Amy Carpenter
June 2026—The immunostaining landscape for lung neuroendocrine tumors continues to evolve, and Humberto Trejo Bittar, MD, in a CAP25 lung cancer session last fall, captured where it stands.
“This is probably the field that is growing the fastest in thoracic pathology, particularly diagnosis of small cell carcinomas and molecular subclassification, both of the small cell carcinoma and large cell neuroendocrine carcinoma,” said Dr. Trejo Bittar, chair of the CAP/NSH Histotechnology Committee. “It’s a field that is changing as we speak.” This group of neoplasms has not had much targeted therapy, he noted, and the prognosis has changed little in the past 20 to 30 years.
By definition, neuroendocrine neoplasms of the lung are considered malignant entities, but there is a spectrum of lesions and behavior varies. “Their histologic appearance can be similar,” said Dr. Trejo Bittar, associate professor, Department of Pathology, H. Lee Moffitt Cancer Center and Research Institute, and Department of Pathology, University of South Florida Morsani College of Medicine.
Expression of neuroendocrine markers does not equate to a neuroendocrine neoplasm diagnosis. “When a tumor is positive for synaptophysin or INSM1, that doesn’t mean you are dealing automatically with a neuroendocrine neoplasm,” he noted.
“Use with caution” is his advice on neuroendocrine markers. “Do not use neuroendocrine markers routinely,” he said, and don’t use them as part of a panel. The 2021 WHO classification of thoracic tumors advises their application in cases with clear morphologic evidence of neuroendocrine differentiation.
“Part of the problem is that we don’t know what it is to be a non-small cell lung carcinoma that expresses synaptophysin,” Dr. Trejo Bittar said. “Is this a worse type of tumor? Is this some sort of better or worse behavior? We don’t know, and that is an area of a lot of research.” The clinical relevance of NSCLC with neuroendocrine features remains uncertain, he said.
Small cell lung carcinoma is the only neuroendocrine neoplasm that can be diagnosed on morphology alone, but “that is easier to say than actually practice,” Dr. Trejo Bittar said, noting the inclination to do a synaptophysin or CAM 5.2. “I still want to see something, but technically they are not necessary.”
Lung neuroendocrine neoplasms are always related to the airways because they originate from Kulchitsky cells, and many Kulchitsky cells in an airway are neuroendocrine cell hyperplasia (Fig. 1, right). “When that is diffuse,” he said, “we can call that diffuse idiopathic pulmonary neuroendocrine cell hyperplasia,” or DIPNECH.
Typical and atypical lung carcinoids will have classic neuroendocrine morphology: salt-and-pepper chromatin and an organoid nested appearance. The difference between the two is the mitotic count, at less than two mitoses for a typical carcinoid and between two and 10 for an atypical carcinoid. “They always should have neuroendocrine morphology,” Dr. Trejo Bittar said.
The Ki-67 proliferation index is not part of the WHO classification criteria and should not be used to classify a tumor as typical or atypical, he said, though it can be helpful to exclude higher-grade lesions.
The typical and atypical carcinoids are thought to have a lot of TTF-1 expression (mostly positive in peripheral, mostly negative in central tumors). “But I don’t use that anymore,” he said, because carcinoid tumors often are TTF-1-negative, particularly peribronchial or central types of tumors.
As with other markers, he noted, TTF-1 should not be used to indicate the site of origin. A TTF-1-positive tumor is not always a carcinoid tumor of lung origin. “They can have expressions of TTF-1 in other locations.”
On a biopsy, the terminology should be carcinoid tumor or neuroendocrine tumor not otherwise specified. “I see often we try to push—we call it typical, we call it atypical. We don’t have to because we need a resection specimen to do the mitotic count.” But if five mitotic figures are seen already, “you can say it is at least an atypical carcinoid tumor and a higher-grade lesion cannot be excluded.”
Large cell neuroendocrine carcinoma (LCNEC) and small cell lung carcinoma (SCLC) are the higher-grade lesions of lung neuroendocrine neoplasms. LCNEC has an organoid nested pattern, basally located nuclear palisading, and comedo-type necrosis. In SCLC, classic smudgy chromatin, nuclear molding, crush artifact, mitosis, and necrosis are seen.
For LCNEC, at least one positive neuroendocrine marker is recommended in the current WHO classification. SCLC can be diagnosed on morphology alone. “Diagnosing a small cell carcinoma in a never smoker is quite challenging and should be done with a lot of caution,” Dr. Trejo Bittar said.
The Ki-67 proliferation index (30 to 100 percent) is an important marker for excluding the higher-grade lesions in small biopsy specimens with crush artifact.
Technically, LCNEC and SCLC have a higher positivity for TTF-1 (LCNEC, 70 percent; SCLC, 85 percent), he said. “But, again, I’ve lost my hope with TTF-1. You don’t often see TTF-1 expression on higher-grade neuroendocrine carcinomas.”
In a 2024 review and update of IHC utility in the diagnosis of pleuropulmonary and mediastinal cancers, the author writes that synaptophysin is the most sensitive marker, especially for LCNEC, and has good specificity (Deng H. Arch Pathol Lab Med. 2024;148[3]:267–283). The author also writes that INSM1 demonstrates high sensitivity and specificity for labeling an entire spectrum of lung neuroendocrine neoplasms (“It depends on the laboratory,” Dr. Trejo Bittar said). And CD56 was said to have the highest sensitivity but very low specificity. “Not everything that is CD56 positive is a neuroendocrine neoplasm. It might not even be a carcinoma,” Dr. Trejo Bittar noted, as many sarcomas and lymphoproliferative processes have CD56 positivity.
Chromogranin is best for carcinoid tumors.
The author suggests that CD56 with INSM1 and/or synaptophysin should be the first choice to confirm pulmonary high-grade neuroendocrine neoplasms or tumors. “I agree,” Dr. Trejo Bittar said.
P40 staining should be used to exclude basaloid squamous cell carcinomas from small cell lung carcinomas, the author writes. “They can look just like it,” Dr. Trejo Bittar said. “So when you’re starting to have these tumors that look like a small cell carcinoma but they’re not staining with any neuroendocrine marker, do your p40. You’ll be surprised, particularly if they are TTF-1-negative, how many of these are basaloid squamous cell carcinoma.”
Napsin A is strongly and diffusely positive in about 10 percent of LCNEC. It’s positive in a large number of large cell neuroendocrine carcinomas and other tumors—thyroid, ovarian, and others, Dr. Trejo Bittar said, “so don’t only rely on napsin A to call something adenocarcinoma.”
In a review published early last year, Sen and coauthors write that there has been substantial progress in understanding the genomic underpinnings of SCLC, with additional advances in lung carcinoids and LCNEC in the past five years (Sen T, et al. Lancet Oncol. 2025;26[1]:e13–e33).
“At the molecular level,” they write, “the defining feature of SCLC is an extremely high frequency of inactivating mutations in TP53 and RB1. Another common event appears to be inactivating mutations in genes regulating histone modification….” Carcinoids are known to have a low mutation burden, they write, and a predisposition to alterations in chromatin remodeling genes.
Based on genomic signatures, LCNEC can be categorized as SCLC-like, NSCLC-like, or carcinoid-like, with tumor aggressiveness highest in SCLC-like LCNEC and lowest in carcinoid-like LCNEC. SCLC has been subtyped depending on the expression of neuroendocrine biomarkers ASCL1, NeuroD1, POU2F3, and YAP1.
POU2F3 can be a useful IHC marker in SCLC with low or absent expression of neuroendocrine markers, Dr. Trejo Bittar said. Baine and coauthors screened 254 SCLC for POU2F3 expression and analyzed histopathologic, genomic, and clinical characteristics of POU2F3-positive tumors (Baine MK, et al. J Thorac Oncol. 2022;17[9]:1109–1121). POU2F3 was expressed in 30 (12 percent) of 254 SCLC and was strongly associated with low expression of standard neuroendocrine markers. “Notably,” the authors write, “POU2F3 was expressed in 75% of SCLC with entirely negative or minimal neuroendocrine marker expression (n=20) and was helpful in supporting the diagnosis of SCLC in such cases.”
Says Dr. Trejo Bittar, “I now do this more routinely to try to prove that a tumor is a small cell carcinoma when everything fails.”
Grading neuroendocrine neoplasms helps to risk stratify patients, classify tumors based on behavior, and guide appropriate treatment. “But Ki-67 should not be done unequivocally on biopsy specimens,” Dr. Trejo Bittar said. “We shouldn’t use a Ki-67 expression to say something must be a small cell carcinoma because there are obviously variabilities.”
The 2021 WHO classification says Ki-67 is not required but is desirable because a high Ki-67 proliferation rate (more than 20 to 30 percent) will favor high-grade neuroendocrine carcinoma over carcinoid tumors. Mitoses can be difficult to appreciate if there is severe crush artifact, Dr. Trejo Bittar noted. “We all have had that case: We don’t really know what’s going on, it looks neuroendocrine, you do a Ki-67 and it’s five percent. You won’t want to call that a small cell carcinoma, even though it looks like it. But that’s where Ki-67 becomes very helpful.” Carcinoid tumors can be identified with elevated mitoses and an elevated proliferation rate.
“You sometimes will have a tumor that looks like a small cell carcinoma, maybe with some areas that look more crushed,” he said. In Fig. 2, the tumor on the left has very low Ki-67. “You can favor here a carcinoid tumor.” Ki-67 immunostaining of the tumor on the right would raise concern for a higher-grade tumor, he said, particularly a small cell carcinoma.
“As we do more Ki-67 in our tumors, we realize there are typical carcinoid tumors and atypical carcinoid tumors, and depending on the case, on their proliferation rate, they will have different survival rates.” Typical carcinoid tumors with a Ki-67 proliferation of less than five percent behave much better than typical carcinoid tumors with a Ki-67 proliferation rate greater than five percent, he said, almost at the level of an atypical carcinoid tumor.
Marchevsky and coauthors wrote in their 2018 article that their results support the concept that by combining WHO diagnosis and Ki-67 labeling index, pulmonary carcinoids can be stratified into three grades: G1 (typical carcinoids with Ki-67 labeling index <5 percent), G2 (typical carcinoids with Ki-67 labeling index ≥5 percent), and G3 (atypical carcinoids), with different prognoses (Marchevsky AM, et al. Mod Pathol. 2018;31[10]:1523–1531).
An emerging category has gained attention, Dr. Trejo Bittar said: carcinoid tumor with elevated mitoses and/or Ki-67 proliferation rate (Fig. 3) (Pelosi G, et al. Histopathology. 2024;85[4]:535–548). It is not fully defined, he said, though it is in the 2021 WHO classification of thoracic tumors.
“Sometimes we get carcinoid tumors that almost look like large cell neuroendocrine carcinoma. They have a little very focal comedo-type necrosis, very focal elevated Ki-67 proliferation rate. What do we do with these tumors?” he asked. “If you call it large cell neuroendocrine carcinoma, you won’t be incorrect. That is an option.” But the category of carcinoid tumor with elevated mitoses and/or Ki-67 proliferation rate also can be used. Tumors in this category resemble a carcinoid tumor microscopically but exhibit increased mitotic activity (>10 per 10 high-power field) and/or a high Ki-67 proliferation index (>30 percent) or notable necrosis, Dr. Trejo Bittar said, noting they’re rare but diagnostically significant. When sequenced, these tumors align more closely with carcinoid tumors than with neuroendocrine carcinomas.
They correspond to grade three pancreatic neuroendocrine tumors in classification terms.
Though often labeled as LCNEC, this tumor has a prognosis that is similar to that of carcinoids, he said, warranting individualized interpretation. Thus, these distinctions should be outlined and fully explained in the pathology report to inform clinical management.
More research is needed to determine “what the next step is for these tumors.”
For now, Dr. Trejo Bittar said, there is no definitive role for Ki-67 in lung neuroendocrine neoplasms. “Future work is needed using various methods,” among them eyeball estimates, point counting, and image analysis, and “probably now with AI.”
PHH3 is a surrogate mitosis marker used in the neuroendocrine neoplasm world. Laflamme and coauthors found that histologic assessments of carcinoid tumors using PHH3 staining provide benefits in terms of scoring times, mitosis detection, and reproducibility of mitotic counts (Laflamme P, et al. Hum Pathol. 2020;106:74–81). “Just as with Ki-67, we haven’t come up with any specific guidelines,” Dr. Trejo Bittar said. Laflamme and coauthors found the benefits to be greater for less experienced pathologists. “If you’re already good at it,” Dr. Trejo Bittar said of finding mitotic figures, “it probably doesn’t add benefits.”
Amy Carpenter is CAP TODAY senior editor.