Endoscopic vs. microscopic transsphenoidal pituitary surgery: a single centre study

Endoscopic pituitary surgery has shown promising results. This study reports the experiences of experienced microscopic pituitary surgeons changing to the endoscopic technique, and the beneficial effects on the postoperative outcomes. 45 transsphenoidal endoscopic-assisted surgeries performed in 2016–2017 were compared with 195 microscope-assisted surgeries performed in 2007–2017 for pituitary adenoma. Tumour size, hormonal status and vision were assessed preoperatively and 3–5 months postoperatively. Cases were identified through electronic patient records. GTR was achieved in 39% of the endoscopic operations vs. 22% of microscopic operations, p = 0.018. Mean duration of surgery was 86 min (77–95) with the endoscopic technique vs. 106 min (101–111) with the microscopic technique, p < 0.001. New hypothalamus–pituitary–adrenal axis deficiencies were observed after 3% of endoscopic vs. 34% microscopic operations, p = 0.001, and overall fewer postoperative pituitary deficiencies were observed in the endoscope-assisted group. Complications within 30 days of surgery occurred in 17% of endoscopic operations vs. 27% of microscopic operations (p > 0.05). Normalization of visual impairment occurred in 37% of the cases with preoperative visual impairment in the endoscopic group vs. 35% of those in the microscopic group (p > 0.05). The endoscopic technique performed better as a surgical procedure for pituitary adenomas. We found no statistically significant differences in complication rate or visual improvement between the two techniques.


Scientific Reports
| (2020) 10:21942 | https://doi.org/10.1038/s41598-020-78823-z www.nature.com/scientificreports/ clinical practice and the associated learning curve for surgeons using it as a new technique for transsphenoidal pituitary surgery. These studies all revealed the endoscopic technique to show promising results on gross tumour resection 18,19 , postoperative pituitary function 18,20 , visual field changes 21 and duration of surgery 22 except in two studies, where no difference between endoscopic and microscopic pituitary surgery could be detected 23,24 . The present retrospective study was conducted to further examine the learning curve of neurosurgeons experienced in the microscopic approach who changed to the endoscopic approach. This, combined with detailed studies on the postoperative status, have not previously been thoroughly investigated. We thus compared pre-and postoperative MRI scans, complications within 30 days of surgery, biochemistry, and computer perimetry for patients treated with either the endoscopic or the microscopic technique.

Results
Inclusion. Of the 64 transsphenoidal PA procedures performed in 2016 and 2017, four cases had unavailable pre-or postoperative MRI scans. Fifteen operations used the microscopic technique, giving 45 cases in the endoscope-assisted group ( Table 1). As previously reported 8 , 180 procedures using the microscopic technique had been performed during 2007-2015. These 180 were added to the 15 cases from 2016-2017 to give 195 in the microscope-assisted group (Table 1).
Resection. Gross  Pituitary function. The HPA-axis was intact preoperatively in 67% of patients in the endoscopic group vs.
64% in the microscopic group (Table 1). New HPA-axis deficiencies developed after surgery in 3% of patients in the endoscopic group and 34% in the microscopic group, p = 0.001 (Table 3). HPA-axis function normalized after surgery in 27% of patients in the endoscopic group and 17% in the microscopic group, p = 0.444 (Fig. 2). The HPT-axis was intact preoperatively in 55% of patients in the endoscopic group vs. 61% in the microscopic group (Table 1). New TSH-dependent deficiencies developed after surgery in 15% of patients in the endoscopic group compared to 38% in the microscopic group, p = 0.030 (Table 3). HPT-axis function normalized after surgery in 16% and 5% respectively, p = 0.092.
The HPG-axis was intact preoperatively in 44% of patients in the endoscopic group vs. 43% in the microscopic group (Table 1). While no patients in the endoscopic group developed new deficiencies after surgery, 38% of patients in the microscopic group did, p = 0.011 (Table 3). HPG-axis function normalized after surgery in 32% of patients in the endoscopic group compared to 19% in the microscopic group, p = 0.16.
Visual field impairment. Visual field assessment by computed perimetry showed that 40% of patients in the endoscopic group had intact vision before surgery compared to 31% in the microscopic group, p = 0.03 (Table 1). Surgery-induced visual field impairment was observed in 3% of patients in the endoscope-assisted www.nature.com/scientificreports/ group and 7% in the microscope-assisted group, p = 0.27 (Table 3). Postoperative visual improvement was observed in 37% of patients in the endoscopic group and 35% in the microscopic group, p = 0.836 (Table 3).

Discussion
This retrospective study of transsphenoidal surgery for pituitary adenomas examined the introduction of an endoscope-assisted technique and the results from the first 45 procedures. The findings suggest that introduction of the endoscopic transsphenoidal approach was associated with a relatively steep learning curve for the surgeons, by improving the degree of surgical resection, reducing the duration of the surgical procedure and reducing the overall complication rate compared to the standard microscope-assisted technique. The present study adds to the literature in favour of using the endoscopic technique for transsphenoidal pituitary surgery. We found that significantly fewer patients undergoing endoscopic surgery had tumour remnants (i.e. a higher rate of gross total resection) compared to those undergoing the microscopic technique. Our findings are similar to previous studies 15,19,20 , but we document for the first time that while the surgeon's learning curve is steep, the benefits of using the endoscope are apparent early after changing technique.
Duration of surgery was shorter for the endoscopic technique (mean 86 min, 95% CI 77-95) compared to the microscopic technique (mean 106 min, 95% CI 101-111). The learning curve illustrated a steady decrease in duration of surgery from the introduction of the endoscopic technique in 2016. These findings are comparable to previous studies showing shorter duration of surgery with the endoscopic technique 21 .
The mean resected volume was smaller in the endoscopic group (2.75 cm 3 (2.08-3.43) vs. 4.90 cm 3 (4.25-5.54), p = 0.002). There was, however, no difference in mean postoperative tumor volume. The reason for this lower  www.nature.com/scientificreports/ resected volume in the endoscopic group compared to the microscopic group is probably, that the mean preoperative tumor volume in the microscopic group were larger than in the endoscopic group. We do not have a solid explanation for this except that (1) surgery is recommended on adenomas larger than 20 mm in diameter and (2) there is a general tendency to operate on smaller adenomas after the introduction of the endoscope, especially since the endoscopic procedure at our institution proved more tolerable by the patients. Only 2% of patients in the endoscopic group were observed with postoperative rhinoliquorrhea compared to 11% in the microscope group. These results are comparable to the findings of Messerer et al. 19 and Eseonu et al. 21 and, while not quite statistically significant, they further support the advantage of using the endoscopic technique for transsphenoidal pituitary surgery.
We found that fewer patients operated with the endoscopic technique developed postoperative anterior pituitary gland deficiency compared to the microscopic group. These results are comparable to other studies 19,22 . Postoperative pituitary function is most often omitted in systematic reviews or meta-analyses 15,23 . However, one meta-analysis combined the results from six previous studies to investigate postoperative pituitary function. The results failed to show statistically significant differences between different surgical procedures 16 , only reporting postoperative hypopituitarism in 3% (8/263) in the endoscopic group vs. 6% (16/262) in the microscopic group. This makes us doubt the assessment of pituitary deficiency in these studies. Our data support the tendencies previously reported, i.e. that the endoscopic procedure was gentler and had less effect on pituitary function than the microscopic technique.
The present study did not find a significant difference in the rate of curative surgery for patients with acromegaly or Cushing's syndrome. Like previous studies, our sample size of secreting adenomas was too low to draw conclusions on this parameter 23 , and further studies focusing on secreting adenomas are needed to clarify this.
We found no difference between the endoscopic and microscopic groups in the proportion of patients with complete recovery of vision after surgery. These findings are comparable to the findings of Muskens et al. 24 . Results vary between publications, however, possibly because postoperative improvement of visual field impairment is greatly influenced by the preoperative duration of optic apparatus compression. As shown by Anik et al. 25 , there are two important phases in visual improvement and its assessment, and the type of examination www.nature.com/scientificreports/ is important for the results. Furthermore, the timing of evaluation of visual field impairment is important when determining improvements after surgery.

Conclusion
We introduced the endoscopic transsphenoidal approach to replace the traditional microscopic transsphenoidal approach for pituitary surgery in 2016. The endoscopic approach shortened the duration of surgery and minimized the complication rate including postoperative pituitary dysfunction. In non-functioning pituitary macroadenomas, the endoscopic technique also improved gross total resection rates. Surgical technique. Endoscopic technique. After induction of general anaesthesia, cottonoids with Moffett's solution (a combination of cocaine, sodium bicarbonate and adrenaline given as standard practice in many rhinological procedures to provide local anaesthesia, vasoconstriction and decongestion 26 ) were placed for 10-15 min in each nostril. Neuronavigation with magnetic resonance imaging (MRI) and computed tomography (CT) was used to achieve endoscopic access to the sphenoid sinus (0° and 30° endoscopes coupled to a camera, Storz). Using a bi-nostril technique, the adenoma was removed using blunt curettes. Haemostasis was achieved with temporary placement of Spongostan and, if necessary, Surgiflo or FloSeal. The dura was sealed using Adherus (Stryker). Lumbar drainage of cerebrospinal fluid was not used. All patients were admitted to a semiintensive neurosurgical ward for postoperative monitoring. If there were no signs of early postoperative surgical complications, patients were routinely transferred to the endocrinology unit the following day.

Methods
Microscopic technique. As described in 8 , after induction of general anaesthesia, a submucosal paraseptal transsphenoidal microsurgical technique was used in all patients using a standard microscope (Zeiss, Pentoro) and neuronavigation (Medtronic) guiding. Adenomas were removed using blunt curettes, and haemostasis was achieved with temporary placement of Spongostan and, if necessary, Surgiflo or FlowSeal. The dura was closed using Tachoseal and, in some cases, the sellar floor was reconstructed using titanium mesh or septal bone. The nasal septum was repositioned and fixed with nasal packing for 12-24 h. Lumbar drainage was not used routinely.
Rhinoliquorrhea. Treatment of postoperative CSF leakage followed the institution local guidelines: in case of suspicion of early postoperative CSF leakage, the patients were kept in bed with up to 45° elevated head rest for additional three days. If a CSF leakage was still suspected, a lumbar drain was inserted and kept in place with the patient in bed for additional five days. In case the rhinoliquorrhea persisted after removal of the lumbar drain, operative closure was performed using muscle and fat graft.  27 classifications were used to classify adenoma extension based on the patients preoperative MRI (1.5 T or 3 T). Tumour volume was calculated using 3D volumetric analysis in Horos on axial sections of T1-weighted MRI images with contrast. Similarly, tumour remnant volume was calculated on routine follow-up MRI scans performed 4-6 months after surgery. In case of uncertainty about tumour delineation, scans were discussed with senior neurosurgeons. Duration of surgery and complications (rhinoliquorrhea, infection, intracranial haemorrhage and the need for additional surgery) within 30 days were registered. Complications were classified according to Ibañez et al. 18 .
Biochemical data. Pituitary hormones were assessed preoperatively and six weeks postoperatively using clinical biochemical data on the hypothalamic-pituitary-somatotropic axis (s-GH and s-IGF-1), the hypothalamicpituitary-gonadal (HPG) axis (s-FSH, s-LH, s-oestradiol for pre-menopausal women, s-total testosterone for men and s-sex hormone-binding globulin (s-SHBG)), the hypothalamic-pituitary-thyroid (HPT) axis (s-TSH, s-T4 and s-thyroid-binding globulin (s-TBG)) and the hypothalamic-pituitary-adrenal (HPA) axis (s-ACTH and Synacthen test). Deficiency in any axis was defined as biochemical data outside reference values for the specific hormone or if the patient was already on substitution therapy. Intact function was defined as hormone levels within the normal range, whereas increased values of IGF-1, cortisol or T4 led to specific tests for secreting adenomas. During the postoperative hospital stay, patients were monitored for the development of diabetes insipidus (DI) and hypopituitarism.