Magnetic resonance imaging characteristics of residual pituitary tissues following transsphenoidal resection of pituitary macroadenomas
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.325377
Source of Support: None, Conflict of Interest: None
Keywords: Diabetes insipidus, immunohistochemistry, magnetic resonance imaging, pituitary adenomas, pituitary tissues, posterior pituitary bright spotKey Message: Posterior pituitary bright spot on T1W MRI is an important predictor of the type of the adenoma as well as the occurence of postoperative DI.
Pituitary adenomas constitute 15–20% of all primary intracranial tumors, which show different morphology and propagation direction, and 10% of which having invasive growth., Pituitary adenomas originate in adenohypophysis, and with the normal pituitary tissue by deformation and displacement. Transsphenoidal microsurgery is the common initial treatment of pituitary macroadenoma because it is minimally invasive and has reasonable efficacy with low morbidity. The most frequent complications associated with transsphenoidal surgery are cerebrospinal fluid leakage and normal pituitary tissue damage. Magnetic resonance imaging (MRI), especially with gadolinium-enhanced MRI, has become the definitive radiological procedure for the diagnosis and follow-up of pituitary tumors, which provides a detailed anatomic view of the sellar region and can help identify the precise anatomical location of a pituitary tissue or of tumor recurrence.
However, few reports have investigated the MR appearance of the normal pituitary tissues in patients with pituitary macroadenomas, including size, shape, location pre- and post-operatively, as well as the signal intensity of neurohypophysis and its clinical correlation. The normal pituitary tissues, including adenohypophysis and neurohypophysis, can be identified easily on MR T1-weighted images (T1WI) in healthy adults or microadenoma patients., Previous studies are mainly focused on the posterior pituitary.,, Some authors believed that neurohypophysis exhibits high signal intensity on MR T1WI, which is related to the neuro-secretory granules containing antidiuretic hormone (ADH), and the reduction of ADH released by neurohypophysis may cause diabetes insipidus (DI).,,
The purpose of this study was to explore the pre- and post-operative magnetic resonance imaging of pituitary tissues following transsphenoidal resection of pituitary macroadenomas. The morphology and reconfiguration of residual pituitary tissues following transsphenoidal resection were observed on contrast-enhanced MR images, and its relationship with pituitary adenomas was analyzed. The appearance of posterior pituitary signal was also investigated as well as its clinical significance.
A total of 108 patients (55 males and 53 females; age range 23–72 years; mean age at surgery 45 years) were prospectively studied between September 2012 and September 2014 [Table 1]. During this period, 293 patients diagnosed with pituitary adenoma were treated, but only those who met the follow-up criteria listed below were included in the study. The inclusion criteria included: (i) surgery via a transsphenoidal approach for pituitary macroadenomas with pathological confirmation, (ii) adequate MR imaging was performed preoperatively and postoperatively (within 48 hours and at 4th to 6th month after surgery), (iii) available clinical and endocrinological data were obtained. The exclusion criteria were: (i) for repeated surgery, (ii) preoperative bromocriptine or radiotherapy history.
All 108 of these patients underwent a standard surgical adenoma excision via a transsphenoidal approach by the corresponding author (Pro. Wang). Transsphenoidal removal of the adenoma was accomplished through the standard microsurgical endonasal transseptal approach. The anterior wall of the sphenoid sinus was opened bilaterally, the sellar floor was opened, and the tumor was removed with standard instruments, such as curettes, dissectors, suction and micro-forceps. The pituitary gland was well identified, and care was taken to minimize damage to the pituitary gland and the pituitary stalk. The cavity created by resecting the tumor was packed with spongia gelatinosa, and the sellar floor was reconstructed. Intersellar tumor was total removed in all cases operatively. Totally 67 patients (62%) had total tumor resection, 41 (38%) had subtotal tumor resection, which was documented on postoperative MR images.
In addition to surgery, each patient was evaluated and treated endocrinologically. The diagnosis of DI followed the criteria published previously.,
The approval from the Ethics Committee of Fujian Medical University was obtained on June 13, 2012. The prior written and informed consent was obtained from every patient.
All of the MRI scanning was performed with 3.0T magnetic resonance scanner (Tim Trio, Siemens Medical Solutions, Erlangen, Germany). T1-weighted images (400-500/8-15, TR/TE) were obtained before and after the Gd-DTPA injection. The parameters used were as follows: scanning field, 180 mm × 180 mm; matrix, 320 × 240; Section thickness, 2. 5 mm. GD-DTPA was used as contrast agent with the dose of 0. 1 mmol/kg and with a velocity of 2 ml/s. In all the patients, the axial, sagittal, and coronal series were evaluated, to specifically identify the residual pituitary issues including the posterior pituitary bright spot (PPBS) and the residual adenohypophysis (RAH) on the pre- and post-gadolinium images, respectively. All the Pre- and postoperative MR images were independently evaluated by at least two observers, including a neuroradiologist and a neurosurgeon, and the data were ultimately classified by consensus.
PACS medical imaging system (INFINITT SH Co., Ltd., Shanghai, China) was used to measure the diameter of pituitary adenoma, as well as the volumes of pituitary adenoma and pituitary tissue. The “platform-like volume calculation formula” was used to calculate tumor volume in coronal enhanced T1WI, as we reported previously. In MRI plain and enhanced scanning images, pituitary tissues were presented as PPBS and pituitary gland tissues. PPBS is located at the posterior part of sella turcica, while adenohypophysis is usually surrounding the tumor. According to the classification criteria reported by Saeki et al., pituitary adenomas shapes were classified into hourglass type with indentation and barrel type without indentation at the diaphragmatic level. On coronal post-contrast T1-weighted MR imaging, the preoperative distortion patterns of the pituitary gland were classified as superior, superolateral and lateral. Pituitary adenomas are divided into suprasellar extension group and non-suprasellar extension group depending on whether the upper part of the tumor extends bulges upwards into the suprasellar cistern. The extent of postoperative restitution of the normal residual pituitary tissue was subdivided into the following four groups by Di Maio based on the extent of morphological restitution: Group 1-normal residual gland or almost normal, Group 2-more than 50% restitution, Group 3-less than 50% of the normal residual gland, Group 4- barely visible or absent residual gland. In the present paper, on the postoperative (4th-6th month) MR images, the appearance of the residual pituitary gland were also subdivided into four groups: G1, measured residual volume = 75-100% of adult pituitary volume; G2, measured volume = 50-75% of adult pituitary volume; G3, measured volume = 20-50% of adult pituitary volume; G4, measured volume < 20% of adult pituitary volume. Pituitary gland volume (PGV) estimates were calculated by summing all voxels within the traced region on each slice. For each individual, pituitary volume change was measured as deviation from the predicted (or average) pattern of volume change. The average adult gland measures 6 mm superior-inferior × 9 mm antero-posterior × 13 mm medio-lateral and weighs 0. 6 g.,
Pathological examination of tumor tissues
Histological sections of the surgical specimens were examined under standard hematoxylin and eosin preparation, and all 108 cases were also studied immunohistochemically by the senior pathologist of our hospital. The tissues were classified according to the criteria published by the World Health Organization in 2004, which showed that 18 tumors were prolactin-secreting, 28 were growth hormone-secreting, 27 were gonadotropin (luteinizing hormone or follicle-stimulating hormone)-secreting, 14 were null cell adenomas, 5 were thyroid-stimulating hormone-secreting, 7 were adrenocorticotrophic hormone-secreting, and 9 were multiple hormone-secreting.
All the statistical analyses were performed using SPSS 19.0 (IBM, Armonk, NY, USA) and P value less than 0. 05 was considered as statistically significant. Measurement data were expressed as mean ± standard deviation (SD) and the t test or variance analysis were performed. Enumeration data were expressed by percentage and Chi-square test or Fisher's exact test was adopted.
PPBS (+) shows no relation to gender or age of patients with pituitary adenoma
On preoperative MR T1-weighted imaging, 86 patients (79.6%) showed PPBS(+), while 22 patients (20.4%) showed PPBS(-). Patients with PPBS(+) group had an mean age of 46.0 ± 13.0 years (range, 16-69 years), and male to female ratio of 21:22; patients with PPBS(-) group had an mean age of 38.1 ± 15.2 years (range, 19–73 years), and male to female ratio of 7:4 (P > 0. 05). The results suggest that the presence of PPBS is not related to gender or age of the patients. While on the 4th-6th months postoperative MR imaging, 103 patients (95. 4%) showed PPBS (+), which was increased than that on preoperative MRI.
PPBS (+) shows more in hourglass shape of pituitary adenoma
To determine how PPBS is related to adenoma height, volume, and shape, we examined the adenomas from all 108 patients. In sagittal images of MRI-T1WI from patients with tumor height >20 mm (n = 63), the average height of adenoma in patients with PPBS(+) was 29 ± 7 mm (range, 20-52 mm, n = 47), which was not significantly different from that in PPBS(-) (average, 32 ± 7 mm; range, 21-43 mm, n = 16). In addition, the average volume of adenoma in patients (tumor height >20 mm) with PPBS(+) was 9100 ± 5800 mm3 (range, 3300-21600 mm3), which was not significantly different from that in patients with PPBS(-) (average, 11700 ± 900 mm3; range, 3000–25700 mm3) (P > 0. 05) [Table 2]. The presence of PPBS(+) was 86.7% (39/45) in patients with pituitary adenomas height ≤20 mm, while 74.6% (47/63) in patients with tumor height >20 mm (P < 0. 05; Chi-square test) [Table 2]; [Figure 1]. According to pituitary adenoma shapes, 53 patients among the 108 cases in the present study had hourglass shape, 16 patients had barrel shape, and the remaining 39 cases cannot be classified according to shapes. The sign of PPBS(+) was visible in 90.6%(48/53) of the hourglass shape group, wherever 43.8% (7/16) in the barrel shape group (P < 0. 05, Fisher's exact test) [Table 2]. These results indicate that PPBS(+) on preoperative MR images may be shown more in patients with hourglass shape of pituitary adenoma, and may be reduced when pituitary adenoma height increases to some degree.
Extension direction of pituitary adenoma affects the presence of residual pituitary tissues (RPT) on pre-operative MRI
To identify the relationship between PPBS(+) and RPT(+) shown on preoperative MRI, T1WI and contrast-enhanced MRI were employed. On preoperative MRI, 86 out of 108 patients showed PPBS(+) on T1WI, and 65 out of 108 patients showed RPT(+) on enhanced MRI. Among the 108 patients with pituitary adenoma, 25 had no suprasellar extension and 83 had suprasellar extension. The rate of PPBS(+) in patients with suprasellar extension (77. 1%) was significantly lower than that in patients without suprasellar extension (88. 0%) (P < 0. 05). In addition, the percentage of RPT(+) in patients with suprasellar extension (55. 4%) was also significantly lower than that in patients without suprasellar extension (76. 0%) (P < 0. 05). Furthermore, the rate of RPT(+) in patients with PPBS(+) was significantly higher than that in patients PPBS(-) (P < 0. 05). The results suggest that suprasellar extension of pituitary adenoma may cause residual pituitary tissue (adenohypophysis and neurohypophysis) to receive more severe compression.
Growth hormone adenoma tends to extend towards undersellar direction, while nonfunctioning adenoma tends to extend towards suprasellar direction
To examine the relationship between pituitary tissue location and immunohistochemical types, the 65 patients with RPT(+) on preoperative MRI were classified into growth hormone type (18 cases), prolactin type (12 cases), gonadotropin type (25 cases), adrenocorticotropic hormone type (5 cases), and null cell type (5 cases). The data showed that growth hormone adenoma tended to grow and extend under the sella, making pituitary tissues being above the tumor [Figure 2]a and [Figure 2]b. Gonadotropin adenomas grew towards suprasellar direction, making pituitary tissues being located on one side of the tumor [Figure 2]c and [Figure 2]d. Totally, 13 out of the 18 growth hormone adenoma cases had RPT superior and superolateral displacement (72. 2%), which was significantly different from that among the 25 gonadotropin adenoma cases (13 cases; 52%) (P < 0. 05, Fisher's exact test) [Table 3]. In addition, study on the relationship between pituitary locations of 65 pituitary adenoma patients and Knosp-Steiner grades showed that patients with lateral, superior, and superolateral pituitary location had significantly different Knosp-Steiner grades (χ2 = 2. 564, P < 0. 05). Of note, patients with lateral pituitary location had relatively low Knosp-Steiner grade [Table 4]. These results indicate that growth hormone adenoma may tend to extend towards undersellar direction, while gonadotropin adenomas tends to extend towards suprasellar direction.
Patients with lower PPBS intensity are more likely to have postoperative DI
To study the relationship between DI and PPBS, the postoperative DI data were reviewed. Among 86 patients with PPBS (+), 9 cases had postoperative DI (10. 5%), which was significantly lower than that among 19 patients with PPBS (-) (7 cases of postoperative DI; 36. 8%) (P < 0. 05) [Table 2]. PPBS signal intensity was measured by comparing to the pontine region. In the 13 iso-intensity signal cases, 4 had postoperative DI. In the 20 cases of high intensity (medium signal intensity) PPBS, 6 cases had postoperative DI. In the 53 patients with higher signal intensity PPBS, 9 had postoperative DI. The three groups of data were significantly different from each other (P < 0. 05) [Table 5]. These results suggest that patients with lower PPBS intensity are more likely to have postoperative DI.
Lateral displacement of RPT usually has the best recovery according to MRI
To understand the relationship between pituitary location on preoperative MRI and postoperative pituitary recovery of the 65 patients who showed pituitary tissues on preoperative enhanced MRI, the recovery status of pituitary tissue after surgery was followed up. All 65 patients showed visible pituitary tissues in T1WI before and after surgery. Among the 65 patients, there were 24 cases of G1, 24 cases of G2, 11 cases of G3, and 6 cases of G4. For G1 and G2, 24 cases were lateral type (85. 7%), 18 cases were superior type (66. 7%), and 6 cases were superolateral type (60%), with significant differences among the 3 types (P = 0. 032) [Table 6]. Furthermore, the postoperative recovery of lateral type of pituitary was better than that of superior and superolateral type [Figure 3]. These results indicate that lateral type of pituitary usually has the best recovery on MRI.
As the tumor compression release after operation, some cases of PPBS(-) on preoperative MR images may show PPBS(+). In the present study, 86 cases (79. 6%) show PPBS(+) preoperatively, while 103 cases (95. 4%) showed PPBS(+) postoperatively, which may resulted in the rebuilding of neurohypophysis pathway and recovery of posterior pituitary function. To some extent, the signal intensity of neurohypophysis is related to compression it receives. The signal intensity of PPBS may also associate with the amount of ADH stored in posterior pituitary. Higher compression may reduce the transportation and storage of ADH in neurohypophysis. Careful MR imaging observations of the hour-glass type adenoma by Saeki et al. showed that the distal end of PPBS was constantly located above the indentation. In barrel type, PPBS was demonstrated in varying locations, such as in the sella, at the distal pituitary stalk above the diaphragm sella, along the whole pituitary stalk, or in a combination of these areas. Therefore, the degree of development of indentation may affect the intensity of transportation interruption and the resulting blockage of neuro-secretory granules at the pituitary stalk. In addition, the percentage of patients with PPBS in the group with hourglass shape adenomas is significantly higher than that in the group with barrel shape adenomas. Giant adenomas can compress supraoptic nucleus and paraventricular nucleus, as well as the pituitary stalk, which may result in reduction of ADH secretion and transmission that leads to the absence of PPBS.
Pre-operative DI is rare in pituitary adenoma. The incidence of postoperative DI in this study is 14.8%, which is consistent with the reported incidence of 0.5–15% in pituitary adenomas. As reported in the literature, 96.8% of patients with central DI, show low signal of the posterior pituitary on MR T1-weight images. The occurrence of postoperative long-term DI may have some relation with the absence of PPBS, indicating that dysfunction in the secretion and storage of ADH leads to DI. However, some patients with idiopathic DI still have PPBS, leading to doubts about using the absence of PPBS as the diagnostic criterion for central DI. In the present study, patients without pre-operative PPBS have higher incidence of postoperative DI, which suggest dysfunction of posterior pituitary, and we should pay more attention to the protection of posterior pituitary in pituitary adenoma surgery.
Zada et al. showed that growth hormone adenomas tend to grow below the sella, and nonfunctioning pituitary adenomas tend to extend above the sella. Salvatore et al. studied 79 cases of MRI of pituitary adenomas before surgery, and classified the displacement of residual pituitary tissue as superior, superolateral or lateral in location. However, these authors find no statistically significant difference between these preoperative pituitary tissue locations and tumor types. In the present study, we discover that the preoperative locations of pituitary tissues in MRI are significantly different between patients with pituitary adenomas of different tissue types. Growth hormone adenoma patients tend to have superior pituitary tissues before surgery, while gonadotropin adenoma patients tend to have lateral pituitary tissues before surgery. This observation is consistent with the report by Zada et al. Certainly, further study should be needed because of the limit of cases in this study.
In the present study, 60.2% of the 108 patients show pituitary tissue on preoperative enhanced MR images, while 95% on the postoperative enhanced MR images. This change may be due to the re-expansion and repositioning of pituitary tissues after surgery. The re-expansion degrees are related to the size and location of adenomas. Bonneville et al. showed that nearly all pituitary tissues in patients with pituitary adenomas of diameter more than 20 mm were pushed out of the sella. Adenomas that expand superiorly tend to push pituitary tissues above the sella, the normal pituitary tissues in about 49% nonfunctioning adenomas are pushed outside of the sella due to the large size of the tumor, and 76% secreting adenomas push the pituitary tissues aside or outside of the sella. Preoperative evaluation of cavernous sinus invasion has important effect on surgery and prognosis. However, the medial wall of cavernous sinus cannot generate its image in T1WI at all times., Pituitary tissues with lateral displacement can indirectly indicate that ipsilateral cavernous sinus is not affected by tumor invasion,, because it is unlikely that cavernous sinus is invaded by adenomas across pituitary tissues. Consistent with this, the present study shows that patients with lateral pituitary tissues before surgery do not have ipsilateral cavernous sinus invasion by tumors.
The present study shows that the postoperative pituitary tissues expand and recover to the 32–85% of normal pituitary size. Pituitary tissues with lateral location have better recovery compared with other types. In most cases, the process of recovery and re-expansion of pituitary tissues after surgery is a slow and progressive process. Patel et al. found that post-operative continued imaging of residual tumor or resection cavity could last for up to 3 months. We suggest that the time point of 4th to 6th months after surgery is ideal for the evaluation of pituitary MRI, which is consistent with the results by Di Maio et al. The postoperative mass may represent a combination of residual tumor, edema, postoperative hemorrhage and hemostatic material. At the 4 to 6 months post operation, the hemostatic material such as gelatin sponge and the packed autologous tissue will be absorbed, the effect of which on imaging can be reduced or avoided. While in the early stage after surgery (within 1 week post operation), residual hemostatic materials may cover the remaining pituitary tissues.
Although the residual pituitary tissue and residual tumor can be distinguished from post-operative changes by experienced practitioners using high-resolution MRI, finding a small residual tumor remains a challenge for us because there is a dilemma whether the surgeon should re-explore or follow up conservatively. Functional imaging modalities in combination with MRI may improve identification of residual tumor and pituitary tissue. For example, it is reported that F-18-FDG and Ga-68-DOTATATE PET/MRI is useful for distinguishing pituitary adenoma and normal pituitary tissue from post post-operative changes. When diagnosis is difficult using MRI alone, PET/MRI may be an ideal tool to distinguish between pituitary microadenomas and pituitary gland.
In conclusion, the present study demonstrates that the rate of PPBS on T1WI is decreased as the increase of pituitary adenoma height. Preoperative PPBS signal intensity is negatively related to DI. Pituitary tissues in patients with growth hormone adenomas are usually located above the pituitary adenomas, while pituitary tissues in patients with gonadotropin adenomas are mostly located aside the adenomas. The postoperative recovery of lateral type of pituitary tissue is better than that of upper and superolateral types.
This work was supported by Fujian Provincial Natural Science Foundation of China [grant number 2015D014] and Fujian Provincial Key Project of Science and Technology Plan of China [grant number 2018Y0067].
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]