Association of different pathologic subtypes of growth hormone producing pituitary adenoma and remission in acromegaly patients: a retrospective cohort study

Acromegaly is a clinical syndrome caused by excess growth hormone (GH) production almost always from a GH-secreting pituitary adenoma. Despite low prevalence, the disorder has received more attention in the clinical setting because of its high morbidity and mortality [1‐5]. Therefore, early diagnosis and appropriate treatment approaches have an especial role to reduce complications, and mortality can be decreased to the level of normal population [6]. The predictive role of age, tumor size, as well as pre-operative GH and IGF-1 levels on remission outcomes are reported in previous studies, and some studies suggested that the clinicopathological information of acromegaly patients may be able to predict the success rate of treatment which provide individualized modalities in therapeutic process [7]. Hence, one of the possible factors which are sometimes considered as remission predictor is pathologic subtypes. Based on the GH granulation pattern in immunohistochemistry (IHC), two pathologic subtypes of GH-producing adenomas are densely granulated (DG) and sparsely granulated (SG) adenomas. About 30–50% of acromegaly patients have DG-GH adenomas with a diffuse positivity for GH [8‐10]. They are diagnosed at older ages (> 50 years) which slowly grow [11‐15]. It was reported that DG-GH adenomas have an indolent course and are more responsive to surgery and somatostatin analogs than SG-one [12, 16‐18]. It is reported that 15–35% of acromegaly patients have SG-GH adenomas with a focal and weak positivity for GH, and they are more common in younger ages (< 50 years) [7]. Some studies indicated that these tumors are more invasive which have a higher ki-67 level [19], larger size, lower GH and IGF-1 levels which are more frequently clinically subtle [12, 13, 15, 20‐23].

These two pathological subtypes considered single staining adenomas, approximately a quarter of somatotropic adenomas also prolactin (PRL)-positive, so-named dual staining (or mixed GH-PRL) adenomas [24]. These adenomas are composed of mixed somatotroph-lactotroph tumors, mammosomatotroph tumors, and more primitive and aggressive tumors such as acidophil stem cells [25, 26]. The new important topic at recent clinicopathological discussion is that dual staining tumors compared with pure somatotropic-ones may have a different treatment and risk of invasion to surrounding tissues [7, 10].

As far as we know, limited previous studies assessed the pathological diversities among different pituitary somatotroph adenomas, and dual staining tumor etiologies are reported to some extent, but their clinical impact is not clear [27]. Thus, the role of pathology, as a predictor factor, is inconclusive and needs more investigation. We investigated this knowledge gap by comprehensive comparison among different pathologic subtypes of pituitary adenomas in terms of clinical manifestations, laboratory, imaging findings, and especially treatment outcomes.

Acromegaly is a slowly progressive disease, and identifying the predictors in the clinical course of disease determines the next step in decision making. We investigated the correlation of histological parameters as a possible predictor, on biochemical remission, for knowing the role of pathology in personalized treatment plan in acromegaly. In our series of acromegaly patients, different pathological subtypes of pituitary adenomas appear do not have a predictive role to estimate biochemical remission, both in binary and triple comparisons. After comparing the various predictors, our results showed that the cavernous sinus invasion and preoperative and one-day postoperative GH have a significant association with remission in acromegaly patients. After adjusting all variables, the cavernous sinus invasion as key factor decreased the remission rate by 91%, followed by one-day postoperative GH, about 64%. After adjusting all variables, the cavernous sinus invasion and one-day postoperative GH levels showed a significant predictive role in the patient’s remission.

In evaluating various predictive factors on remission, most studies emphasize the importance of tumor invasion and GH level, our study is in line with them [34, 35]. Moreover, a systematic review by Agrawal et al. [32] showed that among predictors of surgical outcomes, some of them consistently reduce remission rate in almost all studies such as cavernous sinus invasion, tumor size, and higher GH levels, some of them occasionally reduce remission rates. In our study, after adjusting for covariates, the best single preoperative predictor was cavernous sinus invasion, followed by GH level in the first few days postoperatively. On the other hand, although tumor sizes were higher in SG group, remission rate was not statistically significant among three groups.

Similar to the findings of Rick et al. [27] study, in our results, tumor size showed no significant association with disease remission, also in Kreutzer et al. [36] study, the significance of tumor size on disease remission was borderline.

Studies about prolactin co-expression of GH-producing pituitary adenomas as a remission predictor have had heterogeneous results; some studies showed prolactin co-expression is a negative remission predictor [27, 36, 37], some showed it has not predictive role [38‐40], and some mentioned prolactin co-expression is positive remission predictor [41]. In this regard, our findings are similar to Sun et al. [38], their study was done on 59 acromegaly patients that pathologically classified into three groups SG, DG, and dual staining pituitary adenomas, and they found that surgical remission was similarly about 50% in each group. Minniti et al. [40] also found similar results indicating no prognostic value for dual staining adenoma. However, the results of Rick et al. study were different [27], they reported that single-staining tumors have8.6 times more remission rate than dual-staining tumors, they emphasized that prolactin staining is a major predictor of surgical remission in acromegaly. We think their results need to interpret cautiously because the odds ratio has a wide confidence interval (1.09–109.7). On the other hand, they did not evaluate the different subtypes of single staining group, including the percentage of SG adenoma, also their single staining group had fewer acromegalic features, lower IGF-1, prolactin levels, and smaller tumor size compared with dual staining one. These findings were different from our study which indicated an earlier diagnosis of acromegaly in a single staining group may influence their conclusion. Marinis et al. [37] found that high serum prolactin levels and dual staining pathology negatively affect remission outcomes. However, half of the patients in the remission group (controlled patients) had microadenomas, and on the opposite side, sellar floor perforation and destruction were more frequent (16/25) among non-remission (poorly controlled) patients, which may influence their conclusion.

As we mentioned, the results of different studies contradict the remission rate in dual staining compared with single staining adenomas. There are several hypotheses to justify this result heterogeneity; first, it is possible that single staining identified falsely as dual staining adenomas, and vice versa. Dual staining adenomas are identified when prolactin expression is present in a higher percentage of cells (more than 5%) than pure single staining adenomas [42], but there is uncertainty in practice due to the semi-quantitative nature of IHC method. On the other hand, like any other neuroendocrine tumor, somatotroph adenomas, may intermittently produce prolactin, and they may indicate a wide spectrum of differentiation and variably express prolactin that causes pathological subtypes misclassification, too [43]. Second, studies which used electron microscopy and molecular analysis showed that dual staining adenomas compose of mammosomatotroph tumors and Mixed Somatotroph-Lactotroph, which may have different behavior. Mammosomatotroph tumors look like DG adenomas in IHC method, but they also express estrogen receptor (ER) and prolactin [44]. They are very similar to DG adenomas clinically but they may have more significant hyperprolactinemia [10]. A mixed Somatotroph-Lactotroph tumor consists of two separate cell populations, somatotrophs, and lactotrophs. Each cell population can be DG or SG, so various combinations may occur. The characteristics of these tumors are dependent on how to combine the relative proportions of tumor cells [44]. Hence, depending on which tumors and cell populations cause prolactin immune-reactivity in dual staining adenomas, clinical features, and remission response may be different in various studies. Third, remission criteria in acromegaly didn’t have the same references in various studies, and this caused a patient in different studies to be placed in different remission categories.

Our study indicated no statistically significant difference in remission rate among DG, SG, and dual staining groups.

Medical responsiveness results are less heterogeneous than surgical remission among various pathological subtypes, considering a biological rationale about different SSTR2 expressions and medical responsiveness in acromegaly. SRL responsiveness was assessed among various pathological subtypes, the response rate in SG was lower than in two other groups; however, a non-significant association may be due to the small sample size. In Brzana et al. [13] study, like ours, SG group was less responsive to the medical treatment. Therefore, the tumor behavior and clinical characteristics of dual-stained adenomas are more similar to sparsely granulated adenomas in contrast with our study, this can be due to the different composition of cell composition in dual staining groups. Recurrence, re-surgery, and radiation were higher in SG compared to two other groups; however, only recurrence rate was statistically significant. This finding is also in contrast with Brzana study.

Regarding the strengths of present study, it was a single-center study which was the referral hospital of pituitary diseases. Two well-experienced pathologists reported the results of pathology specimens, separately with an acceptable agreement. Besides, patient’s responses to surgery and medical therapy were assessed according to the newest criterion [31]. Ultimately, this is the first time that in a single study, the pathology of pituitary adenoma was analyzed in two separate strata, single and dual staining in binary comparison, on the one hand, and on the other, triple comparison among DG, SG, and dual staining. However, the study has some limitations. The retrospective nature of study which is often of limited value and patients, did not follow up actively, so some data were missed due to un-accessibility. Second, the sample size was small which can affect our result in various pathological subtypes. Third, the type of cell population, the mammosomatotroph or mixed-somatotroph-lactotroph, in dual staining adenomas had not been determined separately by the electronic microscope because these two subtypes may have different behavior and response to treatment.

In conclusion, our findings do not support a major predictive role of different pathological subtypes of pituitary adenomas to estimate treatment response. However, among all predictors we analyzed, the cavernous sinus invasion and one-day GH level post-operative seem strongly associated with treatment outcomes. Inconclusive results of previous studies about pathologic subtypes, maybe originate from cell population misclassification, different remission criteria. Further studies with more accurate tools such as electron microscopes as well as newer molecular markers are needed to certain the prognostic role of the pathology on the treatment response in acromegaly patients.