Cytologic Diagnosis Of Lung Cancer Research Articles

Comparison of Fine-Tuned Deep Convolutional Neural Networks for the Automated Classification of Lung Cancer Cytology Images with Integration of Additional Classifiers.

Objective:It is essential to accurately diagnose and classify histological subtypes into adenocarcinoma (ADC), squamous cell carcinoma (SCC), and small cell lung carcinoma (SCLC) for the appropriate treatment of lung cancer patients. However, improving the accuracy and stability of diagnosis is challenging, especially for non-small cell carcinomas. The purpose of this study was to compare multiple deep convolutional neural network (DCNN) technique with subsequent additional classifiers in terms of accuracy and characteristics in each histology. Methods:Lung cancer cytological images were classified into ADC, SCC, and SCLC with four fine-tuned DCNN models consisting of AlexNet, GoogLeNet (Inception V3), VGG16 and ResNet50 pretrained by natural images in ImageNet database. For more precise classification, the figures of 3 histological probabilities were further applied to subsequent machine learning classifiers using Naïve Bayes (NB), Support vector machine (SVM), Random forest (RF), and Neural network (NN). Results:The classification accuracies of the AlexNet, GoogLeNet, VGG16 and ResNet50 were 74.0%, 66.8%, 76.8% and 74.0%, respectively. Well differentiated typical morphologies were tended to be correctly judged by all four architectures. However, poorly differentiated non-small cell carcinomas lacking typical structures were inclined to be misrecognized in some DCNNs. Regarding the histological types, ADC were best judged by AlexNet and SCC by VGG16. Subsequent machine learning classifiers of NB, SVV, RF, and NN improved overall accuracies to 75.1%, 77.5%, 78.2%, and 78.9%, respectively. Conclusion:Fine-tuning DCNNs in combination with additional classifiers improved classification of cytological diagnosis of lung cancer, although classification bias could be indicated among DCNN architectures.

Evaluating the Genomic Yield of a Single Endobronchial Ultrasound-guided Transbronchial Needle Aspiration in Lung Cancer: Meeting the Challenge of Doing More With Less

BackgroundMinimally invasive techniques, including endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA), yield small specimens that are adequate for cytologic diagnosis of lung cancer, but also need to provide material for molecular analysis to guide treatment. The number of EBUS-TBNA passes needed for mutation testing remains unclear. We sought to assess the adequacy of a single pass for genomic profiling of actionable mutations. MethodsIn a prospective observational study, paired samples from the same lesion were obtained from patients undergoing EBUS-TBNA for lung cancer diagnosis/staging. Following tumor cell confirmation by rapid on-site evaluation, a “reference” sample comprising ≥ 3 passes was obtained and formalin-fixed paraffin-embedded. A “study” sample comprising a single pass was taken and snap-frozen. The primary outcome was DNA yield and quality from a single pass. The secondary outcome was diagnostic accuracy of a single pass for detecting actionable mutations. ResultsIn 40 patients, single-pass specimens yielded a mean 3.98 μg of highly intact DNA, well above the minimum threshold for targeted sequencing, which was performed in adenocarcinoma cases (n = 24). In 23 cases, there was 100% agreement in mutation status between reference and study samples. In 1 case, the reference sample failed to generate a molecular diagnosis owing to insufficient tumor cells; however, the study specimen identified a KRAS mutation. Tumor cell percentage in mutation-positive specimens was 1% to 70%, suggesting that single-pass samples detect mutations even when tumor cell content is low. ConclusionSingle EBUS-TBNA passes yield DNA of high quantity and quality with high accuracy for molecular profiling, irrespective of tumor cell content.

Role of FNAC, fluid specimens, and cell blocks for cytological diagnosis of lung cancer in the present era.

Cytoblocks prepared from residual tissue fluids and fine-needle aspirations can be useful adjuncts to smears for establishing a more definitive cytopathologic diagnosis. These paraffin embedded cytoblocks have been popular since these can be handled like any other histologic specimen. Rapid on-site evaluation (ROSE) can help in attaining adequate material in the cytoblock, which is a major concern to the cytopathologists. Ancillary studies can be done using cytoblocks including IHC and various molecular techniques. The opportunities for cytopathologists to influence therapy, and uncover strategies in the complex field of lung cancer are exciting and limitless especially in the presence of an adequate cytoblock

Expression and clinical significance of lung‐specific X protein mRNA in bronchial brushing specimens from patients with or without lung cancer

The aim of this study was to evaluate the diagnostic utility of lung-specific X protein (LUNX) mRNA expression in bronchial brushing specimens from patients with lung cancer. LUNX mRNA levels were assessed by performing RT-PCR on liquid-based cytology bronchial brushing specimens from patients with lung cancer (n=104) or benign lung disease (n=91). LUNX mRNA expression was significantly more frequent in patients with all carcinomas, squamous cell carcinomas, adenocarcinomas, as well as patients with central, peripheral and diffuse carcinomas (P<0.01), and non-small cell lung carcinomas (P<0.05) compared with patients with benign disease. The diagnostic performance of RT-PCR analysis of LUNX mRNA was significantly better than that of cytology in terms of sensitivity (93.3±4.8% vs 64.4±9.2%), negative predictive value (91.6±6.0% vs 71.1±7.9%) and accuracy (88.7±4.4% vs 81.0±5.5%), whereas specificity (83.5±7.6%) and positive predictive value (86.6±6.3%) were lower than those of cytology (100%). Liquid-based cytology and RT-PCR can be performed to detect LUNX mRNA expression in bronchial brushing specimens, and this technique may be a useful adjunct to cytological diagnosis of lung cancer. The sensitivity of the technique was greater than that of cytology but its lower specificity should be taken into account.

Factors associated with histologic versus cytologic diagnosis of lung cancer.

6048 Background: In the era of molecularly targeted agents, optimal lung cancer treatment decision making requires high quality pathology, including histologic rather than cytologic results. If there are differences in the likelihood of histologic diagnosis by race, ethnicity, or socioeconomic status (SES), it could contribute to disparities in lung cancer treatment and outcomes. Methods: We evaluated the nature of the pathologic diagnosis among 3760 lung cancer patients enrolled in CanCORS (the Cancer Care Outcomes and Research Consortium), a multi-region population- and health system-based prospective cohort study. Patients were diagnosed in 2002-2005. Patients were classified as having a histologic or a cytologic diagnosis, on the basis of procedures performed from 3 months before to 3 months after diagnosis, abstracted from medical records. A multivariable generalized logistic regression model was fitted to evaluate the association of clinical, sociodemographic, and health system factors on type of diagnosis, conditional on patients’ not having primary surgery for lung cancer. Results: Overall, 44% of patients had a procedure yielding histology (30% with primary surgery and 14% without). Race, SES, and gender were not significantly associated with having a histologic diagnosis. Older patients were significantly less likely to have histology (in reference to age < 65: odds ratio [OR] 0.70 (95% CI 0.51-0.96) for age 65-74 and OR 0.65 (95% CI 0.46, 0.91). for age 75+) In a secondary analysis substituting geographic region for data collection site, we found that patients treated in the South and Midwest (compared to the West) and those receiving care in staff model HMOs were more likely to have cytology only. Conclusions: In the first several years after the approval of targeted agents for lung cancer, the majority of patients were not receiving histologic diagnoses. We did not identify disparities in care by race or ethnicity; but other non-clinical factors including region of the country and health system characteristics appeared to play a role. Future studies should consider the quality of the pathologic information in evaluating disparities in lung cancer treatment and outcomes.

Cytological Diagnosis of Lung Cancer: The Diagnostic Accuracy and Pitfalls

The role of respiratory cytology is to detect and classify pulmonary disease, with an emphasis of neoplastic disease, so that proper therapy can be instituted. As in many branches of cytology, the recognition of malignancy in the cells obtained from the respiratory tract is more straightforward than identifying the type of tumor cell. It is important to accurately determine the true cytopathological cell type in cases of primary lung cancer and to know the accuracy of the diagnosis achieved by the cytological procedures. The well differentiated tumors have characteristic cytoplasmic and nuclear abnormalities that enable physicians to firmly categorize these lesions, as in squamous cell or adenocarcinoma, but some moderately and most poorly differentiated tumors show few distinctive features. This article reviews the malignant and reactive pulmonary cytologic findings and we also report on some of their pitfalls and the cytologic criteria.

EGFR expression as an ancillary tool for diagnosing lung cancer in cytology specimens

Lung cancer evolves in a multistep process, and its early detection portends a better prognosis. Bronchial washings/brushings and fine-needle aspirations are often used as early screening and cytological diagnosis of lung cancer. In some cases, it is difficult to differentiate morphologically malignant from reactive cells. Epidermal growth factor receptor (EGFR) is a transmembrane receptor overexpressed in high percentage lung cancers, and contributes to tumor growth. Assessing EGFR expression levels by fluorescence in situ hybridization (FISH) and immunohistochemistry (IHC) may provide critical information of tumor marker abnormalities, assist in the cytological diagnosis, and stratify patients for EGFR inhibitor therapy. Fifty patients with bronchial washings/brushings or fine-needle aspiration specimens, and corresponding histologically confirmed lung biopsies, were studied for EGFR expression with FISH and IHC. Copy numbers of the EGFR gene locus were analyzed with those of chromosome 7 by FISH. EGFR and FISH results were compared to our FISH data with combined EGFR, c-myc, 5p15.2, and chromosome 6 probes in selected cases. Cell blocks, if available, and tissue biopsy sections were used for EGFR IHC. The intensity of IHC was scored, and quantified. Only balanced aneuploidy of EGFR was identified by FISH. Gene amplification was not detected. The chromosomal abnormalities of EGFR were often accompanied by other chromosomal aneuploidies demonstrated in c-myc (8q24), 5p15.2 or 6p, indicating a general genomic instability. About half of the specimens with confirmed malignancy showed EGFR balanced aneuploidy by FISH, and gene copy number was not coupled with protein expression in many cases. The benign or reactive cytology specimens confirmed by biopsies had high specificity by FISH (96%) and IHC (88%). FISH and IHC analysis of EGFR, possibly along with other tumor markers, may be a useful ancillary tool to classify difficult cytology cases and inform clinicians arranging targeted chemotherapy.

516 POSTER Cap43/NDRG1 is a molecular marker of angiogenesis and prognosis in cervical adenocarcinoma

Lung cancer specimens display recurrent copy number aberrations in distinguished chromosomal regions as compared with normal lung cells. Such alterations have been utilized in design of fluorescence in situ hybridization (FISH) probe sets in attempts to improve the cytological diagnosis of lung cancer. One of such probe sets, LAVysion™, detects copy number changes in the centromeric region of chromosome 6 (CEP6), and regions 5p15, 8q24, and 7p12, often gained in lung cancer.We evaluated the feasibility of the LAVysion™ multi-color probe set in detection of individuals at high risk of lung cancer by applying the FISH probe set on smears prepared of induced sputa obtained from 20 lung cancer patients, 43 asbestos-exposed workers, 21 heavy tobacco smokers, and 15 healthy never-smokers. The hybridized sputum smears were examined using fluorescence microscopy and the number of signals in epithelial cells was examined throughout the hybridized area. Additionally, we review here the previous studies using LAVysion™ probe set.The FISH probe set was slightly more sensitive than cytology alone in detecting lung cancer. No significant differences in copy number gain were found between high-risk individuals and healthy never-smokers. The proportions of individuals with copy number gains in sputa among the lung cancer patients, asbestos-exposed workers, tobacco smokers, and never-smokers were 50, 20, 12, and 27%, respectively, when three or more cells with a copy number gain detected by at least two different probes was used as the cut-off point. In comparison, the sensitivity of cytology in detecting lung cancer was 44%. In the lung cancer patients the number of abnormal cells by FISH correlated well with the cytologic atypia class (Spearman rank correlation coefficient 0.77, p < 0.01). Using multivariant variance analysis, gains in CEP6, 5p15, 8q24 and 7p12 were not associated with smoking or asbestos exposure.FISH did not significantly exceed the sensitivity of sputum cytology in finding lung cancers. Significant differences were not found between sputa of asbestos-exposed individuals, heavy-smokers and never-smokers. More sensitive methods are needed for the follow-up of populations at high risk of contracting lung cancer.

Reliability of cytologic diagnosis of early lung cancer

Baseline screening for lung cancer of 2968 high-risk men and women utilizing HRCT enrolled in ELCAP (Early Lung Cancer Action Project) was performed between 1993-2002. Among them, 65 people had surgical resection of their screen-diagnosed lung cancer, 53 of them on the basis of a diagnosis of malignancy or atypical bronchioloalveolar proliferation (ABP) on fine needle aspiration (FNA) biopsy at Weill Medical College of Cornell University (WMC) prior to surgery. The authors compared the diagnosis obtained from the FNA with the subsequent diagnosis from the surgical specimen to assess the reliability of a cytologic diagnosis of lung cancer on FNA of these screen-diagnosed lung cancers. The FNA biopsies were performed with a 22-gauge Wescott needle by 1 radiologist (D.Y.), with preliminary on-site as well as final diagnosis rendered by a cytologist (M.V., J.K.). These results were correlated with histologic diagnoses obtained as a result of consensus diagnosis by a panel of 5 expert pulmonary pathologists. Of the 53 cases of lung cancer resected following FNA, 4 were diagnosed as atypical bronchioloalveolar proliferation (ABP), 14 as adenocarcinoma with bronchioloalveolar features (ADC-BAC), 28 as adenocarcinoma, not otherwise specified (ADC-NOS), 1 as squamous cell carcinoma (SQCC), 4 as nonsmall-cell carcinoma (NSCC), and 2 as typical carcinoid. In the 49 cases with a malignant cytology and 4 cases of ABP, lung cancer was confirmed histologically. The tumor sizes ranged from 4 mm to 40 mm, mean size 13mm. The final expert panel histologic diagnosis was adenocarcinoma in 47 cases; of these, 42 were invasive (mixed subtype or acinar subtype), and 5 were a noninvasive (bronchioloalveolar carcinoma, BAC). Three of the 42 invasive adenocarcinoma that had a predominant BAC component and 1 case of BAC were diagnosed as ABP on FNA; all were sampled at the periphery of the tumor. Three of 4 cases of invasive adenocarcinoma of high nuclear grade were diagnosed as NSCC, and 1 was inaccurately classified as SQCC on FNA. One case of high-grade noninvasive BAC associated with a scar was diagnosed as NSCC on FNA. Two cases classified as nonkeratinizing SQCC and 2 cases of large cell neuroendocrine carcinoma on histology were misclassified as ADC-NOS by FNA. Two cases of typical carcinoid on histology were also diagnosed as typical carcinoid on FNA. Preoperative diagnosis of lung cancer detected by screening with HRCT could be reliably made by FNA. Difficulty in classification occurs in carcinomas of high nuclear grade with prominent nucleoli, including poorly differentiated SQCC and large cell neuroendocrine carcinoma. These are best diagnosed as NSCC on cytomorphology with further subclassification based on immunohistochemistry, which these authors generally perform on cell-block material. A diagnosis of ABP on FNA may be indicative of noninvasive BAC or an invasive adenocarcinoma with prominent BAC features, usually sampled at its periphery.

Misclassification of Bronchioloalveolar Carcinoma with Cytologic Diagnosis of Lung Cancer

Cytology is commonly used to diagnose non-small cell lung cancer (NSCLC) but is an inaccurate means of diagnosis of bronchioloalveolar carcinoma (BAC). The aims of this study were to calculate the sensitivity and specificity of cytologic diagnosis of BAC and to estimate the misclassification of BAC as other subtypes of NSCLC. Preoperative fine-needle aspiration cytology diagnoses were compared to histology diagnoses in 222 patients, including 51 patients with pure or mixed BAC, who underwent lung resection for NSCLC at our institution since 1999. The sensitivity and specificity of a cytologic diagnosis of BAC were 12% and 99%, respectively. Based on cytologic diagnosis, 63% of BAC was misclassified as adenocarcinoma, and 18% was misclassified as undifferentiated NSCLC. In this cohort, 35% of adenocarcinomas and 12% of undifferentiated NSCLC diagnosed by cytology had BAC histology. Diagnosis of NSCLC by cytology alone results in significant misclassification of BAC, most commonly as adenocarcinoma or undifferentiated NSCLC. Because patients with BAC respond differently to certain treatments such as endothelial growth factor receptor inhibitors and surgical resection of multifocal lung cancer, misclassification of BAC may have important therapeutic implications.

Misclassification of Bronchioloalveolar Carcinoma with Cytologic Diagnosis of Lung Cancer

Misclassification of Bronchioloalveolar Carcinoma with Cytologic Diagnosis of Lung Cancer

Copy number gains on 5p15, 6p11-q11, 7p12, and 8q24 are rare in sputum cells of individuals at high risk of lung cancer

Lung cancer specimens display recurrent copy number aberrations in distinguished chromosomal regions as compared with normal lung cells. Such alterations have been utilized in design of fluorescence in situ hybridization (FISH) probe sets in attempts to improve the cytological diagnosis of lung cancer. One of such probe sets, LAVysion, detects copy number changes in the centromeric region of chromosome 6 (CEP6), and regions 5p15, 8q24, and 7p12, often gained in lung cancer. We evaluated the feasibility of the LAVysion multi-color probe set in detection of individuals at high risk of lung cancer by applying the FISH probe set on smears prepared of induced sputa obtained from 20 lung cancer patients, 43 asbestos-exposed workers, 21 heavy tobacco smokers, and 15 healthy never-smokers. The hybridized sputum smears were examined using fluorescence microscopy and the number of signals in epithelial cells was examined throughout the hybridized area. Additionally, we review here the previous studies using LAVysion probe set. The FISH probe set was slightly more sensitive than cytology alone in detecting lung cancer. No significant differences in copy number gain were found between high-risk individuals and healthy never-smokers. The proportions of individuals with copy number gains in sputa among the lung cancer patients, asbestos-exposed workers, tobacco smokers, and never-smokers were 50, 20, 12, and 27%, respectively, when three or more cells with a copy number gain detected by at least two different probes was used as the cut-off point. In comparison, the sensitivity of cytology in detecting lung cancer was 44%. In the lung cancer patients the number of abnormal cells by FISH correlated well with the cytologic atypia class (Spearman rank correlation coefficient 0.77, p<0.01). Using multivariant variance analysis, gains in CEP6, 5p15, 8q24 and 7p12 were not associated with smoking or asbestos exposure. FISH did not significantly exceed the sensitivity of sputum cytology in finding lung cancers. Significant differences were not found between sputa of asbestos-exposed individuals, heavy-smokers and never-smokers. More sensitive methods are needed for the follow-up of populations at high risk of contracting lung cancer.

O-047 International multicenter randomized double-blind placebo-controlled phase 3 clinical trial of INS316 as an adjunct for acquiring deep-lung sputum specimens for the cytological diagnosis of lung cancer

O-047 International multicenter randomized double-blind placebo-controlled phase 3 clinical trial of INS316 as an adjunct for acquiring deep-lung sputum specimens for the cytological diagnosis of lung cancer

Molecular cytologic diagnosis of lung cancer on bronchial aspirates using promoter hypermethylation as tumor marker

Molecular cytologic diagnosis of lung cancer on bronchial aspirates using promoter hypermethylation as tumor marker

Improved Sputum Expectoration Following a Single Dose of INS316 in Patients with Chronic Bronchitis

Uridine 5'-triphosphate (UTP) is a naturally occurring agonist for P2Y(2) receptors on the apical surface of ciliated respiratory epithelium. UTP stimulates salt and water transport and cilia beat frequency in human airway epithelium in vitro. Single, inhaled doses of UTP stimulate mucociliary clearance in conscious, intubated sheep and in patients with mild chronic bronchitis (smokers and former smokers), suggesting that UTP may be useful for obtaining deep-lung sputum specimens suitable for diagnostic purposes. UTP is being developed for the cytologic diagnosis of lung cancer under the compound number INS316 Solution for Inhalation (Inspire Pharmaceuticals; Durham, NC). Its ability to improve the quality of expectorated sputum was tested in the current study. Placebo-controlled, double-blind, escalating two-period cross-over study. Outpatient volunteers. Twenty-six patients with mild chronic bronchitis. Patients attempted to expectorate a specimen spontaneously, following a single inhaled dose of INS316 (10 to 180 mg), and following placebo. Sputum weight, sputum cell content, spirometry, and oxyhemoglobin saturation. Only 28% of these patients were able to expectorate a macrophage-containing, deep-lung specimen spontaneously or following inhalation of placebo. In contrast, 85% of the patients were able to produce a specimen following inhalation of INS316. The average weight of the sputum expectorated was increased fourfold by placebo and 10-fold by INS316. A mild transient decrease in pulmonary function was observed following INS316 administration. A single dose of INS316 safely improves the ability of patients with mild chronic bronchitis to expectorate a deep-lung sputum specimen suitable for cytologic evaluation.

Immunocytochemical expression of tissue specific transcription factor-1 in lung carcinoma.

AIMS: To investigate the immunocytochemical expression of the tissue specific transcription factor-1 (TTF-1) on cytological specimens of small cell lung carcinoma (SCLC) and to establish its value in the cytological...

Induced sputum and cytological diagnosis of lung cancer : Khajotia RR, Mohn A, Pokieser L, Schalleschak J, Vetter N. 2nd Department of Chest Medicine, Pulmologisches Zentrum, Sanatoriumstrasse 2, A-1145 Vienna. Lancet 1991;338:976–7

Induced sputum and cytological diagnosis of lung cancer : Khajotia RR, Mohn A, Pokieser L, Schalleschak J, Vetter N. 2nd Department of Chest Medicine, Pulmologisches Zentrum, Sanatoriumstrasse 2, A-1145 Vienna. Lancet 1991;338:976–7

Induced sputum and cytological diagnosis of lung cancer

Diagnosis of lung cancer by examination of induced sputum specimens for malignant cells may be a valuable alternative to diagnosis by bronchoscopy. Patients suspected of having lung cancer were recruited and randomly distributed into two groups, one group (n=34) having sputum induced by use of an ultrasonic nebuliser before bronchoscopy, and the other (n=33) undergoing ordinary expectoration before bronchoscopy. 25 patients in the induced-sputum group were diagnosed as having primary lung cancer; induced sputum was positive for malignant cells in 21 of these patients (84%), whereas bronchoscopy was positive in 23 (92%) (not significantly different). In comparison, ordinary sputum was positive in 15 of 29 patients (52%) diagnosed as having primary lung cancer, whereas bronchoscopy was positive in 28 (97%) (p<0·001). Induction of sputum by an ultrasonic nebuliser was an effective procedure for diagnosis of primary lung cancer.

The possibility of applying a monoclonal tumor-restricted antibody in cytologic diagnosis of lung cancer in serous effusions

Fifty serous effusions resulting from a lung disease have been examined. A tumor-restricted anticarcinoma of the lung (rot a small cell lung carcinoma) MAb, termed HuTu-m3, has been used together with conventional cytology to examine the cells. The results achieved by immunocytochernical examination have been compared with cytologic results and the data obtained have been evaluated in terms of sensitivity, specificity, predictive value and test efficiency. The final analysis suggests that this technique might be helpful in the diagnosis of serous effusions by increasing the sensitivity of cancer cell cytologic detection.

Relationship between patient characteristics and the sputum cytologic diagnosis of lung cancer

In 421 patients with a malignant lung process, from whom samples of sputum of satisfactory quality were received, patient characteristics relevant to the cytologic diagnosis of malignancy were investigated. In patients with primary lung cancer, the presence of blood in the sputum was highly significant from the point of view of its association with a correct positive cytologic diagnosis on sputum. The same relationship was noted in patients with metastatic lung cancer. In patients producing bloody sputum, the examination of at least three sputum samples gave a proportion of correct positive diagnoses of 0.88 in primary lung cancer patients and of 0.77 in patients with metastatic lung disease. Furthermore, a high sensitivity of the sputum cytology diagnosis of malignancy was found in primary lung cancer patients with low forced expiratory volume values (less than 50% of the vital capacity), with large tumors (greater than 24 mm in diameter) and with squamous-cell cancers. A central location of the tumor correlated with significantly better cytodiagnostic results in patients with both primary and metastatic cancers.