Background: Lung cancer is the leading cause of cancer-related deaths worldwide, accounting for more than 40,000 deaths each year in Germany. Twenty-five percent of all lung tumours belong to the group of neuroendocrine tumours, encompassing typical (TC) and atypical carcinoids (AC), large-cell neuroendocrine carcinoma (LCNEC) and small-cell lung cancer (SCLC). Even though all four subgroups share some morphological and molecular properties, it is still unclear whether they are different biological entities or form one distinct tumour entity displaying different grades of differentiation. They were previously reported to arise from serotonin producing pulmonary neuroendocrine cells (PNEC). It is suggested that LCNEC and SCLC originate from the same or similar precursor cells. However, it remains unclear how carcinoids can be integrated into this scheme. Material and Methods: In order to address this question, 74 representative tumour specimens were used for sequencing analysis performed on a MiSeq instrument covering the 221 most important mutation hotspots related to human neoplasias. Additionally, mRNA-expression profiles of 80 tumour samples were determined for 91 selected genes using NanoString nCounter technology. Furthermore, three unequivocal samples of each tumour entity were chosen to be tested for their miRNA signature via 384 well TaqMan low-density array real-time qPCR for the expression of 768 unique miRNAs. Finally, the pulmonary neuroendocrine tumour samples were analysed via TaqMan qPCR and immunohistochemistry for expression of the five most important subunits of the 26S proteasome. Results: 44 miRNAs were identified which showed a significantly different miRNA expression between the subtypes of pulmonary neuroendocrine tumours. For 12 miRNAs, the difference was highly significant (p<0.01). Eight of these were negatively correlated with the grade of tumour biology, the other four were positively correlated. Six miRNAs are significantly associated with survival. The mRNA expression analysis showed 48 significant correlations with tumour type. Sixteen most significant (p≤0.001), 20 highly significant (p≤0.01) and 12 significant (p≤0.05) candidates were identified. Nine significant correlations with progression-free survival were found. Overall survival correlated significantly with nine genes. Of the 643 variants that passed the applied filter processing as described above, 122 variants were related to TC, 150 to AC, 164 to LCNEC and the final 207 variants were detected in the SCLC samples. We found a number of molecular features associated with pulmonary neuroendocrine tumours. Conclusion: In summary, we hypothesise that there is adequate evidence that neuroendocrine tumours should be considered an overall biological family of neoplasms. The family of pulmonary neuroendocrine tumours can be biologically well distinguished from all other forms of lung cancers. Additionally, better biological understanding of carcinoids and neuroendocrine lung carcinomas will increase the chance of finding new drug targets, maybe resulting in a more personalized therapeutic concept for this group of malignancies.