Divergences in neuroendocrine prostate cancer frequency as recognized by anatomic pathologists, clinicians, and basic scientists

As a practicing anatomic pathologist specialized in urologic pathology, a vast difference may be observed between what pathologists designate as neuroendocrine (or small cell) carcinoma of the prostate, and what clinicians or basic scientists define as such. This opinion will explore some examples of how these criteria differ, in an effort to create a bridge between urologic pathologists, clinicians who conduct clinical trials, and basic scientists with a focus in prostate cancer research. It is not meant to be a comprehensive review of the topic.

An anatomic pathologist uses narrow criteria to call any tumor neuroendocrine (small cell). There are two main types of criteria. First, the morphology (hematoxylin-eosin stain) should feature cells with minimal cytoplasm, inconspicuous cell borders, and small nuclei often with nuclear molding. This molding refers to the nuclear shape being changed by proximity to other cells, so some nuclei have a triangle shape, boat shape, or crescent moon shape due to close apposition to neighboring cells (Figure 1a). Second, strong immunoreactivity with synaptophysin (Figure 1b) and/or chromogranin is required at least in a preponderance of cells. CD56 and neuron-specific enolase are two other common neuroendocrine markers but are considered less reliable for the diagnosis because they are relatively less specific. Another finding can be increased p53 staining due to mutation of p53 in pure small cell cancer,1 although this stain is not often performed on prostate biopsies. Thus, only a fraction of a percentage of diagnosed prostate cancer cases fit these criteria, probably fewer than 1 in 300, even including mixed acinar-neuroendocrine forms which have been defined.2 Morphology plus staining is needed to rule in definite neuroendocrine features, for treatment choice. This is a critical distinction since therapy for small cell cancer is different, which mainly consists of chemotherapy, and is less amenable to surgery.

Figure 1.

Testing of high-grade prostatic carcinoma by immunostains. (a) Prostatic carcinoma with characteristic small cell morphology: absence of gland formation, nuclear molding, and smudged chromatin (20× objective). (b) Synaptophysin with strong reactivity in prostatic pure small cell carcinoma (4× objective). (c) High-grade prostatic carcinoma without strong morphologic small cell features (10× objective). (d) An immunostaining of this case for synaptophysin shows only weak, rare reactivity, of questionable significance for treatment (10× objective). These figures are originally from the authors, and it is approved by the Institutional Review Board, University of California - Davis (Sacramento, CA, USA; Approval No. 222924), which also approved the exempt from patient consent.

Basic scientists are well aware that lineage plasticity leads to neuroendocrine differentiation since this has received much attention in the literature.4 Thus, they might anticipate a higher frequency of pathologists diagnosing either outright small-cell cancer or features of such differentiation in high-grade prostate cancer cases, compared to the reality. This is supported by the concept of neuroendocrine differentiation as the final common pathway in the evolution of castration-resistant prostate cancer (CRPC).5 The benign prostate has a sparse population of neuroendocrine cells and a much more abundant population of cells with luminal phenotype. Neuroendocrine cells do not significantly express androgen receptor, and when such cells are abundant, tumors show resistance to hormonal therapy.5 Photomicrographs of acinar prostate cancer (non-small cell) that is high-grade sometimes can indeed show an increase in chromogranin staining consistent with elevated proportions of neuroendocrine cells,3 but still these cells are in the minority compared to luminal nonreactive cells. On the other hand, when an anatomic pathologist encounters primary or even metastatic, cancer with Gleason grade 5 single cells, he/she feels obliged to order chromogranin and synaptophysin immunostains, in order to rule out a neuroendocrine component. Usually, both turn out negative. Sometimes, rare and weakly positive tumor cells of uncertain significance are seen and this is reported on the pathology report (Figure 1c and 1d).

The use of mouse models of neuroendocrine cancer may or may not translate to humans. For example, the efficacy of DRP-104 in vivo was evaluated in several cell lines. NCI-H660, derived from human neuroendocrine prostate cancer, was completely inhibited by the drug in vitro and in a mouse xenograft model. “[Histologic] neuroendocrine prostate cancer” was said to be “very commonly observed after hormone therapy”.7 To an anatomic pathologist, this seems debatable because even most biopsies that we see of high-grade prostate cancer in primary or metastatic sites are not overtly small cell.

Certain therapies can induce neuroendocrine-like differentiation of prostate cancer. It develops in the setting of androgen deprivation therapy (ADT)-induced changes to the tumor microenvironment (TME). TME can activate the cholinergic receptor muscarinic 4/protein kinase B (AKT)/N-myc signaling, so that cholinergic receptor, muscarinic 4 (CHRM4) may be a target for therapy.6 Radiation therapy can also induce neuroendocrine-like differentiation in prostate cancer cell lines, as measured by an increase in the nuclear content of phospho-cyclic AMP-response element binding protein (CREB) and cytoplasmic accumulation of activating transcription factor 2. Interestingly, the radiation-induced neuroendocrine-like morphology in cell lines was reversible, whereas there is no evidence that clinical small cell cancer in a patient is reversible in its differentiation.8

Moreover, certain neuroendocrine differentiation markers can be increased in routine morphologic acinar or non-small cell prostate cancer. For example, calcitonin acts as an intracrine–paracrine factor in the prostate that increases tumorigenicity and metastasis. Calcitonin induces zinc finger protein like 1 (ZFPL1), which localizes along with chromogranin in exosomes. ZFPL1 was greatly increased in serum samples from an unselected set of men with generic prostate cancer, suggesting its possible use as a screening marker.9

In a study of metastatic treatment-emergent small-cell neuroendocrine prostate cancer (failed abiraterone or enzalutamide), the incidence of neuroendocrine differentiation was surprisingly high.10 Of 160 patients, 27 (17%) had at least partial neuroendocrine differentiation in their metastatic tumor.10 This determination was made by pathologists based on morphology alone. Patients with treatment-emergent neuroendocrine cancer did have higher serum neuron-specific enolase but not chromogranin, compared to patients without it. Furthermore, this category of cancer was mutually exclusive with genomic alterations in DNA repair, and a transcriptional signature for it was developed and validated in external databases.

Clinical trials are often based on genomic markers to identify patients who might benefit from platinum-based therapy. At University of California-Davis Health, a National Cancer Institute (NCI)-sponsored clinical trial is enrolling men with “prostate cancer with neuroendocrine differentiation”. This includes not only those with small cell histology but also cases with molecular features of genetic alterations in 2 of 3 out of tumor protein p53 (TP53), RB transcriptional corepressor 1 (RB1), or phosphatase and tensin homolog (PTEN), visceral metastases without prostate-specific antigen (PSA) progression, or elevated serum chromogranin or neuron-specific enolase, in order to study the activity of lutetium Lu 177 dotatate (National Clinical Trial [NCT] 05691465). Another study, the Chemoimmunotherapy for the Treatment of Men with Neuroendocrine or Aggressive Variant Metastatic Prostate Cancer (CHAMP) study, merges eligibility for men with neuroendocrine prostate cancer with those having “aggressive variants” to evaluate the efficacy of the combination of nivolumab, ipilimumab, carboplatin, and cabazitaxel (NCT 04709276). Eligibility criteria include not only men with small cell or “small cell-like carcinoma” but also those with “intermediate atypical” carcinoma growing as sheets; also included are those with non-small cell adenocarcinoma but who have any of the following high-risk features including: low serum PSA, high serum lactate dehydrogenase (LDH), calcium or carcinoembryonic antigen (CEA), genetic alterations in 2 of 3 out of TP53, RB1, or PTEN, lytic bone lesions, or short interval to progression to castrate resistance after hormonal therapy.

Because these studies cast a wider net for inclusion of “neuroendocrine” prostate carcinoma, including incorporation of aggressive variant prostate cancers, many patients are included that lack histopathological uniformity. In summary, it is worth remembering that pathologists, basic science researchers, and clinical trialists are embracing different criteria for prostatic tumors or cell lines that with neuroendocrine differentiation biomarkers versus histologic neuroendocrine cancer. As next-generation genomic sequencing is further incorporated in defining prostate cancer variants, genomic overlap in nonneuroendocrine and neuroendocrine tumors from the same individuals may be seen,11 so these criteria may continue to evolve. Everyone in the prostate cancer field would benefit from understanding these differences, especially as it may complicate the interpretation of results across clinical trials. Eventually, a harmonized definition of neuroendocrine prostate cancer may be reached.

AUTHOR CONTRIBUTIONS

Both authors gathered the data and contributed to writing of the manuscript, and read and approved the final manuscript.

COMPETING INTERESTS

Both authors declare no competing interests.