Approximately half of neuroendocrine tumors (NETs) are metastatic at the time of diagnosis,1 but in up to 15% of cases, the anatomic subtype (eg, site of origin) is unknown even after a diagnostic workup.1-4 Accurate identification of the anatomic subtype is important for NET tumor staging, direct monitoring of hormone markers, and selection of the optimal surgical approach and medical treatment. Indeed, the 2017 National Comprehensive Cancer Network (NCCN) Clinical Practice Guidelines in Oncology: Neuroendocrine Tumors recommend different therapies depending on NET anatomic subtype, including somatostatin analogs for well-differentiated midgut and pancreatic tumors, peptide receptor radionuclide therapy for well-differentiated gastric NETs, everolimus for well-differentiated gastroendopancreatic (GEP) and lung NETs, pan-tyrosine kinase inhibitors for pancreatic NETs, and immunotherapy for small-cell NETs.5 Where the anatomic subtype of the primary tumor is not known, NCCN Guidelines recommend intensive tumor-directed localizing studies, including computed tomography (CT), somatostatin receptor (SSR)-based scintigraphy or 68Ga dotatate positron emission tomography (PET)/CT, fluorodeoxyglucose PET/CT, and invasive tests such as endogastroduodenoscopy or colonoscopy.5
What are the best approaches for anatomic characterization of NETs of unknown primary, and how can the current NCCN-recommended standards of care be improved? Characterization of the NET subtype should leverage multimodal approaches, including clinical presentation, pathology evaluation, imaging modalities, biochemical characterization, and gene-expression profiling for NET subtyping, with the sequence of testing dependent on the clinical presentation. A working group consisting of Drs Lowell B. Anthony, Andrew E. Hendifar, Eric Liu, and Robert Ramirez has reviewed the best technologies and proposed an integrated multimodal algorithm for accurate NET subtyping. They concluded that although biochemical testing can inform the NET subtype in well-differentiated NETs when correlated with the clinical presentation, it has questionable utility in patients with poorly differentiated NETs that are unlikely to secrete hormones. Moreover, with respect to pathologic workup, a limited number of immunohistochemistry (IHC) stains should be attempted on chromogranin A- and/or synaptophysin-positive tumors to identify the subtype, but alternative methods should be considered to identify the NET subtype before trying numerous IHC panels that may exhaust the biospecimen.
Regarding diagnostic imaging, 68Ga-PET/CT should be considered to identify the anatomic subtype in cases where CT, magnetic resonance imaging, and OctreoScan are unable to localize the tumor. However, there remains an unmet need for improved imaging modalities in NET subtyping, because 68Ga-PET/CT did not identify the correct subtype in validation studies where the NET subtype was known by other imaging or pathology, functional imaging is not useful in patients with SSR-negative and poorly differentiated NETs, and patients living in a community setting have limited access to 68Ga-PET/CT.
The best technology for gene-expression profiling was thought to be the 92-gene assay (CancerTYPE ID); a 92-gene panel is composed of transcription factors and signal transduction molecules that work together with an algorithm for tumor type and subtype classification.6 The assay works by comparing the expression of 92 genes to a reference database of gene expression from >2000 known tumors, identifies 28 main tumor types and 50 tumor subtypes using RNA extracted from tissue, and has been validated in a multi-institutional study demonstrating 87% sensitivity and >99% specificity.7 A prospective, single-arm clinical trial demonstrated a 37% increase in overall survival for patients with carcinoma of unknown primary who received assay-directed therapy.8
A proposed algorithm for molecular tumor classification (MTC) of NETs in cases of unknown or uncertain diagnosis was proposed and discussed. In brief, the algorithm suggests that in cases where imaging is negative, the patient may not be able to gain access to 68Ga-DOTA-PET, or if the tumor is known to be SSR-negative, MTC provides a tissue-sparing approach to get information on NET subtype. Perhaps most important is that additional staining beyond first tier (eg, chromogranin A, synaptophysin) may be unwarranted if the biopsy is small. In cases where only cytology or limited tissue samples (eg, core needle) can be obtained, where pathology cannot determine the anatomic subtype for well-differentiated NETs, and in cases with a high-grade tumor that is metastatic, go directly to MTC to identify the tumor type and subtype.
- Yao JC, et al. J Clin Oncol. 2008;26:3063-3072.
- Sadowski SM, et al. J Clin Oncol. 2016;34:588-596.
- Zuetenhorst JM, Taal BG. Oncologist. 2005;10:123-131.
- Srivastava A, et al. Am J Surg Pathol. 2009;33:626-632.
- NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines®). Neuroendocrine Tumors. Version 1.2017.
- Ma XJ, et al. Arch Pathol Lab Med. 2006;130:465-473.
- Kerr SE, et al. Clin Cancer Res. 2012;18:3952-3960.
- Hainsworth JD, et al. J Clin Oncol. 2013;31:217-223.