Nuclear Imaging of Neuroendocrine Tumors

Somatostatin receptor (SSTR) imaging has been of interest since the 1980s when somatostatin receptor overexpression specifically subtype 2 on neuroendocrine tumor (NET) cells was identified.(1) SSTR imaging has been in clinical use for over 20 years since the initial approval of ¹¹¹In-pentetreotide (OctreoScan®) for whole body planar and SPECT/CT in 1994. Recently ¹¹¹In-pentetreotide (OctreoScan®) has been largely superseded by the ⁶⁸Ga-labeled somatostatin receptor analogues (⁶⁸Ga-DOTA-SSTR) PET/CT agents. ⁶⁸Ga-DOTATATE was approved in 2016 and ⁶⁸Ga-DOTATOC has been approved recently in August 2019. In addition to the higher resolution offered by PET over SPECT, ⁶⁸Ga-DOTA-SSTR agents have significantly higher affinity for SSTR subtype 2 compared to ¹¹¹In-DTPA-Octreotide, which results in significantly higher detection rates of NETs with ⁶⁸Ga-DOTA-SSTR compared to ¹¹¹In-pentetreotide (2, 3). A prospective intra-patient comparison of ⁶⁸Ga-DOTATATE and ¹¹¹In-DTPA-Octreotide in 130 patients with known or suspected NET demonstrated a sensitivity of 95.1 % for ⁶⁸Ga-DOTATATE versus 30.9% for ¹¹¹I pentetreotide (4). The radiation dose is also significantly lower from ⁶⁸Ga-labeled DOTA-SSTR agents compared to ¹¹¹In-DTPA-Octreotide, with an adult dose of 4.3–4.8 mSv for 185 MBq of ⁶⁸Ga-DOTA-SSTR agents versus 17.7 mSv for a typical 222 MBq administration of ¹¹¹In-DTPA-Octreotide (5–7). ⁶⁸Ga-DOTA-SSTR PET/CT is best suited for well-differentiated (Grade 1 and Grade 2) gastroenteropancreatic neuroendocrine tumors (GEP-NET) given their high expression of SSTRs, while ¹⁸F-fluorodeoxyglucose (FDG) PET is better suited for imaging less well-differentiated and poorly differentiated (high Grade2 and Grade 3) GEP-NETs due to higher glucose metabolism and relatively lower or absent SSTRs.(8, 9)

Guidelines for the appropriate clinical use of ⁶⁸Ga-DOTA-SSTRs have been published. Appropriate indications for SSTR PET imaging include: 1) initial staging after histological confirmation of NET; 2) identification of unknown primary tumor in patients with metastatic NET or elevation of certain serum biomarkers when a primary tumor cannot be identified on conventional imaging; 3) accurate staging prior to planned surgery; 4) detection, characterization, and staging in the setting of disease recurrence; 5) identification of patients who are likely to benefit from peptide receptor radiotherapy (PRRT); 6) confirming the presence of SSTRs (presumptive diagnosis) in patients with anatomic lesions suspicious for NETs on conventional imaging or lesions that cannot undergo biopsy.(8–11)

⁶⁸Ga-DOTA-SSTR PET/CT in diagnosis of GEP-NETs

Definitive diagnosis of GEP-NETs is based on histology, as imaging and circulating biomarkers alone are inadequate for diagnosis. Imaging, however, plays a critical role in detection and localization of primary and metastatic GEP-NETs for histologic diagnosis. The sensitivity, specificity, and accuracy of ⁶⁸Ga-DOTA-SSTRs have been studied. Four meta-analyses evaluating diagnostic performance of the ⁶⁸Ga-DOTA-SSTRs in NETs (all types) that have been published within the last 10 years are summarized here. All of these meta-analyses suggest a very high level of sensitivity and specificity. Yang et al. performed a meta-analysis of 10 individual studies (n=415 patients) evaluating the diagnostic performance of ⁶⁸Ga-DOTATOC and ⁶⁸Ga-DOTATATE PET and with results based on histology or histology plus follow-up. The pooled sensitivity and specificity in the diagnosis of NETs on a per patient basis was 93% and 85% for ⁶⁸Ga-DOTATOC, and 96% and 100% for ⁶⁸Ga-DOTATATE. The ROC for ⁶⁸Ga-DOTATOC and ⁶⁸Ga-DOTATATE was 0.96 and 0.98, respectively. (12) Geijer et al. performed a larger meta-analysis of 22 individual studies (n=2105 patients) each evaluating one of the three ⁶⁸Ga-DOTA-SSTR radiotracers and a variety of NETs (such as GEP, lung, carcinoma with unknown primary [CUP], thymus). Results were based on histology and/or imaging follow-up. The pooled sensitivity and specificity of ⁶⁸Ga-DOTA-SSTR was 93% and 96% with area under the ROC 0.98.(13) Treglia et al. evaluated 16 studies each evaluating one of the three ⁶⁸Ga-DOTA-SSTR (n=567 patients) including a variety of NET (such as GEP, lung, CUP, thymus, and others). Results were based on either histology or clinical/imaging follow-up. Pooled sensitivity and specificity was 93% and 91% with area under the ROC 0.96.(14) Johnbeck et al. reviewed the small number of head-to-head studies comparing one ⁶⁸Ga-DOTA-SSTRs to another, as well as 21 other studies each focusing on an individual ⁶⁸Ga-DOTA-SSTR. They found relatively similar performance among the radiotracers, such that overall no one radiotracer could be advanced as outperforming the others. Summarizing their findings, ⁶⁸Ga-DOTATOC showed sensitivity 92–100% and specificity 83–92%; ⁶⁸Ga-DOTATATE showed sensitivity 72–96% with specificity mostly unreported except for one study where it was reported at 100%; and ⁶⁸Ga-DOTANOC showed sensitivity 68–100% and specificity 93–100%.(15).

Another situation in which ⁶⁸Ga-DOTA-SSTR PET may be used is suspected NET in the setting of symptoms (e.g. flushing and diarrhea) and abnormal elevation of biomarkers suggestive of an underlying NET when a lesion cannot be identified on conventional imaging.(9) The yield for this modality in this patient group is low. Graham et al. reviewed several papers reporting on ⁶⁸Ga-DOTATOC PET for this indication. Pooled results included 57 patients with 7 true positive lesions (12%) and 1 false positive.(16) The yield is higher in patients with positive conventional imaging(17). In spite of the low yield this indication for ⁶⁸Ga-DOTA-SSTR PET is considered appropriate by consensus guidelines, as the results would substantially alter patient management.(9)

Despite its high specificity there are several interpretive pitfalls with ⁶⁸Ga-DOTA-SSTR PET/CT imaging. Normal uptake of ⁶⁸Ga-DOTA-SSTR is seen in the pancreatic uncinate process and occasionally in the pancreatic tail and therefore pancreatic findings in these areas should be generally correlated with cross sectional imaging.(18) Areas of inflammation or infection and reactive lymphadenopathy, recent surgery and radiotherapy, and granulomatous lesions may also show increased uptake of ⁶⁸Ga-DOTA-SSTR due to somatostatin receptor expression on activated lymphocytes and macrophages.(19) Benign bone lesions such as vertebral hemangiomas, degenerative changes, and fractures also have been reported to show increased uptake on ⁶⁸Ga-DOTA-SSTR PET/CT scan.(20)

Multiple neoplasms besides NETs are known to express SSTRs and show variable ⁶⁸Ga-DOTA-SSTR uptake. These include meningioma and multiple other types of brain tumors, breast carcinoma, melanoma, lymphoma, prostate carcinoma, non-small cell lung cancer, esophageal cancer, colorectal cancer, head and neck cancer, sarcoma, renal cell carcinoma, ovarian cancer, testicular cancer, and differentiated thyroid carcinoma.(8, 21)

⁶⁸Ga-DOTA PET/CT in Staging of GEP-NETs

Complete and accurate staging of GEP-NETs is important for therapeutic considerations. Five-year survival is directly related to the localization of disease and development of metastases: 96% (localized), 77% (nodal), 73% (liver), and 50% (extrahepatic metastases).(22) Surgery is the only curative treatment for NETs. Patients with Stage 1-III NET undergo small bowel resection and lymphadenectomy, however, most patients (75%) have metastatic liver disease at presentation.(23) Patients with Stage IV NET may be eligible to undergo additional debulking of hepatic and peritoneal metastases depending on the extent of disease (24).

Conventional imaging with contrast enhanced CT or MRI is the first line staging modality for NETs. Advantages of CT are its wide availability, short scan time, and capability of scanning extended fields of view. CT is relatively low sensitivity for detection of bone metastases, because there may be no accompanying osteosclerosis or osteolysis. Liver/pancreas MRI demonstrates higher rates of detection for liver metastases and pancreatic primary lesions as compared to CT. Detection of bone metastases by MRI is also superior to CT. However, due to the time intensive nature of this scanning modality and its small field of view, MRI is used more often as a tool for evaluation of a specific body part rather than whole body skeletal staging. Whole body protocols have been developed, but to achieve the necessary extended field of view within an acceptable time frame, scan time is shortened at the expense of image detail. Both CT and MRI are limited in their detection of small nodal metastases based on short axis size criteria. MRI is also less sensitive for lung lesions.(25)

Some of the shortcomings of conventional imaging for staging NET may be addressed with whole body PET. Multiple studies have demonstrated the superiority of ⁶⁸Ga-DOTA-SSTR PET over conventional imaging in staging of NETs. Figure 1 demonstrates a patient with NET with widespread metastases in liver, bone, and lungs; the extended field of view and sensitivity of the ⁶⁸Ga-DOTATATE PET/CT are essential to identifying all sites of metastatic tumor.

Figure 1.

Whole body ⁶⁸Ga-DOTATATE PET/CT in a patient with NET of unknown primary. Maximum intensity projection image (A) shows numerous liver and bone metastases as well as bronchopulmonary involvement (black arrow). In addition, the fused coronal PET/CT image (B) shows the primary pancreatic tumor (white arrow).

For detection of neuroendocrine tumor in the liver, bowel, lymph node and bones, Sadowski et al. reported a detection rate between 94–97% for ⁶⁸Ga-DOTATATE, all of which were significantly better compared to ¹¹¹In- pentetreotide and CT scan (4). Frilling et al. were able to identify larger numbers of liver and nodal metastases in a group of NET patients with ⁶⁸Ga-DOTATOC PET/CT as compared to conventional imaging with CT or MRI.(26) Albanus et al. reported a higher sensitivity and specificity on a per patient basis for ⁶⁸Ga-DOTATATE PET plus contrast enhanced CT (ceCT) versus ceCT alone with bone lesions (sensitivity 100% versus 47%; specificity 89% versus 49%) and nodal lesions (sensitivity 92% versus 64%; specificity 83% versus 59%). For lung lesions both PET/ceCT and ceCT showed equal sensitivity (100%) and slightly better specificity with PET/ceCT (95% versus 82%).(27) Abrosini et al. reported higher sensitivity (100% versus 80%) and similar specificity (100% versus 98%) for bone metastases on a per patient basis for ⁶⁸Ga-DOTANOC PET/CT versus CT.(28) Whole body PET has also proven useful in uncovering sites of NET metastasis beyond liver, bone, and lymph nodes including atypical sites such as cardiac and breast metastases.(29)

An important question to ask in assessing performance of ⁶⁸Ga-DOTA-SSTR PET/CT in diagnosis and staging compared to conventional imaging, is whether detection of more lesions is clinically meaningful, that is, does it affect patient care. Several studies have assessed clinical impact of this imaging modality. Barrio et al. performed a meta-analysis of studies dealing with change in management for ⁶⁸Ga-DOTATATE, -TOC, and -NOC, with the conclusion that change in management was seen in 44% of patients undergoing the ⁶⁸Ga-DOTA-SSTR PET/CT. Based on five studies evaluating change in management in patients in whom ¹¹¹In-pentetreotide (OctreoScan®) SPECT/CT was previously done, ⁶⁸Ga-DOTA-SSTR PET/CT resulted in change in management an average of 39% of the time (range of 16–71%). (30) Recognizing that change in management is not the same thing as implementation of those changes, Calais et al. examined the rate of clinical follow-through. Their prospective study of ⁶⁸Ga-DOTATATE PET/CT (n=96 patients) showed an intended change in management in 50% of cases and a high implementation rate (75%).(31)

In summary, ⁶⁸Ga-DOTA-SSTR PET/CT is clearly superior to conventional imaging and SSTR scintigraphy for detection of hepatic and extrahepatic metastases. Therefore ⁶⁸Ga-DOTA-SSTR PET/CT plays an important role in localizing all sites of disease for the purposes of guiding the appropriate course of treatment, and its use for this indication is supported by published guidelines.

⁶⁸Ga-DOTA-SSTR PET/CT in Identification of Unknown Primary NET in Patients Presenting with Metastases

Between 9–19% of neuroendocrine tumor patients present with metastatic disease with an unknown primary tumor site. (32) Localization of the primary tumor is important in the management of this patient population because resection of the primary tumor and metastases significantly improves survival and is the treatment goal for patients with well-differentiated neuroendocrine tumor metastases. (33) Even if the disease is not completely resectable, debulking surgery can significantly improve symptom control in patients with severe endocrine symptoms. (33) In patients with metastatic liver disease from midgut NET, resection of the primary tumor was found to be an independent predictor of survival with a median survival of 9.92 years for patients who underwent resection of the primary NET versus 4.68 years for patients with no resection of the primary tumor.(34) The benefit of resection of the primary NET was also demonstrated in pancreatic NETs with metastases, with a median survival of 5.42 years for patients who had resection of the nonfunctional pancreatic primary lesion versus 0.83 years for patients without resection.(35)

Several studies have demonstrated the utility of ⁶⁸Ga-DOTA-SSTR PET in localization of the site of unknown primary NET in patients with liver metastasis. In a prospective study of patients with negative conventional imaging, Menda et al. found ⁶⁸Ga-DOTATOC PET/CT detected 38% of confirmed unknown primary tumors with a false positive rate of 7%. (36) In another study by Schreiter et al., 33 patients with metastatic NET underwent ⁶⁸Ga-DOTATOC PET/CT with a detection rate of 46% of the unknown primary tumor site. (37) Alonso et al. reported a detection rate of 59% of primary tumor in 29 patients with metastatic NETs, with a change in management in 24% of patients. (38) ⁶⁸Ga-DOTA-SSTR PET/CT has been recommended in this patient population as treatment options vary depending on the origin of the tumor.(9) Figure 1 and 2 show patients in whom ⁶⁸Ga-DOTA-SSTR PET/CT was helpful for identification of the primary NET lesion. Figure 1 shows a patient with widespread metastatic disease with primary lesion identified in the pancreas. Figure 2 shows a patient with liver metastases with a small primary tumor found in the distal small bowel with mesenteric nodal metastases.

Figure 2.

Whole body ⁶⁸Ga-DOTATOC PET/CT in a patient with NET of unknown primary. Maximum intensity projection image (A) and fused transaxial PET/CT images (B-D) show multiple liver metastases and foci of extrahepatic uptake in the abdomen corresponding to a primary tumor in the distal small bowel (dashed arrow) and calcified mesenteric nodal metastases (solid arrow).

⁶⁸Ga-DOTA-SSTR PET/CT in detection of recurrence and restaging

⁶⁸Ga-DOTA-SSTR PET/CT is highly sensitive in detection of recurrent NETs (39, 40) Merola et al. have followed 143 patients with metastatic NETs with CT every 6 month and somatostatin receptor imaging every 12 months. In this multicenter retrospective study, addition of somatostatin receptor imaging significantly improved the sensitivity for detection of recurrence with a change in management observed in 73% of patients as a result of somatostatin receptor imaging. (39) Haug et al. also have reported a high accuracy for ⁶⁸Ga-DOTATATE PET/CT in the setting of recurrent NET. Their series of 63 patients consisted of individuals undergoing scans for routine surveillance (n=30), increased serum tumor markers (n=27), or clinical suspicion of recurrence (n=6). Diagnosis was based on histopathologic confirmation (n=25) and clinical follow-up (n=38). Overall sensitivity and specificity for the modality in NETs was 90% and 82%, respectively.(40)

Despite these promising results for somatostatin receptor PET imaging in detection of recurrent NETs, there is no consensus on precise interval and duration of follow-up. Singh et al. have reviewed and summarized guidelines from five different international organizations.(41) Groups based in the United States (such as National Comprehensive Cancer Network [NCCN] and North American Neuroendocrine Tumor Society [NANETS]) recommend interval surveillance imaging to be done with conventional imaging (CT or MRI), but recommend against routine interval surveillance with ⁶⁸Ga-DOTA-SSTR PET/CT.(8) (40, 42). Somatostatin receptor PET/CT is not recommended routinely for restaging after curative surgery until evidence of biochemical or radiological progression because it is unlikely to change management. Somatostatin receptor PET/CT is suggested for restaging at time of clinical or biochemical progression and for monitoring disease that is seen predominantly on somatostatin receptor PET/CT(9).

⁶⁸Ga-DOTA-SSTR PET/CT to Determine Eligibility for Peptide Receptor Radionuclide Therapy (PRRT)

Somatostatin receptor based PRRT with ¹⁷⁷Lu-DOTATATE is indicated for treatment of metastatic and unresectable GEP-NETs and has been shown to significantly improve progression free survival and overall survival compared to octreotide treatment. (43) Ideal candidates for PRRT should exhibit high somatostatin receptor expression, which is assessed with somatostatin receptor imaging. Lesions with higher uptake (maximum standardized uptake value [SUV_max] of 16.4 or higher) of ⁶⁸Ga-DOTATOC PET/CT have been shown to have more favorable response to PRRT. (44) In another study, a high proportion of patients responded to ¹⁷⁷Lu- DOTATATE chemoradionuclide therapy with concomitant 5-FU chemotherapy when selection criteria included tumor uptake of ⁶⁸Ga-DOTA-SSTR greater than liver.(45)

For ⁶⁸Ga-DOTA-SSTR PET, a 3-point qualitative assessment scoring has been proposed to measure SSTR-expression based on ⁶⁸Ga-DOTA-SSTR PET/CT imaging, with Level 3 uptake indicating tumor uptake higher than normal liver uptake, Level 2 uptake describing uptake intensity between blood pool and liver and Level 1 uptake assigned to lesions with uptake equal or less than blood pool. (46) Patients with NET lesions with Level 3 uptake are considered candidates for PRRT(46).

⁶⁸Ga-DOTA-SSTR PET/CT in monitoring response to therapy

Conventional imaging techniques that focus on tumor size based on response evaluation criteria in solid tumors (RECIST) and Southwest Oncology Group (SWOG) criteria risk underestimating response to treatment in GEP-NETs. GEP-NETs tend to show modest morphologic change in response to pharmacologic therapy or PRRT, even in those patients in whom these therapies improve symptoms and progression free survival.(47, 48) Haug et al. evaluated the change in tumor-to-spleen SUV (SUV_T/S) in 33 patients at baseline and after a single cycle of PRRT. They found that the 23 patients in whom SUV_T/S decreased any amount after the first cycle had longer progression-free survival than the 8 patients in whom SUV_T/S was stable or increased after the first cycle. They noted that SUV_max alone did not demonstrate the same predictive value.(47) Gabriel et al. compared ⁶⁸Ga-DOTATOC PET/CT with conventional imaging in the monitoring effects of therapy. They found that PET was useful in identifying sites of early progression based on the appearance of new lesions, as some lesions that were inconspicuous on CT were readily identifiable on PET. As a quantitative imaging technique, however, the intrinsic tumor activity on PET as measured by SUV_max at baseline or during therapy was not useful in predicting patient outcome.(49) This may have been due to inherent heterogeneity of tumor avidity both at baseline and follow-up and other factors affecting repeatability and reproducibility of the radiotracer. In the setting of non-radiologic targeted therapies such as the tyrosine kinase inhibitor everolimus and the mTOR inhibitor sunitinib data on response to therapy has been assessed with conventional imaging using lesion size, but not with receptor-based imaging such as ⁶⁸Ga-DOTA-SSTR PET/CT.(50)

At present there is a need to identify optimal methods for monitoring of response to treatment in NETs. ⁶⁸Ga-DOTA-SSTR PET/CT shows utility in detection of disease progression based on development of new lesions, however, change in intrinsic tumor activity based on SUV measures has not yet shown reliability in monitoring response to therapy. Non-imaging biomarkers based on the circulating mRNA transcriptome (e.g. the liquid biopsy NETtest) may be useful as a method of response assessment. The NET liquid biopsy has shown utility in predicting response to somatostatin analogs, and when integrated with tumor grade information, also predicts and monitors PRRT efficacy.(51)

⁶⁸Ga DOTA-SSTR PET/MRI in NETs

Currently ⁶⁸Ga-DOTA-SSTR PET/MRI in NETs is not widely available for clinical use. MRI exhibits better tissue contrast than CT for organs such as liver, pancreas, bone, and brain. MRI is particularly helpful in detecting liver metastases and given the tendency of NETs to metastasize to the liver, combination of ⁶⁸Ga-DOTA-SSTR PET with MRI may be useful. However, at this time there is no data to suggest superiority of ⁶⁸Ga-DOTA-SSTR PET/MRI versus ⁶⁸Ga-DOTA-SSTR PET/CT in regard to disease detection. Two prospective studies comparing ⁶⁸Ga-DOTATOC PET/MRI (n=8) and ⁶⁸Ga-DOTANOC PET/MRI (n=28) to PET/CT with the same radiotracers showed comparable performance between the PET/MRI and PET/CT in terms of lesion detection.(52, 53) Whether the PET/MRI modality will play a role in prediction of response to therapy and assessment of treatment response is an ongoing topic of research interest.(48)

Challenges of PET/MRI that slow its adoption include length of time for acquisition of multiple MRI pulse sequences (which can add up to an hour or more), attenuation correction, claustrophobia from longer narrower gantry, lack of familiarity with MRI acquisition and interpretation, issues with reimbursement, and lack of consensus about where PET/MRI adds value.(54) MRI is also limited in its evaluation of lung tissue. Sawicki et al. observed that ¹⁸F-FDG PET/MRI missed more pulmonary nodules <5–10 mm compared to PET/CT in a series of oncologic patients. Although most of these missed nodules turned out to be benign some relevant metastatic lesions were missed, suggesting that PET/MRI with current available techniques may be problematic for accurate staging.(55) Conventional MRI is the most sensitive technique for identification of liver metastases.(48) These scans are performed with multiphase contrast enhancement, which some imagers also advocate for PET/MRI for these patients.(56) Given that it is the MRI rather than PET that contributes the most to the long imaging times of PET/MRI, there are pressures to minimize MRI pulse sequences to shorten the overall time of imaging acquisition. To that end, Seith et al. reported that eliminating the IV contrast and performing nonenhanced ⁶⁸Ga-DOTA-SSTR PET/MRI comprised of just three MRI sequences (T2 Half Fourier Acquisition Single Shot Turbo Spin Echo [T2 HASTE], T2-weighted fast spin-echo sequence [T2 TSE], and diffusion weighted imaging [DWI]) resulted in a shorter acquisition time (35 minutes) and comparable performance in lesion detection as compared to PET/CT.(57)

Summary.

• Multiple consensus guidelines acknowledge the role of ⁶⁸Ga-DOTA-SSTR PET at multiple points along the management path of NET patients.

• ⁶⁸Ga-DOTA-SSTR PET demonstrates superior performance to conventional imaging in the setting of initial detection, staging, detection of recurrent tumor, and detection of unknown primary in the setting of known metastatic disease.

• ⁶⁸Ga-DOTA-SSTR PET has a low yield in diagnosis of NETs in the setting of symptoms and biochemical suspicion of NET. However, it may still play a role in guiding management in this patient population.

• Familiarity with normal distribution and causes of false positives on ⁶⁸Ga-DOTA-SSTR PET, such as inflammatory conditions and other neoplasms, are essential for optimal imaging.

• The role of ⁶⁸Ga-DOTA-SSTR PET/CT is not established in monitoring response to systemic therapy but may identify suspected progression through detection of new sites of disease.