Association between Neuroendocrine Tumors Biomarkers and Primary Tumor Site and Disease Type based on Total ⁶⁸Ga-DOTATATE-Avid Tumor Volume Measurements

Abstract

Objective

To determine the association between neuroendocrine tumor (NET) biomarker levels and extent of disease as assessed by ⁶⁸Ga DOTATATE PET/CT imaging.

Design

A retrospective analysis of a prospective database of patients with NETs.

Methods

Fasting plasma chromogranin A (CgA), neuron-specific enolase (NSE), gastrin, glucagon, vasoactive intestinal peptide (VIP) and pancreatic polypeptide (PP), and 24-hour urinary 5-hydroxyindoleacetic acid (5-HIAA) levels were measured. Correlation between biomarkers and total ⁶⁸Ga-DOTATATE-avid tumor volume (TV) was analyzed.

Results

The analysis included 232 patients. In patients with pancreatic NETs (n=112), ⁶⁸Ga-DOTATATE TV correlated with CgA (r=0.6, p=0.001, Spearman). In patients with multiple endocrine neoplasia type 1 (n=39), ⁶⁸Ga-DOTATATE TV correlated with glucagon (r=0.5, p=0.02) and PP levels (r=0.5, p=0.049). In patients with von Hippel-Lindau (n=24), plasma VIP (r=0.5, p=0.02) and PP levels (r=0.7, p<0.001) correlated with ⁶⁸Ga-DOTATATE TV. In patients with small intestine NET (SINET, n=74), ⁶⁸Ga-DOTATATE TV correlated with CgA (r=0.5, p=0.02) and 5-HIAA levels (r=0.7, p<0.001), with 5-HIAA ≥8.1 mg/24h associated with metastatic disease with high positive (81.8%) and negative (85.7%) predictive values (p=0.001). ⁶⁸Ga-DOTATATE TV in patients with NET of unknown primary (n=16) and those with NET of other primary location (n=30) correlated with 5-HIAA levels (r=0.8, p=0.002, and r=0.7, p=0.02, respectively).

Conclusions

Our data supports the use of specific NET biomarkers based on the site of the primary NET and the presence of hereditary syndrome-associated NET. High urinary 5-HIAA levels indicate the presence of metastatic disease in patients with SINET.

Introduction

Most neuroendocrine tumors (NETs) originate from endocrine cells spread throughout the respiratory and gastrointestinal tracts. Although these tumors tend to have an indolent natural course, NETs present with distant metastases in 40–50% of patients [1–3], usually to the liver and regional lymph nodes (LN), and less often to the lungs, brain, bones and peritoneal cavity[1].

Tumor biomarkers are important for diagnosis and follow-up of patients with NETs, especially after treatment of locally advanced and/or metastatic disease. Several biomarkers are used for NET surveillance, including plasma levels of chromogranin A (CgA), pancreatic polypeptide (PP), neuron-specific enolase (NSE), and urinary 24-hour 5-hydroxyindoleacetic acid (5-HIAA) levels depending on the primary tumor location [4,5]. The sensitivity and specificity of different biomarkers depend on the site of the primary tumor, with CgA mainly used for patients with pancreatic and midgut NET, NSE being most useful in patients with poorly differentiated NET, and 5-HIAA in patients with carcinoid syndrome. The accuracy of biomarker measurement depends on other factors unrelated to the tumor, such as kidney function and presence of gastric atrophy for CgA, and dietary factors and medical treatment for 5-HIAA[6]. Furthermore, the determination of whether elevated NET biomarkers accurately reflect the presence of disease has been limited by the “gold standard test” used for assessing the presence of disease; pathologic analysis, and/or the type of imaging modality. As a result of these limitations, there is a lack of consensus on the utility of these biomarkers for estimating the total tumor burden in patients with NETs [7].

Well-differentiated NETs typically express somatostatin receptors (SSTRs), thus radiolabeled somatostatin analogues have been used for diagnosis and follow-up of patients with NET [5]. ¹¹¹In-pentetreotide (Octreoscan) has been widely used for NET imaging, but is limited by its low image resolution and long scan duration and relatively lower SSTRs affinity[8]. In contrast, the ⁶⁸-Gallium (⁶⁸Ga)-DOTA-compounds recently introduced into clinical practice (⁶⁸Ga-DOTATATE, ⁶⁸Ga-DOTANOC, ⁶⁸Ga-DOTATOC), have been shown to have higher sensitivity for detecting NET than Octreoscan[9] and significantly impact the management of patients with NET [10,11].

To our knowledge, no study has yet assessed the utility of biochemical biomarkers for evaluating NET burden using the new generation DOTA-peptide based PET/CT imaging [12,13]. Thus, in this study we analyzed the correlation between biochemical biomarkers of NETs and total ⁶⁸Ga-DOTATATE-avid tumor volume (TV), as a measure of total tumor burden in patients with NET.

Subjects and Methods

Study population

Patients known to have NETs based on imaging (CT, MRI, ¹⁸F-FDG PET), biochemical evidence of NETs, and/or pathologically confirmed NET were included in this study. The study was performed under an investigational new drug approval from the United States Food and Drug Administration. The full study eligibility criteria were as previously reported [10]. The study was approved by the National Cancer Institute Institutional Review Board and the National Institutes of Health Radiation Safety Committee (NCT01967537). Written informed consent was obtained from all study participants.

NETs were sub-grouped according to the primary tumor location based on conventional anatomic imaging, ⁶⁸Ga-DOTATATE PET/CT and/or pathological evaluation. Patients were subdivided into those with pancreatic NETs (PNETs) or small intestine NETs (SINETs). Subjects with metastatic NET and with pathologic ⁶⁸Ga-DOTATATE uptake but no clear primary lesion were defined as NET of unknown primary (NEToUP, n=16). Thirty subjects with NET of other origins were grouped as NET of other primary location (NEToOL) due to their small number. Duodenal NETs were included in NEToOL due to their distinct characteristics compared with other SINET [14]. Tumor grade was determined according to the 2010 World Health Organization (WHO) definitions as G1, G2 or G3 according to Ki-67 Index (<3%, 3–20% and >20%) and mitotic rate (<2, 2–20 and >20 per 10 high-power microscopic fields) in patients who had their tumor resected or biopsied, respectively[15].

Biochemical and Imaging Evaluation

All patients underwent testing for fasting plasma CgA, PP, NSE, vasoactive intestinal polypeptide (VIP), gastrin and glucagon, and 24-hour 5-HIAA urinary collections. Thirty one patients receiving chronic treatment with proton-pump inhibitors and one patient with a plasma creatinine >2 mg/dl were excluded from the analysis of plasma CgA levels, as they may have an increase in plasma CgA levels that does not reflect disease burden. Plasma glucagon levels (LINCOplex Kit, Luminex 200, Missouri), plasma VIP levels (Peninsula Laboratories, California), and plasma PP levels (Mayo Medical laboratories, Rochester, Minnesota)[3] were measured using an immunoradiometric assay; whereas immunochemiluminometric assays were used for measuring plasma chromogranin A (Mayo Medical laboratories, Rochester, Minnesota), gastrin (Immulite 2000, Siemens, California), and NSE levels (NSE Kryptor, BRAHMS, Germany). Urinary 5-HIAA levels were measured using liquid chromatography-tandem mass spectrometry (Mayo Medical laboratories, Rochester, Minnesota) [4]. The cut-offs for positivity were determined by the kit manufacturer or by the laboratory performing the tests.

Both ⁶⁸Ga-DOTATATE PET/CT scans and biomarkers level measurements were performed concurrently at study inclusion. The biochemical evaluation for patients with a clinical suspicion of insulinoma included measurements of fasting glucose, insulin, proinsulin, and C-peptide during hypoglycemia. Supervised fasting test was conducted if necessary.

For ⁶⁸Ga-DOTATATE PET/CT imaging, 185 MBq (5 mCi) ⁶⁸Ga-DOTATATE was administered through a peripheral intravenous line. After approximately 60 minutes, the patient was positioned supine in a PET/CT scanner, and images were obtained from mid thighs to the top of the skull. A low-dose, non-contrast enhanced CT was used for attenuation correction and anatomic localization. Maximum standardized uptake values (SUVmax) of visualized lesions were calculated based on patient total body weight. Patients treated with long-acting somatostatin analogues were scanned before the next scheduled monthly dose while those on short-acting somatostatin analogues discontinued the drug 24 hours before imaging.

Quantification of tumor volume by ⁶⁸Ga-DOTATATE PET/CT imaging

Disease burden was assessed by quantifying ⁶⁸Ga-DOTATATE uptake using the MIM Vista workstation (version 6.5.9). A VOI (volume of interest) encompassing the entire scan was drawn and subsequently a SUVmax threshold-based approach was applied in order to include all sites of non-physiologic uptake. The SUV-threshold based technique for segmentation is commonly used and was tested in many clinical studies [16]. The software used in the current analysis enables automatic generation of separate VOIs encircling all areas above the SUVmax threshold set by the user. This software has been tested and reported before [17]. Afterwards, an experienced nuclear medicine physician who was blinded to the clinical and laboratory data manually excluded areas with physiologic ⁶⁸Ga-DOTATATE uptake. The volumes of all lesions with a pathological 68Ga-DOTATATE uptake were obtained automatically. The sum of these volumes reflects the entire tumor tissue expressing somatostatin-receptor type 2, and was defined as total volume (TV).

Statistical analysis

Statistical analyses were performed using SPSS 20.0 software (SPSS Inc., Chicago, IL, USA). Results are expressed as mean ± standard deviation (SD) unless otherwise indicated. For group comparisons, the independent Student’s t-test were used to analyze differences in continuous variables and the chi-square test was used to analyze differences in categorical variables. The Pearson product was used for analysis of correlations between variables. Receiver operating characteristic (ROC) curve analysis was performed to assess the accuracy of 24-hour urinary 5-HIAA and plasma CgA levels for indicating metastatic SINET. Both TV and biochemical biomarker levels were logarithmically transformed to approximate normality. Non-parametric tests were used as appropriate, including Mann-Whitney U for continuous variables, and Fisher’s exact test for categorical variables. The p value for significance was set at less than 0.05.

Results

The study cohort consisted of 232 patients with a mean age of 54.3±14.1 years, mean body-mass index of 28.9±6.8 Kg/m², and 130 (56.0%) were females. The sites of NETs were PNET (n=112), SINET (n=74) and tumors of unknown primary (n=16). NEToOL consisted of gastric (n=6), duodenal (n=10), rectal (n=6), appendiceal (n=1), large intestine (n=2), bronchial (n=3), thymic (n=1) and bladder (n=1) NETs. The study cohort clinical characteristics are summarized in Table 1. Mean biomarkers levels, normal reference ranges and percentage of patients with elevated biomarkers levels are summarized in Table 2.

Table 1.

Study cohort clinical characteristics (n=232)

Primary tumor n(%)
PNET	112 (48.3%)
SINET	74 (31.9%)
Unknown primary	16 (6.9%)
Other *	30 (12.9%)
Genetic syndromes n(%)
MEN-1	39 (16.8%)
VHL	28 (12.1%)
NF-1	1 (0.4%)
Cowden	1 (0.4%)
Functional status n(%)
Nonfunctional	121 (52.2%)
Insulinoma	12 (5.2%)
Gastrinoma	16 (6.9%)
Carcinoid syndrome	79 (34.1%)
Other hormone production	4 (1.7%)
WHO Tumor grade^
G1	51 (52.0%)
G2	42 (42.9%)
G3	5 (5.1%)
Metastases n(%)
Any	134 (57.8%)
Lymph nodes	84 (36.2%)
Liver	83 (35.8%)
Bones	32 (13.8%)
Lungs	12 (5.2%)
Treatment
SSA n(%)	79 (34.3%)
History of surgical resection of NETs before 68Ga-DOTATATE PET imaging	133 (57.3%)
Pathological uptake on 68Ga-DOTATATE PET	185 (79.7%)

PNET, pancreatic neuroendocrine tumor; SINET, small-intestine neuroendocrine tumor; MEN-1, multiple endocrine neoplasia syndrome type 1; VHL, von Hippel-Lindau disease; NF-1, neurofibromatosis type 1; SSA, somatostatin analogues

^*consisted of gastric (n=6), duodenal (n=10), rectal (n=6), appendiceal (n=1), large intestine (n=2), bronchial (n=3), thymic (n=1) and bladder (n=1) NETs

^{^}Only a subset of patients had WHO tumor grade status available.

Table 2.

Neuroendocrine tumor biomarker levels distribution in study cohort

Biomarker (n) Upper limit of normal range	Mean±SD	Number (%) of patients with elevated biomarkers*
24-hour urinary 5-HIAA levels (n=208)
≤ 8 mg/24h	10.5±22.9	48 (23.1%)
Plasma chromogranin A levels+ (n=83)
≤ 93 ng/mL	237±712	24 (28.9%)
Plasma NSE levels (n=81)
≤ 15 ng/mL	9.4±1.9	6 (7.4%)
Plasma gastrin levels (n=219)
≤ 100 pg/mL	176±661	42 (19.2%)
Plasma glucagon levels (n=220)
≤ 80 pg/mL	41.4±42.0	25 (11.4%)
Plasma VIP levels (n=169)
≤ 75 pg/mL	48.0±117.0	8 (4.7%)
Plasma PP levels (n=199)
≤ 291pg/mL	228±493	28 (14.1%)

SD, standard deviation; 5-HIAA, 5-hydroxyindoleacetic acid; NSE, neuron-specific enolase; VIP, vasoactive intestinal peptide; PP, pancreatic polypeptide

^*Based on upper limit of the reference range.

⁺Patients with chronic kidney injury (n=1) and/or treatment with proton-pump inhibitors (n=31) were excluded from the analysis of chromogranin A.

Tumor burden measurement by ⁶⁸Ga-DOTATATE TV

The mean ⁶⁸Ga-DOTATATE TV was 69.2±152.1 ml and mean SUVmax was 69.1±50.9. ⁶⁸Ga-DOTATATE TV positively correlated with 24-hour urinary 5-HIAA levels (r=0.5, p<0.001) and plasma CgA levels (r=0.4, p=0.001). Patients with metastatic disease had higher 24-hour 5-HIAA levels (14.8±29.7 vs. 4.8±2.2 mg/24h, p<0.001), plasma gastrin levels (252±863 vs. 74±120 pg/ml, p=0.02), and plasma CgA levels (382±1030 vs.112±255 ng/ml, p=0.002) compared to patients without metastatic disease, respectively.

Tumor burden measurement in patients with PNET

Patients with PNETs (n=112) had mean ⁶⁸Ga-DOTATATE TV of 71.3±167.5 ml and mean SUVmax of 85.5±58.3. Sixty-two patients with PNETs (55.4%) had metastatic disease at evaluation. Among patients with PNETs, tumor burden as measured by ⁶⁸Ga-DOTATATE TV positively correlated with plasma CgA levels (r=0.6, p=0.001, Spearman’s rho), and plasma NSE levels had a similar trend (r=0.4, p=0.05). Similar trend was found for plasma NSE levels among patients with metastatic PNETs (r=0.5, p=0.07).

Among patients with multiple endocrine neoplasia type-1 (MEN-1) with PNETs, ⁶⁸Ga-DOTATATE TV positively correlated with plasma glucagon (r=0.5, p=0.01, Figure 1A) and PP levels (r=0.5, p=0.049). In patients with von Hippel-Lindau (VHL), ⁶⁸Ga-DOTATATE TV positively correlated with plasma VIP (r=0.5, p=0.02, Figure 1B) and PP levels (r=0.7, p<0.001, Figure 1C).

Figure 1.

Correlation analysis between ⁶⁸Ga-DOTATATE-avid tumor volume of pancreatic neuroendocrine tumors to plasma glucagon levels among patients with multiple endocrine neoplasia syndrome type-1 (A), plasma pancreatic polypeptide and vasoactive intestinal-peptide levels among patients with von Hippel-Lindau disease (B and C, respectively)

Patients with WHO 2010 G1 PNETs had positive correlation between TV and plasma glucagon levels (n=22, r=0.5, p=0.02) that was not significant in patients with G2 PNETs (n=16). In 16 patients with gastrinomas of either pancreatic or duodenal location, ⁶⁸Ga-DOTATATE TV had strong positive correlation with plasma gastrin levels (r=0.8, p=0.001).

Tumor burden measurement in patients with SINET

In patients with SINET (n=74), the mean ⁶⁸Ga-DOTATATE TV and SUVmax were 74.1±133.5 ml and 49.1±25.9, respectively. 24-hour urinary 5-HIAA (p<0.001) levels were higher among patients with distant metastases compared with patient without metastases. The burden of SINET as measured by ⁶⁸Ga-DOTATATE TV correlated positively with plasma CgA levels (r=0.5, p=0.02; Figure 2A) and with 24-hour urinary 5-HIAA levels (r=0.7, p<0.001; Figure 2B). Similarly, among the subgroup of patients with metastatic SINET, ⁶⁸Ga-DOTATATE TV had positive correlation with CgA levels (r=0.5, p=0.02) and with 24-hour urinary 5-HIAA levels (r=0.7, p<0.001). In patients with carcinoid syndrome (n=32), ⁶⁸Ga-DOTATATE TV correlated with 24-hour urinary 5-HIAA (r=0.7, p<0.001) and plasma CgA levels (r=0.5, p=0.02), whereas in patients without carcinoid syndrome ⁶⁸Ga-DOTATATE TV correlated only with 24-hour urinary 5-HIAA levels (n=15, r=0.9, p<0.001).

Figure 2.

Correlation analysis between ⁶⁸Ga-DOTATATE-avid tumor volume to plasma chromogranin A and 24-hour 5-hydroxyindoleacetic acid (5-HIAA) in patients with small intestine neuroendocrine tumors (A and B, respectively).

Urinary 5-HIAA levels ≥8.1 mg/24h (upper limit of normal range, 8 mg/24h) predicted metastatic disease based on ⁶⁸Ga-DOTATATE PET/CT among subjects with SINET that were not treated with somatostatin analogues, with a sensitivity of 81.8% and a specificity of 85.7%, and positive and negative predictive value of 81.8% and 85.7%, respectively (p=0.001, Figure 3A). This was true also among patients treated with somatostatin analogues (Figure 3B) and among all subjects with SINET (Figure 3C). In contrast, ROC curve analysis revealed limited utility of plasma CgA levels for predicting metastatic SINET (area under the curve = 0.668, 95% confidence interval 0.548–0.788, p=0.008).

Figure 3.

Receiver Operating Characteristic (ROC) analysis for 24-hour urinary 5-hydroxyindoleacetic acid (5-HIAA) levels for predicting metastatic disease in subjects with small-intestine neuroendocrine tumors (SINETs). Patients with (A) and without (B) somatostatin analogues treatment, and including all SINET patients (C).

Tumor burden measurement in NEToUP and NEToOL

Among sixteen patients with NEToUP correlation analysis revealed strong correlation between ⁶⁸Ga-DOTATATE TV and 24-hour urinary 5-HIAA levels (r=0.8, p=0.002; Figure 4). Subjects with NEToOL (n=30) had positive correlation between ⁶⁸Ga-DOTATATE TV and 24-hour 5-HIAA (r=0.7, p=0.02, Spearman’s rho).

The analyses were repeated after excluding patients treated with somatostatin analogues with findings similar to those described for the whole study cohort analysis. Among subjects with NEToOL, excluding patients with somatostatin analogue treatment revealed a correlation between ⁶⁸Ga-DOTATATE TV and plasma PP levels (n=7, r=0.8, p=0.03).

Discussion

In this study, we analyzed the utility of seven biomarkers for assessing NET burden using a novel tumor burden measurement approach based on ⁶⁸Ga-DOTATATE TV in a large cohort of patients with NET. ⁶⁸Ga-DOTATATE TV correlated strongly with urinary 5-HIAA and with plasma CgA among patients with SINETs. In addition, MEN-1 patients with PNETs had correlation between tumor burden and plasma glucagon levels, and subjects with VHL disease had correlation between ⁶⁸Ga-DOTATATE TV and plasma PP and VIP levels. Finally, urinary 5-HIAA levels ≥8.1 mg/24h was associated with metastatic disease in patients with SINET, independent of somatostatin analogue treatment.

Tumor biomarkers measurement are the cornerstone of NET diagnosis[18], and are essential for surveillance after various treatment modalities are carried out. Plasma CgA levels reflect tumor burden in non-functioning PNETs[19], and high urinary 5-HIAA was found mainly in patients with metastatic SINET[20], and is associated with carcinoid heart disease [21]. Indeed, the current guidelines for NET management recommend the use of urinary 5-HIAA for SINET diagnosis, and the measurement of CgA and NSE for the diagnosis of PNET [4,5]. The current study provides data to support these recommendations of using specific NET biomarkers based on the primary tumor site and that elevated values of urinary 5-HIAA for SINET often indicates the presence of metastatic disease which may not be seen by conventional anatomic imaging studies.

Somatostatin receptor based PET/CT is a very sensitive imaging modality to detect NETs [10,22,23]. ⁶⁸Ga-labeled somatostatin analogue PET/CT is more sensitive for NET localization[5,10], is faster, and requires lower radiation dose[24] compared with Octreoscan. Hence, in the current study, disease extent was assessed by quantification of whole body ⁶⁸Ga-DOTATATE pathological uptake. Thus, we believe our analysis of the levels of a comprehensive panel of NET biomarkers and NET disease burden more accurately reflects the diagnostic and surveillance clinical utility of biomarkers in patients with NET.

Total ⁶⁸Ga-DOTATATE uptake in NETs was reported in the past in a series of patients, but not in specific subtypes of NET, and no correlation analysis with tumor biomarkers was reported [12,13]. Moreover, the correlations between tumor biomarkers and tumor burden reported in the past have been based on anatomical imaging[19], counting the number of lesions[10], presence/absence of liver metastases[25] and not based on accurate total tumor volume measurements, as performed in the current analysis. The strong correlation between 24-hour urinary 5-HIAA levels and ⁶⁸Ga-DOTATATE TV in SINET is not surprising[20], and might explain the strong correlation between total volume measurements and 5-HIAA among patients with NEToUP, as these tumors arise mainly from the small intestine but are not detected[26].

Our findings have important clinical implications. First, baseline 5-HIAA levels ≥8.1 mg/24h should prompt evaluation for metastatic disease in patients with SINET. Second, among patients with NET associated with hereditary syndromes, PP may be used as a biomarker for PNETs, as well as, glucagon and VIP in patients MEN-1 and VHL, respectively.

Our analysis has some limitations. First, the vast majority of the patients in the current study had gastro-entero-pancreatic (GEP) NETs, limiting the ability to make strong conclusions about biomarkers utility among patients with primary tumors at other sites. Second, although the analysis was performed by an experienced nuclear medicine specialist that was blinded to the clinical context, interobserver variation was not tested for in the current study, and should be evaluated in future studies. Third, plasma CgA levels may be elevated in patients with atrophic gastritis and can lead to positive results and every patient in the current study did not have endoscopy and biopsy to determine the presence of this. Finally, the biochemical biomarkers in the current analysis were performed using a single method per biomarker. Since there might be a difference in the measured levels between different analytical methods, the study results may only apply to the methods used in our study and other analytical methods would need to be studied in the future.

In conclusion, our data supports the use of specific NET markers based on the site of the primary NET, and the presence of hereditary syndrome associated NET. High levels of 24-hour urinary 5-HIAA commonly indicate the presence of metastatic disease in patients with SINET. The heterogeneity of patients with NET necessitates careful matching of biochemical biomarker for each patient according to both tumor and patient characteristics, and might justify personalized biomarker panel evaluation for tumor burden assessment.