and MRI study. Seven
of these patients had presented with the complaints of galactorhhea, amenorrhea, or
infertility; however, the serum prolactin levels remained <26 ng/mL. One male patient
was being evaluated for bulky pituitary reported from an outside institute and was
subsequently diagnosed as hypothyroidism.
MRI in six of these patients was reported as normal study. In two of these patients,
MRI had reported a focus of differential enhancement on dynamic contrast study suspicious
for adenoma.
MR protocol
MRI was done in all patients on a 1.5 Tesla (Siemens Avanto, Erlangen, Germany) system
with an actively shielded whole body superconducting magnet. Imaging was done using
a 20-channel head–neck coil. A coronal precontrast VIBE sequence was obtained with
TR/TE of 497/10 ms, 320 × 320 matrix, 230 mm Field-of-view (FOV), and 2.5 mm slice
thickness without gap for assessment of pituitary and brain morphology. After precontrast
sequence, coronal dynamic contrast scan was done using fast SE sequence, 15 mm FOV,
3 mm slice thickness, 0.2 mm slice gap. Contrast was injected when first dynamic ended
and second began. Total duration of the dynamic sequence was 225 s which included
1 dynamic of precontrast study and six cycles during and after contrast injection.
10 mL of 0.01-mmol/kg gadopentetate dimeglumine (Magnevist; Bayer HealthCare Pharmaceuticals,
Wayne, NJ) was injected at a rate of 2 mL/s.
MRI analysis
For the group 1 patents, the final MRI report of our institute was referenced for
obtaining the tumor location. The first author confirmed the presence of the lesion
by analyzing the MR images. For group 2 patients, a senior radiologist with 15 years
of experience was required by forced choice to outline a focus of decreased signal
in the dynamic contrast images and to specify whether an adenoma could be present
in the images. The radiologist was blinded to the MRI report but was not blinded to
the final clinical diagnosis. Of these eight patients, he indicated the presence of
suspicious foci of differential enhancement in two of the patient, which was present
in two or more of the sequential postcontrast enhancement. In other six patients,
an area of differential enhancement (lowest signal) was outlined for analysis, which
was present in single image.
Region of interest (ROI) was drawn on the focus of differential enhancement (zone
a) in all the patients of group 1 and group 2. Another ROI was drawn on all the patients
in normal appearing tissue of pituitary gland (zone b). Signal intensity time curves
were generated in all the patients at both these locations as per the standard institutional
protocol.
Three parameters were recorded for each patient:
-
Baseline T1 SIR at 0 s at suspicious zone (zone a) of differential enhancement (SIR
T) and at normal pituitary (zone b) (SIR P)
-
SIR difference: SIRP− SIR T
-
Relative SIR difference: The ratio of the SIR difference divided by the SIR of normal
pituitary = (SIRP− SIR T)/SIR P.
The negative values, if any, for SIR difference and relative SIR difference was ignored
for calculation of mean.
Statistical analysis
Statistics was performed using SPSS software (IBM Corp 2013. Version 22.0. Armonk,
NY). The first part was using the independent sample Mann–Whitney U-test to compare mean of all these three parameters between group 1 and group 2. Second
part was drawing receiver-operated characteristic (ROC) curve for these three parameters
to predict the presence of microadenoma.
Results
Evaluation of STC from ROI placed over the normally enhancing anterior pituitary and
suspicious area of differential enhancement in group 1 [Figure 1] and group 2 patients [Figure 2] revealed that 14 out of 15 cases of group 1 (with microadenoma) demonstrated lower
baseline T1 signal intensity compared with the normal anterior pituitary. While one
patient, who was on bromocriptine therapy for prolactinoma, showed mildly higher T1
signal than normal pituitary. In group 2, in six of eight patients, the area under
consideration showed lower T1 signal, whereas two showed higher T1 signal than the
normal pituitary.
Figure 1: Demonstrates placement of ROI within the normal appearing anterior pituitary gland
(yellow) and microadenoma (red). Corresponding signal-time curve shows that there
is significant difference enhancement between the microadenoma and normal pituitary
and difference in precontrast signal at t = 0 s
Figure 2: Dynamic MRI of a 24-year-old female patient in group 2, presenting with galactorrhea.
ROI placed in normally enhancing anterior pituitary (red) and an area of differential
decreased signal (yellow) is shown and the corresponding signal-time curves show that
the suspicious area shows almost similar (and slightly higher) baseline T1 signal
at t = 0 s
The mean values, standard deviation, and results of Independent-sample Mann–Whitney
U-test to compare the mean values between the two groups for the three parameters
are summarized in [Table 1]. The mean baseline T1 SIR was lower in patients with the diagnosis of microadenoma
(group 1), although the difference was not significant (P = 0.065). The difference of baseline T1 SIR between the normal pituitary and the
zone of concern (SIR difference) was significantly higher in group 1 (P = 0.003). The relative SIR difference was also significantly higher in cases compared
with controls (P = 0.005).
Table 1
Mean, standard deviation, and comparison of the mean values of three parameters between
two groups using Independent-sample Mann-Whitney (t-test
|
Parameters
|
Group 1 (n=15)
|
Group 2 (n=8)
|
P
|
|
Mean
|
SD
|
Mean
|
SD
|
|
SIR = Signal intensity ratio
|
|
Baseline T1 SIR value
|
285.13
|
48.6
|
331.13
|
50.79
|
0.065
|
|
SIR difference
|
95.07
|
111.84
|
12.25
|
8.97
|
0.003
|
|
Relative SIR difference ratio
|
0.215
|
0.205
|
0.037
|
0.028
|
0.005
|
ROC curve drawn to predict the presence of microadenoma demonstrates high area under
curve for all the three parameters [Figure 3] and [Figure 4]. [Table 2] summarizes the results of ROC curve analysis for the three parameters. Of these,
the SIR difference showed highest area under curve closely followed by the Relative
SIR difference.
Table 2
Receiver-operated characteristic curve drawn to predict the presence of micro adenoma
for baseline T1 SIR values, its difference from normal pituitary SIR, ratio (difference/SIR
of normal pituitary)
|
Test parameter
|
Area under curve
|
Cut off value
|
Sensitivity (%)
|
Specificity (%)
|
Cut-off with 100% specificity
|
Sensitivity of cut off with 100% specificity (%)
|
|
SIR=Signal intensity ratio
|
|
Baseline T1 SIR value
|
0.738
|
312
|
66.7
|
62.5
|
-
|
-
|
|
SIR difference
|
0.863
|
21
|
73.3
|
75
|
26.5
|
66.7
|
|
Relative SIR difference ratio
|
0.850
|
0.057
|
80
|
75
|
0.107
|
60
|
Figure 3: Receiver-operated characteristic curve of the baseline T1 signal if the suspicious
lesions in prediction of presence of microadenoma
Figure 4: Receiver-operated characteristic curve of the signal intensity ratio (SIR) difference
(blue) and relative SIR difference ratio (green) in prediction of presence of microadenoma
Based on ROC curves, we found that a value of 312 for baseline T1 signal predicts
microadenoma with moderate sensitivity and specificity. A SIR difference of 21 and
relative SIR of 0.057 between the normal pituitary and the area showing differential
enhancement predicted microadenoma with high sensitivity and specificity. More importantly,
we could obtain a specificity of 100%, for cut-off values of 26 and 0.107 of SIR difference
and the relative difference, respectively, with reasonable sensitivities.
Discussion
DCE MRI has currently become the most utilized technique for the detection of pituitary
microadenoma. Most of the previous studies evaluating DCE MRI in detection of microadenoma
have relied on subjective visual assessment in delineating these tumors. An area of
low signal in anterior pituitary on dynamic post contrast images measuring 3–10 mm
is diagnosed as microadenoma. The problem with this technique is increased rate of
false positivity.[2],[3] Pituitary often shows heterogeneous contrast enhancement owing to variable blood
supply. An area of lower contrast enhancement on a single image can be misinterpreted
as an adenoma. Moreover, the accuracy of the diagnosis is dependent upon the experience
of the interpreting radiologist and there are no objective criteria for the diagnosis.
Few studies have attempted to evaluate STCs on DCE MRI to assess time course of enhancement
of tumors and the normal pituitary, enabling an increased diagnostic confidence. One
such study was conducted by Yuh et al., who concluded that pituitary adenoma enhanced 9.3 ± 1.5 s after straight-sinus
enhancement and significantly (12.0 s) before anterior pituitary enhancement.[10] Rossi Espagnet et al., were first to study STCs of the pituitary in 52 patients to establish optimum acquisition
time for microadenoma detection and concluded that 120 s is ideal time for imaging.[11] They also found out that there was significant difference in peak enhancement of
microadenoma and normal anterior pituitary. Moreover, pituitary microadenoma showed
mean time-to-peak of 90 s, whereas normal anterior pituitary showed an earlier peak
enhancement with mean time-to-peak of 80 s.
Stadnik et al. compared DCE MRI and precontrast T1 images in 12 patients with microadenoma and
found that both T1 sequence and DCE MRI was able to detect 84% (10/12) lesions, whereas
the combination of both dynamic MRI and precontrast T1 sequence was able to detect
100% of cases.[12] Ma et al. in his study observed that a significant correlation between tumor consistency and
expression of collagen IV was seen with signal intensity on precontrast T1 SE sequences.[13]
In our study, 14 out of 15 cases of microadenoma revealed lower signal intensity compared
with the normal anterior pituitary. While in one patient, who was on bromocriptine
therapy for prolactinoma, showed mildly higher T1 signal than normal pituitary. This
result was in concordance with the previous studies that showed that most of the microadenoma
demonstrates increased T1 and T2 relaxation times. One of the initial studies has
reported atypical T1 and T2 relaxation signals in patients with bromocriptine and
attributed this to loss of cell volume owing to medical therapy.[14]
Our study demonstrated that difference of T1 SIR of the normal pituitary and the lesion
under consideration is significantly high in patients with pituitary microadenoma
than those without. The SIR difference and the relative SIR difference were able to
predict the presence of microadenoma with 100% specificity with reasonable sensitivities.
A difference of T1-SIR of 26.5 between the lesion and the normal pituitary and a relative
SIR difference ratio of 0.028 were able to predict the presence of microadenoma in
all the cases. The findings of the present study have great implication while interpretation
of DCE MRI and we speculate that addition of these parameters can reduce false positivity
rate that has been one of the major criticisms of DCE MRI.
Our study utilizes the combined merits of dynamic contrast properties as well as internal
relaxation properties of microadenomas for their diagnosis. Moreover, this study provides
quantitative parameters that can be reliably used to increase the diagnostic confidence
of DCE MRI in diagnosing pituitary microadenoma.
Small sample size and lack of surgical and histopathological evidence were the major
limitations of this study. Moreover, owing to its retrospective nature, we could not
perform T1 mapping of pituitary, which could further validate our concept. Also, presence
of nonfunctioning adenomas in the control group could not be excluded as the clinical
diagnosis was considered gold standard in the present study and surgical evidence
was lacking. Last, interpreting radiologist in our study was not blinded to the clinical
picture that can lead to patient selection bias. We recommend future multicentric
prospective studies having larger sample sizes with an effort to reduce selection
bias and histopathological confirmation to validate the results of our study.
Conclusion
In conclusion, our study demonstrated that assessment of baseline precontrast SIR
derived through STC of the pituitary microadenoma, suspected on dynamic contrast MRI,
can increase diagnostic confidence in their diagnosis and localization. Moreover,
we have shown that the quantitative assessment would be more meaningful if interpreted
with the SIR of an internal reference, i.e., normal appearing pituitary gland.
Acknowledgements
The authors would also like to acknowledge their team of MR technicians led by Mrs
Raichel Luyees and Dr B N Maurya for their role in this study.
Declaration of patient consent
The authors certify that they have obtained all appropriate patient consent forms.
In the form the patient(s) has/have given his/her/their consent for his/her/their
images and other clinical information to be reported in the journal. The patients
understand that their names and initials will not be published and due efforts will
be made to conceal their identity, but anonymity cannot be guaranteed.
