EUROPEAN JOURNAL OF RADIOLOGY

ELSEVIER European Journal of Radiology 18 (1994) 185-190

Pituitary rn$roadenomas: experience with Gd-DOTA-enhanced MR imaging at 0.5 Telsa

Michkle Hamon-Kkrautreta*, Xavier Leclerca, Didier Dewaillyb, Jean-Pierre Pruvoa, Pierre Fossatib, Jacques Clarissea

‘Department of Neuroradiology. ‘Department of Endocrinology. H6pital B. C.H. R. U. Lille, 2 Avenue Oscar Lambret, 59037 Lille Ceakx. France

Received 25 March 1994; revision received 30 June 1994; accepted 4 July 1994

Abstract

Purpose: We report our experience with magnetic resonance imaging at 0.5 Tesla in the radiological diagnosis of pituitary adeno- mas. Methods: Over 2 years we performed a prospective study in 38 patients with pituitary microadenomas to assess the potential additional benefit of gadolinium in the detection of small intrasellar lesions. The protocol included three corona1 Tl-weighted sequences: precontrast, early postcontrast (obtained less than 2 min after injection) and late postcontrast (obtained 5 min after injec- tion). For each sequence lesions were classified according to their visibility into three categories: definite lesion, probable lesion or absent lesion. Results: Lesions were classified as well-defined in 55% of patients on the precontrast study, 89.5% on the early postcontrast study and 60.5% on the late postcontrast study. Of nine microadenomas not detected on the precontrast scans, all were clearly seen on the early post contrast scans; only four were well-defined on the late postcontrast study. Only one lesion was not seen on the early postcontrast study but was well-defined on precontras study as a spontaneous high-intensity focal area: it was a hemorrhagic microadenoma. Conclusion: Our results suggest that the early postcontrast study is the most useful sequence for the detection of microadenomas. Precontrast image is necessary to detect hemorrhagic lesions; the late postcontrast sequence has a low additionai diagnostic yield and seems unnecessary unless the other sequences are inconclusive.

Keywords: Pituitary, neoplasms; Pituitary, MRI; Magnetic resonance (MR), contrast media; Magnetic resonance (MR), technique

1. Introduction

Magnetic resonance imaging (MRI) has proved to be useful in the evaluation of patients with suspected pitui-

tary disease; the technique allows excellent delineation of the contour of the gland due to the intrinsic differ- ences in intensity between soft-tissue, cerebrospinal fluid, and vessels [1,2]. However, the lack of sufficient contrast between normal gland tissue and microadeno- matous tissue sometimes precludes a definitive diagnosis [3,4]. By optimizing the interface between normal and abnormal tissue, gadolinium-enhanced MRI may prove to be useful in the detection of small intrasellar lesions [2,5,6]. This report describes our experience in the radiological diagnosis of pituitary adenomas, using a 0.5-T MRI. To assess the usefulness of intravenously ad-

* Corresponding author.

ministered gadolinium (Gd) in enhancing the contrast between normal and adenomatous tissue, imaging was prospectively performed in consecutive patients with clinical and biochemical evidence of a pituitary adeno- ma. The aim of this study was to evaluate the diagnostic value of gadolinium imaging, using precontrast, early postcontrast and late postcontrast scans.

2. Materials and methods

2.1. Patients

Over a 2-year period, we prospectively studied the ad- ditional diagnostic yield of postcontrast Gd-enhanced magnetic resonance imaging in the detection of small in- trasellar lesions in patients referred for evaluation of pituitary microadenomas.

Forty-four patients with biochemical evidence of a

0720-048x/94/$07.00 0 1994 Elsevier Science Ireland Ltd. All rights reserved SSDI 0720_048X(94)W55 1 -M

M. Hamon-KPrautret et al. /Eur. J. Radio/ 18 (1994) 185-190

A - Precontrast seauence R - Earlv oostcontrast sequence

C - Late mstcontrast sequence

Fig. I. Visibility of lesions on: precontrast sequence (A), early postcontrast sequence (B) and late postcontrast sequence(C).

hormonally active pituitary adenoma were examined by pre and postcontrast Gd-enhanced MR imaging. In six patients no intrasellar lesions were detected on any MRI sequence.They were presumed to have primary idio- pathic hyperprolactinemia.

Of the remaining 38 patients, MR images revealed abnormalities consistent with a diagnosis of an intra- sellar focal lesion on at least one of the sequences obtained. They were 37 women and one man, with a mean age of 29 years, range 16-49 years. Serum prolac- tin levels were elevated in 35 patients; two patients had clinical and biological signs of Cushing’s disease, and one was acromegalic. Microadenomas ranged from 3 to 10 mm in diameter (mean diameter, 5.5 f 1.9). The majority, 31, were localated laterally in the sella turcica; seven were in the midline.

For all patients, the diagnosis of microadenoma was based on a combination of clinical, biological and radiological data; in three patients anatomic conlirma- tion was obtained after surgical resection of the adeno- ma. Of the tumours removed, two were prolactin- secreting microadenomas; the other was a corticotropin- secreting microadenoma.

2.2. Technique of magnetic resonance imaging

All patients were imaged on a 0.5-T system (MR Max, General Electric), using a circular head coil, with a 250- mm field of view (FOV). The protocol comprised three coronal Tl-weighted sequences (fast gradient recalled echo, TR: 250 ms repetition time, TE: 12 ms echo time, flip angle: 90, 224 x 224 matrix): one before adminis- tration of contrast media and two consecutive sequences immediafely after intravenous injection of 0.1 mmolkg

a b C

Fig. 2. (a) Precontrast coronal TI-weighted GE image: pituitary gland appears homogeneous. No lesion is detected. (b) Early postcontrast Tl- weighted GE image: microadenoma is clearly seen as a focal area of focal hypointensity in the left part of the sella turcica. (c) Late postcontrast Tl-weighted GE image: microadenoma enhances, becomes isointense with adjacent normal tissue and is no longer detectable.

hf. Hamon-Khaurrer et al. /Eur. J. Radiol. 18 (1994) 185-190 187

a b

Fig. 3. (a) Precontrast coronal Tl-weighted GE image: spontaneous focal area of high intensity in the left lateral part of the sella turcica (hemor- rhagic microadenoma). (b) Early postcontrast Tl-weighted GE image: normal pituitary gland enhances, becomes isointense with the lesion which is no longer detectable.

of Gd-DGTA (Guerbet). Early postcontrast images (250/12/2 exe) were obtained at 1.48 min after injection, and late postcontrast scans (250/U/4 exe) at 5 min. Cut thickness was 3 mm, without a gap. Images were reconstructed with a two-dimensional Fourier trans- form on a 224 x 224 imaging matrix.

2.3. Analysis

All images were evaluated by two neuroradiologists by consensus. The presence or absence of a lesion, its lo- cation and size were evaluated. For each sequence (precontrast, early postcontrast and late postcontrast), lesions were classified as definite, probable or absent. Associated abnormalities of the contour of the gland (focal upward convexity of the gland, abnormalities of the sellar floor), and deviation of the infundibulum were also assessed.

3. Results

3.1. Lesion detection

The patient group consisted of 37 women and one man, with a mean age of 29 years, range 16-49 years. Serum prolactin levels were elevated in 35 patients; two patients had clinical and biological signs of Cushing’s disease, and one was acromegalic. Microadenomas ranged from 3 to 10 mm in diameter (mean diameter 5.5 f 1.9). The majority, 31, were located laterally in the sella turcica; seven were in the midline.

Definite lesions were detected in 34 cases (89.5%) on the early postcontrast sequence, in 23 cases (60.5%) on the late postcontrast sequence and in 21 cases (55%) on the precontrast sequence (Fig. 1).

In most cases, adenomas were identified by the presence of low-intensity areas on unenhanced scans

188 A4. Hamon-K&at&et et al. /Eur. J. Radiol. 18 (1994) 18%190

and which were subsequently visualized as focal areas of relative non-enhancement on images obtained after Gd administration, Nine microadenomas were not detected on the unenhanced images. Of these, all nine were clear- ly defined on the early postcontrast scans but only four lesions were identified on the late postcontrast study (Fig. 2). In eight patients, a focal non-enhancing area considered to be a probable lesion on the unenhanced scan was clearly identified as such after administration of Gd: in seven cases on the early postcontrast scans, and in three cases on the late sequence, Finally, one lesion, well seen before injection as a spontaneous high- intensity focal area, was not identified on early and late postcontrast sequences: it was thought to be a hemor- rhagic microadenoma (Fig. 3). Three areas were classified as probable lesions on the early postcontrast scans: one was well-defined on precontrast and late postcontrast sequences, another was well-defined on the precontrast scan but was classified as a probable lesion on late postcontrast scans; the third was only clearly seen on the late sequence.

3.2. Ancillary MR findings

Ancillary findings such as deviation of the pituitary stalk, abnormalities of the superior contour of the gland and of the sellar floor were assessed in each case. In seven patients, no ancillary findings were present. The infundibulum was deviated to the side opposite to the suspected lesion in 14 patients. Focal superior contour abnormality was present in 17 cases. Asymmetry of the sellar floor (more than 2 mm) was observed in 19 cases.

4. Discussion

The reported experience in the evaluation of MRI de- tection ofmicroadenomas is mixed, reflecting the large dependence of results on field strength and technique (choice of the pulse sequence; time interval between Gd intravenous administration and data acquisition) [1,5-8,10,11]. In recent studies performed at 1.5 T, 71-91% of pituitary microadenomas were demonstrated without the use of paramagnetic contrast agents [1,2,4,12]. Thus, using high field MRI, precontrast Tl- weighted images appear adequate to detect most adeno- mas; the use of Gd contrast agent improves sensibility and sensitivity of MRI by 5-10% and paramagnetic agents play a minor role. At low or moderate field strength, noncontrast MRI studies detect micro- adenomas with a sensitivity of 25-80%, and improve- ment in lesion detection after Gd infusion seems to be more important than at high field strength [3,13].

Our study was performed to evaluate the visibility of microadenomas on Tl -weighted images, respectively before, immediately after and late after Gd injection, and to assess the utility of employing Tl series with

short acquisition time beginning as soon as possible after Gd injection (based on the hypothesis that the microadenomas take up contrast agent more slowly that the hypophysis and that, therefore, the contrast differ- ence between lesion and pituitary gland is greatest for a rather short period of time directly after contrast media administration).

Our data suggest that iv. administration of Gd appears to have a useful role in that it facilitates visualization of some lesions on Tl-weighted images: lesions were well-defined in 55% on unenhanced scans and in 89.5% on early postcontrast scans. This benefit of Gd is due to different enhancement characteristics of lesion tissue compared to normal pituitary tissue. Pat- terns of enhancement with Gd appear to mirror the CT experience [ 14- 191. The pituitary gland enhances quick- ly and homogeneously after intravenous administration of Gd-DOTA: this is due to the abundant blood supply of the anterior lobe of the pituitary gland from the pitui- tary portal system and to its lack of blood-brain barrier. The degree of enhancement (about 60%) is intermediate between that of the cavernous sinus (about 141%) and that of the brain [20-231. Enhancement of the pituitary gland persists for approximately 1 h but is typically maximal or nearly maximal at 3 min [ 10,111. Adenomas also take up contrast but generally enhance more slowly and for a longer period than normal gland. Therefore, selective enhancement of the normal pituitary relative to the adenoma immediately after contrast improves lesion visibility. On the contrary, Doppman et al. [lo] and Dwyer et al. [1 1] have shown that on later images, delayed enhancement of the lesion relative to normal gland may lead to a reversal in signal intensities (adeno- mas appear hyperintense because of slower wash-out of Gd from the abnormal tissue). This phenomenon has been described in soft or semi-liquid lesions which are dark on unenhanced and immediate postcontrast scans and become bright on very delayed images. So after in- jection, there is an intermediate period during which a lesion may be isointense with adjacent normal tissue, and Gd might obscure rather facilate detection.

Based on these observations, it appears that acquiring Tl-weighted images as soon as possible following Gd- infusion is a critical factor for the detection of micro- adenomas [9,24]. In our study, the gradient-echo MR pulse sequence was chosen instead of the spin-echo tech- nique in order to obtain the required temporal resolu- tion and the best contrast between lesion and pituitary gland. This increase in contrast would be expected to compensate for the lower signal to noise ratio that oc- curs at intermediate field strength using short TR and only two excitations. Our results support this hypothe- sis: the most useful sequence was the early postcontrast sequence obtained less than 2 min after injection (89% lesions were well-defined, and of nine lesions not visualized on precontrast all were classified as definite

M. Hamon-Khautret et al. /Eur. J. Radiol. 18 (1994) 185-190 189

lesions on early postcontrast scans, whereas only four were classified as definite on the late postcontrast scan). At high field strength, gradient-echo images through the skull base are markedly degraded by differences in mag- netic susceptibility between the gland and the adjacent air-containing sphenoidal sinus leading to artifacts in the region of the sella turcica; for this reason, most stud- ies at 1.5 T are performed using spin-echo pulse [9,18]. Hence, with respect to high field, the use of moderate fields in MRI allows one to obtain images by gradient- echo pulse with less artifacts and better temporal resolu- tion [25]. Although this is obtained at the cost of a lower signal/noise ratio, it nevertheless appears to be suffi- ciently sensitive to detect small intrasellar lesions such as microadenomas.

5. Conclusions

Recent evaluations of MRI for the diagnosis of pitui- tary microadenomas report variable diagnostic accuracy and variable ‘most useful’ pulse sequence parameters: the performance of MR equipment (field strength) and the choice of sequences are of importance for the accurate detection of microadenomas.

According to the recent literature it seems that, for a 1.5-T high signal to noise MR unit, the Tl-weighted SE precontrast sequences permit confident diagnosis in most cases, and injection of Gd should be performed only in cases where the diagnosis remains in doubt after a precontrast scan.

Our study was performed at 0.5 T, using gradient- echo pulse sequence in order to obtain good temporal resolution. Our results suggest that a precontrast study is always necessary to detect hemorrhagic lesions; the early postcontrast study appears to be the most useful sequence for the detection of microadenomas: the late postcontrast study does not usually improve the diag- nostic yield and seems to be unnecessary unless the other sequences are inconclusive.

In conclusion, for a 0.5-T unit, we advocate the following series of sequences for the diagnosis of pitui- tary microadenomas: a coronal Tl-weighted unenhanc- ed GE sequence followed by the same sequence immediately after Gd injection. If the diagnosis remains dubious, repetition of a Tl-weighted enhanced EG series later after contrast injection should be considered.

References

Kucharczyk W, Davis DO, Kelly WM. Sze G, Norman D, Newton TH. Pituitary adenomas: High-resolution MR imaging at 1.5 T. Radiology 1986; 161: 761-765. Kulkami MV, Francis Lee K, Mac Ardle CB, Yeakley JW,, Haar FL. 1.5-T MR imaging of pituitary microadenomas: tech- nical considerations and CT correlation. AJNR 1988; 9: 5-1 I. Davis PC, Hoffman JC, Spencer T, Tindall GT, Braun IF. MR

4

5

6

7

8

9

10

II

12

13

14

15

16

17

18

19

20

21

22

imaging of pituitary adenoma: CT, clinical and surgical correla- tion. AJNR 1987; 8: 107-l 12. Pack WW, Dillon WP, Norman D, Newton TH, Wilson CB. High-resolution MR imaging of pituitary microadenomas at 1.5 T: Experience with Cushing disease. AJR 1989; 152: 145-151. Pojunas KW, Daniels DL, Williams AL, Haughton VM. MR imaging of prolactin-secreting microadenomas. AJNR 1986; 7: 209-213. Davis PC, Hoffman JC, Malko JA, Tindall GT, Takei Y, Avruch L, Braun IF. Gadolinium-DTPA and MR imaging of pituitary adenoma: a preliminary report. AJNR 1987; 8: 817-823. Davis PC, Gockhale KA, Joseph G, Peterman SB, Adams DA, Tindall GT, Hudgins PA, Hoffman JC. Pituitary adenoma: cor- relation of half-dose gadoliniumaenhanced MR imaging with surgical findings in 26 patients. Radiology 1991; 180: 779-784. Elster AD. Modem imaging of the pituitary. Radiology 1993; 187: I-14. Newton DR, Dillon WP, Norman D, Newton TH, Wilson CB. Gd-DTPA Enhanced MR imagifig of pituitary adenomas. AJNR 1989; lo: 949-954. Doppman JL, Frank JA, Dwyer AJ, Oldfield EH, Miller DL, Nieman LK, Chrousos GP, Cutler GB, Loriaux DL. Gadolinium DTPA enhanced MR imaging of ACTH-secreting microadenomas of the pituitary gland. J Comput Assist Tomogr 1988; ‘12: 728-735. Dwyer AJ, Frank JA, Doppman JL, Oldfield EH, Hickey AM, Cutler GB, Loriaux DL, Schiabl TF. Pituitary adenomas in pa- tients with Cushing disease: initial experience with Gd-DTPA- enhanced MR imaging. Radiology 1987; 163: 421-426. Stadnick T, Stevenaert A, Beckers A, Luypaert R, Buisseret T, Osteaux M. Pituitary microadenomas: diagnosis with two and three-dimensional MR imaging at 1.5 T before and after injec- tion of gadolinium. Radiology 1990; 176: 419-428. Mikhael MA, Ciric IS. MR imaging of pituitary tumors before and after surgical and/or medical treatment. J Comput Assist Tomogr 1988; 12: 441-445. Bonneville JF, Cattin F, Moussa-Bacha K, Portha C. Dynamic computed tomography of the pituitary gland: the “tuft sign”. Radiology 1983; 149: 145-148. Bonneville JF, Cattin F, Gorczyca W, Hardy J. Pituitary microadenomas: early enhancement with dynamic-CT implica- tions of arterial blood supply and potential importance. Radiol- ogy 1993; 187: 857-861. Finelli DA, Kaufman B. Varied microcirculation of pituitary adenomas at rapid dynamic, contrast-enhancemet MR imaging. Radiology 1993; 189: 205-210. Miki Y, Matsuo M, Nishizawa S, Kuroda Y, Keyaki A, Makita Y, Kawamura J. Pituitary adenomas and normal pituitary tis- sue: enhanced patterns on gadopentate-enhanced MR imaging. Radiology 1990, 177: 35-38. Sakamoto Y, Takahashi M, Korogi Y, Bussaka, H, Ushio Y. Normal and abnormal pituitary glands: gadopentate dime- glumine-enhanced MR imaging. Radiology 1991; 178: 441-445. Tien RD. Sequence of enhancement of various portions of the pituitary gland on Gadolinium-enhances MR images: Correla- tion with regional blood supply. AJR 1992; 158: 651-654. Berry 1, Brant-Zawadzki M, Osaki L, Brasch R, Murovic J, Newton TH. Gd-DTPA in clinical MR of the brain: 2. Extraax- ial lesions and normal structures. AJR 1986; 147: 1231-1235. Breger RK, Papke RA, Pojunas KW, Haughton VM, Williams AL, Daniels QL. Benign extraaxial tumors: contrast enhance- ment with Gd-DTPA. Radiology 1987; 163: 427-429. Kilgore DP, Breger RK, Daniels DL, Pojunas KW, Williams AL, Haugton VM. Cranial tissues: normal MR’appearance after

190 IU. Hamon-KPrautret et al. /Eur. J. Radiol. 18 (1994) 185-190

intravenous injection of Gd-DTPA. Radiology 1986; 160: 757-761.

23 Macpherson P, Hadley DM, Teasdale E, Teasdaie G. Pituitary adenomas: does gadolinium enhance their demonstration? Neuroradiology 1989; 3 1: 293-298.

24 Stadnick, Stevenaert A, Beckers A, van Herzele D, Luypaert R, Buisseret T, Osteaux M. Contrast behavior between micro-

adenoma and normal pituitary gland after gadolinium injection as a function of time at 1.5 T. Neuroradiology 1992; 34: 184-189.

25 Wilms G, Van Hecke P, Marchal G, Decrop E, Baert AL. MR imaging of the brain: preliminary results with a gradient echo sequence after paramagnetic contrast enhancement. BTR/JBR 1989; 12: 89-94.