Malignancy in patients with acromegaly: a multicenter study across the Gulf region

Abstract

BACKGROUND:

Chronic growth hormone and insulin-like growth factor 1 excess may promote tumorigenesis, but the magnitude of malignancy risk in acromegaly remains uncertain and regional data from the Gulf are limited.

OBJECTIVE:

To describe the frequency, spectrum, detection, outcomes, and clinical correlates of malignancy among patients with acromegaly treated across Gulf tertiary centers.

DESIGN:

Multicenter retrospective observational study

SETTING:

Five tertiary centers across Saudi Arabia, Qatar, and the United Arab Emirates.

PATIENTS AND METHODS:

Medical records of patients aged >16 years with acromegaly who were evaluated from 2010 through 2024 were reviewed. Demographic, disease, treatment, cancer-testing, and malignancy data were collected. Group comparisons were descriptive and exploratory because only 10 malignancies occurred.

MAIN OUTCOME MEASURES:

Documented malignancy frequency, cancer type, method of detection, stage, treatment, outcome, and clinical correlates.

SAMPLE SIZE:

157 patients

RESULTS:

The mean age was 47.1 years, 101 patients (64.3%) were male, and the mean disease duration was 9.4 years. Malignancy was documented in 10 patients (6.4%), most commonly papillary thyroid carcinoma (8/10). Two of the eight papillary thyroid carcinomas were classified as T1a (≤10 mm); exact pathology measurements were not consistently available. One patient had colorectal cancer, and one had prostate cancer. Exploratory comparisons showed that patients with malignancy were older than those without malignancy (mean age, 56.6 vs 46.5 years; P=.039). The remaining comparisons were inconclusive. All patients with malignancy were in remission or had an excellent response at the last documented follow-up.

CONCLUSION:

Malignancy was documented in 6.4% (10 out of 157) in this multi-center Gulf acromegaly cohort, predominantly papillary thyroid carcinoma. The findings support individualized surveillance based on population guidance, symptoms, family history, and clinical findings rather than an acromegaly-specific screening program.

LIMITATIONS:

Retrospective design, small number of malignancy events, missing data, lack of a population comparator, and nonstandardized cancer testing may introduce selection and detection biases.

INTRODUCTION

Acromegaly is a chronic multisystem disorder caused by sustained growth hormone (GH) hypersecretion and increased insulin-like growth factor 1 (IGF-1), usually from a somatotroph pituitary adenoma. In addition to its characteristic somatic manifestations, prolonged hormonal excess contributes to cardiovascular, metabolic, respiratory, skeletal, and neoplastic complications. Because the phenotype evolves gradually, diagnosis is often delayed, extending exposure to GH/IGF-1 excess and increasing cumulative morbidity.^1,2

A biologic association between acromegaly and malignancy is plausible. GH/IGF-1 signaling promotes cellular proliferation and survival, inhibits apoptosis, and interacts with pathways involved in angiogenesis and genomic stability.^1,3 Nevertheless, the clinical magnitude of cancer risk remains uncertain. Meta-analyses and cohort studies have produced heterogeneous estimates, ranging from no material increase in overall cancer incidence to modestly elevated risk, with the most consistent signals involving thyroid and colorectal malignancies.^4-8 Differences in age, duration of follow-up, biochemical control, comparator selection, and screening intensity may explain part of this variation. Surveillance bias is particularly important because patients with acromegaly undergo frequent clinical assessment and may be more likely than the general population to receive thyroid imaging or colonoscopy.

This uncertainty has direct clinical consequences. Current expert recommendations do not support indiscriminate cancer testing solely because acromegaly is present. Thyroid ultrasonography is generally reserved for patients with a palpable nodule or other clinical indication, whereas colorectal screening should consider age, family history, local population guidance, and baseline colonoscopic findings.⁹ Defining region-specific patterns is therefore important to avoid both missed clinically relevant disease and overdiagnosis of indolent tumors.

Most evidence has originated from Europe, North America, and, more recently, East Asia. Data from Gulf Cooperation Council (GCC) countries remain limited despite regional differences in population age structure, background cancer epidemiology, referral patterns, and access to screening. Gulf acromegaly cohorts have documented substantial disease burden and frequent multimodal treatment,^10,11 but published malignancy data have focused mainly on thyroid disease, with a Saudi single-center study reporting no clear excess of thyroid cancer.¹² Evidence regarding other cancer types and the circumstances leading to their detection is sparse.

Accordingly, this multicenter study aimed to describe the frequency, spectrum, method of detection, stage, and clinical correlates of malignancy among patients with acromegaly treated at five tertiary centers in Saudi Arabia, Qatar, and the United Arab Emirates. By characterizing cancer diagnoses and local screening practices, we sought to provide regional data that could inform clinically proportionate surveillance and future controlled studies.

METHODS

This multicenter, retrospective observational study included patients treated at five tertiary care centers across three Gulf countries: King Fahad Medical City, King Abdulaziz Medical City, and King Saud University Medical City in Saudi Arabia; Hamad Medical Corporation in Qatar; and Tawam Hospital in the United Arab Emirates (UAE). Medical records of patients evaluated between 2010 and 2024 were reviewed. This study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Tawam Hospital, Al Ain, UAE (approval no. DOH/ADHRTC/2022/1377; October 22, 2024) and King Fahad Medical City, Riyadh, Saudi Arabia (approval no. 26–530; August 13, 2026). The requirement for individual informed consent was waived owing to the retrospective, noninterventional nature of the study.

Participants and definitions

Patients aged >16 years with a diagnosis of acromegaly were eligible. Patients with insufficient records to confirm the diagnosis or evaluate the prespecified outcomes were excluded. Acromegaly was diagnosed on the basis of compatible clinical features and an IGF-1 level above the age-specific normal range.¹³

Biochemical remission at last follow-up was defined as an age-adjusted IGF-1 concentration within the reference range. This cross-sectional measure was not treated as cumulative hormonal exposure. Disease duration was measured from the estimated onset of symptoms to the last follow-up. Active-disease duration was measured from the estimated symptom onset to the first documented biochemical remission. Neither measure was calculated relative to cancer diagnosis because the timing of malignancy in relation to acromegaly was not consistently documented.

Data collection

Collected variables included demographics, anthropometrics, acromegaly duration and activity, adenoma characteristics, biochemical results, treatments, thyroid ultrasonography, colonoscopy, malignancy type, method of detection, stage, treatment, and outcome. Cancer diagnoses were abstracted from pathology and clinical records. Testing was not standardized across centers. Records did not consistently identify whether thyroid ultrasonography or colonoscopy was performed as routine asymptomatic screening or because of symptoms, a palpable nodule, goiter, family history, or another clinical indication. Exact papillary thyroid carcinoma (PTC)pathology measurements were also not consistently available; when exact size was unavailable, size ranges were reported only when they could be inferred from the recorded American Joint Committee on Cancer (AJCC) tumor (T) category.

Statistical analysis

Statistical analysis was performed using R version 4.4.1. Continuous variables were reported as mean±standard deviation (SD). Categorical variables were reported as counts and percentages (%). Fisher's exact test was used for categorical variables and Welch's two-sample t test was used for continuous variables. Statistical significance was defined as a P<.05. Because there were only 10 malignancy events, all comparisons were exploratory and intended for descriptive hypothesis generation; P values were not interpreted as evidence for or against an association.

RESULTS

A total of 157 patients were included (Table 1). Mean age was 47.1±13.0 years, 101 patients (64.3%) were male, the mean body mass index (BMI) was 31.7±6.1 kg/m², and the mean disease duration was 9.4±6.8 years. Among 123 patients with available adenoma-size data, 107 (87.0%) had a macroadenoma. Mean active-disease duration was 6.4±5.4 years. Diabetes mellitus was present in 42/145 (29.0%) patients with available data. Surgery was performed in 140/155 (90.3%) patients. Medical therapies included dopamine agonists in 50/154 patients (32.5%), first-generation somatostatin receptor ligands—octreotide or lanreotide—in 87/154 (56.5%), the second-generation somatostatin receptor ligand pasireotide in 14/154 (9.1%), and the growth hormone receptor antagonist pegvisomant in 15/154 (9.7%). Radiotherapy was delivered to 32/156 (20.5%). At last follow-up, disease-control status was available for 154 patients: 64 (41.6%) had active disease and 90 (58.4%) were in biochemical remission.

Table 1.

Cohort characteristics, treatment, and outcomes.

Variables	Results
Age (years)	47.1±13.0
Body Mass Index (kg/m2)	31.7±6.1
Disease duration (years)	9.4±6.8
Active-disease duration (years)	6.4±5.4
Sex (%)
Female	56/157 (35.7)
Male	101/157 (64.3)
Adenoma size (%)
Macroadenoma	107/123 (87.0)
Microadenoma	16/123 (13.0)
Disease control at last follow-up (%)
Active disease	64/154 (41.6)
Biochemical remission	90/154 (58.4)
Thyroid ultrasonographya	98/142 (69.0)
Colonoscopya	91/147 (61.9)
Treatment (%)
Surgery	140/155 (90.3)
Radiotherapy	32/156 (20.5)
First-generation somatostatin receptor ligand (octreotide and lanreotide)	87/154 (56.5)
Second-generation somatostatin receptor ligand pasireotide	14/154 (9.1)
Dopamine agonist	50/154 (32.5)
Pegvisomant	15/154 (9.7)
Malignancy	10/157 (6.4)

Note: Data presented as mean±SD or n/N (%), where N is the available-case denominator;

^aIndicates that the test was documented but the clinical indication was not consistently available.

Cancer status was documented for all 157 patients; 10 patients (6.4%) had a malignancy (Table 2). Eight had papillary thyroid carcinoma (PTC). Two PTCs were recorded as T1a (≤10 mm); exact pathology measurements were not consistently available for the other tumors. All patients with documented thyroid cancer underwent surgery, two received radioactive iodine, and all had an excellent response at the last documented follow-up.

Table 2.

Characteristics and outcomes of patients with malignancy.

No.	Age, sex	Cancer type	TNM classification	Tumor-size information	Stage	Treatment	Outcome
1	64, F	PTC	T2 N0 M0	>20–≤40 mm by T2 category; exact size unavailable	I	Surgery + RAI	Excellent response
2	80, F	PTC	T1a N0a M0	≤10 mm (microcarcinoma)	I	Surgery	Excellent response
3	42, M	PTC	T2 N0a M0	>20–≤40 mm by T2 category; exact size unavailable	I	Surgery	Excellent response
4	42, M	PTC	T1b N0a M0	>10–≤20 mm by T1b category; exact size unavailable	I	Surgery + RAI	Excellent response
5	44, M	PTC	T2 N0 M0	>20–≤40 mm by T2 category; exact size unavailable	I	Surgery	Excellent response
6	62, M	PTC	T3b N1a M0	Exact size unavailable; T3b denotes gross extrathyroidal extension	II	Surgery	Excellent response
7	55, M	PTC	T1b N0 M0	>10–≤20 mm by T1b category; exact size unavailable	I	Surgery	Excellent response
8	45, M	PTC	T1a N0 M0	≤10 mm (microcarcinoma)	I	Surgery	Excellent response
9	68, F	Colorectal cancer	NA	NA	I	Surgery + chemotherapy	Remission
10	64, M	Prostate cancer	NA	NA	I	Radiotherapy	Remission

F, female; M, male; PTC, papillary thyroid carcinoma; RAI, radioactive iodine; NA, not applicable. TNM classifications were transcribed from the source medical records; PTC prognostic stage groups were reassigned according to the AJCC 8th edition using age at cancer diagnosis. T1a corresponds to a tumor ≤10 mm; T1b to >10–≤20 mm; and T2 to >20–≤40 mm. T3b denotes gross extrathyroidal extension and does not specify tumor size. N1a denotes metastasis to central-compartment lymph nodes. Precise tumor size in millimeters was not uniformly available in the source pathology reports; size ranges shown in the table are derived from the recorded T category and are not imputed exact measurements.

One patient had colorectal cancer. This 68-year-old woman presented with diarrhea and was diagnosed with stage I disease after previously declining screening colonoscopy. She underwent surgery and chemotherapy and was in remission at the last follow-up. Her father, mother, and sister had gastrointestinal malignancies.

One patient had prostate cancer. This 64-year-old man presented with obstructive lower urinary tract symptoms and was found to have a suspicious prostatic nodule during evaluation. He was diagnosed with stage I prostate cancer and was treated with radiotherapy alone. At last follow-up, he remained in remission.

Formal tumor size in millimeters was not uniformly available in the pathology reports; the T-classification already recorded for each case (Table 2) provides an approximate size correlate under the AJCC system (T1a, ≤1 cm [microcarcinoma]; T1b, >1–2 cm; T2, >2–4 cm; T3, >4 cm and/or minimal extrathyroidal extension). Using this correlate, 2/8 PTCs (25%) were in the T1a microcarcinoma range, 2/8 (25%) were T1b, 3/8 (38%) were T2, and 1/8 (12%) was T3b with regional nodal involvement (N1a). The predominance of lowstage, node-negative disease is consistent with a contribution from ultrasound-driven detection of subclinical disease, although most carcinomas exceeded the strict microcarcinoma threshold and one had nodal spread, so overdiagnosis of indolent microcarcinoma does not fully account for the thyroid cancer cases observed.

Table 3 presents descriptive, exploratory comparisons by malignancy status. Patients with malignancy were older than those without malignancy (mean age, 56.6 vs. 46.5 years; P=.039). The remaining comparisons were inconclusive because of the small malignancy group and missing data. Among patients with available data, thyroid ultrasonography had been performed in 7/8 patients with malignancy (88%) and 91/134 without malignancy (68%; P=.435); the corresponding proportions for colonoscopy were 7/10 (70.0%) and 84/137 (61%), respectively (P=.742). The indications for these tests were not consistently recorded, so comparisons by testing status cannot estimate screening effectiveness or cancer risk.

Table 3.

Exploratory comparison according to malignancy status.

Characteristics	Overall N=157a	No N=147a	Yes N=10a	P valuec
Sex (%)
Female	56 (35.7)	53 (36.1)	3 (30)	>.999
Male	101 (64.3)	94 (63.9)	7 (70)
Age (years)	47.1±13.0	46.5±12.8	56.6±13.1	.039
Body mass index (kg/m2)	31.7±6.1	31.7±6.2	31.9±6.1	.940
Body mass index category (%)
Normal	17 (10.8)	16 (10.9)	1 (10)	.888
Overweight	49 (31.2)	45 (30.6)	4 (40)
Obese	91 (58.0)	86 (58.5)	5 (50)
Adenoma size (%)
Macroadenoma	107 (87.0)	100 (87.0)	7 (88)	>.999
Microadenoma	16 (13.0)	15 (13.0)	1 (12)
Disease duration (years)	9.4±6.8	9.5±6.9	8.1±4.9	.398
Active-disease duration (years)	6.4±5.4	6.4±5.4	5.8±6.2	.770
Disease control at last follow-up (%)
Active disease	64 (41.6)	59 (41.0)	5 (50)	.742
Biochemical remission	90 (58.4)	85 (59.0)	5 (50)
Glycemic status (%)
No diabetes	56 (38.6)	53 (39.3)	3 (30)	.917
Prediabetes	47 (32.4)	43 (31.9)	4 (40)
Diabetes	42 (29.0)	39 (28.9)	3 (30)
Thyroid ultrasonographyb	98 (69.0)	91 (67.9)	7 (87)	.435
Colonoscopyb	91 (61.9)	84 (61.3)	7 (70)	.742

^aData presented as n (%), using available-case denominators, or mean±SD.

^bIndicates that the test was documented but the clinical indication was not consistently available.

^cP values are from Fisher's exact test for categorical variables and Welch's two-sample t test for continuous variables. No adjustment was made for multiple comparisons. All P values are exploratory and severely underpowered because only 10 malignancy events occurred. Nonsignificant results do not demonstrate absence of association, and the age result should be considered hypothesis-generating.

DISCUSSION

In this multicenter GCC cohort, malignancy was documented in 10 out of 157 (6.4%) patients, most commonly PTC. Recorded outcomes were favorable. Because only 10 events occurred, all subgroup comparisons are exploratory and should be read as descriptive. In addition, the study lacked a population comparator and used nonstandardized testing; therefore, 6.4% represents the observed proportion in this cohort and not an estimate of excess cancer risk attributable to acromegaly.

The observed proportion was lower than that reported in several European and US referral cohorts but similar to the 6.8% reported in a Turkish two-center study.^6,14-16 Cross-study comparisons require caution because cancer ascertainment varies with cohort age, disease duration, referral setting, screening intensity, and background cancer incidence.

The relatively young mean age (47.1 years) and mean disease duration of (9.4 years) in our cohort may have limited the detection of malignancies with long latency. Recent large, controlled studies have also yielded inconsistent findings. Korean and Swedish nationwide cohorts reported higher overall cancer incidence in acromegaly,^7,8 whereas other population-based analyses found smaller or site-specific associations.^5,17 These differences reinforce the need for matched regional comparators and time-to-event analyses before causal conclusions are drawn.

PTC accounted for 8/10 malignancies and occurred in 8/157 (5.1%) patients. Thyroid nodules are common in acromegaly, and a meta-analysis estimated thyroid cancer prevalence at 4.3%, although estimates were heterogeneous and susceptible to ascertainment bias.¹⁸ Two PTCs were T1a tumors (≤10 mm), meeting the size criterion for papillary thyroid microcarcinoma; exact tumor measurements were not consistently available for the remaining cases. The predominance of early-stage, low-risk PTC in our cohort, together with the frequent use of thyroid ultrasonography, makes detection bias and overdiagnosis plausible. Because ultrasound indications were not consistently documented, the study cannot separate asymptomatic routine imaging from imaging prompted by a nodule, goiter, or other clinical finding. This does not exclude a biologic effect of GH/IGF-1 signalling, but it cautions against interpreting every detected microcarcinoma as clinically consequential. Our results do not support routine thyroid ultrasonography for all patients with acromegaly; examination-directed imaging remains consistent with consensus recommendations.⁹

Only one colorectal cancer was identified. It was stage I and diagnosed during evaluation of diarrhea in a 68-year-old woman with a strong family history of gastrointestinal cancer; no asymptomatic colorectal cancer was detected among patients who underwent colonoscopy. This finding should not be interpreted as evidence that colorectal risk is absent or that screening is ineffective, given the single event and incomplete, nonprotocolized screening. Contemporary studies continue to report variable colorectal cancer estimates, while colorectal adenomas appear consistently increased.^4,8,17,19,20 A pragmatic approach is therefore to follow local age- and risk-based screening, with additional attention to family history, symptoms, disease activity, and prior colonoscopic findings, rather than assuming that this cohort validates or refutes universal colonoscopy.⁹

Apart from age, no statistically significant association was found between malignancy and sex, adenoma size, disease duration, biochemical status, or treatment. These analyses were exploratory and severely underpowered; a nonsignificant P value must not be interpreted as evidence of no association. The age finding itself is hypothesis-generating and may reflect confounding. In a prospective US cohort, cumulative exposure to IGF-1 excess predicted cancer more clearly than isolated measurements,⁶ whereas the recent Swedish study found little difference in cancer incidence by biochemical control.⁸ Our control variable was based only on the last recorded IGF-1 value and could not represent cumulative GH/IGF-1 exposure. Disease duration was measured to last follow-up and active-disease duration to first documented remission, not to the date of cancer diagnosis. The records did not consistently establish whether malignancy preceded, coincided with, or followed acromegaly diagnosis; these variables therefore cannot be interpreted as cancer-latency measures. Residual confounding from age, smoking, family history, obesity, and environmental factors remains possible.

Most patients underwent pituitary surgery, and 21% received radiotherapy. No secondary intracranial malignancy was documented; however, the cohort size and follow-up were insufficient to evaluate this rare, long-latency outcome. Large studies indicate a small increased risk of second brain tumors after pituitary radiotherapy, which warrants appropriate long-term clinical follow-up without implying that radiotherapy should be withheld when indicated.²¹ The effects of acromegaly medications on cancer incidence likewise cannot be determined from these data.

Medical therapy was heterogeneous, with first-generation somatostatin receptor ligands used most frequently, followed by dopamine agonists, pegvisomant, and pasireotide. Treatments commonly overlapped or changed over time. Because cumulative drug exposure was unavailable and treatment subgroups were small and confounded by disease severity, agent-specific associations with malignancy could not be assessed. Accordingly, these findings should not be interpreted as evidence that any acromegaly medication modifies malignancy risk.

Strengths include the multicenter design, inclusion of five tertiary centers across three GCC countries, and characterization of cancer type, detection, stage, and outcome. Limitations include the retrospective design, modest sample size, only 10 cancer events, missing data, lack of a matched control group or standardized cancer ascertainment, variable testing between centers, uncertain indications for thyroid ultrasonography and colonoscopy, and potential referral, survival, selection, surveillance, and detection bias. The study could not reliably establish the temporal relationship between cancer diagnosis and acromegaly onset or biochemical control, and a single last-follow-up IGF-1 value did not measure cumulative hormonal exposure. Overall, the findings add regional descriptive evidence but do not justify an acromegaly-specific screening program. Larger prospective Gulf studies using population comparators, standardized surveillance, cumulative GH/IGF-1 exposure, and time-to-event cancer incidence rates are needed.

CONCLUSION

Among 157 patients with acromegaly treated at five Gulf tertiary centers, malignancy was documented in 10 cases (6.4%), most commonly PTC, with favorable recorded outcomes. The absence of a population comparator, low number of cancer events, nonstandardized testing, and likely detection bias preclude conclusions about excess cancer risk, subgroup associations, or screening effectiveness. These findings support individualized surveillance based on established population guidance, symptoms, family history, and clinical findings, while underscoring the need for larger controlled regional studies.

Funding Statement

None.

AUTHORS' CONTRIBUTIONS

K.M.A. and M.H.A. conceived and designed the study. K.M.A., M.M., T.E., A.E., T.J., Z.D., and M.H.A. contributed to data acquisition and clinical interpretation. F.A. performed the statistical analysis and contributed to interpretation of the results. K.M.A. drafted the initial manuscript. M.H.A., M.M., T.E., F.A., A.E., T.J., and Z.D. critically revised the manuscript for important intellectual content. All authors reviewed and approved the final manuscript and agreed to be accountable for all aspects of the work.

FUNDING

This study received no external research funding. Recordati Rare Diseases provided support for the article publication fee only and had no role in study design; data collection, analysis, or interpretation; manuscript preparation; or the decision to submit.

DATA AVAILABILITY

Deidentified data supporting the findings may be available from the corresponding author, Mussa H. Almalki, upon reasonable request and subject to institutional approvals and applicable data-protection requirements.

GENERATIVE AI STATEMENT

During preparation of this manuscript, the authors used ChatGPT (OpenAI) to assist with language editing, organization, and consistency checking. The tool was not used to generate or analyze study data. All authors reviewed and revised the output, verified the accuracy of the manuscript, and take full responsibility for its content.