In the present study, we reported outcomes for 42 patients undergoing repeat TSS for recurrent and persistent disease, achieving an overall remission rate of 69.0%. According to literatures, the immediate remission rate of the reoperation after recurrence is reported up to 87%[3, 9, 21], which is not lower than other second-line therapy such as radiation therapy and medical treatment. The recurrence rate of CD after the initial TSS is 10–25% with a follow-up time of 10 years[22–24]. Ram et al reported that the surgeon applied a second TSS right after the first once the postoperative serum cortisol level didn’t meet the standard of remission. With an interval time of 1 to 6 weeks, 71% of those patients with persistent disease achieved immediate remission and 53% (9/17) are in a long-term remission[19]. The other authors showed a remission rate of 70% with the reoperation within 10 days[25]. Second TSS leads an additional 8% of the patients to long-term CD remission[6]. The recurrence groups have a slightly higher remission rate without significance comparing to the persistence groups in the present study. Similar finding is shown in a study of Ram implicating that the failure of the initial surgery suggested being more difficult to treat successfully with surgery than most patients with recurrence[19]. Therefore, the criteria of selecting the potential patients and the strategy of the reoperation are still worth discussing.
Surgical strategy
The surgical strategy for the first-time operation of CD varies depending on the major concern of different pituitary surgeons. Some surgeons intend to preserve more normal gland in the surgery while others chasing higher remission rate. Selective adenectomy is a reasonable choice for the visible tumor. Several authors adopt a slight extended resection with a rim or sometimes 2-3mm like-normal tissue around the tumor, which could also be called as a partial hypohypsectomy[26, 27]. Hemi-hypohypsectomy is more common in the case that no tumor was identified in the operation, while the MRI or BIPSS indicating remarkable lateralization of the tumor origin[28]. The widely exploration on contralateral side should also be conducted in case of whose BIPSS results is inconsistent with those of MRI, as it might help to find the very tiny tumor. Regarding some author, more extend procedures including subtotal or sometimes total pituitary gland resection will be done to maximize the remission rate, which could be up to 75.9–81.8%[27, 29]. It might be a reasonable recommendation where imaging/intraoperative findings are not definitive, considering that the negative impact on these reoperated patients with persistent hypercortisolism rather than hypopituitarism. It is interesting that the pathology confirm is pretty low in these cases with extended resection even though it showed a high remission rate. Now there seems to be a trend that surgeons performing less total hypohypsectomy as this can lead to hypopituitarism[10, 29, 30], given that it might not apparently increase remission rate but decrease the quality of life[31].
MRI finding
When it comes to radiology, we need to emphasize that negative MRI doesn’t necessarily mean the inexistence of pituitary adenoma or a negative pathological result. A number of cases from Wagenmakers et al showed that the remission after the repeated transsphenoidal surgery was not predictable by having a positive finding on MRI before either the first or second operation[14]. Preoperative MRI provides reference for diagnosis of pituitary adenoma but has limited predictive function of prognosis of the patients[8], especially for the repeat operation in which the original anatomy structure is more or less destroyed in the initial surgery. A positive result of MR imaging before the second operation is supposed to provide confidence for the surgeons. The reported remission rate of the redone operation with a positive MRI respectively was up to 72.7%[14]. According to our study, the two positive-MRI groups with different initial surgical outcome show higher remission rate albeit non-significantly. Positive MRI finding suggests that better endocrinological outcomes are achieved by second operation in both recurrent and sustained group compared to those with negative imaging. Excellent remission rate (more than 80%) was achieved in the recurrent group with positive-MRI, and thus encourages a repeat TSS. Acceptable remission rate (over 60%) which is close to the alternative treatment options in the recurrence group with negative-MRI as well as persistence group with positive-MRI. It is noted that one patient with persisting CD and negative-MRI achieved remission after reoperation. Therefore, a second surgical treatment for these patients needs to be carefully considered whether it will be beneficial.
Generally, as for the recurrent or persistent cases of CD patients underwent an initial surgery, we regard MRI as a possible assistant method for decision making. A second operation is considered with visible lesion remain in the MRI under assumption that removal of residual tumor leads to remission of the disease. Meanwhile, some recurrent and persistent patients with negative MRI findings also benefit from a reoperation. Furthermore, the MRI has its limitations on revealing accurate structures of the original operated area. The distortion and cicatrization of the previous operation and material packing in sellar region lead to confusion[16, 17]. Unlike the considerable remission rate achieved after the initial operation despite a negative MRI, a reoperation without certain lesion detected on MRI has a dissatisfactory remission rate[2], parallel to the result in our study. Nevertheless, U.J. Knappe and D.K. Lüdecke[8] put forward different opinion about the significance of MRI that it was not usually helpful in determine the therapeutic strategies due to its low incidence of detecting of the existence of microadenomas (missed diagnosis in 38%-70% cases). However, the BIPSS in those cases where MRI provided no certain information of tumor is therefore critical for the surgeons to ascertain the pituitary origin of the disease[15, 32], although there’s also a voice suggesting that MRI and BIPSS are not helpful in locating the recurrent tumor[14]. It might not help to identify the tumor in cavernous sinus or other parasella regions.
Pathology
By comparing the pathology results and remission situations of those recurrent patients as well as persisting patients, we failed to find any relation between pathology results and remission expectations. These findings are supported by series of Ram[19], with 11 of the 17 patients had no tumor found at the second procedure, and 6 of the 11 achieved remission. In the series of Marco Locatelli[15], 8 out of 12 patients had no tumor found at the second operation, and 5 had surgical remissions. Even in those remission cases, the positive rate of pathological exams is not as high as expected. There’s no significant difference of remission rate between patients grouped by pathological results or one-to-one correspondence between histopathological confirmation and the surgical outcome[15]. So far, less evidence supports the prediction of the repeat operation outcome by either of two pathology results.
Other consideration/factors
For the initial preoperative assessment with the BIPSS in those MRI negative cases is critical in providing confidence of increasing the remission rate as reconfirming the pituitary origin[15, 32], since prediction accuracy was high and remission rates were acceptable[33]. In the patients with recurrence and persistent hypercortisolism after their first operation, it is not easy to identify a solid lesion on MRI compared with the initial preoperative scanning. It is worth mentioning that BIPSS might provide more information especially for those who did not underwent this test before the first operation. Also, it might avoid unnecessary transsphenoidal surgery twice in patients with persistent hypercortisolism, by revealing a false positive for pituitary ACTH overproduction. The results of BIPSS have the potential value not only to confirm the pituitary origin in the condition that the first histology did not show an ACTH-positive staining, but also to guide the exploration and decision making for hemi-hypohypsectomy or accessing into the cavernous sinus, especially for those without obvious tumor identified intraoperatively. A careful dissection is highly recommended on the side of obviously lateralized BIPSS result, which sometime also indicating the cavernous sinus invasion not shown on the MRI and the necessity of opening the medical wall to achieve an extended exploration. The predictive value of BIPSS lateralization in repeated surgery need further investigated although it is not optimal in native patients with Cushing’s disease[34, 35].
In Burke’s study, there was a comparation between the historic cohorts with microscope and current cohorts with endoscope[10]. Although the result did not show statistically significant difference in either remission or recurrence, a trend for better outcome was observed more common after the operation with endoscopy. On the basis of a study of Lonser et al[36], over 20% of CD patients have cavernous sinus invasion confirmed histologically. The authors advocated a complete resection including invaded sella dura and the medial cavernous sinus wall in experienced surgeons’ hand. In fact, it is worth noting that endoscopy with magnification and lighting provide a panoramic view to perform extended exploration of the sella including the cavernous sinus, compared to the microscope-based approach. Micko et al demonstrated that endoscope allows radical inspection of the entire medial wall of the cavernous sinus[37]. It increased the lateral angle of visualizations and facilitate the differentiation between tumor and other tissues. These advantages over microscopic transsphenoidal approach are critical for recurrent and unremitted cases, which still require larger size to verify the conclusion.
Other treatments adjunctive to the repeat surgery
Medications being used for CD include centrally acting inhibitors of ACTH secretion including cabergoline and pasireotide, adrenal steroidogenesis inhibitors to block several steps in cortisol synthesis, as well as glucocorticoid-receptor antagonists[38]. They are typically only considered as methods of disease control prior to surgical resection, as well as adjunctive methods for the radiation therapy of persistent and recurrent disease. On the other hand, the literature has noted that Ketoconazole may also contribute to enhance tumor appearance on MRI to facilitate pituitary resection in some circumstance[39]. Castinetti et al. described that visible lesions may be identified on the MRI in one third of patients who undertook a Ketoconazole administration[40].
In literature, the reoperation for persistence cases without visible lesion on MRI is rarely satisfactory[41], whereas these patients may benefit from the radiosurgery by using the entire sellar region as a therapeutic target[5]. The hormonal normalization was achieved after radiosurgery in half of the cases, including those with negative MR finding[42]. In general, the outcome of radiosurgery and less commonly radiotherapy is more favorable particularly in MRI-negative cases with persistent hypercortisolism compared to the repeat surgery. It is plausible that complications may be less and length of stay in hospital may be shorter[43, 44]. Salvage TSS for refractory CD after radiation therapy has been reported rarely[45], owing to the difficulty of disruption of surgical landmarks, the formation of scar tissue, and the effects of preoperative radiotherapy[43].
Bilateral adrenalectomy is generally considered the ultima ratio in patients who fail to respond to all other treatment options. However, patients undergoing bilateral adrenalectomy will require lifelong surveillance of the corticotroph tumor progression leading to Nelson’s syndrome via MRI and ACTH measurements. Most experts agree that selective transsphenoidal adenomectomy should be recommended as first-line therapy in patients with Nelson’s syndrome before extrasellar expansion of the tumor occurs[46].
Limitations
Similar to previous studies, our sample size was not large enough to conduct powerful statistical analyses. Some patient lost in follow-up very soon made the current study has limited long-term outcome. There is a trend of predictable value of positive preoperative MRI findings, which is not solid to support an apparent relationship. Some potential weakness of the present study also includes that the outcome is only focused on the biochemical benefits of the remission with the surgical intervention. And also, it might underestimate the risks of hypopituitarism and decreased quality of life. We admit that larger series are needed to further investigate the potential predict factors and the best surgical strategy.