A Gene Expression Classifier from Whole Blood Distinguishes Benign from Malignant Lung Nodules Detected by Low-Dose CT

Serum laser Raman spectroscopy as a potential diagnostic tool to discriminate the benignancy or malignancy of pulmonary nodules

It has been proved that Raman spectral intensities could be used to diagnose lung cancer patients. However, the application of Raman spectroscopy in identifying the patients with pulmonary nodules was barely studied. In this study, we revealed that Raman spectra of serum samples from healthy participants and patients with benign and malignant pulmonary nodules were significantly different. A support vector machine (SVM) model was developed for the classification of Raman spectra with wave points, according to ANOVA test results. It got a good performance with a median area under the curve (AUC) of 0.89, when the SVM model was applied in discriminating benign from malignant individuals. Compared with three common clinical models, the SVM model showed a better discriminative ability and added more net benefits to participants, which were also excellent in the small-size nodules. Thus, the Raman spectroscopy could be a less-invasive and low-costly liquid biopsy.

Combinatorial Blood Platelets-Derived circRNA and mRNA Signature for Early-Stage Lung Cancer Detection

Despite the diversity of liquid biopsy transcriptomic repertoire, numerous studies often exploit only a single RNA type signature for diagnostic biomarker potential. This frequently results in insufficient sensitivity and specificity necessary to reach diagnostic utility. Combinatorial biomarker approaches may offer a more reliable diagnosis. Here, we investigated the synergistic contributions of circRNA and mRNA signatures derived from blood platelets as biomarkers for lung cancer detection. We developed a comprehensive bioinformatics pipeline permitting an analysis of platelet-circRNA and mRNA derived from non-cancer individuals and lung cancer patients. An optimal selected signature is then used to generate the predictive classification model using machine learning algorithm. Using an individual signature of 21 circRNA and 28 mRNA, the predictive models reached an area under the curve (AUC) of 0.88 and 0.81, respectively. Importantly, combinatorial analysis including both types of RNAs resulted in an 8-target signature (6 mRNA and 2 circRNA), enhancing the differentiation of lung cancer from controls (AUC of 0.92). Additionally, we identified five biomarkers potentially specific for early-stage detection of lung cancer. Our proof-of-concept study presents the first multi-analyte-based approach for the analysis of platelets-derived biomarkers, providing a potential combinatorial diagnostic signature for lung cancer detection.

Clinical Scores, Biomarkers and IT Tools in Lung Cancer Screening-Can an Integrated Approach Overcome Current Challenges?

As most lung cancer (LC) cases are still detected at advanced and incurable stages, there are increasing efforts to foster detection at earlier stages by low dose computed tomography (LDCT) based LC screening. In this scoping review, we describe current advances in candidate selection for screening (selection phase), technical aspects (screening), and probability evaluation of malignancy of CT-detected pulmonary nodules (PN management). Literature was non-systematically assessed and reviewed for suitability by the authors. For the selection phase, we describe current eligibility criteria for screening, along with their limitations and potential refinements through advanced clinical scores and biomarker assessments. For LC screening, we discuss how the accuracy of computerized tomography (CT) scan reading might be augmented by IT tools, helping radiologists to cope with increasing workloads. For PN management, we evaluate the precision of follow-up scans by semi-automatic volume measurements of CT-detected PN. Moreover, we present an integrative approach to evaluate the probability of PN malignancy to enable safe decisions on further management. As a clear limitation, additional validation studies are required for most innovative diagnostic approaches presented in this article, but the integration of clinical risk models, current imaging techniques, and advancing biomarker research has the potential to improve the LC screening performance generally.

Gene-Function-Based Clusters Explore Intricate Networks of Gene Expression of Circulating Tumor Cells in Patients with Colorectal Cancer

Colorectal cancer (CRC) is a complex disease characterized by dynamically deregulated gene expression and crosstalk between signaling pathways. In this study, a new approach based on gene-function-based clusters was introduced to explore the CRC-associated networks of gene expression. Each cluster contained genes involved in coordinated regulatory activity, such as RAS signaling, the cell cycle process, transcription, or translation. A retrospective case-control study was conducted with the inclusion of 119 patients with histologically confirmed colorectal cancer and 308 controls. The quantitative expression data of 15 genes were obtained from the peripheral blood samples of all participants to investigate cluster-gene and gene-gene interactions. DUSP6, MDM2, and EIF2S3 were consistently selected as CRC-associated factors with high significance in all logistic models. CPEB4 became an insignificant factor only when combined with the clusters for cell cycle processes and for transcription. The CPEB4/DUSP6 complex was a prerequisite for the significance of MMD, whereas EXT2, RNF4, ZNF264, WEE1, and MCM4 were affected by more than two clusters. Intricate networks among MMD, RAS signaling factors (DUSP6, GRB2, and NF1), and translation factors (EIF2S3, CPEB4, and EXT2) were also revealed. Our results suggest that limited G1/S transition, uncontrolled DNA replication, and the cap-independent initiation of translation may be dominant and concurrent scenarios in circulating tumor cells derived from colorectal cancer. This gene-function-based cluster approach is simple and useful for revealing intricate CRC-associated gene expression networks. These findings may provide clues to the metastatic mechanisms of circulating tumor cells in patients with colorectal cancer.

Development of a Molecular Blood-Based Immune Signature Classifier as Biomarker for Risks Assessment in Lung Cancer Screening

BACKGROUND

Low-dose CT (LDCT) screening trials have shown that lung cancer early detection saves lives. However, a better stratification of the screening population is still needed. In this respect, we generated and prospectively validated a plasma miRNA signature classifier (MSC) able to categorize screening participants according to lung cancer risk. Here, we aimed to deeply characterize the peripheral immune profile and develop a diagnostic immune signature classifier to further implement blood testing in lung cancer screening.

METHODS

Peripheral blood mononuclear cell (PBMC) samples collected from 20 patients with LDCT-detected lung cancer and 20 matched cancer-free screening volunteers were analyzed by flow cytometry using multiplex panels characterizing both lymphoid and myeloid immune subsets. Data were validated in PBMC from 40 patients with lung cancer and 40 matched controls and in a lung cancer specificity set including 27 subjects with suspicious lung nodules. A qPCR-based gene expression signature was generated resembling selected immune subsets.

RESULTS

Monocytic myeloid-derived suppressor cell (MDSC), polymorphonuclear MDSC, intermediate monocytes and CD8+PD-1+ T cells distinguished patients with lung cancer from controls with AUCs values of 0.94/0.72/0.88 in the training, validation, and lung cancer specificity set, respectively. AUCs raised up to 1.00/0.84/0.92 in subgroup analysis considering only MSC-negative subjects. A 14-immune genes expression signature distinguished patients from controls with AUC values of 0.76 in the validation set and 0.83 in MSC-negative subjects.

CONCLUSIONS

An immune-based classifier can enhance the accuracy of blood testing, thus supporting the contribution of systemic immunity to lung carcinogenesis.

IMPACT

Implementing LDCT screening trials with minimally invasive blood tests could help reduce unnecessary procedures and optimize cost-effectiveness.

Multiplex Analysis of CircRNAs from Plasma Extracellular Vesicle-Enriched Samples for the Detection of Early-Stage Non-Small Cell Lung Cancer

BACKGROUND

The analysis of liquid biopsies brings new opportunities in the precision oncology field. Under this context, extracellular vesicle circular RNAs (EV-circRNAs) have gained interest as biomarkers for lung cancer (LC) detection. However, standardized and robust protocols need to be developed to boost their potential in the clinical setting. Although nCounter has been used for the analysis of other liquid biopsy substrates and biomarkers, it has never been employed for EV-circRNA analysis of LC patients.

METHODS

EVs were isolated from early-stage LC patients (n = 36) and controls (n = 30). Different volumes of plasma, together with different number of pre-amplification cycles, were tested to reach the best nCounter outcome. Differential expression analysis of circRNAs was performed, along with the testing of different machine learning (ML) methods for the development of a prognostic signature for LC.

RESULTS

A combination of 500 μL of plasma input with 10 cycles of pre-amplification was selected for the rest of the study. Eight circRNAs were found upregulated in LC. Further ML analysis selected a 10-circRNA signature able to discriminate LC from controls with AUC ROC of 0.86.

CONCLUSIONS

This study validates the use of the nCounter platform for multiplexed EV-circRNA expression studies in LC patient samples, allowing the development of prognostic signatures.

Controversies and challenges in lung cancer screening

Two large randomized controlled trials have shown mortality benefit from lung cancer screening (LCS) in high-risk groups. Updated guidelines by the United State Preventative Service Task Force in 2020 will allow for inclusion of more patients who are at high risk of developing lung cancer and benefit from screening. As medical clinics and lung cancer screening programs around the country continue to work on perfecting the LCS workflow, it is important to understand some controversial issues surrounding LCS that should be addressed. In this article, we identify some of these issues, including false positive rates of low-dose CT, over-diagnosis, cost expenditure, LCS disparities in minorities, and utility of biomarkers. We hope to provide clarity, potential solutions, and future directions on how to address these controversies.

Lung cancer risk prediction models based on pulmonary nodules: A systematic review

BACKGROUND

Screening with low-dose computed tomography (LDCT) is an efficient way to detect lung cancer at an earlier stage, but has a high false-positive rate. Several pulmonary nodules risk prediction models were developed to solve the problem. This systematic review aimed to compare the quality and accuracy of these models.

METHODS

The keywords "lung cancer," "lung neoplasms," "lung tumor," "risk," "lung carcinoma" "risk," "predict," "assessment," and "nodule" were used to identify relevant articles published before February 2021. All studies with multivariate risk models developed and validated on human LDCT data were included. Informal publications or studies with incomplete procedures were excluded. Information was extracted from each publication and assessed.

RESULTS

A total of 41 articles and 43 models were included. External validation was performed for 23.2% (10/43) models. Deep learning algorithms were applied in 62.8% (27/43) models; 60.0% (15/25) deep learning based researches compared their algorithms with traditional methods, and received better discrimination. Models based on Asian and Chinese populations were usually built on single-center or small sample retrospective studies, and the majority of the Asian models (12/15, 80.0%) were not validated using external datasets.

CONCLUSION

The existing models showed good discrimination for identifying high-risk pulmonary nodules, but lacked external validation. Deep learning algorithms are increasingly being used with good performance. More researches are required to improve the quality of deep learning models, particularly for the Asian population.

The potential of using blood circular RNA as liquid biopsy biomarker for human diseases

Circular RNA (circRNA) is a novel class of single-stranded RNAs with a closed loop structure. The majority of circRNAs are formed by a back-splicing process in pre-mRNA splicing. Their expression is dynamically regulated and shows spatiotemporal patterns among cell types, tissues and developmental stages. CircRNAs have important biological functions in many physiological processes, and their aberrant expression is implicated in many human diseases. Due to their high stability, circRNAs are becoming promising biomarkers in many human diseases, such as cardiovascular diseases, autoimmune diseases and human cancers. In this review, we focus on the translational potential of using human blood circRNAs as liquid biopsy biomarkers for human diseases. We highlight their abundant expression, essential biological functions and significant correlations to human diseases in various components of peripheral blood, including whole blood, blood cells and extracellular vesicles. In addition, we summarize the current knowledge of blood circRNA biomarkers for disease diagnosis or prognosis.

Evaluation of an RNAseq-Based Immunogenomic Liquid Biopsy Approach in Early-Stage Prostate Cancer

The primary objective of this study is to detect biomarkers and develop models that enable the identification of clinically significant prostate cancer and to understand the biologic implications of the genes involved. Peripheral blood samples (1018 patients) were split chronologically into independent training (n = 713) and validation (n = 305) sets. Whole transcriptome RNA sequencing was performed on isolated phagocytic CD14+ and non-phagocytic CD2+ cells and their gene expression levels were used to develop predictive models that correlate to adverse pathologic features. The immune-transcriptomic model with the highest performance for predicting adverse pathology, based on a subtraction of the log-transformed expression signals of the two cell types, displayed an area under the curve (AUC) of the receiver operating characteristic of 0.70. The addition of biomarkers in combination with traditional clinical risk factors (age, serum prostate-specific antigen (PSA), PSA density, race, digital rectal examination (DRE), and family history) enhanced the AUC to 0.91 and 0.83 for the training and validation sets, respectively. The markers identified by this approach uncovered specific pathway associations relevant to (prostate) cancer biology. Increased phagocytic activity in conjunction with cancer-associated (mis-)regulation is also represented by these markers. Differential gene expression of circulating immune cells gives insight into the cellular immune response to early tumor development and immune surveillance.

Genome-wide gene-smoking interaction study identified novel susceptibility loci for non-small cell lung cancer in Chinese populations

Gene-smoking interactions play important roles in the development of non-small cell lung cancer (NSCLC). To identify single nucleotide polymorphisms (SNPs) that modify the association of smoking behavior with NSCLC risk, we conducted a genome-wide gene-smoking interaction study in Chinese populations. The genome-wide interaction analysis between SNPs and smoking status (ever- versus never-smokers) was carried out using genome-wide association studies (GWAS) of NSCLC, which included 13,327 cases and 13,328 controls. Stratified analysis by histological subtypes was also conducted. We used a genome-wide significance threshold of 5×10 -8 for identifying significant gene-smoking interactions and 1×10 -6 for identifying suggestive results. Functional annotation was performed to identify potential functional SNPs and target genes. We identified three novel loci with significant or suggestive gene-smoking interaction. For NSCLC, the interaction between rs2746087 (20q11.23) and smoking status reached genome-wide significance threshold (OR = 0.63, 95%CI: 0.54-0.74, P = 3.31×10 -8), and the interaction between rs11912498 (22q12.1) and smoking status reached suggestive significance threshold (OR = 0.72, 95%CI: 0.63-0.82, P = 8.10×10 -7). Stratified analysis by histological subtypes identified suggestive interactions between rs459724 (5q11.2) and smoking status (OR = 0.61, 95%CI: 0.51-0.73, P = 7.55×10 -8) in the risk of lung squamous cell carcinoma. Functional annotation indicated that both classic and novel biological processes, including nicotine addiction and airway clearance, may modulate the susceptibility to NSCLC. These novel loci provide new insights into the biological mechanisms underlying NSCLC risk. Independent replication in large-scale studies is needed and experimental studies are warranted to functionally validate these associations.

Incorporating Machine Learning into Established Bioinformatics Frameworks

The exponential growth of biomedical data in recent years has urged the application of numerous machine learning techniques to address emerging problems in biology and clinical research. By enabling the automatic feature extraction, selection, and generation of predictive models, these methods can be used to efficiently study complex biological systems. Machine learning techniques are frequently integrated with bioinformatic methods, as well as curated databases and biological networks, to enhance training and validation, identify the best interpretable features, and enable feature and model investigation. Here, we review recently developed methods that incorporate machine learning within the same framework with techniques from molecular evolution, protein structure analysis, systems biology, and disease genomics. We outline the challenges posed for machine learning, and, in particular, deep learning in biomedicine, and suggest unique opportunities for machine learning techniques integrated with established bioinformatics approaches to overcome some of these challenges.

Multiplex Detection of Clinically Relevant Mutations in Liquid Biopsies of Cancer Patients Using a Hybridization-Based Platform

BACKGROUND

With the advent of precision oncology, liquid biopsies are quickly gaining acceptance in the clinical setting. However, in some cases, the amount of DNA isolated is insufficient for Next-Generation Sequencing (NGS) analysis. The nCounter platform could be an alternative, but it has never been explored for detection of clinically relevant alterations in fluids.

METHODS

Circulating-free DNA (cfDNA) was purified from blood, cerebrospinal fluid, and ascites of patients with cancer and analyzed with the nCounter 3 D Single Nucleotide Variant (SNV) Solid Tumor Panel, which allows for detection of 97 driver mutations in 24 genes.

RESULTS

Validation experiments revealed that the nCounter SNV panel could detect mutations at allelic fractions of 0.02-2% in samples with ≥5 pg mutant DNA/µL. In a retrospective analysis of 70 cfDNAs from patients with cancer, the panel successfully detected EGFR, KRAS, BRAF, PIK3CA, and NRAS mutations when compared with previous genotyping in the same liquid biopsies and paired tumor tissues [Cohen kappa of 0.96 (CI = 0.92-1.00) and 0.90 (CI = 0.74-1.00), respectively]. In a prospective study including 91 liquid biopsies from patients with different malignancies, 90 yielded valid results with the SNV panel and mutations in EGFR, KRAS, BRAF, PIK3CA, TP53, NFE2L2, CTNNB1, ALK, FBXW7, and PTEN were found. Finally, serial liquid biopsies from a patient with NSCLC revealed that the semiquantitative results of the mutation analysis by the SNV panel correlated with the evolution of the disease.

CONCLUSIONS

The nCounter platform requires less DNA than NGS and can be employed for routine mutation testing in liquid biopsies of patients with cancer.

Analysis of extracellular vesicle mRNA derived from plasma using the nCounter platform

Extracellular vesicles (EVs) are double-layered phospholipid membrane vesicles that are released by most cells and can mediate intercellular communication through their RNA cargo. In this study, we tested if the NanoString nCounter platform can be used for the analysis of EV-mRNA. We developed and optimized a methodology for EV enrichment, EV-RNA extraction and nCounter analysis. Then, we demonstrated the validity of our workflow by analyzing EV-RNA profiles from the plasma of 19 cancer patients and 10 controls and developing a gene signature to differentiate cancer versus control samples. TRI reagent outperformed automated RNA extraction and, although lower plasma input is feasible, 500 μL provided highest total counts and number of transcripts detected. A 10-cycle pre-amplification followed by DNase treatment yielded reproducible mRNA target detection. However, appropriate probe design to prevent genomic DNA binding is preferred. A gene signature, created using a bioinformatic algorithm, was able to distinguish between control and cancer EV-mRNA profiles with an area under the ROC curve of 0.99. Hence, the nCounter platform can be used to detect mRNA targets and develop gene signatures from plasma-derived EVs.

Lung Cancer and Immunity Markers

An in-depth understanding of lung cancer biology and mechanisms of tumor progression has facilitated significant advances in the treatment of lung cancer. There remains a pressing need for the development of innovative approaches to detect and intercept lung cancer at its earliest stage of development. Recent advances in genomics, computational biology, and innovative technologies offer unique opportunities to identify the immune landscape in the tumor microenvironment associated with early-stage lung carcinogenesis, and provide further insight in the mechanism of lung cancer evolution. This review will highlight the concept of immunoediting and focus on recent studies assessing immune changes and biomarkers in pulmonary premalignancy and early-stage non-small cell lung cancer. A protumor immune response hallmarked by an increase in checkpoint inhibition and inhibitory immune cells and a simultaneous reduction in antitumor immune response have been correlated with tumor progression. The potential systemic biomarkers associated with early lung cancer will be highlighted along with current clinical efforts for lung cancer interception. Research focusing on the development of novel strategies for cancer interception prior to the progression to advanced stages will potentially lead to a paradigm shift in the treatment of lung cancer and have a major impact on clinical outcomes.See all articles in this CEBP Focus section, "NCI Early Detection Research Network: Making Cancer Detection Possible."

Noninvasive Diagnostics for Early Detection of Lung Cancer: Challenges and Potential with a Focus on Changes in DNA Methylation

Lung cancer remains the leading cause of cancer deaths in the United States and the world. Early detection of this disease can reduce mortality, as demonstrated for low-dose computed tomography (LDCT) screening. However, there remains a need for improvements in lung cancer detection to complement LDCT screening and to increase adoption of screening. Molecular changes in the tumor, and the patient's response to the presence of the tumor, have been examined as potential biomarkers for diagnosing lung cancer. There are significant challenges to developing an effective biomarker with sufficient sensitivity and specificity for the early detection of lung cancer, particularly the detection of circulating tumor DNA, which is present in very small quantities. We will review approaches to develop biomarkers for the early detection of lung cancer, with special consideration to detection of rare tumor events, focus on the use of DNA methylation-based detection in plasma and sputum, and discuss the promise and challenges of lung cancer early detection. Plasma-based detection of lung cancer DNA methylation may provide a simple cost-effective method for the early detection of lung cancer.See all articles in this CEBP Focus section, "NCI Early Detection Research Network: Making Cancer Detection Possible."

Statistical inference for net benefit measures in biomarker validation studies

Referral strategies based on risk scores and medical tests are commonly proposed. Direct assessment of their clinical utility requires implementing the strategy and is not possible in the early phases of biomarker research. Prior to late-phase studies, net benefit measures can be used to assess the potential clinical impact of a proposed strategy. Validation studies, in which the biomarker defines a prespecified referral strategy, are a gold standard approach to evaluating biomarker potential. Uncertainty, quantified by a confidence interval, is important to consider when deciding whether a biomarker warrants an impact study, does not demonstrate clinical potential, or that more data are needed. We establish distribution theory for empirical estimators of net benefit and propose empirical estimators of variance. The primary results are for the most commonly employed estimators of net benefit: from cohort and unmatched case-control samples, and for point estimates and net benefit curves. Novel estimators of net benefit under stratified two-phase and categorically matched case-control sampling are proposed and distribution theory developed. Results for common variants of net benefit and for estimation from right-censored outcomes are also presented. We motivate and demonstrate the methodology with examples from lung cancer research and highlight its application to study design.

Lung Cancer and Immunity Markers

An in-depth understanding of lung cancer biology and mechanisms of tumor progression has facilitated significant advances in the treatment of lung cancer. There remains a pressing need for the development of innovative approaches to detect and intercept lung cancer at its earliest stage of development. Recent advances in genomics, computational biology, and innovative technologies offer unique opportunities to identify the immune landscape in the tumor microenvironment associated with early-stage lung carcinogenesis, and provide further insight in the mechanism of lung cancer evolution. This review will highlight the concept of immunoediting and focus on recent studies assessing immune changes and biomarkers in pulmonary premalignancy and early-stage non-small cell lung cancer. A protumor immune response hallmarked by an increase in checkpoint inhibition and inhibitory immune cells and a simultaneous reduction in antitumor immune response have been correlated with tumor progression. The potential systemic biomarkers associated with early lung cancer will be highlighted along with current clinical efforts for lung cancer interception. Research focusing on the development of novel strategies for cancer interception prior to the progression to advanced stages will potentially lead to a paradigm shift in the treatment of lung cancer and have a major impact on clinical outcomes.See all articles in this CEBP Focus section, "NCI Early Detection Research Network: Making Cancer Detection Possible."

Stem signatures associated antibodies yield early diagnosis and precise prognosis predication of patients with non-small cell lung cancer

BACKGROUND

This study was designed to detect patients with early NSCLC with tentatively using the stem signatures associated autoantibodies (AAbs), and to evaluate its latent values in the early diagnosis and precise prognosis prediction.

METHODS

The serum concentrations of selective antibodies were quantitated by enzyme-linked immunosorbent assay (ELISA), and a total of 458 cases were enrolled (training set = 401; validation set = 57). TCGA databases were used to analyze the distinct expressions and prognostic values of related genes. The optimal cut-off values were 11.60 U/ml for P53, 4.90 U/ml for MAGEA1, 3.85 U/ml for SOX2, and 7.05U/ml for PGP9.5.

RESULTS

We found that the stem signatures associated antibodies of MAGEA1, PGP9.5, SOX2, and TP53 exhibited high expressions in NSCLC, negatively correlating with the overall survival (OS) (P < 0.05). In the test groups, the diagnosis sensitivity of P53, PGP9.5, SOX2, and MAGEA1 reached to 21.5%, 39.0%, 50.3%, and 35.0%, respectively, and the specificity reached to 98.7%, 99.4%, 92.2%, and 97.4%. The four candidates' panel gave a sensitivity of 71.8% with a specificity of 89%. In the validation group, the detection of the four antibodies in early diagnosis of NSCLC also exhibited high specificity and sensitivity, further consolidating their potential application.

CONCLUSIONS

The detection regarding stem signatures associated antibodies could be used as effective tools in early NSCLC diagnosis, but not for localized screening of cancers, and their abnormal expression was in accordance with poorer survival.

Management of incidental nodules in lung cancer screening: ready for prime-time?

Current clinical management of lung nodule patients is inefficient and therefore causes patient misclassification, which increases healthcare expenses. A precise and robust lung nodule classifier could minimise healthcare costs and discomfort for patients. http://bit.ly/2oMIEwQ.

Development and Validation of a 18F-FDG PET/CT-Based Clinical Prediction Model for Estimating Malignancy in Solid Pulmonary Nodules Based on a Population With High Prevalence of Malignancy

PURPOSE

To develop a prediction model based on ¹⁸F-fludeoxyglucose (¹⁸F-FDG) positron emission tomography/computed tomography (PET/CT) for solid pulmonary nodules (SPNs) with high malignant probability.

PATIENTS AND METHODS

We retrospectively reviewed the records of CT-undetermined SPNs, which were further evaluated by PET/CT between January 2008 and December 2015. A total of 312 cases were included as a training set and 159 as a validation set. Logistic regression was applied to determine independent predictors, and a mathematical model was deduced. The area under the receiver operating characteristic curve (AUC) was compared to other models. Model fitness was assessed based on the American College of Chest Physicians guidelines.

RESULTS

There were 215 (68.9%) and 127 (79.9%) malignant lesions in the training and validation sets, respectively. Eight independent predictors were identified: age [odds ratio (OR) = 1.030], male gender (OR = 0.268), smoking history (OR = 2.719), lesion diameter (OR = 1.067), spiculation (OR = 2.530), lobulation (OR = 2.614), cavity (OR = 2.847), and standardized maximum uptake value of SPNs (OR = 1.229). Our AUCs (training set, 0.858; validation set, 0.809) was better than those of previous models (Mayo: 0.685, P = .0061; Peking University People's Hospital: 0.646, P = .0180; Herder: 0.708, P = .0203; Zhejiang University: 0.757, P = .0699). The C index of the nomogram was 0.858. Our model reduced the diagnosis of indeterminate nodules (26.4% vs. 79.2%, 53.5%, 39.6%, and 34.0%, respectively) while improved sensitivity (81.3% vs. 16.4%, 49.2%, 62.5%, and 68.0%, respectively) and accuracy (65.4% vs. 16.4%, 39.6%, 52.8%, and 58.5%, respectively).

CONCLUSION

Our model could permit accurate diagnoses and may be recommended to identify malignant SPNs with high malignant probability, as our data pertain to a very high-prevalence cohort only.