Lung Cancer Prediction Models Research Articles

Radiomics in Early Lung Cancer Diagnosis: From Diagnosis to Clinical Decision Support and Education.

Lung cancer is the most frequent cause of cancer-related death around the world. With the recent introduction of low-dose lung computed tomography for lung cancer screening, there has been an increasing number of smoking- and non-smoking-related lung cancer cases worldwide that are manifesting with subsolid nodules, especially in Asian populations. However, the pros and cons of lung cancer screening also follow the implementation of lung cancer screening programs. Here, we review the literature related to radiomics for early lung cancer diagnosis. There are four main radiomics applications: the classification of lung nodules as being malignant/benign; determining the degree of invasiveness of the lung adenocarcinoma; histopathologic subtyping; and prognostication in lung cancer prediction models. In conclusion, radiomics offers great potential to improve diagnosis and personalized risk stratification in early lung cancer diagnosis through patient–doctor cooperation and shared decision making.

A Classifier for Improving Early Lung Cancer Diagnosis Incorporating Artificial Intelligence and Liquid Biopsy.

Lung cancer is the leading cause of cancer-related deaths worldwide and in China. Screening for lung cancer by low dose computed tomography (LDCT) can reduce mortality but has resulted in a dramatic rise in the incidence of indeterminate pulmonary nodules, which presents a major diagnostic challenge for clinicians regarding their underlying pathology and can lead to overdiagnosis. To address the significant gap in evaluating pulmonary nodules, we conducted a prospective study to develop a prediction model for individuals at intermediate to high risk of developing lung cancer. Univariate and multivariate logistic analyses were applied to the training cohort (n = 560) to develop an early lung cancer prediction model. The results indicated that a model integrating clinical characteristics (age and smoking history), radiological characteristics of pulmonary nodules (nodule diameter, nodule count, upper lobe location, malignant sign at the nodule edge, subsolid status), artificial intelligence analysis of LDCT data, and liquid biopsy achieved the best diagnostic performance in the training cohort (sensitivity 89.53%, specificity 81.31%, area under the curve [AUC] = 0.880). In the independent validation cohort (n = 168), this model had an AUC of 0.895, which was greater than that of the Mayo Clinic Model (AUC = 0.772) and Veterans’ Affairs Model (AUC = 0.740). These results were significantly better for predicting the presence of cancer than radiological features and artificial intelligence risk scores alone. Applying this classifier prospectively may lead to improved early lung cancer diagnosis and early treatment for patients with malignant nodules while sparing patients with benign entities from unnecessary and potentially harmful surgery.Clinical Trial Registration NumberChiCTR1900026233, URL: http://www.chictr.org.cn/showproj.aspx?proj=43370.

Assessing the association of diabetes with lung cancer risk.

BackgroundDiabetes is a well-established risk factor for many cancers, but its relationship with lung cancer incidence remains unclear. In this study, we aimed to assess if diabetes is independently associated with lung cancer risk and histology subtype among participants in a screening study.MethodsIn a retrospective cohort study using data from the Prostate, Lung, Colorectal, and Ovarian (PLCO) study, we assessed the association of self-reported diabetes with lung cancer incidence using Poisson regression while adjusting for other established risk factors in the PLCOM2012, a validated lung cancer prediction model. The adjusted association of diabetes and lung cancer cell type was evaluated using nominal regression. Stratified analyses were also conducted according to sex, smoking history, and body mass index categories.ResultsOverall, 140,395 participants were included in our analysis. Diabetes was not significantly associated with lung cancer incidence [incidence rate ratio (IRR): 1.03, 95% confidence interval (CI): 0.91–1.17]. Similarly, stratified analyses also did not show significant associations between diabetes and lung cancer risk (all P values >0.05). We found no significant difference in the distribution of lung cancer histology in participants with vs. without diabetes (P=0.30).ConclusionsDiabetes was not an independent risk factor for lung cancer in a large cohort of PLCO participants. We did not observe differences in histology according to diabetes status. These results suggest that patients with diabetes do not need more aggressive lung cancer screening. Future research including more detailed metabolic parameters may further elucidate the relationship between metabolic disease and lung cancer risk.

Performance of various risk prediction models in a large lung cancer screening cohort in Gdańsk, Poland-a comparative study.

BackgroundOptimal selection criteria for the lung cancer screening programme remain a matter of an open debate. We performed a validation study of the three most promising lung cancer risk prediction models in a large lung cancer screening cohort of 6,631 individuals from a single European centre.MethodsA total of 6,631 healthy volunteers (aged 50–79, smoking history ≥30 pack-years) were enrolled in the MOLTEST BIS programme between 2016 and 2018. Each participant underwent a low-dose computed chest tomography scan, and selected participants underwent a further diagnostic work-up. Various lung cancer prediction models were applied to the recruited screenees, i.e., (I) Tammemagi’s Prostate, Colorectal, and Ovarian Cancer Screening Trial 2012 (PLCOm2012), (II) Liverpool Lung Project (LLP) model, and (III) Bach’s lung cancer risk model. Patients (I) with 6-year lung cancer probability ≥1.3% were considered as high risk in PLCOm2012 model, (II) in LLP model with 5-year lung cancer probability ≥5.0%, and (III) in Bach’s model with 5-year lung cancer probability ≥2.0%. The particular model cut-off values were employed to the cohort to evaluate each model’s performance in the screened population.ResultsLung cancer was diagnosed in 154 (2.3%) participants. Based on the risk estimates by PLCOm2012, LLP and Bach’s models there were 82.4%, 50.3% and 19.8% of the MOLTEST BIS participants, respectively, who fulfilled the above-mentioned threshold criteria of a lung cancer development probability. Of those detected with lung cancer, 97.4%, 74.0% and 44.8% were eligible for screening by PLCOm2012, LLP and Bach’s model criteria, respectively. In Tammemagi’s risk prediction model only four cases (2.6%) would have been missed from the group of 154 lung cancer patients primarily detected in the MOLTEST BIS.ConclusionsLung cancer screening enrollment based on the risk prediction models is superior to NCCN Group 1 selection criteria and offers a clinically significant reduction of screenees with a comparable proportion of detected lung cancer cases. Tammemagi’s risk prediction model reduces the proportion of patients eligible for inclusion to a screening programme with a minimal loss of detected lung cancer cases.

Multi-Criteria Decision Support System for Lung Cancer Prediction

Lung cancer is one of the most common deadly malignant tumours, with the most rapid morbidity and death worldwide. Cancer risk prediction is a challenging and complex task in the field of healthcare. Many studies have been carried out by researchers to analyse and establish lung cancer symptoms and factors. However, further improvements are vital and required to be conducted in order to overcome the persistent challenges. In this study, a multi-criteria decision support system for lung cancer risk prediction based on a web-based survey data has been presented and realised. The proposed framework aims to incorporate the powerful of analytical hierarchy process (AHP) with artificial neural network for constituting lung cancer prediction model. The multiple criteria decision-making strategy (AHP) assigns a weight to each individual cancer symptom feature from survey data. The weighted features are then used to train multi-layer perceptron artificial neural network (ANN) to build a disease prediction model. Experimental analysis and evaluation performed on 276 subjects revealed promising prediction performance of developed lung cancer prediction framework in terms of various classification metrics.

A prediction model to evaluate the pretest risk of malignancy in solitary pulmonary nodules: evidence from a large Chinese southwestern population.

Lung cancer is the leading cause of cancer death and there have been clinical prediction models. This study aimed to evaluate the diagnostic performance of published models and create new models to evaluate the probability of malignant solitary pulmonary nodules (SPNs) in Chinese population. We consecutively enrolled 2061 patients with SPNs from West China Hospital between January 2008 and December 2016, each SPN was pathologically confirmed. First, four published prediction models, Mayo clinic model, Veterans Affairs (VA) model, Brock model and People's Hospital of Peking University (PEH) model were validated in our patients. Then, utilizing logistic regression, decision tree and random forest (RF), we developed three new models and internally validated them. Area under the receiver operating characteristic curve (AUC) values of four published models were as follows: Mayo 0.705 (95% CI 0.658-0.752, n = 726), VA 0.64 6 (95% CI 0.598-0.695, n = 800), Brock 0.575 (95% CI 0.502-0.648, n = 550) and PEH 0.675 (95% CI 0.627-0.723, n = 726). Logistic regression model, decision tree model and RF model were developed, AUC values of these models were 0.842 (95% CI 0.778-0.906), 0.734 (95% CI 0.647-0.821), 0.851 (95% CI 0.789-0.914), respectively. The four published lung cancer prediction models do not apply to our population, and we have established new models that can be used to predict the probability of malignant SPNs.

Urinary fluorescent metabolite O-aminohippuric acid is a useful biomarker for lung cancer detection.

Urine contains diagnostically important metabolites that can act as natural fluorophores. However, whether these fluorescent metabolites can be used in lung cancer diagnosis is unknown. This study was conducted to determine whether fluorescent urinary metabolites could be useful biomarkers for lung cancer detection. A total of 46 lung cancer patients and 185 volunteers without cancer were evaluated between November 2013 and November 2014. Samples of the first urine of the day were collected from lung cancer patients and diagnosed at the Hamamatsu University School of Medicine and the Hamamatsu Medical Center prior to cancer treatment, and from volunteers without cancer at the Hamamatsu Medical Imaging Center. Fluorescent urinary metabolites were screened by high-performance liquid chromatography and select effective fluorescent substances for distinguishing cancer from non-cancer status. The fraction of patients at each stage of cancer severity were: 41.3% stage I, 8.7% stage II, 19.6% stage III, and 30.4% stage IV. A robust predictive biomarker for lung cancer was selected by the multivariate logistic analysis of fluorescent metabolites and identified to be O-aminohippuric acid (OAH). The area under the curve (AUC) data for OAH was 0.837 (95% CI 0.769-0.898, P < 0.001). We identified a fluorescent urinary metabolite that can predict lung cancer. OAH exceeds the AUC (0.817) of lung cancer detection by AminoIndex® cancer screening, can be analyzed non-invasively without additional sample processing, and may be a valuable addition to existing lung cancer prediction models.

Korean National Lung Cancer Screening

The results of large-scale clinical studies have shown that the lung cancer mortality rate can be reduced by lung cancer screening using low-dose computed tomography (LDCT) in high-risk populations. Lung cancer screening requires rigorous quality control to ensure that imaging can be introduced into evidence-based medical systems and that results can be effectively delivered to examinees. Cessation of smoking is indispensable for reducing mortality in parallel with lung cancer screening. Pulmonary nodules found in LDCT during the Korean National Lung Cancer Screening are categorized according to their characteristics, size, and time of discovery based on the Lung Imaging Reporting And Data System (Lung-RADS); management guidelines are followed according to categorization. To improve the efficiency of lung cancer screening, studies are currently ongoing to enable selection of high-risk groups using lung cancer prediction models and biomarkers. Based on the risk estimation classification of lung cancer, it is expected that the selection of screening subjects and the screening cycle can be differentiated, which will increase the efficiency of screening, reduce the risk of unnecessary radiation exposure, and reduce the cost of screening.

Predictive model for the diagnosis of benign/malignant small pulmonary nodules.

There is some doubt that all nodules <8 mm are really mainly benign and that simple follow-up is adequate in all cases. The purpose of this study is to create a predictive model for the diagnosis of benign and malignant small pulmonary nodules.This was a retrospective case–control study of patients who had undergone pulmonary nodule resection at the Zhejiang University Jinhua Hospital. Patients with pulmonary nodules of ≤10 mm in size on chest high-resolution computed tomography were included. Patients’ demographic characteristics, clinical features, and high-resolution computed tomography findings were collected. Logistic regression and receiver-operating characteristic analysis were used to create a predictive model for malignancy.A total of 216 patients were included: 160 with malignant and 56 with benign nodules. Nodule density (odds ratio [OR] = 0.996, 95% confidence interval [CI]: 0.993–0.998, P = .001), vascular penetration sign (OR = 3.49, 95% CI: 1.39–8.76, P = .008), nodule type (OR = 4.27, 95% CI: 1.48–12.29, P = .007), and incisure surrounding nodules (OR = 0.18, 95% CI: 0.04–0.84, P = .03) were independently associated with malignant nodules. These factors were used to create a mathematical model that had an area under the receiver-operating characteristic curve of 0.744. Using a cut-off of 0.762 resulted in 63.1% sensitivity and 75.0% specificity.This study proposes a pulmonary nodule prediction model that can estimate benign/malignant lung nodules with good sensitivity and specificity. Mixed ground-glass nodules, vascular penetration sign, density of lung nodules, and the absence of incisure signs are independently associated with malignant lung nodules.

Determinants of impaired lung function and lung cancer prediction among never-smokers in the UK Biobank cohort.

BackgroundImpaired lung health represents a significant burden on global health, including chronic obstructive pulmonary disease (COPD) and lung cancer. Given its global health impact, it is important to understand the determinants of impaired lung function and its relation to lung cancer risk independent of smoking. However, to date, no study has evaluated determinants of impaired lung function in a cohort exclusively of never-smokers, who also represent a growing proportion of all lung cancers.MethodsA total of 222,274 never-smokers with reproducible spirograms were identified in the UK Biobank population-based cohort and included in the analysis. Baseline volumetric measures of lung function, including forced expiratory volume in 1-s (FEV1) and forced vital capacity (FVC), were used to define lung function impairment. Determinants of impaired lung function were evaluated using Poisson regression with robust variance estimation. The added value of lung function in lung cancer prediction was evaluated using Fine and Gray regression accounting for the competing risk of all-cause mortality.FindingsLung function impairment was associated with low birthweight, ambient air pollution (PM2·5 μg/mm3), and overweight, after adjustment for other important risk factors. We observed modest improvement in discrimination by adding lung function to our lung cancer prediction model for never-smokers. The highest optimism-corrected AUC at 3 (0·700, 95% CI: 0·654–0·734) and 5 years (0·694, 95% CI: 0·658–0·736) included FEV1 (% of GLI predicted FEV1), while the highest AUC at 7 years was based on the inclusion of FEV1/FVC (0·722, 95% CI: 0·687–0·762).InterpretationWe identified several modifiable risk factors associated with increased risk of lung function impairment among lifetime never-smokers in UKB. We achieved moderate discrimination for lung cancer risk-prediction for never-smokers, and found modest improvement with the inclusion of lung function.FundThis study was supported by a Canada Research Chair to RJH.

Accuracy Enhanced Lung Cancer Prognosis for Improving Patient Survivability Using Proposed Gaussian Classifier System

Statistical classifier and good accuracy is an essential part of the research in medical data mining. Accurate prediction of lung cancer is an essential step for making effective clinical decisions. After identifying the lung cancer, minimum scopes are available in the medications for patient living in the world. Hemoglobin level and TNM stage wise patients survival period has to be varied. Some group of people survival period is minimal and another group of people survival time is lengthy. This study is aimed to develop a prediction model with new clinical variables to predict lung cancer patients. It's based on revised 8th edition study of TNM in lung cancer. These new attributes are collected from SEER databases, Indian cancer hospitals and research centers. The collected new attributes are classified using supervised machine learning algorithms of linear regression, Naïve Bayes classifier and proposed algorithms of Gaussian K-Base NB classifier. In particular, for TNM stage 1 group of people with normal hemoglobin level (NHBL), that group of lung cancer patient quality of life is highly enhanced. Which proved by using supervised machine learning algorithms. The proposed algorithm classified the database in terms of with respect to tumor size and HB level and the results are confirmed in the R environment. The continuous attribute classification method to prove first level of TNM in lung cancer patient along with standard hemoglobin has to be maintained that the people survivability rate is higher than the smaller level of hemoglobin people survival rate. The Gaussian K-Base NB classifier is more effective than the existing machine learning algorithms for lung cancer prediction model. The proposed classification accuracy has measured using ROC methods.

Lung cancer prediction using machine learning and advanced imaging techniques.

Machine learning based lung cancer prediction models have been proposed to assist clinicians in managing incidental or screen detected indeterminate pulmonary nodules. Such systems may be able to reduce variability in nodule classification, improve decision making and ultimately reduce the number of benign nodules that are needlessly followed or worked-up. In this article, we provide an overview of the main lung cancer prediction approaches proposed to date and highlight some of their relative strengths and weaknesses. We discuss some of the challenges in the development and validation of such techniques and outline the path to clinical adoption.

Highly Sensitive Marker Panel for Guidance in Lung Cancer Rapid Diagnostic Units

While evidence for lung cancer screening implementation in Europe is awaited, Rapid Diagnostic Units have been established in many hospitals to accelerate the early diagnosis of lung cancer. We seek to develop an algorithm to detect lung cancer in a symptomatic population attending such unit, based on a sensitive serum marker panel. Serum concentrations of Epidermal Growth Factor, sCD26, Calprotectin, Matrix Metalloproteinases −1, −7, −9, CEA and CYFRA 21.1 were determined in 140 patients with respiratory symptoms (lung cancer and controls with/without benign pathology). Logistic Lasso regression was performed to derive a lung cancer prediction model, and the resulting algorithm was tested in a validation set. A classification rule based on EGF, sCD26, Calprotectin and CEA was established, able to reasonably discriminate lung cancer with 97% sensitivity and 43% specificity in the training set, and 91.7% sensitivity and 45.4% specificity in the validation set. Overall, the panel identified with high sensitivity stage I non-small cell lung cancer (94.7%) and 100% small-cell lung cancers. Our study provides a sensitive 4-marker classification algorithm for lung cancer detection to aid in the management of suspicious lung cancer patients in the context of Rapid Diagnostic Units.

Application of Bayesian classifier in diagnosis of lung cancer by multiple autoantibody biomarkers

Objective To establish a Bayesian classifier-based lung cancer prediction model,and to discuss its predictive efficiency.Methods Using the reaction data of previously screened 6phage peptide clones with the sera of 90lung cancer patients and 90healthy controls,we established a Bayesian classifier-based lung cancer prediction model,with the data analyzed by BinReg 2.0software.The predictive efficiencies of different models(Bayesian classifier-based prediction model, Logistic regression,principal component regression,and support vector machine)were evaluated by receiver operating characteristic(ROC)curves.Results The sensitivity and specificity of Bayesian classifier-based lung cancer prediction model were 92.00%and 96.00%,respectively.And the model satisfactorily distinguished lung cancer patients and healthy controls. Conclusion Our Bayesian classifier-based lung cancer prediction model can accurately predict the risk of lung cancer.

Abstract 3634: A new model improves lung cancer prediction in the preoperative evaluation of patients with suspicious lung lesions.

Abstract Lung cancer predictive models, such as the Mayo model (Swensen, 1997), that are used for screening and referral to surgery are based on only traditional variables and are outdated. Clinical information such as FEV1, FDG-PET and lesion growth increasingly are used in treatment decisions. This information needs to be incorporated into our clinical prediction models. We developed a lung cancer risk model utilizing the more extensive clinical information available at the point of a decision for surgery that predicts cancer better than existing models. We evaluated a lung cancer prediction model using multivariable logistic regression in a population being evaluated for lung surgery at a single academic institution in an IRB approved study. The model included non-linear relationships between continuous variables and lung cancer and used multiple imputation to handle missing data. Internal validation was conducted using the bootstrap method with 500 iterations. Area under the curve (AUC) and Brier score were calculated for the training model and internal bootstrap validation model. The characteristics of the new model were compared to the Mayo model. 492 individuals were recruited at who had been evaluated for surgery with known or suspected lung cancer. Lung cancer prevalence was 72%. Diagnosis was determined pathologically (92%) or by greater than 18 months of followup among those who didn't undergo surgery. Missing data occurred with FDG-PET scan (22%), growth on serial CT scans (13%), predicted FEV1 (10%) and pre-operative disease symptoms (7%). The remaining variables of interest had less than 5% missing data. Age (OR 1.05; 95%CI: 1.03-1.08), pack years (OR 1.03; 95%CI: 1.00-1.05), pre-operative lesion maximum diameter (OR 1.06; 95%CI: 1.04-1.08), lesion growth (OR 2.92 95%CI: 1.10-5.65), previous cancer (OR 1.86 95%CI: 1.05-3.32) and FDG-PET avidity (OR 4.56 95%CI: 2.17-9.57) predicted lung cancer (p&amp;lt;0.05). AUC for the initial model was 0.87 (95%CI: 0.84 - 0.91) and Brier score was 0.12. Bootstrap sampling estimated AUC of 0.85 and Brier score of 0.13 demonstrating internal validity of the model. The Mayo model was evaluated for those with complete data (93%) and had an AUC of 0.80 (95%CI: 0.75 - 0.85) which was significantly less (P&amp;lt;0.001) than that observed for the new model. The Mayo model generally underestimated risk and its Brier score was 0.17, showing poorer calibration than the new model. Our internally validated model performed better in distinguishing low risk from high risk patients than the Mayo model in a surgical population being evaluated for lung cancer. The Mayo model, with its more limited clinical information, underestimated risk. Future work will validate this model in an external dataset and prospectively evaluate the impact of the model on patient safety and resource utilization. Citation Format: Stephen A. Deppen, Melinda C. Aldrich, Ronald Walker, Catherine A. Necessary, Chiu-lan Chen, Huiyun Wu, Jeffery Blume, Pierre P. Massion, Theodore Speroff, Robert S. Dittus, Bill J. Putnam, Eric L. Grogan. A new model improves lung cancer prediction in the preoperative evaluation of patients with suspicious lung lesions. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3634. doi:10.1158/1538-7445.AM2013-3634

Estimating Individual Risk for Lung Cancer

Lung cancer risk prediction models hold the promise of improving patient care and streamlining research. The ultimate goal of these models is to inform clinicians as to which interventions their individual patients should receive to reduce lung cancer-associated morbidity and mortality. In this paper, we discuss the history and current state of lung cancer prediction models, focusing on three models: the Bach model, the Spitz model, and the Liverpool Lung Project (LLP) model. We also discuss the prospects for further development of improved prediction models for lung cancer risk. Although current models can identify those smokers at highest risk for lung cancer, these models are presently of limited use in the clinical setting. Nevertheless, lung cancer risk prediction models can be used during study enrollment to select more appropriate study subjects, and may eventually be useful in identifying patients for lung cancer screening or to receive chemoprevention.

Diagnosis and prediction of lung cancer through different classification techniques with tumor markers

To study which classification model was most suitable for establishing a multi-tumor markers lung cancer prediction model, through established logistic regression model, decision trees model and artificial neural network model. RIA analysis, ELISA, spectrophotometry, high-performance liquid chromatography (HPLC) and atomic absorption spectrometry were used to measure the serum CEA, CA125, gastrin, NSE, beta2-MG, Sil-6 receptors, sialic acid, nitric oxide, Cu, Zn, Ca and the pseudo-urine nucleoside of urine samples in lung cancer patients, benign lung disease patients and healthy controls. The lung cancer diagnosis models were established by logistic regression analysis, decision tree analysis and artificial neural network training. The diagnosis sensitivities of the logistic regression analysis, decision tree analysis and artificial neural network model with 12 tumor markers in lung cancer were 94.00%, 100.00% and 100.00%; the specificity were 100.00%, 98.89% and 100.00%; the total accurate 94.29%, 95.00% and 90.00%, respectively. The results of three classification models with 12 tumor markers in diagnosis of lung cancer are ideal. Especially the C5.0 decision tree model and the artificial neural network model are more suitable for the prediction and diagnosis of the lung cancer.

Development and Validation of a Lung Cancer Risk Prediction Model for African-Americans

Because existing risk prediction models for lung cancer were developed in white populations, they may not be appropriate for predicting risk among African-Americans. Therefore, a need exists to construct and validate a risk prediction model for lung cancer that is specific to African-Americans. We analyzed data from 491 African-Americans with lung cancer and 497 matched African-American controls to identify specific risks and incorporate them into a multivariable risk model for lung cancer and estimate the 5-year absolute risk of lung cancer. We performed internal and external validations of the risk model using data on additional cases and controls from the same ongoing multiracial/ethnic lung cancer case-control study from which the model-building data were obtained as well as data from two different lung cancer studies in metropolitan Detroit, respectively. We also compared our African-American model with our previously developed risk prediction model for whites. The final risk model included smoking-related variables [smoking status, pack-years smoked, age at smoking cessation (former smokers), and number of years since smoking cessation (former smokers)], self-reported physician diagnoses of chronic obstructive pulmonary disease or hay fever, and exposures to asbestos or wood dusts. Our risk prediction model for African-Americans exhibited good discrimination [75% (95% confidence interval, 0.67-0.82)] for our internal data and moderate discrimination [63% (95% confidence interval, 0.57-0.69)] for the external data group, which is an improvement over the Spitz model for white subjects. Existing lung cancer prediction models may not be appropriate for predicting risk for African-Americans because (a) they were developed using white populations, (b) level of risk is different for risk factors that African-American share with whites, and (c) unique group-specific risk factors exist for African-Americans. This study developed and validated a risk prediction model for lung cancer that is specific to African-Americans and thus more precise in predicting their risks. These findings highlight the importance of conducting further ethnic-specific analyses of disease risk.

Commentary: Early Diagnosis of Lung Cancer: Where Do We Stand?

Lung cancer is the leading cause of cancer deaths worldwide [1]; approximately 85% of all lung cancers occur in current or former cigarette smokers [2]. Recently, two important studies [3, 4] received worldwide press coverage for the discordant results reported [5–8]. The critical issue was whether spiral computed tomography (CT) scan screening could reduce lung cancer deaths in heavy cigarette smokers and could thus be considered the ultimate test of any effective cancer screening. The first study [3] reported that low-dose CT screening could lead to a therapeutic strategy that resulted in a 10-year survival rate of 88% for patients with stage I disease. The authors, in a large collaborative study, screened 31,567 asymptomatic persons at risk for lung cancer using lowdose CT; 27,456 repeated screenings were performed 7–18 months after the previous screening. They estimated the 10year lung cancer–specific survival rate among participants with clinical stage I lung cancer that was detected on CT screening and diagnosed by biopsy, regardless of the type of treatment received, and among those who underwent surgical resection of clinical stage I cancer within 1 month. Screening resulted in a diagnosis of lung cancer in 484 participants. Of these participants, 412 (85%) had clinical stage I lung cancer, and the estimated 10-year survival rate was 88% in this subgroup (95% confidence interval [CI], 84%–91%). Among the 302 participants with clinical stage I cancer who underwent surgical resection within 1 month after diagnosis, the survival rate was 92% (95% CI, 88%– 95%). The eight participants with clinical stage I cancer who did not receive treatment died within 5 years after diagnosis. The second study reported that there is no evidence that screening with chest CT scan does anything to reduce lung cancer deaths [4]. Those investigators used a validated lung cancer prediction model to estimate the expected numbers of various lung cancer outcomes among the combined cohort of 3,246 participants. To assess the effectiveness of CT screening, they then compared the observed numbers of lung cancer outcomes with the numbers of expected cases. They observed a more than threefold higher number of new lung cancer cases (144 observed versus 44.5 expected) and a tenfold higher number of lung cancer resections (109 observed versus 10.9 expected). However, there were no fewer advanced lung cancer cases (42 observed versus 33.4 expected) or lung cancer deaths (38 observed versus 38.8 expected). Several limitations of the first study [3] have to be highlighted: (a) the data used were the baseline examination and first CT repeat; (b) the specific survival rate was reported for 412 (1.3%) clinical stage I patients; (c) survival in the vast majority of subjects after 2 years is not valuable; (d) deaths related to surgery were reported within 4 weeks, whereas the standard is 30 days [9]; and (e) biopsies were performed in 535 of the 5,646 participants (9.5%) with positive results on chest CT. The limitation of the second study [4] is that the investigators used a validated lung cancer prediction model to estimate the expected numbers of various lung cancer

Elevated risk of lung cancer among people with AIDS

Lung cancer is a common malignancy among people with AIDS (PWA). Lung cancer risk was compared between PWA and the general population and its relationship with immunosuppression was assessed. The likelihood that excess risk is explained by a high prevalence of smoking was also investigated. Records on adolescent and adult PWA (N = 397 927) were linked with cancer registries in 11 US regions. Cancer risk was assessed for the period 60 months before to 60 months after AIDS onset, with specific emphasis on the period 4-27 months after onset. Observed incidence was compared with general population rates and rates from a lung cancer prediction model for smokers. Compared with the general population, lung cancer risk among PWA was elevated overall [n = 1489 cases; standardized incidence ratio (SIR), 3.8; 95% confidence interval (CI), 3.6-4.1] and in the 4-27 months after AIDS (n = 393 cases; SIR, 2.9; 95% CI, 2.6-3.2). In the 4-27 months after AIDS, risk was significantly elevated for all demographic subgroups, and was especially high among young PWA (SIRs for ages 15-29 years, 10.4; 30-39 years, 6.3; 40-49 years, 3.7). Lung cancers generally presented at an advanced stage. Risk was not associated with CD4 cell counts at AIDS (Ptrend = 0.36). Under plausible smoking assumptions, observed incidence was significantly higher than predicted among 40-49 and 50-59-year-old men with AIDS (observed/predicted = 5.03 and 1.43, respectively) and 40-49-year-old women with AIDS (observed/predicted = 1.88), but not among older PWA. Lung cancer risk was substantially elevated among PWA. Smoking could not entirely account for the observed elevation, especially among younger adults, suggesting a role for additional co-factors.