Predictors of Remission of Acromegaly following Surgical Treatment in Growth Hormone-Secreting Pituitary Adenoma

Abstract

Background and Objective  Surgery is the treatment of choice for growth hormone (GH)-secreting pituitary adenoma. The remission of random GH depends on various factors. We aimed to evaluate the predictors related to remission of random GH following surgical treatment.

Methods  We collected the data retrospectively from the chart review from a single unit of neurosurgery. The diagnostic criteria for remission were a random GH < 1 ng/mL or nadir GH < 0.4 ng/mL after an oral glucose tolerance test.

Results  Data from a total of 110 (females 62 [56.4%]) patients were available for follow-up and were analyzed. The mean age was 36.5 years (14–69 years). Vision impairments were seen in 39 (35.5%) patients. The mean duration of symptoms before surgery was 34 months. The mean volume of the tumor was 7.2 mL (0.44–109.8 mL). Knosp grade 3 and 4 tumors were seen in 41.5% of cases. The mean preoperative random GH level was 68.9 ng/mL. Transsphenoidal surgery was done in 107 (97.3%) cases. The gross total resection could be done in 36 (32.7%) cases. At 3 months, 25 (26%) patients had a biochemical remission. In univariable analysis, lower Knosp grade, preoperative GH level < 40 ng/mL, gross total resection, and male gender were associated with remission at 3 months. In regression analysis, preoperative GH and male gender were related to remission at 3 months.

Conclusion  The preoperative GH level < 40 ng/mL is associated with higher chances of remission after surgery for GH-secreting pituitary adenoma.

Introduction

Acromegaly is endocrinopathy with excess growth hormone (GH) in the circulation. It affects mainly the central nervous system, cardiovascular system, and skeletal system. ¹ The severity of the disease is based on the serum level of GH and the duration of the disease. Surgery is the standard of care followed by adjuvant radiotherapy, either gamma knife radiosurgery (GKRS) or another modality of radiotherapy. Medical therapy plays an important role in residual disease or primary cases with a high risk for surgery. ² The goal of the surgery is to remove the tumor as much as possible, preserve normal pituitary function, and restore vision. The normalization of the GH level varies depending on the type of surgery, tumor character, and extent of resection (EOR). The ideal time to assess the remission is at 12 weeks after surgery. ³ There are various factors like tumor size, surgical approach, and cavernous sinus extension related to the remission of GH after intervention. ^{4 5 6} We aimed to evaluate the predictors related to the remission of random GH levels following treatment.

Materials and Methods

Data

We retrospectively reviewed the medical records, neuroimaging studies, and pathology reports of all patients who underwent surgery for GH pituitary adenoma from 2010 to 2020 in a single unit of neurosurgery in our hospital. As the study involved a review of records retrospectively and no patients were contacted for the study, ethics committee approval was not sought. All patients were evaluated with neuro-ophthalmological evaluation, hormonal assessment, and gadolinium-enhanced magnetic resonance imaging (MRI) before surgery. The diagnosis of GH-secreting pituitary adenoma was made based on serum GH levels and serum insulin-like growth factor-1 (IGF-1) levels and confirmed with the sellar and suprasellar mass lesion. Later, all the cases underwent surgery either endoscopic/microscopic transsphenoidal surgery (TSS) or transcranial approach depending on the location, extent of the tumor, and surgeon's preference. Histopathology data and immunohistochemistry (IHC) data were collected from the pathology report.

Analysis

The remission rate was assessed 3 months after surgery before the initiation of adjuvant therapy. The criteria for remission were random GH < 1 ng/mL or nadir GH < 0.4 ng/mL after an oral glucose tolerance test. The IGF-1 levels were not considered for defining remission. The EOR was determined by reading MRI studies performed at 3 months of follow-up after surgery. The gross total resection (GTR) was defined as no residual tumor reported independently by a neuroradiologist. Predictive factors were selected based on clinical, radiological, and intraoperative findings. For the statistical analysis purpose, we categorized age (< 40 vs. > 40 years), gender (male vs. female), vision involvement (yes vs. no), the volume of the tumor (< 7 vs. > 7 mL, based on the mean volume of the cohort), T2-weighted (T2W) character in MRI of the tumor (hyper vs. hypo or iso), Knosp grade (grade 1, 2 vs. grade 3, 4), preoperative GH level (< 40 vs. > 40 ng/mL), the surgical approach (endoscopic TSS vs. microscopic TSS), and the EOR (GTR vs. subtotal resection). The EOR was confirmed with MRI done at follow-up after 3 months of surgery. The following histopathology variables were analyzed: granularity (dense or sparse), immune staining (GH or both GH and prolactin [PRL]), and Mindbomb Homolog-1 (MIB) index (≤ 3 vs. > 3). MCh the criteria for densely/sparsely were subjective as reported by the pathologist. The statistical analysis was done utilizing SPSS software (version 20, SPSS, Inc.). The statistical analyses of categorical variables were conducted utilizing the chi-square and Fisher's exact test, as appropriate. Univariable and regression analyses were performed to search for any relationship between individual predictive factors and remission rate at 3 months. A p -value of < 0.1 in the univariable analysis was considered for regression analysis. The variables with missing data were not included in the regression model. For regression analysis, the outcome was surgical remission. A p -value of < 0.05 was considered significant.

Results

Clinical, Imaging, and Biochemical Findings

A total of 110 patients underwent surgery for GH-secreting pituitary adenoma. All patients were symptomatic; none of the patients was diagnosed incidentally. Acromegaly features were present in 88.2% of cases. Vision impairment either in the form of reduction in visual acuity, visual field, or both were seen in 39 (35.5%) patients. The mean duration of symptoms before surgery was 34 months (range: 1–144 months). The mean maximum diameter of tumor was 2.28 cm (range: 0.8–6.1 cm). Only three cases were microadenomas. The mean volume of the tumor was 7.37 mL (0.44—109.8 mL). The mean tumor volume in males was 7.20 mL and in females was 7.58 mL, which was significantly different ( p  = 0.007). The most common Knosp grade was grade 2 (49.1%). The T2W character of the tumors was analyzed, and 76 (69.1%) of the tumors had a hyperintense character ( Table 1 ). The mean preoperative random GH level was 68.9 ng/mL and the mean preoperative IGF-1 level was 747.4 ng/mL ( Table 2 ). The mean GH level in males was 66.77 ng/mL and in females was 70.47 ng/mL, which was significantly different, p  = 0.002.

Table 1. Clinical and imaging features ( n  = 110) .

Clinical presentation
Age in years, mean (range)	36.5 (14–69)
Gender	Male 48 (43.6%), Female 62 (56.4%)
Duration of symptoms in months prior to surgery, mean (range)	34 (1–144)
Vision impairment	39 (35.5%)
Acromegaly features	97 (88.2%)
• Skin and soft tissue changes	• 89 (81%)
• Acral changes	• 69 (62.7%)
• Menstrual irregularity a	• 18 (29.0%)
• Amenorrhea a	• 23 (37.1%)
• Galactorrhea a	• 5 (8.1%)
• Nerve entrapment	• 6 (5.5%)
Headache	16 (14.5%)
Magnetic resonance imaging findings
Tumour volume in mL, mean (range)	7.37 (0.44–109.8)
Knosp grade
• Grade: 1	10 (9.1%)
• Grade: 2	54 (49.1%)
• Grade: 3	29 (26.4%)
• Grade: 4	17 (15.5%)
T2W signal
• Hyper	76 (69.1%)
• Hypo	7 (6.4%)
• Iso	27 (24.5%)
Operative details
Surgical approach
• Endoscopic TSS	90 (81.8%)
• Microscopic TSS	17 (15.5%)
• Transcranial	3 (2.7%)
Extent of resection
• Gross total	36 (32.7%)
• Subtotal	74 (67.3%)

Abbreviations: TSS, transsphenoidal surgery; T2W, T2-weighted.

^aAmong 62 females.

Table 2. Biochemistry and pathology.

Serum hormones ( n  = 110)
Preoperative serum GH level, ng/mL, mean (range) a	68.9 (4–558)
Preoperative IGF-1 level, ng/mL, mean (range) a b	747.5 (101–1431)
Pathology
Granularity ( n  = 36)
• Sparse	32 (88.9%)
• Dense	4 (11.1%)
Immunohistochemistry ( n  = 51)
• GH positive	33 (64.7%)
• GH and PRL positive	18 (35.3%)
MIB index ( n  = 55)
• ≤ 3	39 (35.5%)
• > 3	16 (14.5%)

Abbreviations: GH, growth hormone; IGF, insulin-like growth factor; PRL, prolactin.

^aChemiluminescent immunoassay.

^bZ scores not available, however, the values were higher than the range provided by the laboratory for the given age and gender.

Surgery

The most common surgical approach was the endoscopic TSS in 107 (81.8%). The GTR rate was 36 (32.7%). The granules status in the pathology report was available in 36 patients, and predominantly were sparse 32 (88.9%) type. IHC was performed in 51 cases, and 33 (64.7%) of cases showed only GH immune stain positivity whereas 18 (35.3%) had both GH and PRL staining. The MIB index was more than 3 in 16 (14.5%) cases. Though 7% of patients developed diabetes insipidus (DI) in the postoperative period, only one patient had permanent DI. One patient underwent reexploration due to a hematoma in the operative site. Four cases succumbed during the perioperative period: three due to meningitis and one due to cardiomyopathy ( Table 3 ).

Table 3. Complication ( n  = 110) .

Diabetes insipidus	9 (7%)
Cerebrospinal fluid leak	6 (5%)
Syndrome of inappropriate antidiuretic hormone	2 (2%)
Meningitis	1 (0.8%)
Vision deterioration	2 (2%)
Reexploration	1 (0.8%)
Seizures	1 (0.8%)
Mortality	4 (4%)

Follow-Up

Follow-up was available for 102 patients and the median duration of follow-up was 6 months (3–84 months). Twenty-five (26%) patients had remission at 3 months ( Table 4 ). Radiotherapy was given to 46 (45%) cases. Five patients with mammosomatotroph adenoma received cabergoline. Pasireotide was given to eight patients and long-acting octreotide was given to one patient. The remission rate at the last available follow-up was 53%.

Table 4. Follow-up evaluation and treatment ( n  = 102) .

Three-month remission rate	25 (26%)
Radiotherapy
• Gamma knife radiosurgery	20 (20%)
• Intensity-modulated radiotherapy	26 (25%)
Medical therapy
• Cabergoline	5 (5%)
• Octreotide	1 (0.9%)
• Pasireotide	8 (7.8%)

Outcome Analysis

We performed statistical analysis for remission at 3 months. We considered predictive factors from the demography, radiology, treatment, and histopathology. There were only three cases of transcranial surgery hence not included in the analysis. In univariable analysis, low Knosp grade (odds ratio [OR]: 6, confidence interval [CI]: 1.91, 18.83, p  = 0.0007), lower preoperative GH level (< 40 ng/mL) (OR: 7.07, CI: 2.57, 19.44, p  = 0.00005), the EOR (GTR, OR: 3.35, CI: 1.37, 8.18, p  = 0.009), and male gender (OR: 3.85, CI: 1.55, 9.61, p  = 0.003) were related with remission at 3 months ( Table 5 ). We considered factors with a p -value less than 0.1 for further analysis. On regression analysis, lower preoperative GH (< 40 ng/mL) (OR: 7.35, CI: 2.23–24.21, p  = 0.001) and male gender (OR: 4.86, CI: 1.61–14.67, p  = 0.005) were related to remission at 3 months ( Table 6 ).

Table 5. Univariate analysis.

Variable	Odds ratio	95% confidence interval	p -Value
Age (≤ 40 vs. > 40 years)	0.77	0.30, 1.97	0.60
Gender (male vs. female)	3.85	1.55, 9.61	0.003
Vision impairment (yes vs. no)	0.40	0.15, 1.10	0.07
T2W (hyper vs. iso or hypo)	1.09	0.44, 2.73	0.85
Tumour volume (< 7 vs. > 7 mL)	2.11	0.77, 5.79	0.14
Knosp grade (1 and 2 vs. 4 and 5)	6	1.91, 18.83	0.0007
Granules (sparse vs. dense)	0.18	0.02, 1.63	0.18
IHC (GH vs. GH + PRL)	0.94	0.23, 3.78	0.92
MIB index (≤ 3 vs. > 3)	1.03	0.27, 3.95	0.97
Preop GH (≤ 40 vs. > 40)	7.07	2.57, 19.44	0.00005
Surgery (endoscopic TSS vs. microscopic TSS)	1.03	0.34, 3.14	0.98
EOR (GTR vs. STR)	3.35	1.37, 8.18	0.009

Abbreviations: EOR, extent of resection; GH, growth hormone; GTR, gross total resection; IHC, immunohistochemistry; PRL, prolactin; STR, subtotal resection; TSS, transsphenoidal. p -value typed in bold indicates significant values.

Table 6. Regression analysis.

Variable	Odds ratio	95% confidence interval	p -Value
Vision impairment (yes vs. no)	0.41	0.12–1.38	0.15
Knosp grade (1 and 2 vs. 4 and 5)	3.27	0.88–12.11	0.07
Preop GH (≤ 40 vs. > 40)	7.35	2.23–24.21	0.001
EOR (GTR vs. STR)	2.50	0.84–7.40	0.98
Gender (male vs. female)	4.86	1.61–14.67	0.005

Abbreviations: EOR, extent of resection; GH, growth hormone; GTR, gross total resection; STR, subtotal resection. p -value typed in bold indicates significant values.

Discussion

Ours is a relatively large surgical series of GH-secreting pituitary adenomas from a single neurosurgical unit. The mean tumor size and the mean GH levels were much higher in our series. There were only three cases of microadenoma. Moreover, the tumor size and GH levels were significantly higher in males compared with females. In the initial years, microscopic TSS was done, however, later most of the patients underwent endoscopic TSS approach. Due to a large proportion of patients with higher Knosp grade and larger size tumors, only one-third of patients underwent GTR. The biochemical remission rate was also less in our series due to the larger size of tumors. Though some factors could be associated with biochemical remission we found that lower GH levels were associated with higher chances of remission after surgery.

Surgery is the treatment of choice for GH-secreting adenoma. The assessment of GH level is recommended at 12 weeks after surgery as per the guidelines for defining remission. ³ We did clinical, imaging, and biochemical evaluation 3 months after surgery to look for the EOR and remission. The surgical remission rate varies in published series and the range is 31.9 to 84.9%. ^{1 7 8 9} The remission rate was only 26% in our present series. The remission rate is high in microadenoma (80%) compared with macroadenoma (50%). ¹⁰ We had only three cases of microadenoma. The mean tumor volume of 7.37 mL in our series is much larger than reported in other series. ^{7 8} A systematic review was published of 972 patients. The remission rate of invasive macroadenoma was 47.6%, of noninvasive macroadenoma 76.4%, and of noninvasive microadenoma 74.2%. ¹¹ Some studies have also shown an association between poor remission rates and cavernous sinus invasion ^{7 12 13 14 15 16} In our series cavernous sinus invasion as suggested by Knosp grades 3 and 4 were seen in more than 40% of cases. The odds for remission were less with a higher Knosp grade in univariable analysis in our series.

Phan et al published a meta-analysis comparing the two surgical approaches for acromegaly and they found the remission rate is better in the endoscopy group compared with microscopy (83.8% vs. 66.9%, p ≤ 0.001) for the noninvasive macroadenoma. ¹² The remission rate was high with endoscopy (20.9%) compared with microscopy (4.5%) in our series but it was not statistically significant. The result of remission is more dependent on the extent of tumor resection rather than the surgical approach. The GTR rate in our study was 32.7%. In our present series, we had a lower rate of GTR and remission as we had a higher proportion of patients with high Knosp grades tumor, some patients in initial years did not undergo endoscopic surgery, and intentional nonremoval cavernous sinus part of tumor because of high morbidity. Moreover, we have an in-house GKRS facility available and we prefer to refer patients with residual tumors to the GKRS center.

Some studies showed that the preoperative random GH levels and IGF-1 levels were predictors of surgical remission ^{10 16 17 18 19 20 21} and the almost same number of studies established no correlation. ^{14 22 23 24} In our present study, we found that a lower preoperative GH level (< 40 ng/mL) was significantly associated with remission. Sun et al found that the female gender and young age were associated with surgical remission. ²⁵ In our study, we found that the male gender was associated with higher odds of remission. The preoperative tumor volume and preoperative GH levels were significantly lower in males compared with females in our cohort, which could have contributed to higher remission in males in our study.

Limitation

The present study is retrospective in nature, so the uniformity of data based on the time frame, surgical approach, and experience of the surgeon was lacking. In the univariable analysis, the EOR was associated with remission but in regression analysis, it was not. Only one-third of patients underwent GTR due to the relatively larger size of tumors in our cohort probably. Hence, the EOR did not retain significance in regression analysis. We did not consider the IGF-1 level as a predictor as follow-up IGF-1 assessment was not possible due to financial constraints in all cases. Medical management and postsurgery management were not uniform, hence not considered for analysis.

Conclusion

Surgery fulfills the goal of treatment, that is, biochemical remission for acromegaly patients. We found the male gender and preoperative GH level were associated with the 3-month remission of the GH level.