Introduction

The operative treatment of pituitary adenomas causing acromegaly is a widely accepted first line standard therapy [1–3]. The pertinent algorithms developed in consensus conferences or by scientific societies suggest surgery as a first procedure for most patients. After all, the mechanical extraction of growth hormone-secreting tumour tissue acts rapidly and leads to normal biochemical parameters in the majority of the patients who undergo initial surgery. Standard treatment implies the attempt to most aggressively resect the tumoural tissue selectively, by identifying and preserving the normal pituitary gland in an attempt of reaching remission of acromegaly without sacrificing pituitary function [4–6]. This is most easily achieved during transsphenoidal operations, which allow direct visualization of normal and pathological tissues.

To date, more than 95 % of patients operated upon for acromegaly undergo transsphenoidal surgery [4, 7, 8]. The biochemical parameters used to define remission changed over time [9, 10] and became more stringent. Nevertheless, with development of surgical techniques, improved imaging and visualization techniques, accumulated experience and interdisciplinary co-operations one gets the impression that the amount of tumour tissue resected during the operations and the rate of “normal” growth hormone and IGF-1 levels achieved, increased considerably.

Surgical techniques

As already mentioned, the vast majority of pituitary tumour operations in patients with acromegaly are to date performed by the transsphenoidal approach [4]. There are many variations possible, which start with the positioning of the patient. While some surgeons, like us, prefer to operate on a patient in supine position with the head slightly extended, others favor a semi-sitting position, thus facing the patient [6, 7, 11]. Some surgeons still use an image intensifier for navigation purposes; others prefer the routine use of neuronavigation. This operation can be performed with and without dissection of the nasal septal mucosa. Dissection of a submucosal tunnel is mostly performed following a medial nasal incision. A nasal speculum keeps the mucosal tunnel open. At this stage, the operating microscope is usually brought into place. The vomer is used as a midline orientation. A sphenoidotomy is performed. Septations of the sphenoid sinus are resected so one has an unobstructed visualization of the sellar floor through the sphenoid sinus (transsphenoidal). In patients with incomplete pneumatisation of the sphenoid sinus, extensive drilling is required to expose the posterior basal surface of the tumour. Thereafter, the basal dura of the sella is incised. Frequently soft tumours produce spontaneously through the dural opening. Variously shaped curettes and microforceps are available for tumour extraction. The normal pituitary, which is often deformed and compressed against the cavernous sinus or into the posterior portion of the fossa is identified and every possible attempt is made to preserve it.

The extent of tumour resection is estimated by inspection and palpation of the tumour cavity. Frequently the arachnoid that covers the superior surface of the tumour decends into the intrasellar space. It has a smooth surface and can thus be easily recognized. However, whenever it descends in several folds, each of them has to be probed separately since tumour may be hidden within the pouches. Generally speaking, the wider the connection between intra- and extra-sellar tumour portions, the easier is the resection of a large adenoma [11, 12].

Invasive nature of a tumour restricts its resectability. Localized invasion, may be dealt with by resection of the invaded mucosa of the sphenoid sinus and drilling invaded bone of the skull base. Parasellar tumour is usually traced along the perforations that have been created by the tumour. Tumour, that is located lateral from the carotid artery can usually not be resected. The term “enlarged adenomectomy” refers to a small layer of normal gland (“periadenoma”), that is resected around the adenoma. The most perfect overview of an adenoma cavity after selective adenomectomy is usually achieved in microadenomas and moderately sized macroadenomas.

Even with utmost experience and optimal technical equipment there are still pituitary adenomas that cannot be resected completely. Mirrors and particularly endoscopes allow improved visualization of tumour portions, particularly those localized outside of the midline. There are several methods to reconstruct the sellar floor by implants. Many neurosurgeons apply a regimen of external lumbar CSF drainage for a few days in case of intraoperative CSF leaks. Often a nasal tamponade is required [4, 7].

As an increasingly frequently practiced variant, a direct peri-nasal approach to the sphenoid sinus can be chosen. In this case the speculum is guided by radiofluoroscopic control or navigation techniques to the rostrum. Mucosal incision and sphenoidotomy are then performed just in front of the vomer. The operation is then continued as just described [11]. Endoscopic transsphenoidal surgeons use this direct peri-nasal approach and do not need a nasal speculum [5]. There are technical advantages and disadvantages of endoscopic versus microsurgical techniques [5, 13, 14] Still, many data, particularly on the long-term follow-up results largely derive from microsurgery series, in some of which the endoscopy was additionally utilized when the surgeon felt that it could be useful [8]

Extrasellar tumour extension of the adenoma is no more a contraindication for transsphenoidal surgery. With the extended transsphenoidal approaches, which imply opening to the tuberculum sellae or drilling through the clivus, tumour location and extensions, which in the past presented contraindications for transsphenoidal surgery, can to date also be treated via the nasal route. Unfortunately, one of the disadvanteges of the extended transsphenoidal approach is a high frequency of postoperative CSF leaks [5].

In certainly less than 10 % of patients who harbour tumours with extreme asymmetrical intracranial extension which have a poor communication with intrasellar portions, still transcranial approaches are needed to reduce the tumour mass [7, 8, 15, 16]. Mostly, frontal or fronto-temporal, basal craniotomies are used. The tumour tissue is dissected under direct vision from the arachnoid layers of the basal cisterns, blood vessels, cranial nerves and the infundibulum. It is not only much more difficult to preserve pituitary tissue and function but also to resect a tumour completely during transcranial surgery [15].

Surgical results

Expert neurosurgical centers with a high degree of focus on pituitary diseases reach normalization rates between 75 and 90 % in growth hormone-secreting pituitary microadenomas and some 45–70 % in pituitary macroadenomas. In most papers there is no distinction between patients with acromegaly and gigantism. Table 1 provides an overview on a selection of surgical series published between 2000 and 2016. The individual results between the centers are difficult to compare, since the selection of which patient is going to be operated and which one will be exposed to primary medical therapy is not homogenously performed.

Table 1 Results of transsphenoidal surgery for GH-secreting pituitary adenomas using current criteria for remission [2, 9, 10]
Full size table

There are unequivocal prognostic factors such as the size of the tumour, the magnitude of preoperative GH and, respectively, IGF-1 oversecretion. One extremely important factor is whether the tumour harbors an invasive growth pattern, particularly as the cavernous sinus is concerned [17–19]. The degree of parasellar extension can be assessed by preoperative magnetic resonance imaging (Fig. 1).

Fig. 1
figure 1

Invasive pituitary macroadenoma of the left cavernous sinus and encasement of the carotid artery (Knosp Grade IV) in a 41-year-old male patient in coronal T1-weightes MRI sections. The large tumoru (a) is decreased in size (b) but invasive portions remained. Preoperative GH dropped from 12.9 to 3.3 ng/ml 3 months postoperatively and IGF-levels were decreased from 781 to 535 ng/ml

Full size image

However, the frequently employed Knosp classification correlates imaging findings with the percent likelihood of invasion [20, 21]. Generally speaking, once there is massive invasion, the long-term surgical normalization rate approaches zero. A normal postoperative MRI is usually considered a prerequisite for surgical normalization (Fig. 2), but in a few cases persistent tumour has been documented by imaging, with biochemical normalization according to strict criteria. These are, according to the Endocrine Society Practice Guidelines, proposed such as a random GH below 1 ng/ml, suppression to below 0.4 ng/ml, also associated with a normal IGF-1 [2]. Microscopic and endoscopic transsphenoidal operations produced similar results [22, 23]. In one center, a direct comparison between an experienced microsurgeon and an experienced endoscopic surgeon revealed no differences in remission and complication rates, respectively [22]. The assay variabilities between different GH and IGF-1 determinations raise some concern in this connection [24]. Moreover, the excellent results published from a few expert centers are not to be expected in each and any place across a country. An analysis of the British Acromegaly Registry revealed enormous differences in the remission rate reached in 26 different neurosurgical departments [25]. Patient numbers varies from 3 to 79. The rate of normal IGF-1 postoperatively, as determined at least 3 months after the operation ranged from 0 % in one center to 68 % in another one and averaged 39 %. Likewise the complications rates seem to strongly correlate with the experience of the surgeon and the patient load of the neurosurgical center [16, 26, 27]

Fig. 2
figure 2

68-year-old female patient presenting with a non-invasive 9 mm large microadenoma adjacent to the right cavernous sinus depicted in coronal T1-weighted MRI (a). Postoperatively there is no more tumour visible (b). GH-(0.33 ng/m) and IGF-1-levels (127 ng/ml) respectively were normalized

Full size image

Socio-economic considerations

Surgery is an ablative procedure and thus, treatment costs usually occur only once, whereas medical treatments are associated with continuous expenses over the lifetime of the patient. The costs for treatment are thus relatively low and the remission rate is high, particularly in experienced, specialized reference centers. Germany, to date (in 2016), the reimbursement for either transsphenoidal or transcranial surgery, unless this is associated with complications, is 7895 € [28], whereas an annual treatment with octreotide LAR (30 mg/every 4 weeks) costs 33.099 €, lanreotide LAR (120 mg/every 4 weeks). pasireotide LAR (60 mg/4 weeks) 62.745 € and pegvisomant (30 mg daily) 109.950 € [29].

Financially considered, the health system of a country saves a lot of money, particularly of the remission rate of the surgical centers is high, the complication rate is low and if young patients are treated surgically, of whom most do need long-term expensive medical treatments. Resources are needed for these who require combined treatments. In one hypothetical calculation performed in the United States, the savings for one 40 year-old patient with a pituitary microadenoma and a life expectancy of 78 years, amount to 2.5 million US$ [30].

Debulking

Debulking denotes surgery in a patient in whom a biochemical remission of acromegaly is not expected for that the tumour harbors features that would prevent total excision. Parasellar encasement of the carotid artery, giant size and excessively increased growth hormone levels are all unfavorable prognostic parameters. Since normalization is not expected after the operation, postoperative therapies are required. The rationale of debulking is, that one can expect a better out-come in response to subsequent medical treatments after an operation that has been successful at lowering GH levels, even if not to normal levels, by removing some or most of the pituitary tumour mass [31].

The studies which are available in the medical literature which have compared the efficacies of preoperative and postoperative medical treatment with somatostatin analogs have convincingly shown the value of debulking. After surgery, in all four studies, the rate of patients achieving remission of the disease by (in some studies identical) drug treatments were significantly improved [32–35]. The finding that percent reduction of growth hormone levels correlates closely with percent resected tumour volume in acromegaly also supports the value of debulking surgery [36].

Reoperations

In contrast to the initial management of an acromegalic patient, reoperations do not have a fixed place in the treatment algorithms [2]. There are several scenarios (Fig. 3), in which reoperations can be considered, such as an unsatisfactory amount of tumour resection when a much more complete excision was to be expected on the basis of preoperative imaging. Tumour progression or the failure to achieve a remission under medical treatments are other indication for redo procedures. Furthermore, one could imagine that it makes sense to reduce a tumour’s volume to make it suitable for treatments in which size restrictions exist, e.g. radiosurgery.

Fig. 3
figure 3

Modified algorithm of the Endocrine Society’s Clinical Practiceguidelines according to Katznelson et al. [2] with the author’s suggestions in which a re-operation may be considered: 1 if a non-invasive remnant is visible on post-op MRI-scan, 2 to enhance the effect of medical therapy by further debulking, 3 debulking procedure when residual adenoma is increasing in size during medial therapy, 4 ability to apply radiosurgery instead of conventional radiotherapy by further debulking of residual adenoma

Full size image

Scarring resulting from the previous operation, tissue defects which cause a distortion of the anatomy, partial cisternal herniation and other postoperative sequelae make a reoperation technically more difficult and challenging. The results are generally less favourable than in primary operations and the rate of some complications is definitely increased in pituitary tumour reoperations [37–39].

Preoperative medical therapy

The issue of preoperative medical therapy before surgery is still a matter of controversy, at least if a potential improvement of the surgical results is considered. Most neurosurgeons would agree that primary medical therapy as a pretreatment can improve the clinical condition of an acromegalic patient who is severely affected by the potential comorbidity. Whether the normalization rates by surgery are affected by pretreatment with somatostatin analogs is another issue. Whereas some authors [40, 41] find no positive effect on such pretreatments, others claim an improvement of surgical normalization rates [4, 42–44]. It is conspicuous that the groups who found a difference had surprisingly low remission rates in their cohorts of non-pretreated patients which ranged from 18 to 24 % and thus achieved much worse results that the average recorded from 26 centers in Great Britain in a study which has been already quotes above [25].

We know from morphological studies that the visible effects of somatostatin analog treatment on adenoma tissue in acromegaly are minor [45]. Moreover, at least in one study, the finding of improved results was only found 3 months after the operation and no more maintained after 1 year [44]. Thus, in summary, a definitive conclusion as to the usefulness of preoperative treatment with somatostatin analogs is lacking, let alone an appreciation of the effects of dopamine agonist or pegvisomant treatment where as yet there are no data.

Intraoperative imaging

The imaging method of choice for pituitary lesions is magnetic resonance scanning, since it has the best soft tissue contrast. For some time intraoperative MRI became available which allowed to objectively assess the radicality of tumour extraction. Studies in non-functioning pituitary adenomas showed in increased rate of total tumour resections and also a convincing increase in the amount of tumour tissue extracted [46]. When residual tumour was depicted the surgeon could trace and try to remove it, since the patient remained in the operating theatre.

Only a few series with application of intraoperative MR imaging in acromegaly are reported. Fahlbusch et al. [47] improved the amount of adenoma resection in 5 out of an unfavourable cohort of 23 patients using a 1.5T high field system. However, in only two of these radical resection was accomplished. Bellut et al. [48] performed 39 operations in 39 acromegalic patients, using a low field 0.15T system. Intraoperative MRI revealed residual tumour in eight patients, but only two of these achieved a biochemical remission. Tanei et al. [49] treated seven patients with acromegaly and achieved a remission in six of these. In three of these six patients, residual adenoma tissue was identified in the basis of intraoperative MRI and resected in the second step of the operation. In the most recent study, Netuka et al. [50] operated a total of 105 patients with acromegaly with a total remission rate of 60.9 % in their 3T intraoperative MRI suite. There were 16 microadenomas and 89 macroadenomas. Re-resection on the basis of intraoperative MRI were performed in 22 patients, of whom 9 had normal GH and IGF-1 levels postoperatively. Thus, in all of these investigations in patients with acromegaly, an improvement in the surgical results was achieved. The benefit was pronounced in some and only slight in other series.

Moreover there are many other technical supports and gadgets available which help the surgeon in individual situations which details discussion was performed elsewhere [41].

Conclusion

After a long evolution of finding indications, development of appropriate surgical techniques and perioperative managements, particularly the transsphenoidal operation in acromegaly is to date a safe, relatively inexpensive and efficient treatment option. There is a huge bunch of medical literature supporting its place and value. Thus, it is not surprising that its use is frequently recommended as the first step of treatment by expert conferences and management recommendations. It is, however, important to realize that the excellent results and safety data published by expert centers cannot be expected in each and any place across any country.