DOTATATE PET Imaging in Olfactory Neuroblastoma and Association with SSTR Expression

Abstract

Introduction  Olfactory neuroblastoma (ONB), or esthesioneuroblastoma, is a rare neuroectodermal tumor of the nasal cavity and paranasal sinuses. Most of these tumors express somatostatin receptors (SSTRs), providing a potential target for radionuclide imaging with Ga-68 DOTATATE. However, this imaging modality has not been extensively studied in ONB.

Methods  We conducted a retrospective chart review of 96 endoscopic endonasal skull base surgery cases for ONB performed at our institution between 2000 and 2021. Histo (H) scores were assigned to each tumor and normalized DOTATATE standardized uptake values (nSUVs) were measured as well.

Results  Nine patients (5 males and 4 females) with ONB were ultimately included in the study. The average age of the patients was 50 years. All ONBs had a positive SSTR2 expression (H-score > 105; mean: 180). All ONBs showed DOTATATE avidity (mean nSUV for ONB: 6.7). However, there was no correlation between H-score and nSUV, with an r ² of 0.24 ( p  = 0.18).

Conclusion  Our study shows that SSTR2 expression is found in all ONBs with associated DOTATATE avidity, which may serve as a valuable imaging modality to monitor for recurrent and metastatic disease in ONB.

Introduction

Olfactory neuroblastoma (ONB), or esthesioneuroblastoma, is a rare neuroectodermal tumor of the nasal cavity and paranasal sinuses. The incidence of ONB is around 0.4 per million, representing a small fraction of all sinonasal malignancies (around 2%). ¹ ONB usually arises from specialized olfactory neuroepithelium within the superior nasal vault. ² It presents with nonspecific clinical findings including nasal obstruction, epistaxis, and anosmia. Once diagnosis is confirmed, ONB is graded via the Hyams histologic grading system on a range from I to IV. ³ This grading system accounts for several features including mitotic activity, nuclear pleomorphism, and necrosis. Generally, higher-grade tumors are associated with worse prognosis. ⁴

While the optimal treatment paradigm has not been established, these tumors are usually treated with a combination of surgery, radiation, and chemotherapy. Survival and recurrence rates can vary widely, with some patients remaining disease free for over 10 years. ⁵ While only around 10% of patients initially present with nodal disease, over 40% of patients with ONB will develop cervical metastases after local treatment. ^{6 7} This underlines the need for prolonged monitoring of recurrence in this tumor type.

There is no standardized disease surveillance protocol for ONB, but magnetic resonance imaging (MRI) can be employed to monitor for recurrence. However, MRI and fluorodeoxyglucose positron emission tomography (FDG-PET) do not always detect cervical lymph node metastases. ⁸ This underlines the need for a novel imaging modality to monitor for both recurrence and metastases.

Somatostatin receptor 2 (SSTR2) is a membrane-bound G-protein coupled receptor expressed in several types of neural and neuroendocrine cells. ⁹ SSTR2 is present in several neuroendocrine neoplasms, including pancreatic and small bowel malignancy. ¹⁰ Most ONBs show diffuse and moderate SSTR2 expression, with one recent study from our institution showing expression in 99% of ONBs. ¹¹ This provides a potential target for imaging, as mimics of ONB tend not to express SSTR2. ⁹ Several studies have utilized DOTATATE, a somatostatin analog that can be conjugated to a radionuclide, as a targeted imaging modality. ^{12 13 14} When conjugated to gallium, this is known as ⁶⁸ Ga-DOTATATE. Its specificity and efficacy have already been established in SSTR-expressing gastrointestinal neuroendocrine tumors, but its use in ONB is still being investigated. ¹⁵ A recent study performed DOTATATE imaging on three patients with histologically confirmed ONB, all of whom had a positive SSTR2 expression with corresponding DOTATATE uptake. ¹⁶ However, the exact correlation between SSTR2 expression and DOTATATE uptake remains unclear. The objective of this study is to evaluate the use of DOTATATE imaging in ONB and to correlate SSTR2 expression and DOTATATE uptake.

Methods

Patient Selection

With Institutional Review Board approval (no. 20040320), a retrospective chart review was conducted at the University of Pittsburgh Medical Center (UPMC) for ONB cases that underwent endoscopic endonasal skull base surgery (EESBS) between January 2000 and February 2021. This ultimately yielded 96 cases. We only included patients who had received DOTATATE imaging at UPMC, which started in 2018. Only patients who had pathology specimens available for review at our institution were selected. One patient was reclassified as olfactory carcinoma (OC) and excluded from analysis. This yielded nine patients for a detailed review ( Fig. 1 ).

Fig. 1.

Patient selection criteria.

Data Extraction

For our cohort of nine patients, we extracted the following information: demographics, physical examination findings from initial patient visits, tumor location based on radiologic findings, margin status following surgery, and metastasis/recurrence information. For DOTATATE imaging, we extracted radioactive isotope type and dose, scanning modality (combined PET-MRI vs. PET-computed tomography [PET-CT]), as well as the total number of scans and their dates. The Kadish staging and Modified Hyams grading were assigned as previously described. ¹¹

Radiology

DOTATATE imaging was reviewed by a radiologist at UPMC. For each available DOTATATE scan, maximum standardized uptake values (SUVmax) within the radio-avid lesions and scalp SUVmax were calculated. Scalp SUVmax served as an internal control for each scan. If a patient had several DOTATATE scans available demonstrating radio-avid lesions, the SUVmax values were collected for each scan. For eventual analysis, we utilized SUV values from the scan closest to the date of surgery. Normalized SUV ratio (nSUV) was calculated as the ratio of SUVmax (tumor) to SUVmax (scalp).

Pathology

Surgical pathology slides were reviewed by a head and neck pathologist at UPMC. The cases were reviewed by a pathologist (RRS) to confirm diagnosis and modified Hyams grade as previously described. ¹¹ Only cases with 2 mm ² or more on slides were assigned a grade. SSTR2 immunostaining and Histo (H) scoring were performed according to established protocol (Clone: UMB1 [monoclonal rabbit] 1:100 dilution, AbCam, Toronto, ON, Canada). ¹¹ Briefly, H-score is calculated with the following formula: 1 × % weakly stained cells + 2 × % moderately stained cells + 3 × % strongly stained cells. H-score >105 was considered positive for SSTR2 expression.

Statistical Analysis

Categorical variables were reported as counts and continuous variables were reported as means ± standard deviation. The correlation between H-scores and nSUV was calculated utilizing Pearson's correlation coefficient. Statistical analysis was performed utilizing JMP, Version <16.0.0> , SAS Institute Inc., Cary, NC, 1989–2022.

Results

Table 1 summarizes the clinical, demographic, and pathologic features of the patients. The average age was 50.0 years, with a range of 13.6 to 72.7 years. There were five male and four female patients. Five patients had surgery for recurrent metastatic disease, 2 patients for primary disease, 1 patient for nonrecurrent metastatic disease, and 1 patient for locally recurrent disease. The metastatic sites included cervical lymph nodes ( N  = 4), dura ( N  = 2), parapharyngeal space ( N  = 1), and parotid lymph node ( N  = 1). For diagnostic imaging, eight patients received PET-MRI and one received PET-CT. Six of these patients had their DOTATATE imaging preoperatively.

Table 1. Patient demographics, medical history, imaging type, and pathology information.

Patient total ( N )	9
Age, mean (range)	50.0 (13.6–72.7) y
Sex ( N )
Male	5
Female	4
Imaging ( N )
PET-MRI	8
PET-CT	1
DOTATATE timing ( N )
Pre-op	5
Post-op	4
Disease status ( N )
Metastatic	6
Recurrent	6
Previous radiation therapy ( N )
Yes	6
No	3
Grade
II	6
III	1
Not assigned	2

Abbreviations: CT, computed tomography; MRI, magnetic resonance imaging; PET, positron emission tomography.

Physical examination revealed a visible mass on in-office nasal endoscopy in six patients that were confirmed as ONB. Three of the six patients with a visible mass on examination were otherwise asymptomatic. One patient presented with a neck mass that was found to be a metastatic ONB. Other presenting symptoms in our cohort included congestion ( N  = 2), epistaxis ( N  = 1), and dysosmia ( N  = 1).

The H-scores were available for all patients. All ONBs had a positive SSTR2 expression (H-score > 105; mean: 180) with H-scores ranging from 150 to 270 ( Table 2 ). The H-scores for three patients were derived from the tumor biopsied at initial presentation, rather than the recurrence that was imaged via DOTATATE. All ONBs showed DOTATATE avidity (mean nSUV for ONB: 6.7). There was no correlation between H-score and nSUV ( Fig. 2 ), with an r ² of 0.24 ( p  = 0.18).

Table 2. Lesions with associated H-score and normalized standard uptake value (nSUV).

Patient	H-score	nSUV
1	230	8.5
2	150	11.6
3	160	4.1
4	180	2.52
5	230	19.8
6	210	2.9
7	175	3.5
8	270	13.2
9	190	7.7

Fig. 2.

The H-score is plotted for each lesion against its normalized standardized uptake value (nSUV). The r ² value is 0.24 ( p  = 0.18).

Three patients (patients 1, 8, and 9) had metastatic or recurrent disease detected on DOTATATE that was not detected on other imaging modalities. For patient no. 1, FDG-PET showed mucosal thickening with focal FDG avidity in the right maxillary sinus that was considered postinflammatory. DOTATATE imaging 10 months later showed significant radiotracer uptake in that same location concerning for metastasis. The patient underwent resection for suspected disease, which was histologically confirmed as recurrent ONB.

Nine months after surgery for an ONB recurrence, patient no. 8 received a surveillance MRI orbit/face/neck at an outside facility that showed bilateral internal jugular chain lymph node prominence that did not appear pathologic, with no specific mention of level IIb lymphadenopathy. Four months after the MRI, they received a DOTATATE scan that showed a highly avid level IIb node concerning for disease. This was not confirmed via pathology (the patient was managed nonoperatively), but the node decreased in size/avidity after Lutathera therapy.

For patient no. 9, MRI showed stable cervical lymphadenopathy not suspected to be malignant. Five months later, DOTATATE imaging demonstrated significant avidity within those same nodes ( Fig. 3 ). Surgical pathology confirmed metastatic ONB ( Fig. 4 ).

Fig. 3.

These are DOTATATE and magnetic resonance imaging (MRI) images from patient no. 9. ( A ) Level 1a node is noted by arrow on MRI in the left image, with the corresponding DOTATATE avidity on the right image. ( B ) Retropharyngeal nodes cannot be clearly seen on MRI on the left image but are significantly DOTATATE avid on the right image.

Fig. 4.

( A ) Hematoxylin and eosin (H&E) and ( B ) somatostatin receptor 2 (SSTR2) (B) staining of pathology from patient no. 9. H-score = 190.

Discussion

This study evaluates the use of DOTATATE in ONB, where recurrences can occur years after diagnosis, with one study citing a median time to recurrence of 6.9 years. ¹⁷ There is no established protocol for monitoring these patients, and current imaging modalities may be inadequate to detect a recurrence. The advantage of DOTATATE is its specificity for the SSTR2 receptor, which can potentially identify metastatic disease that cannot be seen on other imaging modalities.

Our study confirms that SSTR2 expression is consistently seen in ONB, both in primary and metastatic disease. However, the SSTR2 H-score does not appear to correlate with DOTATATE SUV in this small series. Aside from small sample size, the basis for this lack of correlation is likely multifactorial. Despite a high level of automation and standardization of immunohistochemical staining, technical variability as a result of preanalytical factors (tissue quality, fixation, amenability to epitope retrieval, etc.) is unavoidable on live patient specimens in contrast to in vitro or xenograft studies. ¹⁸ Additionally, other factors such as limited precision of manual scoring of chromogenic immunohistochemistry and dynamic range limits (for the cases with very high H-scores) may also contribute to a lack of correlation. Furthermore, only one section per case is examined for SSTR2 immunoexpression and may not be entirely representative of the entire tumor volume. While there was no significant numerical correlation between DOTATATE uptake and SSTR2 expression, all lesions showed a degree of DOTATATE uptake and had a positive SSTR2 expression. The recent study by Lechner et al reported three patients treated with peptide-receptor nuclide therapy ¹⁷⁷ Lu-DOTATATE. ¹⁶ Only patients with an SSTR2-positive ONB that also had DOTATATE uptake on imaging were treated. In the study, slides were assigned SSTR2 positivity based on staining intensity and extent, with 82.4% of their cohort positive for SSTR2, but numerical thresholds were not reported. DOTATATE uptake higher than liver but less than kidney (grade 2) or uptake more than kidney (grade 3) was required for treatment inclusion. Their finding that DOTATATE uptake is seen in SSTR2-positive ONB aligns with our results, but larger cohorts may be required to confirm if a numerical correlation exists.

An excluded patient with OC is of particular interest regarding SSTR2 expression and prediction of DOTATATE avidity. This tumor was reclassified as OC, showed a subthreshold H-score (75), and yet still demonstrated significant DOTATATE avidity. However, it must be noted that our H-score threshold was derived from our previous analysis, which justified that the H-score value as the most accurate for supporting a diagnosis of ONB ¹¹ ; the threshold expression required to predict detection on imaging has never been defined in ONB. In a study of nonlung neuroendocrine tumors, Yu et al found that an H-score threshold of 160 maximized both sensitivity and specificity for positive DOTATATE imaging (86.1 and 89.3%, respectively). ¹⁹ Instead of SUV, the authors used a semiquantitative scoring system for DOTATATE uptake defined as “0: none; 1: much lower than that of the liver; 2: slightly less than or equal to that of the liver; 3: greater than that of the liver; and 4: greater than that of the spleen.” They also found a correlation between H-score and SUV in their cohort, but this correlation weakened significantly in heterogeneous tumors when compared with homogeneous tumors. Here, the OC case is clearly below this threshold, indicating that if an H-score threshold is applicable, it will likely be tumor specific. The patient's tumor H-score was derived from a strong but small subpopulation of positive cells rather than weak expression in a larger proportion of cells. Aside from aforementioned limitations, this case illustrates that the relative value of staining intensity and the proportion of cell staining is not necessarily captured by H-score alone.

Most importantly, our study identified three patients with metastatic or recurrent ONB detected on DOTATATE that was not detected on PET/CT or MRI. This would indicate that DOTATATE may have superior sensitivity for metastatic or recurrent disease. Failure to recognize metastasis or recurrence in any malignancy can have significant consequences. Unfortunately, our sample size did not allow any head-to-head analysis comparing DOTATATE sensitivity and specificity to other imaging modalities. Furthermore, the time between DOTATATE imaging and other imaging modalities was variable among three patients. Specifically for patient no. 1, DOTATATE imaging and FDG-PET were separated by 10 months. Although imaging was separated by 10 months, a visible lymph node on prior MRI and the slow growth rate of ONB suggest that metastatic disease was present on the first scan.

Conclusion

Our study shows that SSTR2 expression is common to all ONBs with DOTATATE avidity, which may serve as a valuable imaging modality to monitor for recurrent and metastatic disease in ONBs. Even in our small sample size, we were able to detect ONB with an increased sensitivity compared with MRI or PET/CT. This may aid in improved staging of neck disease at presentation and earlier detection of recurrent or metastatic disease. Prospective studies are warranted to investigate its role in improving clinical outcomes in this patient population.