The Importance of Long Term Follow Up After Endoscopic Pituitary Surgery: Durability of Results and Tumor Recurrence
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.287675
Source of Support: None, Conflict of Interest: None
Keywords: Adenomas, endoscopic, pituitary, recurrences, surgery, transsphenoidalKey Message: Long term follow up (median: 144 months) demonstrate that recurrence of pituitary adenomas can be observed in about one third of patients. Previous surgery, cavernous sinus invasion and Ki-67>5% are factors associated with a higher recurrence rate.
Endoscopic endonasal approaches (EEA) have become increasingly popular in the last 15 years.,,,,, In neurosurgery, the most common application of this technique is resection of pituitary adenomas. Retrospective studies, multi-institutional collaborations, and systematic reviews of the literature have demonstrated the efficacy of this technique for pituitary adenomas.,,,,,,,,, At this point, it is widely accepted that EEA is at least as effective as transphenoidal microsurgery for resection of pituitary adenomas, and potentially superior in some cases, including for tumors with cavernous sinus invasion. One potential criticism regarding the results of EEA for pituitary adenomas is that most series have a short follow-up when compared to microsurgery series, and therefore, recurrence rates comparisons may be biased and inaccurate. Indeed, studies reporting long-term follow-up (longer than 5 years) after EEA are lacking but necessary to demonstrate the durability of the results of this technique.,,,
One of the first clinical series about EEA for pituitary adenomas was reported from our center by Dedashti et al., who evaluated the results of this technique in 200 patients treated from July 2004 to March 2007. In that study, excellent results were observed, including gross total removal rate of approximately 90% for tumors without cavernous sinus involvement. Disease control was achieved in 71% of growth hormone secreting, 81% for adrenocorticotropin hormone secreting, and 88% for prolactin-secreting adenomas, with no recurrence at the time of the last follow-up. However, as reported by the authors, those early results were based on a relatively short follow up period (median period 19 months).
The goal of the current study is the evaluation of long-term results of EEA for pituitary adenomas, based on treatment results of patients included in our initial series. Therefore, we aim to review the clinical, radiological and pathology data of patients who underwent EEA for resection of pituitary adenomas at our center, from 2004 to 2007 with follow up of at least 5 years and complete documentation. Additionally, we review the experience of other centers regarding long term management of pituitary adenomas and identification of recurrent adenomas.
We performed a retrospective, observational study, based on review of clinical and radiological data of all patients who underwent EEA for resection of pituitary adenomas at the Toronto Western Hospital, from July 2004 to March 2007, who had at least 5 years of follow-up data available for evaluation. Patients were excluded if: underwent pituitary adenoma resection via a different technique (transcranial microsurgery or transphenoidal microsurgery); other pituitary lesions not classified as pituitary adenomas; incomplete clinical or radiological data.
The database of the Department of Pathology of our institution was used for identification of patients. Clinical, endocrinological and radiological assessments were carried out before and after surgery. Patients were then followed on a yearly basis in a multidisciplinary pituitary clinic.
Recurrence was defined as a newly discovered mass that was diagnosed pathologically or identified radiologically (in the group without residual tumors after gross total removal) or a mass larger than the one on the first postoperative MRI. For those with functioning adenomas, biochemical evidence of disease recurrence was used a criterion for definition of recurrent disease.
Clinical and radiological evaluation
Preoperative radiological assessment consisted of an MRI of the sella using established pituitary protocols. Preoperative assessment consisted of contrast-enhanced MRI scan. A postoperative CT scan was done routinely the day following surgery to rule out any early post-operative complication and postoperative MRI scan was done between 2 and 3 months after surgery to assess the quality of resection. Follow-up MRI scans were then done on a yearly basis.
Tumor size was classified according to maximum tumor diameter in 2 categories: microadenoma (≤10 mm) and macroadenomas (>10 mm). The Knosp classification was applied to evaluate the relation of the tumor with cavernous sinus. Knosp grades 3 and 4, i.e., those with three-fourths or more encasement of the internal carotid artery or crossing of the lateral intercarotid line by the tumor, were considered to present significant invasion of the cavernous sinus. The extent of resection was classified as gross total resection (GTR) when no tumor was observed in the first postoperative scan, near-total resection (NTR) when >90% of the tumor was resected, and partial resection (PR) when <90% of the tumor was resected. For statistical analysis, NTR and PR were grouped into a single class, named subtotal resection (STR).
A full neurological and endocrinological assessment was routinely done before and after surgery. Formal visual assessment (visual field test and visual acuity) was done when the patient had visual symptoms and the tumor abutted the optic apparatus. A full hormonal evaluation (TSH, T3 T4, prolactin, testosterone level, serum cortisol, GH, FSH and LH) was performed before and after surgery and during follow up.
All patients in this study underwent an endonasal endoscopic transsphenoidal resection of the pituitary adenoma with intra-operative stereotactic image guidance and the use of microvascular Doppler probes. The objective of surgery was to achieve maximum safe resection while trying to preserve the normal pituitary gland and surrounding neurovascular structures. All procedures were carried out using a pure endoscopic approach primarily with the aid of a 0° 4-mm endoscope (Karl Storz GmbH & Co. KG). A right middle turbinectomy was done routinely to improve exposure followed by harvesting of a unilateral, vascularized naso-septal flap to be used for skull base reconstruction after tumor removal. A posterior septectomy allowed for a bi-nostril bimanual technique using microsurgical dissection. A wide sphenoidotomy was then performed to allow visualization of the sphenoid roof including the lateral optic carotid recesses bilaterally and antero-posterior visualization from the tuberculum to clivus. All intra sphenoidal septations were drilled flush with the sellar floor. The bony exposure of the sellar floor included the area between the medial walls of the cavernous sinuses bilaterally and from the anterior intercavernous sinus to the posterior intercavernous sinus. Neuronavigation and Doppler ultrasound were used routinely to identify the carotid arteries and guide the extent of the bony sellar and parasellar resection. The dura was opened in a multi-cruciate fashion extending laterally to the medial wall of the cavernous sinus and superiorly to the anterior intercavernous sinus. A combination of microsurgical piecemeal removal using varied angle curettes and suction was used to debulk and reduce the tumor mass, followed by identification and meticulous bimanual dissection to separate the tumor from the normal pituitary gland. On occasion, when feasible, an extracapsular removal was carried out. The bony defect was repaired in a multi-layered fashion with the harvested nasoseptal flap positioned to cover the entire exposed bone. The flap edges were covered with Surgicell (Ethicon) and Tissue glue (Tisseel, Baxter) was applied subsequently.
Statistical analysis was performed using SPSS (version 22.0, IBM Corp.). Categorical values are shown as percentages. Associations between the effect of tumor size (micro or macroadenoma), tumor subtype, age, and invasion of the cavernous sinus on extent of resection (EOR) were calculated using a chi-square test for categorical variables and the Pearson correlation coefficient for continuous variables. Continuous variables were compared using Independent Sample T test. Proportions were compared using the chi-square or Fisher exact test wherever appropriate. Two-sided significance tests were used throughout, and the significance level was kept at P < 0.05.
The progression-free survival was estimated using the Kaplan-Meier method. The progression-free survival curves obtained for partial or total tumor resection, age, Ki-67, occurrence of previous surgery, and cavernous sinus invasion were compared using the Log-rank test. Univariate and multivariate survival analysis were performed using a Cox regression model. Multivariate analysis included factors selected by a backward stepwise selection method with a 0.2 significance level for removal from the model. Statistical analyses were performed using SPSS (version 22.0, IBM Corp.). The significance level was kept at P < 0.05.
A total of 98 patients fulfilled the stringent inclusion criteria for this study. The number of female patients included was slightly higher than males (50 and 48 patients, respectively) and median age of patients included in the study was 50 ±13.1 years. Most patients presented with nonfunctioning pituitary adenomas (n = 66, 67.3%). and visual deficits were commonly reported (n = 45, 46%) prior to treatment. Previous surgery had been performed in 22 cases (22.4%). Regarding size and extension, most tumors were macroadenomas (n = 87, 88.7%) without signs of cavernous sinus invasion (n = 58, 59.2%). The median FU was 144 ± 37.1 months [Table 1].
Initial treatment results
Gross total resection was achieved in 55 patients (56.1%) [Table 2]. Main factors related to the extent of resection were cavernous sinus invasion and previous surgery (P = 0.001 and P = 0.012, respectively, [Table 3]). Tumor size, as a continuous variable, had a tendency to impact EOR but it was not statistically significant (GTR: 2.2 ± 1.01 versus STR: 2.68 ± 1.1, P = 0.055). Pathological evaluation demonstrated gonadotroph adenomas to be the most common tumor subtype in the study population (n = 31, 31.6%) followed by null cell adenomas (n = 23, 23.5%) [Table 4]. For those with secreting adenomas, GTR was achieved in 22 / 32 cases (66.7%) and disease control was initially obtained in 21 cases (n = 63.6%) after surgery.
Surgery led to visual improvement in 42/45 patients (93.2%) that presented with preoperative visual deficits and partial and complete improvement of extrinsic ocular movement function in 3 / 4 cases (75%) and 1/4 cases (25%), respectively. A new postoperative hormonal deficit was observed in 8 patients (8.1%) while diabetes insipidus, with need of DDAVP replacement, was diagnosed in 7 cases (7.1%). Postoperative CSF leaks were observed in 4 patients (4.1%) and were treated with lumbar drain (1 case), reoperation (2 cases), and conservative management (1 case) (bed rest). There was no carotid injury, meningitis or epistaxis observed. The median hospital stay was 4.1 ± 3.0 days [Table 2].
Long-term follow-up and recurrences
Radiological or endocrinological (in cases of secreting adenomas) recurrences were observed in 36 patients (36.7%) [Table 5]. Recurrences were rare in the first 3 years of follow up and became significantly more frequent 50 months after surgery [Figure 1]. Recurrences of non-functioning adenomas were detected in 25/66 cases (37.8%), while secreting adenomas recurred in 11/32 (31.2%) (P = 0.65). Recurrences were observed in 34% of patients who had had GTR while recurrences were observed in 39.5% of cases that underwent subtotal resection (P = 0.25). Additionally, patients who underwent STR had a tendency to present with recurrences sooner than those who underwent GTR, although no statistical significance was observed in Log rank analysis (P = 0.55) [Figure 2]. Specifically, recurrence after GTR of nonfunctioning adenomas, however, had a tendency to occur later than for those who underwent STR (mean: 103.4 ± 13.4 months vs. 62.5 ± 10.1 months, P = 0.15) [Figure 3] and [Figure 4]. Univariate analysis demonstrated previous surgery (P = 0.005), cavernous sinus invasion (P = 0.05) and Ki-67 (P = 0.01) to be factors associated with higher chance of recurrence [Table 5]. These results are better illustrated via Kaplan-Meier Survival curves and log rank analysis [Figure 5], [Figure 6], [Figure 7]. Multivariate Cox-regression analysis demonstrate that previous surgery and Ki-67 >5% are factors associated with recurrences [Table 6].
Management of recurrences
Conservative management, surgery and radiation therapy were utilized for management of recurrent nonfunctioning pituitary adenomas (25 / 66 cases, 37.8%) [Table 7]. The mean interval time between the initial surgery and recurrence was 80.5 ± 45.2 months. Surgery was performed in 19/25 cases (76%), followed by adjuvant radiation therapy in 6/25 cases (24%). Fractionated radiation therapy was used for management of a second time recurrence in 4 cases. Intervention for management of recurrence was not necessary, based on the clinical status of the patient, in 6 / 25 cases (24%) during follow-up and who were then followed with serial MRI scans.
Management of recurrent secreting adenomas varied according to the tumor subtype. Patients with acromegaly presented with recurrent disease in 6/19 cases (31.5%). Adjuvant medical therapy was used in 4/6 cases (66.6%), radiation therapy in 3/6 cases (50%) and reoperation in 4/6 cases (66.6%). At last FU, 2 of those 6 patients had achieved hormonal control. Disease recurrence was observed in 5/7 (71.4%) of patients with Cushing's Disease. Radiation therapy (4 / 5, 80%), medical therapy (3/5, 60%) and reoperation (2/5, 40%) were used in those cases, leading to disease control at last FU in 3 out of 5 cases (60%) that presented recurrences. Surgery led to initial disease control in 3/ 4 cases (75%) of prolactinomas; adjuvant medical therapy with dopamine agonists was used in one case that still had active disease after surgery. No recurrences were observed in prolactinomas or TSH-adenoma patients.
Surgical outcomes after first operation and EOR
The initial results of EEA for resection of pituitary adenomas observed in the study demonstrate the effectiveness of the technique. GTR (n = 55, 56.1%) or NTR (n = 34, 34.7%) were observed in majority of patients (89 patients, 90.8%). Visual improvement occurred in 93.2% and recovery of cranial nerve deficits (partial improvement in 75% and complete recovery in 25%) in most cases.
In a review performed by Esquenazi et al., in which the primary goal was to compare endoscopic endonasal versus microscopic transsphenoidal surgery for recurrent and/or residual pituitary adenomas, the GTR rates in the initial operation from 6 endoscopic series totaling 187 patients were reported.,,,,, The pooled proportion of GTR was 53.5% for the endoscopic studies and 46.6% for the microscopic studies. The respective percentage of visual improvement in the endoscopic cohort after the initial operation was 73.2%. This was superior to the microscopic group, which reached a rate of visual improvement in 49.6%.
The occurrence of previous surgery and cavernous sinus invasion were factors significantly associated with EOR in our study. Such variables have been associated with subtotal resection in other studies.,,,,
An overall recurrence rate of 36.7% (n = 36) was observed in our study. Recurrences were rarely observed in the first 3 years but were progressively more common after 50 months of FU. These results are comparable to those reported by Batista et al., which reported a 29.9% rate of recurrences in their study, and Dallapiazza et al. (26,2%). Gross total resection in the initial operation seems to affect the frequency of recurrences of pituitary adenomas in other series in the literature. Interestingly, Dallapiazza et al. report a recurrence rate of 12% (7/57) when GTR has been initially achieved whereas the radiological progression rate climbed up to 61% (14/23) when the adenoma was only partially resected. Radiological progression after partial tumor resection seems to be an anticipated event, since it has being reported between 39% and 75% of the cases throughout the literature. A meta-analysis published in 2012 totaling 971 non-functioning adenoma patients surgically treated supports these findings. Specifically, that study observed a recurrence rate of 12% in the GTR cohort versus 46% in the cases where a residue has been identified in the post-operative MRI.
In our series, GTR had tendency to present with lower recurrence rate when compared to STR (34% versus 39.5%, P = 0.25). Interestingly, statistical significance was not observed when that factor was assessed through univariate or multivariate analysis. This may be secondary to different reasons, including limited number of patients included in the study, the inclusion of a mixed population of cases (non-functioning and functional adenomas) and the impact of tumor biology/different subtypes in long follow-up. A similar result is observed even when non-functioning tumor (NFPT) are analyzed as a separate group [Figure 3]. Our group is currently carrying out a further investigation of this finding through the analysis of all cases of NFPT treated in our center (manuscript in preparation).
Timing of recurrence and importance for long-term FU
As demonstrated in [Table 6], the mean timing of recurrence varied in the different endoscopic studies, from 49.9 to 86 months.,,,,,,, More importantly, the timing of recurrence seems to differ between the totally removed and partially removed tumors. Specifically, for the partially removed tumors the mean timing to progression is reported in 2 studies as 6,5 and 36 months. However, the mean timing of recurrence for the totally removed adenomas, in the same studies, is significantly longer; 36 and 53 months, respectively., In the same vein, Chen et al. in their meta-analysis report that in the cohort of totally resected adenomas (N = 300), recurrences occurred even after 10 years of follow-up. In our study, subgroup analysis of NFPT who presented with recurrences suggested a similar difference between those who underwent GTR or STR, although not statistically significant (mean timing of recurrence GTR group: 103.4 ± 13.4 months versus STR group: 62.5 ± 10.5 months, P = 0.15).
Predictive factors of recurrences
Factors such as tumor subtype, cavernous sinus invasion, size and previous treatment have been associated with recurrences.,,
Likewise different subtypes of pituitary adenomas have been reported to demonstrate a higher propensity to recur.,, Langlois et al. studied a large cohort of 814 surgically treated non-functioning pituitary tumors and reported on higher recurrence rate among corticotroph and somatotroph tumors when compared to gonadotroph or null cell tumors. Interestingly, 5.1% of silent corticotroph adenomas transformed to secretory adenomas with associated symptomatic hypercortisolism., Moreover, Batista et al. report a higher recurrence rate of silent corticotroph adenomas (57.8%), especially type II, when compared to the recurrence rate of the whole cohort of 137 operated non-functioning pituitary adenomas (29.9%). Null cell adenomas have been associated with a more invasive presentation and a more aggressive clinical course and higher tendency to recur. Corticotroph and null cell adenomas represented 35.7% of cases included in the current study and recurrences were observed in 34.8% and 50% of those tumors – a factor that contributed to the overall recurrence rate observed in this study.
Knosp grade and preoperative visual deficits were predictive factors for recurrence in univariate analysis according to Dallapiazza et al., but in their multivariate analysis only Knosp grade/cavernous sinus invasion was an independent predictive factor. In our study, Knosp grade was associated with recurrence rate in univariate Log Rank analysis (P = 0.03), but the only factors associated with recurrence in Multivariate analysis were previous surgery and Ki-67 >5%.
Secreting pituitary adenomas
In our series, 19.4% of cases were attributable to GH secreting adenomas. Successful short-term remission rates are reported as high as 73%, 1 year after surgery. However, there is evidence supporting the importance of long term follow up in these lesions due to documented delayed recurrence. Biermasz et al. demonstrated an early remission rate of 60%, which dropped to 39%, at 10-year follow-up. Similarly, our results illustrated a recurrence rate of 31.5% (6/19 cases) at 5-year follow-up. Interestingly, both studies demonstrated that approximately a third of patients recurred, suggesting a possible peak recurrence time of 5 years. Further work is required to define the optimum follow up period. Extent of resection, cavernous sinus invasion and preoperative and early postoperative GH and IGF-1 levels (within 24-48 hours after surgery) as well as by clinical grade are all important prognostic factors for tumor recurrence., Management of recurrent secreting adenomas varies according to the tumor subtype, including repeat surgery, radiotherapy, somatostatin analogue and dopamine agonists.,,
The first-line therapy for Cushing's disease is transsphenoidal surgery with selective adenomectomy and preservation of normal pituitary gland function. Recurrent Cushing's disease can be often diagnosed biochemically before clinical symptoms re-develop. Initial surgical remission can be achieved in 76% to 78% of cases., Although recurrence rates have been reported to range from 8% to 66%, the highest risk for recurrent disease is observed in the first five years after surgery, but it can occur as late as several decades after surgery. Disease recurrence was observed in 5/7 (71.4%) in our series with a 5-year remission rate of 60%.
Remission rates after radiation may be delayed in functional adenomas. Thus medical treatment should be continued, and close monitoring of levels are necessary. Radiation therapy was the most common therapy (80%) in recurrent Cushing's disease patients in our series. Moore et al., described SRS in patients with cavernous sinus invasion with persistent or recurrent Cushing's disease and reported a remission rate of 57% and no recurrence after follow-up of 55 months. Medical therapy can be used as a long-term treatment after failure of surgery or radiation therapy or in patients who are not candidates for surgery. It is useful preoperatively to reduce morbidity and as emergency treatment in severe and life threatening hypercortisolism.
There are a number of limitations to our study. Results from our series consist of data acquired from retrospective analysis and have its inherent biases. Although we carefully analyzed the data of all patients included in the study, the relatively small number of patients may limit the generalization of the results reported in our study. The original goal of the study, to analyze all the 200 patients included in the original paper published by the senior author in 2008, could not be fulfilled due to lack of availability of complete data documentation for half of those patients. Indeed, this demonstrates the challenge of obtaining long-term follow-up in patients even in a socialized universal health care system. Finally, our current study included a mix population of nonfunctioning and secreting pituitary adenomas, which may impact the result due to heterogeneity of the population. Our group is performing additional studies to assess long term results of specific pituitary adenomas subtypes, in order to overcome potential limitations of the current study.
Our results demonstrated the importance and need of long-term follow-up after endoscopic endonasal surgery for management of pituitary adenomas. Short-term results (<2 years), as reported by the current study and by our previous publication and others, are important to assess the efficacy and safety of surgery to achieve disease control. However, such short-term analysis fails to provide accurate data of the durability of results of endoscopic pituitary surgery.
Our current results, based on analysis of patients with a mean follow-up of 144 months, suggest an overall recurrence rate of 36%. Such recurrences were more commonly observed after 5 years of follow up and more commonly noted in patient who had previous surgery and in those with tumors with cavernous sinus invasion. Therefore, our data supports the need for long term follow up for all patients who undergo surgical resection of pituitary adenomas regardless of the technique used, especially for patients with more invasive tumors.,,,,,
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6], [Table 7]