The 'grey area' between small cell and non-small cell lung carcinomas. Light and electron microscopy versus clinical data in 14 cases

Large Cell Neuroendocrine Carcinoma of the Lung

Large cell neuroendocrine (LCNE) carcinomas of the lung are a newly recognized, highly aggressive and frequently misdiagnosed entity. We report a case of stage I LCNE lung carcinoma initially misdiagnosed as large cell undifferentiated carcinoma or poorly differentiated adenocarcinoma. The tumor was very extensively necrotic and its neuroendocrine differentiation was only demonstrable with immunohistochemical staining with PHE-5 monoclonal antibody and with antisera against synaptophysin and calcitonin. ACTH, somatostatin and neurofilaments were not demonstrable. The clinical course was ominous and the patient died within 17 months. The reason for this rapid fatal outcome could be ascribed either to the neuroendocrine phenotype of the tumor, or to the extensive necrosis, or both.

Large cell neuroendocrine carcinoma of the lung: pathological study and clinical outcome of 18 resected cases

Large cell neuroendocrine carcinoma of the lung (LCNEC) has been recently redefined by the World Health Organisation (WHO) classification but the appropriate treatment remains unclear. We reviewed 18 consecutive resected cases of LCNEC. Two pathologists assessed diagnosis by applying rigorously the last WHO criteria. We reported the pathological features and the clinical outcome of this particular tumour. All patients were men with a median age of 63 years. Clinicopathologic stages corresponded to stage I (n = 8), II (n = 8) and IIIA (n = 2). All patients were treated as non-small cell lung carcinoma (NSCLC) and underwent surgery without any adjuvant treatment except four post-operative radiotherapy for N2 or T3 disease. The evolution was pejorative for 14 patients: one patient died of post-operative complications and 13 patients relapsed with distant metastases that occurred in 10 cases within 6 months after surgery. One-year survival rate was 27% and survival rate at the end of follow-up was 22%, which were both less than expected for stage-comparable NSCLC. Survival was neither influenced by lymph node status nor by pathological or molecular findings. Among the 10 evaluable patients with metastatic disease that received palliative platin-etoposide chemotherapy only two had partial tumour regressions (20%). Our study suggests that applying to LCNEC the NSCLC standard treatment lead to poor prognosis even in localised disease with a high incidence of early metastatic spread and a low response rate to chemotherapy. This way of relapse underlies the necessity of an efficient chemotherapy in order to improve survival.

Ultrastructural heterogeneity in undifferentiated bronchial carcinoma

Electron microscopy is often suggested as a useful aid to the classification of light microscopically undifferentiated bronchial malignancies, features such as dense-core vesicles, desmosomes or tonofilaments, and microacini, allowing their designation as endocrine, squamous, or adenocarcinomas respectively. However, there is no reason to suppose that the heterogeneity of malignant bronchial tumours so often apparent by light microscopy or on immunolabelling might not occur at the ultrastructural level too. Extensive sampling of all deposits from eight subjects coming to necropsy with undifferentiated bronchial carcinoma revealed ultrastructural features of glandular and squamous differentiation to be widespread and often to occur together, although dense-core vesicles were not seen in any of the tumours studied. Heterogeneity was present within individual tumour deposits and particularly between different deposits of those tumours which had disseminated, such that any ultrastructural diagnosis would have been significantly influenced by sampling. Such variation should be borne in mind when ultrastructural features are used to classify bronchial malignancies.

Comparative ultrastructure of needle aspiration biopsy and surgical resection specimens of lung tumors

Ultrastructural examination affords conclusive evidence for classification of lung tumors. Tissue properly fixed for electron microscopy is not available in many cases, however. Ultrastructural diagnosis of resected specimens obviously follows, rather than directs, the surgical treatment. Fine-needle aspiration (FNA) of lung masses is recommended as a means to obtain lung tumor tissue for electron microscopy. Nevertheless, no comparison has been made between ultrastructural information gained from aspiration specimens and resected specimens. Electron microscopy was performed on transthoracic FNA specimens of 10 lung tumors for which surgical resection was subsequently performed. Glutaraldehyde-fixed specimens from FNA and surgical resection were prepared for electron microscopy according to routine procedures. Fixation of the FNA specimens was equivalent or superior to that of the resected specimens in 9 of the cases. Three of the FNA specimens contained necrotic as well as viable tissue. Features essential for diagnosis such as desmosomes, junctions, neurosecretory granules, intermediate filaments, glycogen, lipid, mucin, and microvilli were identifiable in both FNA and resected specimens. FNA specimens therefore yield a representative sample of the ultrastructural features of lung tumors when adequate cellular material is obtained. Use of a coaxial needle sampling technique with immediate microscopic assessment reduces the likelihood of retrieving only blood or necrotic tissue in the electron microscopy specimens.

Distribution and estimation of nucleolar organizer regions in various human lung tumours

A quantitative study of nucleolar organizer regions in human lung carcinomas was carried out on routinely processed paraffin embedded tissue sections. We examined 104 lung carcinomas including 38 squamous cell carcinomas, 36 adenocarcinomas, 18 large cell anaplastic carcinomas, 6 small cell carcinomas and 6 carcinoids. No significant differences were found in mean number of NORs between squamous, adenocarcinoma and undifferentiated carcinomas including large cell and small cell carcinomas. Carcinoids had comparatively lower means except for one typical carcinoid. Considering the high incidence of overlap between ranges of NOR counts in these groups of tumours and in agreement with the only other study of lung tumours (which comprised only carcinoids and small cell carcinomas), we conclude that this technique cannot be reliably used to discriminate between various histologic types of lung cancers. However, long term follow up of these patients is needed to establish the value of the AgNOR technique for prognostic guidance.

Ultrastructural heterogeneity of lung carcinomas: representativity of samples for electron microscopy in tumor classification

Histologic heterogeneity of tumors is a well-known phenomenon, which has been repeated studied at the light microscopic level. Electron microscopy has been advocated as an adjunct in classification of tumors which pose difficulties on light microscopic classification. However, in view of tumor heterogeneity, it might be anticipated that the problem of sample error could detract from the usefulness of electron microscopy in tumor typing. Ultrastructural heterogeneity of tumors has thus far not been systematically investigated. We performed an ultrastructural randomized and blinded study of superficial and deep samples of 44 resected lung carcinomas, tumors which are notorious for their histologic heterogeneity. Neuroendocrine and squamous differentiation, as well as adenodifferentiation, were assessed separately and semiquantitatively in each sample. Twenty-six tumors showed more than one type of differentiation in at least one sample. However, in only two cases did the main type of differentiation differ between the two samples. A further nine cases showed one predominant differentiation type in both samples, but a similarly pronounced second differentiation type in one of the samples. Thus, in terms of ultrastructural diagnosis, the two samples showed a major discrepancy in two of 44 cases, and a minor discrepancy in nine of 44 cases. We conclude that ultrastructural heterogeneity of lung tumors is a common occurrence, but that it only rarely leads to totally different ultrastructural diagnoses.

Leu-7 in small cell neoplasms. An immunohistochemical study with ultrastructural correlations

In an effort to delineate the distribution of the Leu-7 antigen in small cell neoplasms, 283 paraffin-embedded examples of such tumors were studied immunohistochemically. These included 125 malignant lymphomas, 94 neuroendocrine carcinomas, 13 adenocarcinomas, 14 squamous carcinomas, four malignant melanomas, six neuroblastomas, four nephroblastomas (Wilms' tumors), six primitive neuroectodermal neoplasms, three "Askin" tumors, ten Ewing's sarcomas, and four embryonal rhabdomyosarcomas. Histologic diagnoses were verified by the use of electron microscopic study or independent immunostains. Overall, 44% of small cell neuroendocrine carcinomas expressed Leu-7, whereas nonendocrine carcinomas were uniformly nonreactive for this determinant. All neuroblastomas yielded immunopositivity, as did three primitive neuroectodermal tumors, three rhabdomyosarcomas, two "Askin" tumors, one nephroblastoma, one Ewing's sarcoma, and one malignant melanoma. None of the small cell lymphomas were Leu-7 positive. These results suggest that Leu-7 is a specific neuroendocrine marker in cases of small cell carcinoma, but its sensitivity is lower than that of other "endocrine" determinants. Reactivity patterns for Leu-7 in other small cell tumors demonstrate no specificity for any given line of cellular differentiation. However, the shared expression of this antigen by neuroblastomas, neuroectodermal tumors, Ewing's sarcomas, and Wilm's tumors contributes further to the hypothesis that these neoplasms may be related histogenetically.