Low-Dose Computed Tomographic Screening for Lung Cancer

Deep Learning-Based Slice Thickness Reduction for Computer-Aided Detection of Lung Nodules in Thick-Slice CT

BACKGROUND/OBJECTIVES

Computer-aided detection (CAD) systems for lung nodule detection often face challenges with 5 mm computed tomography (CT) scans, leading to missed nodules. This study assessed the efficacy of a deep learning-based slice thickness reduction technique from 5 mm to 1 mm to enhance CAD performance.

METHODS

In this retrospective study, 687 chest CT scans were analyzed, including 355 with nodules and 332 without nodules. CAD performance was evaluated on nodules, to which all three radiologists agreed.

RESULTS

The slice thickness reduction technique significantly improved the area under the receiver operating characteristic curve (AUC) for scan-level analysis from 0.867 to 0.902, with a p-value < 0.001, and nodule-level sensitivity from 0.826 to 0.916 at two false positives per scan. Notably, the performance showed greater improvements on smaller nodules than larger nodules. Qualitative analysis confirmed that nodules mistaken for ground glass on 5 mm scans could be correctly identified as part-solid on the refined 1 mm CT, thereby improving the diagnostic capability.

CONCLUSIONS

Applying a deep learning-based slice thickness reduction technique significantly enhances CAD performance in lung nodule detection on chest CT scans, supporting the clinical adoption of refined 1 mm CT scans for more accurate diagnoses.

Distinct roles of miR-34 family members on suppression of lung squamous cell carcinoma

miR-34, whose mimic was used on phase I clinical trial, has been extensively reported since its dysfunction in various cancers including non-small-cell lung cancer (NSCLC). However, the roles of miR-34 family members in the progression of lung squamous carcinoma (SCC) in patients who have occupational-exposure experience are unclear yet. Here, we comprehensively investigated the expression levels of miR-34 family members in SCC patients and compared the roles of them in SCC in vitro and vivo. The results showed that the average levels of miR-34a and miR-34b/c were decreased in patients. The analysis of miR-34a to miR-34b/c levels in patients graded different stages or metastases or recurrence showed that miR-34b/c was reduced earlier and more significantly than miR-34a. In vitro assays demonstrated that both miR-34a and miR-34b/c inhibits SCC cells proliferation, migration and invasion via Notch1 pathway, while miR-34b/c effects more than miR-34a does. As miR-34a was significantly decreased in cancer recurrence, the further analysis of relationship between miR-34a and stem cell adhesion molecular CD44 showed that miR-34a was significantly correlated with CD44 levels in patients. Knockdown of CD44 significantly blocked miR-34a mediated inhibition of cell migration and invasion. Treating the purified CD44^hi cells with miR-34 overexpression lentivirus inhibited the tumor outgrowth. By contrast, anti-miR-34 facilitated tumor development of CD44^low cells. Our study showed that miR-34 family members are negative regulator for SCC development, even though the inhibition is mediated by multiple and complicated signal pathways, which provides theoretical basis for SCC treatment and a biomarker candidate for SCC prognosis.

Risk stratification model for patients with stage I invasive lung adenocarcinoma based on clinical and pathological predictors

Background

The aim of this study was to propose a new kind of pathological classification and further establish a prognostic model for resected stage I invasive adenocarcinoma (IADC).

Methods

Clinicopathological data were collected from 2 hospitals. The new proposed pathological reclassification was defined according to certain subtype instead of a predominant one. Survival curves were plotted by Kaplan-Meier analysis. Cox regressions were analyzed for recurrence-free survival (RFS) and overall survival (OS), through which prognostic scores and stratification models were established. The comparison between risk models and the eighth edition of tumor, node, metastasis (TNM) classification was conducted through receiver operating characteristic curves (ROC), as identified by the area under the curve (AUC) and z test.

Results

In all, 1,196 patients were enrolled. At multivariable analysis, solid and micropapillary of the new pathological reclassification, along with stage IA3 and IB were independent predictors for poorer RFS. Stage IB and smoking status significantly indicated worse OS. After normalization and standardization of log-hazard ratio (HR), personalized scores were calculated and the risk stratifications with 3 risk groups were generated. Compared with TNM classification, the risk model of RFS showed advantage over early-recurrence prediction (1-year: 0.653 vs. 0.556, P=0.033; 3-year: 0.663 vs. 0.076, P=0.008). No marked difference was observed in long-term RFS or OS.

Conclusions

Considering the harboring of certain patterns may be a new concept in adenocarcinoma classification. The risk stratification model based on this pathological classification and the eighth TNM classification showed remarkable superiority over TNM alone in predicting early recurrence of stage I adenocarcinoma. However, TNM classification remained valuable for long-term recurrence and survival prediction.

Reviewing Lung Cancer Screening: The Who, Where, When, Why, and How

Lung cancer screening with annual low-dose computed tomography (CT) decreases lung cancer mortality in high-risk patients, as defined by smoking history (> 30 pack-years) and age (55-74 years). Risks to screening include overdiagnosis, anxiety about indeterminate nodules, and radiation exposure. To be effective, lung cancer screening must combine individualized risk assessment, shared decision-making, smoking cessation, structured reporting, high quality and multi-specialty cancer care, and reliable follow-up; a multidisciplinary approach is crucial. Specialty organizations have outlined both the components of high quality lung cancer screening programs and the proposed metrics that programs should track. Long-term outcomes of lung cancer screening in the general population, further refinement of who to screen, and use of biomarkers for early cancer detection are ongoing research questions.

Choice of the surgical approach for patients with stage I lung squamous cell carcinoma ≤3 cm

Background

We tried to explore the surgical procedures for stage I squamous cell carcinoma (SCC) with a size of ≤3 cm by using the Surveillance, Epidemiology, and End Results (SEER) database. Furthermore, we investigated the relationships between the chosen surgical option and the size of SCC.

Methods

In total, 1,147 patient data sets were collected from 2010 to 2011 using the SEER database. Afterwards, 849 patients with a pT1-2aN0M0 SCC with a size of ≤3 cm after a lobectomy or sublobectomy procedure were identified. Kaplan-Meier curves were conducted to compare the overall survival (OS) rates and the lung cancer-specific survival (LCSS) rates between the two surgical approaches. Cox proportional hazards regressions were performed to discover the independent risk factors for both the OS and LCSS rates. Lastly, subgroup analysis was stratified by the size of the SCC and then classified by the 8^th edition T category.

Results

The sublobectomy procedure did not demonstrate a difference for the OS rate. Additionally, it demonstrated a worse LCSS rate when compared with a lobectomy for stage I SCC. In the subgroup analysis, a lobectomy was shown to have a better survival outcome only when the SCC was >2 and ≤3 cm. Multivariable analysis showed that a size of >2 to ≤3 cm, and an age of >60 were independently associated with poorer OS while the sublobectomy procedure and pleural invasions (PI) were related with a poorer LCSS rate. In the stratification of data for the tumor size, the cox proportional analysis still confirmed the protective effects of the lobectomy in subgroups of SCCs with sizes between >2 to ≤3 cm as well as the T1c category.

Conclusions

The choice of the SCC surgery can be recommended based on the tumor size. A lobectomy procedure demonstrated a better LCSS against the sublobectomy in stage I SCC. SCC with sizes of >2 to ≤3 cm could become a pretty good indicator for lobectomy, while a sublobectomy may be an adequate substitute when the SCC size is ≤2 cm, especially for patients who cannot tolerate a lobectomy. T1c category can also suggest a lobectomy instead of sublobectomy for stage I SCC patients.

Predicting prognosis of post-chemotherapy patients with resected IIIA non-small cell lung cancer

Background

Recently, nomogram has been widely used in cancer prognoses. However, the predicting model for post-chemotherapy patients with resected IIIA non-small cell lung cancer (NSCLC) still remains scarce. Here, we tried to develop nomograms for predicting the recurrence and survival of these patients.

Methods

We retrospectively analyzed our database from October 2007 to May 2013 at Fudan University Shanghai Cancer Center. 437 qualified patients were included. Univariable and multivariable analyses of cox regression were performed successively to select prognostic factors and nomograms for recurrence-free survival (RFS) and overall survival (OS) were developed. Concordance indexes (C-index) and calibration curves were created to measure the consistency between predicted and actual survivals. Finally, risk group stratifications according to risk scores calculated from nomograms were delineated.

Results

With a total of 437 patients, five independent prognostic factors related to RFS and two to OS were selected to develop nomograms, respectively. Both 3- and 5-year RFS and OS calibration curves indicated a moderate concordance between the predicted and actual outcomes, consisted with the C-index 0.656 (95% CI: 0.626-0.687) for RFS and 0.651 (95% CI: 0.611-0.691) for OS. Different risk groups showed significant differences in RFS and OS.

Conclusions

We developed nomograms of RFS and OS for predicting recurrence and survival of post-chemotherapy patients with resected IIIA NSCLC. These nomograms could help doctors more easily estimate the prognosis and choose optimal decisions for individual during clinical practices.

Predictors of recurrence and survival of pathological T1N0M0 invasive adenocarcinoma following lobectomy

BACKGROUND

This retrospective research was designed to investigate the relationship between pT1N0M0 invasive adenocarcinoma (IADC) harboring solid (SOL) and/or micropapillary (MIP) components and its prognosis following lobectomy.

METHODS

Clinical data of pT1N0M0 IADC patients were retrospectively collected from Shanghai Chest Hospital. Survival curves were plotted by Kaplan-Meier methods. Multivariable cox regressions were conducted to discover the independent risk factors of recurrence-free survival (RFS) and overall survival (OS), through which nomograms were performed to visualize the risk of recurrences and outcomes in personalized information.

RESULTS

Totally, 1965 patients were enrolled, including 248 harboring SOL/MIP and 1717 not. IADC demonstrated worse 5-year RFS (81.9 vs. 92.2%, p < 0.001) and OS (85.7 vs. 94.4%, p < 0.001) when harboring SOL and/or MIP components. And this status became an independent factor associated with poorer RFS (HR 2.445, 95% CI 1.565-3.821, p < 0.001) and OS (HR 2.139, 95% CI 1.180-3.878, p = 0.012) instead of novel classification of IADC predominant patterns. No difference existed between SOL/MIP predominant and minor patterns. In addition, age > 60, smoking, post-chemotherapy and T1b were all indicating poorer RFS and smoking was also related with worse OS. The c-indexes of nomograms were 0.723 for RFS (95% CI, 0.662-0.784) and 0.703 for OS (95% CI, 0.629-0.777) respectively.

CONCLUSIONS

Once the pT1N0M0 IADC harboring SOL/MIP, it strongly indicated the worse clinical recurrence and survival outcome, no matter whether the SOL and/or MIP was predominant. Smoking was correlated with worse prognosis for those patients. Age > 60 and stage T1b also indicated poorer RFS. Whether post-chemotherapy was harmful to pT1N0M0 IADC patients needed further research.

The indication of completion lobectomy for lung adenocarcinoma ≤3 cm after wedge resection during surgical operation

PURPOSE

This retrospective study was designed to find out the potential indications of completion lobectomy (CL) during wedge resection (WR) operation among patients with lung adenocarcinoma (ADC) ≤3 cm, by the use of Shanghai Chest Hospital Lung Cancer Database.

PATIENTS AND METHODS

There were totally 1938 patients in this study, including 746 WRs and 1192 CLs. The propensity score matching (PSM) was performed to minimize the effect of confounders. Univariable and multivariable cox regressions were analyzed to discover the independent risk factors of recurrence-free survival (RFS) and overall survival (OS). Subgroup analysis and Kaplan-Meier survival curves were performed if necessary.

RESULTS

The 5-year RFS (86.1 vs 91.5%, p = 0.001 for unmatched group; 84 v 92%, p < 0.001 for PSM group) and OS (83.6 vs 91.7%, p < 0.001 for unmatched group; 81.6 vs 88.2%, p < 0.001 for PSM group) all indicated a better prognosis when conducting CL. Subgroup analysis suggested that WR was appropriate for non-invasive ADC. Three prognostic factors (sex, surgical approach and pleural invasion) were correlated with RFS and two (sex and surgical approach) corresponded with OS in invasive ADC through multivariable analysis. Non-lepidic-predominant component showed a better RFS and OS when CL was operated after WR in the subgroup of invasive ADC.

CONCLUSION

CL was an appropriate remediation to WR when the existence of invasive ADC, especially non-lepidic-predominant one. While WR could be applied if non-invasive ADC was confirmed. Whether lepidic-predominant adenocarcinoma was fit for WR needed further study.

Predicting the recurrence risk factors and clinical outcomes of peripheral pulmonary adenocarcinoma ≤3 cm with wedge resection

PURPOSE

This study was designed to investigate the risk factors of recurrence and survival of clinical stage I lung adenocarcinoma underwent wedge resection by the use of Shanghai Chest Hospital Lung Cancer Database.

PATIENTS AND METHODS

A total of 746 patients with clinical stage I adenocarcinoma underwent wedge resection from 2010 to 2015 in our database were included in this study. Univariable and multivariable Cox proportional hazards regression were performed successively to select significant risk factors and then nomograms as well as the concordance indexes for RFS, OS and LCSS were developed, respectively. Kaplan-Meier survival curves were performed if necessary, with the identification of log-rank test.

RESULTS

The 5-year RFS, OS and LCSS of clinical stage I adenocarcinoma underwent wedge resection were 86.1, 83.6 and 85.2%, respectively. There were three independent risk factors related with RFS (sex, pathology, pleural invasion), two related with OS (sex, volume ratio) and two with LCSS (sex, volume ratio) with the analysis of Cox regression and were selected to develop nomograms. The C-indexes of RFS, OS and LCSS were 0.767 (95% CI 0.667-0.867), 0.782 (95% CI 0.660-0.904) and 0.794 (95% CI 0.669-0.919), respectively. Lymphadenectomy did not show differences statistically but had tendencies of better RFS, OS and LCSS among the subgroup of invasive adenocarcinoma.

CONCLUSION

Sex, pathology and pleural invasion could be recommended as criteria for clinical stage I adenocarcinoma undergoing wedge resection. And the larger the wedge volume and/or the smaller the tumor volume was, the better OS and LCSS were. If the volume ratio reached 10:1 or more, the survival rate was approximately 90% for both OS and LCSS. Whether lymphadenectomy was necessary for WR, especially in invasive adenocarcinoma, needed further research.

Associations of Radiomic Data Extracted from Static and Respiratory-Gated CT Scans with Disease Recurrence in Lung Cancer Patients Treated with SBRT

Radiomics aims to quantitatively capture the complex tumor phenotype contained in medical images to associate them with clinical outcomes. This study investigates the impact of different types of computed tomography (CT) images on the prognostic performance of radiomic features for disease recurrence in early stage non-small cell lung cancer (NSCLC) patients treated with stereotactic body radiation therapy (SBRT). 112 early stage NSCLC patients treated with SBRT that had static free breathing (FB) and average intensity projection (AIP) images were analyzed. Nineteen radiomic features were selected from each image type (FB or AIP) for analysis based on stability and variance. The selected FB and AIP radiomic feature sets had 6 common radiomic features between both image types and 13 unique features. The prognostic performances of the features for distant metastasis (DM) and locoregional recurrence (LRR) were evaluated using the concordance index (CI) and compared with two conventional features (tumor volume and maximum diameter). P-values were corrected for multiple testing using the false discovery rate procedure. None of the FB radiomic features were associated with DM, however, seven AIP radiomic features, that described tumor shape and heterogeneity, were (CI range: 0.638-0.676). Conventional features from FB images were not associated with DM, however, AIP conventional features were (CI range: 0.643-0.658). Radiomic and conventional multivariate models were compared between FB and AIP images using cross validation. The differences between the models were assessed using a permutation test. AIP radiomic multivariate models (median CI = 0.667) outperformed all other models (median CI range: 0.601-0.630) in predicting DM. None of the imaging features were prognostic of LRR. Therefore, image type impacts the performance of radiomic models in their association with disease recurrence. AIP images contained more information than FB images that were associated with disease recurrence in early stage NSCLC patients treated with SBRT, which suggests that AIP images may potentially be more optimal for the development of an imaging biomarker.

Timeliness of Recommended Follow-Up After an Abnormal Finding on Diagnostic Chest CT in Smokers at High Risk of Developing Lung Cancer

PURPOSE

Diagnostic chest CT frequently results in abnormal findings that require follow-up. We assessed the timeliness of follow-up after CT abnormalities were identified in symptomatic smokers at high risk for developing lung cancer.

METHODS

In an academic primary care network, we identified current smokers aged 55-79 years who received a diagnostic chest CT to evaluate symptoms during 2012. Medical chart abstraction identified radiologist recommendations and follow-up care. The outcome was the proportion of patients who received timely follow-up (within 30 days of recommendation) after an abnormal chest CT. We assessed for predictors of compliance with recommended follow-up.

RESULTS

Of 3,257 eligible smokers, 446 (14%) had a chest CT during 2012. We excluded 70 patients who already had lung cancer, died, had imaging done elsewhere, or left the practice. Of the remaining 376 patients, 337 (90%) had abnormal chest CT findings, and 184 (55%) had a specific follow-up recommendation. Among those with recommended follow-up, only 102 of 184 (55%) had timely follow-up. Those who had a CT performed to evaluate pulmonary disease and those receiving care in community health centers were more likely to receive timely follow-up. Of 27 patients newly diagnosed with lung cancer, 18 (67%) had their first oncology visit within 30 days of diagnosis.

CONCLUSIONS

Among patients undergoing diagnostic chest CTs, most received follow-up for abnormal findings, but it was often delayed. Systems to support patients in obtaining recommended follow-up are needed to ensure that the benefits of lung cancer screening translate into usual clinical practice.

Blinded Validation of Breath Biomarkers of Lung Cancer, a Potential Ancillary to Chest CT Screening

Background

Breath volatile organic compounds (VOCs) have been reported as biomarkers of lung cancer, but it is not known if biomarkers identified in one group can identify disease in a separate independent cohort. Also, it is not known if combining breath biomarkers with chest CT has the potential to improve the sensitivity and specificity of lung cancer screening.

Methods

Model-building phase (unblinded): Breath VOCs were analyzed with gas chromatography mass spectrometry in 82 asymptomatic smokers having screening chest CT, 84 symptomatic high-risk subjects with a tissue diagnosis, 100 without a tissue diagnosis, and 35 healthy subjects. Multiple Monte Carlo simulations identified breath VOC mass ions with greater than random diagnostic accuracy for lung cancer, and these were combined in a multivariate predictive algorithm. Model-testing phase (blinded validation): We analyzed breath VOCs in an independent cohort of similar subjects (n = 70, 51, 75 and 19 respectively). The algorithm predicted discriminant function (DF) values in blinded replicate breath VOC samples analyzed independently at two laboratories (A and B). Outcome modeling: We modeled the expected effects of combining breath biomarkers with chest CT on the sensitivity and specificity of lung cancer screening.

Results

Unblinded model-building phase. The algorithm identified lung cancer with sensitivity 74.0%, specificity 70.7% and C-statistic 0.78. Blinded model-testing phase: The algorithm identified lung cancer at Laboratory A with sensitivity 68.0%, specificity 68.4%, C-statistic 0.71; and at Laboratory B with sensitivity 70.1%, specificity 68.0%, C-statistic 0.70, with linear correlation between replicates (r = 0.88). In a projected outcome model, breath biomarkers increased the sensitivity, specificity, and positive and negative predictive values of chest CT for lung cancer when the tests were combined in series or parallel.

Conclusions

Breath VOC mass ion biomarkers identified lung cancer in a separate independent cohort, in a blinded replicated study. Combining breath biomarkers with chest CT could potentially improve the sensitivity and specificity of lung cancer screening.

Trial Registration

ClinicalTrials.gov NCT00639067

Diagnostic impact of digital tomosynthesis in oncologic patients with suspected pulmonary lesions on chest radiography

OBJECTIVES

To assess the actual diagnostic impact of digital tomosynthesis (DTS) in oncologic patients with suspected pulmonary lesions on chest radiography (CXR).

METHODS

A total of 237 patients (135 male, 102 female; age, 70.8 ± 10.4 years) with a known primary malignancy and suspected pulmonary lesion(s) on CXR and who underwent DTS were retrospectively identified. Two radiologists (experience, 10 and 15 years) analysed in consensus CXR and DTS images and proposed a diagnosis according to a confidence score: 1 or 2 = definitely or probably benign pulmonary or extrapulmonary lesion, or pseudolesion; 3 = indeterminate; 4 or 5 = probably or definitely pulmonary lesion. DTS findings were proven by CT (n = 114 patients), CXR during follow-up (n = 105) or histology (n = 18).

RESULTS

Final diagnoses included 77 pulmonary opacities, 26 pulmonary scars, 12 pleural lesions and 122 pulmonary pseudolesions. DTS vs CXR presented a higher (P < 0.05) sensitivity (92 vs 15 %), specificity (91 vs 9 %), overall accuracy (92 vs 12 %), and diagnostic confidence (area under ROC, 0.997 vs 0.619). Mean effective dose of CXR vs DTS was 0.06 vs 0.107 mSv (P < 0.05).

CONCLUSIONS

DTS improved diagnostic accuracy and confidence in comparison to CXR alone in oncologic patients with suspected pulmonary lesions on CXR with only a slight, though significant, increase in radiation dose.

KEY POINTS

• Digital tomosynthesis (DTS) improves accuracy of chest radiography (CXR) in oncologic patients. • DTS improves confidence of CXR in oncologic patients. • DTS allowed avoidance of CT in about 50 % of oncologic patients.