Precise and automated lung cancer cell classification using deep neural network with multiscale features and model distillation.

Lung diseases globally impose a significant pathological burden and mortality rate, particularly the differential diagnosis between adenocarcinoma, squamous cell carcinoma, and small cell lung carcinoma, which is paramount in determining optimal treatment strategies and improving clinical prognoses. Faced with the challenge of improving diagnostic precision and stability, this study has developed an innovative deep learning-based model. This model employs a Feature Pyramid Network (FPN) and Squeeze-and-Excitation (SE) modules combined with a Residual Network (ResNet18), to enhance the processing capabilities for complex images and conduct multi-scale analysis of each channel's importance in classifying lung cancer. Moreover, the performance of the model is further enhanced by employing knowledge distillation from larger teacher models to more compact student models. Subjected to rigorous five-fold cross-validation, our model outperforms existing models on all performance metrics, exhibiting exceptional diagnostic accuracy. Ablation studies on various model components have verified that each addition effectively improves model performance, achieving an average accuracy of 98.84% and a Matthews Correlation Coefficient (MCC) of 98.83%. Collectively, the results indicate that our model significantly improves the accuracy of disease diagnosis, providing physicians with more precise clinical decision-making support.

Small cell lung cancer - news in the tumour´s biology.

Small cell lung carcinoma (SCLC) is a high grade neuroendocrinne tumour accounting for approximately 15 % of lung cancers. It is characterised by early relapse and low survival rate. The treatment has remained unchanged for decades. Histological and cytological characteristics are summarised in brief, along with genetic alterations of the tumour. A new molecular subtype classification is presented according to the expression of transciptional factors ASCL1 (SCLC-A), NEUROD1 (SCLC-D), POU2F3 (SCLC-P) and YAP1 (SCLC-Y). These subtypes represent different ways of tumorigenesis, and the distinct genomic alterations may offer new therapeutic strategies.

Classification of subtypes including LCNEC in lung cancer biopsy slides using convolutional neural network from scratch.

Identifying the lung carcinoma subtype in small biopsy specimens is an important part of determining a suitable treatment plan but is often challenging without the help of special and/or immunohistochemical stains. Pathology image analysis that tackles this issue would be helpful for diagnoses and subtyping of lung carcinoma. In this study, we developed AI models to classify multinomial patterns of lung carcinoma; ADC, LCNEC, SCC, SCLC, and non-neoplastic lung tissue based on convolutional neural networks (CNN or ConvNet). Four CNNs that were pre-trained using transfer learning and one CNN built from scratch were used to classify patch images from pathology whole-slide images (WSIs). We first evaluated the diagnostic performance of each model in the test sets. The Xception model and the CNN built from scratch both achieved the highest performance with a macro average AUC of 0.90. The CNN built from scratch model obtained a macro average AUC of 0.97 on the dataset of four classes excluding LCNEC, and 0.95 on the dataset of three subtypes of lung carcinomas; NSCLC, SCLC, and non-tumor, respectively. Of particular note is that the relatively simple CNN built from scratch may be an approach for pathological image analysis.

The impact of reclassifying cancers of unspecified histology on international differences in survival for small cell and non-small cell lung cancer (ICBP SurvMark-2 project).

Survival from lung cancer remains low, yet is the most common cancer diagnosed worldwide. With survival contrasting between the main histological groupings, small-cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC), it is important to assess the extent that geographical differences could be from varying proportions of cancers with unspecified histology across countries. Lung cancer cases diagnosed 2010-2014, followed until 31 December 2015 were provided by cancer registries from seven countries for the ICBP SURVMARK-2 project. Multiple imputation was used to reassign cases with unspecified histology into SCLC, NSCLC and other. One-year and three-year age-standardised net survival were estimated by histology, sex, age group and country. In all, 404 617 lung cancer cases were included, of which 47 533 (11.7%) and 262 040 (64.8%) were SCLC and NSCLC. The proportion of unspecified cases varied, from 11.2% (Denmark) to 29.0% (The United Kingdom). After imputation with unspecified histology, survival variations remained: 1-year SCLC survival ranged from 28.0% (New Zealand) to 35.6% (Australia) NSCLC survival from 39.4% (The United Kingdom) to 49.5% (Australia). The largest survival change after imputation was for 1-year NSCLC (4.9 percentage point decrease). Similar variations were observed for 3-year survival. The oldest age group had lowest survival and largest decline after imputation. International variations in SCLC and NSCLC survival are only partially attributable to differences in the distribution of unspecified histology. While it is important that registries and clinicians aim to improve completeness in classifying cancers, it is likely that other factors play a larger role, including underlying risk factors, stage, comorbidity and care management which warrants investigation.

A deep learning model for the classification of indeterminate lung carcinoma in biopsy whole slide images.

The differentiation between major histological types of lung cancer, such as adenocarcinoma (ADC), squamous cell carcinoma (SCC), and small-cell lung cancer (SCLC) is of crucial importance for determining optimum cancer treatment. Hematoxylin and Eosin (H&E)-stained slides of small transbronchial lung biopsy (TBLB) are one of the primary sources for making a diagnosis; however, a subset of cases present a challenge for pathologists to diagnose from H&E-stained slides alone, and these either require further immunohistochemistry or are deferred to surgical resection for definitive diagnosis. We trained a deep learning model to classify H&E-stained Whole Slide Images of TBLB specimens into ADC, SCC, SCLC, and non-neoplastic using a training set of 579 WSIs. The trained model was capable of classifying an independent test set of 83 challenging indeterminate cases with a receiver operator curve area under the curve (AUC) of 0.99. We further evaluated the model on four independent test sets-one TBLB and three surgical, with combined total of 2407 WSIs-demonstrating highly promising results with AUCs ranging from 0.94 to 0.99.

Distinct transcriptional programs of SOX2 in different types of small cell lung cancers.

SOX2 is recognized as an oncogene in human small cell lung cancer (SCLC), which is an aggressive neuroendocrine (NE) tumor. However, the role of SOX2 in SCLC is not completely understood, and strategies to selectively target SOX2 in SCLC cells remain elusive. Here, we show, using next-generation sequencing, that SOX2 expressed in the ASCL1-high SCLC (SCLC-A) subtype cell line is dependent on ASCL1, which is a lineage-specific transcriptional factor, and is involved in NE differentiation and tumorigenesis. ASCL1 recruits SOX2, which promotes INSM1 and WNT11 expression. Immunohistochemical studies revealed that SCLC tissue samples expressed SOX2, ASCL1, and INSM1 in 18 out of the 30 cases (60%). Contrary to the ASCL1-SOX2 signaling axis controlling SCLC biology in the SCLC-A subtype, SOX2 targets distinct genes such as those related to the Hippo pathway in the ASCL1-negative, YAP1-high SCLC (SCLC-Y) subtype. Although SOX2 knockdown experiments suppressed NE differentiation and cell proliferation in the SCLC-A subtype, they did not sufficiently impair the growth of the SCLC-Y subtype cell lines in vitro and ex vivo. The present results support the importance of the ASCL1-SOX2 axis as a main subtype of SCLC, and suggest the therapeutic potential of targeting the ASCL1-SOX2 axis.

Systems-level network modeling of Small Cell Lung Cancer subtypes identifies master regulators and destabilizers.

Adopting a systems approach, we devise a general workflow to define actionable subtypes in human cancers. Applied to small cell lung cancer (SCLC), the workflow identifies four subtypes based on global gene expression patterns and ontologies. Three correspond to known subtypes (SCLC-A, SCLC-N, and SCLC-Y), while the fourth is a previously undescribed ASCL1+ neuroendocrine variant (NEv2, or SCLC-A2). Tumor deconvolution with subtype gene signatures shows that all of the subtypes are detectable in varying proportions in human and mouse tumors. To understand how multiple stable subtypes can arise within a tumor, we infer a network of transcription factors and develop BooleaBayes, a minimally-constrained Boolean rule-fitting approach. In silico perturbations of the network identify master regulators and destabilizers of its attractors. Specific to NEv2, BooleaBayes predicts ELF3 and NR0B1 as master regulators of the subtype, and TCF3 as a master destabilizer. Since the four subtypes exhibit differential drug sensitivity, with NEv2 consistently least sensitive, these findings may lead to actionable therapeutic strategies that consider SCLC intratumoral heterogeneity. Our systems-level approach should generalize to other cancer types.

Radiomics for Classification of Lung Cancer Histological Subtypes Based on Nonenhanced Computed Tomography.

OBJECTIVES: To evaluate the performance of using radiomics method to classify lung cancer histological subtypes based on nonenhanced computed tomography images. MATERIALS AND METHODS: 278 patients with pathologically confirmed lung cancer were collected, including 181 nonsmall cell lung cancer (NSCLC) and 97 small cell lung cancers (SCLC) patients. Among the NSCLC patients, 88 patients were adenocarcinomas (AD) and 93 patients were squamous cell carcinomas (SCC). In total, 1695 quantitative radiomic features (QRF) were calculated from the primary lung cancer tumor in each patient. To build radiomic classification model based on the extracted QRFs, several machine-learning algorithms were applied sequentially. First, unsupervised hierarchical clustering was used to exclude highly correlated QRFs; second, the minimum Redundancy Maximum Relevance feature selection algorithm was employed to select informative and nonredundant QRFs; finally, the Incremental Forward Search and Support Vector Machine classification algorithms were used to combine the selected QRFs and build the model. In our work, to study the phenotypic differences among lung cancer histological subtypes, four classification models were built. They were models of SCLC vs NSCLC, SCLC vs AD, SCLC vs SCC, and AD vs SCC. The performance of the classification models was evaluated by the area under the receiver operating characteristic curve (AUC) estimated by three-fold cross-validation. RESULTS: The AUC (95% confidence interval) for the model of SCLC vs NSCLC was 0.741(0.678, 0.795). For the models of SCLC vs AD and SCLC vs SCC, the AUCs were 0.822(0.755, 0.875) and 0.665(0.583, 0.738), respectively. The AUC for the model of AD vs SCC was 0.655(0.570, 0.731). Several QRFs ("Law_15," "LoG_Uniformity," "GLCM_Contrast," and "Compactness Factor") that characterize tumor heterogeneity and shape were selected as the significant features to build the models. CONCLUSION: Our results show that phenotypic differences exist among different lung cancer histological subtypes on nonenhanced computed tomography image.

Triple test with tumor markers CYFRA 21.1, HE4, and ProGRP might contribute to diagnosis and subtyping of lung cancer.

BACKGROUND AND AIM: Early diagnosis and histological subtyping are important issues in the management of patients with lung cancer (LC). The aim of this study is to investigate the diagnostic value of a panel of serum tumor markers in newly diagnosed patients with LC. METHODS: Venous blood samples were collected from 99 patients with LC (42 adenocarcinoma, 35 squamous, and 22 small cell carcinoma) and 30 patients with benign lung disease. Progastrin releasing peptide (ProGRP), squamous cell carcinoma antigen (SCCAg), cytokeratin 19-fragments (CYFRA 21.1), human epididymis protein 4 (HE4), Chromogranin A (CgA) and neuron specific enolase (NSE) levels were measured. The diagnostic value of the biomarkers was assessed with ROC curve analyses; the area under the curve (AUC) was calculated. RESULTS: Serum CYFRA 21.1, ProGRP, SCCAg, NSE levels were significantly higher in LC patients. While ProGRP levels were higher (p = 0.009) in SCLC; CYFRA 21.1 and SCCAg levels were higher in NSCLC (p = 0.019 and p = 0.001, respectively). The sensitivity and specificity of tumor markers were 72%, 83% for CYFRA 21.1; 70%, 57% for HE4; 18%, 93% for ProGRP; 43%, 77% for SCCAg; 54%, 53% for CgA; 73%, 50% for NSE. CYFRA 21.1 (p < 0.001, r = 0.394), HE4 (p = 0.014, r = 0.279) and CgA (p = 0.023, r = 0.259) levels were positively correlated with tumor stage in NSCLC. CgA levels were significantly higher in extensive stage SCLC (p = 0.004). CYFRA 21.1 had the highest diagnostic value for LC (AUC = 0.865). When it is combined with HE4, diagnostic value increased (AUC = 0.899). ProGRP had the highest diagnostic value (AUC = 0.875, p < 0.001) for discriminating SCLC from NSCLC. CONCLUSION: A panel of three tumor markers CYFRA 21.1, HE4 and ProGRP may play a role for discriminating LC from benign lung disease and subtyping as SCLC.

Pulmonary Non-Small Cell Carcinoma With Morphologic Features of Adenocarcinoma or "Non-Small Cell Carcinoma Favor Adenocarcinoma" in Cytologic Specimens Share Similar Clinical and Molecular Genetic Characteristics.

OBJECTIVES: Define if the presence of morphologic features of adenocarcinoma (ACA) in non-small cell lung carcinoma (NSCLC) on cytology specimens correlates with clinical and biologic features. METHODS: A total of 209 cases of NSCLC diagnosed on fine-needle aspiration in a 3-year period were included. RESULTS: After morphologic review, the cases were classified as ACA (n = 115), NSCLC favor ACA (n = 43), and NSCLC-not otherwise specified (NOS) (n = 18). Squamous cell (SCC) (n = 14) and NSCLC favor SCC (n = 19) were excluded from further analysis. Patients with EGFR-mutated tumors had longer overall survival than those with EGFR wild-type tumors (P = .01). When comparing cases with morphologic features of ACA, NSCLC favor ACA, and NSCLC-NOS, there were no differences in the presence or absence of tested mutations, clinical stage, or survival. CONCLUSION: Patients diagnosed with pulmonary ACA, NSCLC favor ACA, or NSCLC-NOS in cytology specimens have similar clinical stage, survival, and molecular alterations.

Automated Classification of Lung Cancer Types from Cytological Images Using Deep Convolutional Neural Networks.

Lung cancer is a leading cause of death worldwide. Currently, in differential diagnosis of lung cancer, accurate classification of cancer types (adenocarcinoma, squamous cell carcinoma, and small cell carcinoma) is required. However, improving the accuracy and stability of diagnosis is challenging. In this study, we developed an automated classification scheme for lung cancers presented in microscopic images using a deep convolutional neural network (DCNN), which is a major deep learning technique. The DCNN used for classification consists of three convolutional layers, three pooling layers, and two fully connected layers. In evaluation experiments conducted, the DCNN was trained using our original database with a graphics processing unit. Microscopic images were first cropped and resampled to obtain images with resolution of 256 × 256 pixels and, to prevent overfitting, collected images were augmented via rotation, flipping, and filtering. The probabilities of three types of cancers were estimated using the developed scheme and its classification accuracy was evaluated using threefold cross validation. In the results obtained, approximately 71% of the images were classified correctly, which is on par with the accuracy of cytotechnologists and pathologists. Thus, the developed scheme is useful for classification of lung cancers from microscopic images.

The Use of Immunohistochemistry Improves the Diagnosis of Small Cell Lung Cancer and Its Differential Diagnosis. An International Reproducibility Study in a Demanding Set of Cases.

INTRODUCTION: The current WHO classification of lung cancer states that a diagnosis of SCLC can be reliably made on routine histological and cytological grounds but immunohistochemistry (IHC) may be required, particularly (1) in cases in which histologic features are equivocal and (2) in cases in which the pathologist wants to increase confidence in diagnosis. However, reproducibility studies based on hematoxylin and eosin-stained slides alone for SCLC versus large cell neuroendocrine carcinoma (LCNEC) have shown pairwise κ scores ranging from 0.35 to 0.81. This study examines whether judicious use of IHC improves diagnostic reproducibility for SCLC. METHODS: Nineteen lung pathologists studied interactive digital images of 79 tumors, predominantly neuroendocrine lung tumors. Images of resection and biopsy specimens were used to make diagnoses solely on the basis of morphologic features (level 1), morphologic features along with requested IHC staining results (level 2), and all available IHC staining results (level 3). RESULTS: For the 19 pathologists reading all 79 cases, the rate of agreement for level 1 was 64.7%, and it increased to 73.2% and 77.5% in levels 2 and 3, respectively. With IHC, κ scores for four tumor categories (SCLC, LCNEC, carcinoid tumors, and other) increased in resection samples from 0.43 to 0.60 and in biopsy specimens from 0.43 to 0.64. CONCLUSIONS: Diagnosis using hematoxylin and eosin staining alone showeds moderate agreement among pathologists in tumors with neuroendocrine morphology, but agreement improved to good in most cases with the judicious use of IHC, especially in the diagnosis of SCLC. An approach for IHC in the differential diagnosis of SCLC is provided.

Validation of the SHOX2/PTGER4 DNA Methylation Marker Panel for Plasma-Based Discrimination between Patients with Malignant and Nonmalignant Lung Disease.

INTRODUCTION: Low-dose computed tomography (LDCT) is used for screening for lung cancer (LC) in high-risk patients in the United States. The definition of high risk and the impact of frequent false-positive results of low-dose computed tomography remains a challenge. DNA methylation biomarkers are valuable noninvasive diagnostic tools for cancer detection. This study reports on the evaluation of methylation markers in plasma DNA for LC detection and discrimination of malignant from nonmalignant lung disease. METHODS: Circulating DNA was extracted from 3.5-mL plasma samples, treated with bisulfite using a commercially available kit, purified, and assayed by real-time polymerase chain reaction for assessment of DNA methylation of short stature homeobox 2 gene (SHOX2), prostaglandin E receptor 4 gene (PTGER4), and forkhead box L2 gene (FOXL2). In three independent case-control studies these assays were evaluated and optimized. The resultant assay, a triplex polymerase chain reaction combining SHOX2, PTGER4, and the reference gene actin, beta gene (ACTB), was validated using plasma from patients with and without malignant disease. RESULTS: A panel of SHOX2 and PTGER4 provided promising results in three independent case-control studies examining a total of 330 plasma specimens (area under the receiver operating characteristic curve = 91%-98%). A validation study with 172 patient samples demonstrated significant discriminatory performance in distinguishing patients with LC from subjects without malignancy (area under the curve = 0.88). At a fixed specificity of 90%, sensitivity for LC was 67%; at a fixed sensitivity of 90%, specificity was 73%. CONCLUSIONS: Measurement of SHOX2 and PTGER4 methylation in plasma DNA allowed detection of LC and differentiation of nonmalignant diseases. Development of a diagnostic test based on this panel may provide clinical utility in combination with current imaging techniques to improve LC risk stratification.

Classification and Pathology of Lung Cancer.

Advancement in the understanding of lung tumor biology enables continued refinement of lung cancer classification, reflected in the recently introduced 2015 World Health Organization classification of lung cancer. In small biopsy or cytology specimens, special emphasis is placed on separating adenocarcinomas from the other lung cancers to effectively select tumors for targeted molecular testing. In resection specimens, adenocarcinomas are further classified based on architectural pattern to delineate tissue types of prognostic significance. Neuroendocrine tumors are divided into typical carcinoid, atypical carcinoid, small cell carcinoma, and large cell neuroendocrine carcinoma based on a combination of features, especially tumor cell proliferation rate.

Evaluation of Machine Learning Algorithm Utilization for Lung Cancer Classification Based on Gene Expression Levels.

BACKGROUND: Lung cancer remains one of the most common cancers in the world, both in terms of new cases (about 13% of total per year) and deaths (nearly one cancer death in five), because of the high case fatality. Errors in lung cancer type or malignant growth determination lead to degraded treatment efficacy, because anticancer strategy depends on tumor morphology. MATERIALS AND METHODS: We have made an attempt to evaluate effectiveness of machine learning algorithms in the task of lung cancer classification based on gene expression levels. We processed four publicly available data sets. The Dana-Farber Cancer Institute data set contains 203 samples and the task was to classify four cancer types and sound tissue samples. With the University of Michigan data set of 96 samples, the task was to execute a binary classification of adenocarcinoma and non-neoplastic tissues. The University of Toronto data set contains 39 samples and the task was to detect recurrence, while with the Brigham and Women's Hospital data set of 181 samples it was to make a binary classification of malignant pleural mesothelioma and adenocarcinoma. We used the k-nearest neighbor algorithm (k=1, k=5, k=10), naive Bayes classifier with assumption of both a normal distribution of attributes and a distribution through histograms, support vector machine and C4.5 decision tree. Effectiveness of machine learning algorithms was evaluated with the Matthews correlation coefficient. RESULTS: The support vector machine method showed best results among data sets from the Dana-Farber Cancer Institute and Brigham and Women's Hospital. All algorithms with the exception of the C4.5 decision tree showed maximum potential effectiveness in the University of Michigan data set. However, the C4.5 decision tree showed best results for the University of Toronto data set. CONCLUSIONS: Machine learning algorithms can be used for lung cancer morphology classification and similar tasks based on gene expression level evaluation.

The Impact of the Fourth Edition of the WHO Classification of Lung Tumours on Histological Classification of Resected Pulmonary NSCCs.

INTRODUCTION: Histopathological classification of lung cancer is of central importance in the diagnostic routine, and it guides therapy in most patients. The fourth edition of the World Health Organization (WHO) Classification of Lung Tumours was recently published and includes changes to the diagnostic procedure for non-small cell carcinomas (NSCCs), with more emphasis on immunohistochemical (IHC) staining. METHODS: A total of 656 unselective cases of resected pulmonary NSCC were diagnosed according to the 2004 WHO classification. After IHC staining with cytokeratin 5, p40, p63, thyroid transcription factor 1 (clones 8G7G3/1 and SPT24), and napsin A, the diagnoses were revised in accordance with the new fourth edition of the WHO classification. RESULTS: Reclassification led to a new histological annotation in 36 of the 656 cases (5%). Most notable was the decrease in cases previously classified as large cell carcinomas (56 versus 12 cases). This was partially due to the exclusion of 21 neuroendocrine tumors from this group, with 20 cases ascribed to the adenocarcinoma group on the basis of IHC markers. Only seven cases of adenocarcinoma or squamous cell carcinoma were reclassified after the addition of IHC staining. There was a substantial overlap in staining properties between different markers of squamous and adenocarcinomatous differentiation, respectively, but in 17 to 31 cases (3%-5%), the diagnosis depended on the choice of markers. CONCLUSIONS: The fourth edition of the WHO Classification of Lung Tumours leads to changes in histological type in 5% of resected NSCCs. The incorporation of IHC staining into NSCC diagnostics demands awareness that the choice of ancillary stains has an effect on diagnosis.

The International Association for the Study of Lung Cancer Lung Cancer Staging Project: Proposals for the Revision of the Clinical and Pathologic Staging of Small Cell Lung Cancer in the Forthcoming Eighth Edition of the TNM Classification for Lung Cancer.

INTRODUCTION: Small cell lung cancer (SCLC) is commonly classified as either limited or extensive, but the Union for International Cancer Control TNM Classification of Malignant Tumours seventh edition (2009) recommended tumor, node, and metastasis (TNM) staging based on analysis of the International Association for the Study of Lung Cancer (IASLC) database. METHODS: Survival analyses were performed for clinically and pathologically staged patients presenting with SCLC from 1999 through 2010. Prognosis was compared in relation to the TNM seventh edition staging to serve as validation and analyzed in relation to proposed changes to the T descriptors found in the eighth edition. RESULTS: There were 5002 patients: 4848 patients with clinical and 582 with pathological stages. Among these, 428 had both. Survival differences were confirmed for T and N categories and maintained in relation to proposed revisions to T descriptors for seventh edition TNM categories and proposed changes in the eighth edition. There were also survival differences, notably at 12 months, in patients with brain-only single-site metastasis (SSM) compared to SSM at other sites, and SSM without a pleural effusion showed a better prognosis than other patients in the M1b category. CONCLUSION: We confirm the prognostic value of clinical and pathological TNM staging in patients with SCLC, and recommend continued usage for SCLC in relation to proposed changes to T, N, and M descriptors for NSCLC in the eighth edition. However, for M descriptors, it remains uncertain whether survival differences in patients with SSM in the brain simply reflect better treatment options rather than better survival based on anatomic extent of disease.

The Expression of miR-375 Is Associated with Carcinogenesis in Three Subtypes of Lung Cancer.

Many studies demonstrated unique microRNA profiles in lung cancer. Nonetheless, the role and related signal pathways of miR-375 in lung cancer are largely unknown. Our study investigated relationships between carcinogenesis and miR-375 in adenocarcinoma, squamous cell carcinoma and small cell lung carcinoma to identify new molecular targets for treatment. We evaluated 723 microRNAs in microdissected cancerous cells and adjacent normal cells from 126 snap-frozen lung specimens using microarrays. We validated the expression profiles of miR-375 and its 22 putative target mRNAs in an independent cohort of 78 snap-frozen lung cancer tissues using quantitative reverse-transcriptase PCR. Moreover, we performed dual luciferase reporter assay and Western blot on 6 targeted genes (FZD8, ITGA10, ITPKB, LRP5, PIAS1 andRUNX1) in small cell lung carcinoma cell line NCI-H82. We also detected the effect of miR-375 on cell proliferation in NCI-H82. We found that miR-375 expression was significantly up-regulated in adenocarcinoma and small cell lung carcinoma but down-regulated in squamous cell carcinoma. Among the 22 putative target genes, 11 showed significantly different expression levels in at least 2 of 3 pair-wise comparisons (adenocarcinoma vs. normal, squamous cell carcinoma vs. normal or small cell lung carcinoma vs. normal). Six targeted genes had strong negative correlation with the expression level of miR-375 in small cell lung carcinoma. Further investigation revealed that miR-375 directly targeted the 3'UTR of ITPKB mRNA and over-expression of miR-375 led to significantly decreased ITPKB protein level and promoted cell growth. Thus, our study demonstrates the differential expression profiles of miR-375 in 3 subtypes of lung carcinomas and finds thatmiR-375 directly targets ITPKB and promoted cell growth in SCLC cell line.