Measurement methods and algorithms for the management of solid nodules

Influence of CT dose reduction on AI-driven malignancy estimation of incidental pulmonary nodules

OBJECTIVES

The purpose of this study was to determine the influence of dose reduction on a commercially available lung cancer prediction convolutional neuronal network (LCP-CNN).

METHODS

CT scans from a cohort provided by the local lung cancer center (n = 218) with confirmed pulmonary malignancies and their corresponding reduced dose simulations (25% and 5% dose) were subjected to the LCP-CNN. The resulting LCP scores (scale 1-10, increasing malignancy risk) and the proportion of correctly classified nodules were compared. The cohort was divided into a low-, medium-, and high-risk group based on the respective LCP scores; shifts between the groups were studied to evaluate the potential impact on nodule management. Two different malignancy risk score thresholds were analyzed: a higher threshold of ≥ 9 ("rule-in" approach) and a lower threshold of > 4 ("rule-out" approach).

RESULTS

In total, 169 patients with 196 nodules could be included (mean age ± SD, 64.5 ± 9.2 year; 49% females). Mean LCP scores for original, 25% and 5% dose levels were 8.5 ± 1.7, 8.4 ± 1.7 (p > 0.05 vs. original dose) and 8.2 ± 1.9 (p < 0.05 vs. original dose), respectively. The proportion of correctly classified nodules with the "rule-in" approach decreased with simulated dose reduction from 58.2 to 56.1% (p = 0.34) and to 52.0% for the respective dose levels (p = 0.01). For the "rule-out" approach the respective values were 95.9%, 96.4%, and 94.4% (p = 0.12). When reducing the original dose to 25%/5%, eight/twenty-two nodules shifted to a lower, five/seven nodules to a higher malignancy risk group.

CONCLUSION

CT dose reduction may affect the analyzed LCP-CNN regarding the classification of pulmonary malignancies and potentially alter pulmonary nodule management.

CLINICAL RELEVANCE STATEMENT

Utilization of a "rule-out" approach with a lower malignancy risk threshold prevents underestimation of the nodule malignancy risk for the analyzed software, especially in high-risk cohorts.

KEY POINTS

• LCP-CNN may be affected by CT image parameters such as noise resulting from low-dose CT acquisitions. • CT dose reduction can alter pulmonary nodule management recommendations by affecting the outcome of the LCP-CNN. • Utilization of a lower malignancy risk threshold prevents underestimation of pulmonary malignancies in high-risk cohorts.

Contextualizing the Role of Volumetric Analysis in Pulmonary Nodule Assessment: AJR Expert Panel Narrative Review

Pulmonary nodules are managed on the basis of their size and morphologic characteristics. Radiologists are familiar with assessing nodule size by measuring diameter using manually deployed electronic calipers. Size may also be assessed with 3D volumetric measurements (referred to as volumetry) obtained with software. Nodule size and growth are more accurately assessed with volumetry than on the basis of diameter, and the evidence supporting clinical use of volumetry has expanded, driven by its use in lung cancer screening nodule management algorithms in Europe. The application of volumetry has the potential to reduce recommendations for imaging follow-up of indeterminate solid nodules without impacting cancer detection. Although changes in scanning conditions and volumetry software packages can lead to variation in volumetry results, ongoing technical advances have improved the reliability of calculated volumes. Volumetry is now the primary method for determining size of solid nodules in the European lung cancer screening position statement and British Thoracic Society recommendations. The purposes of this article are to review technical aspects, advantages, and limitations of volumetry and, by considering specific scenarios, to contextualize the use of volumetry with respect to its importance in morphologic evaluation, its role in predicting malignancy in risk models, and its practical impact on nodule management. Implementation challenges and areas requiring further evidence are also highlighted.

Lung nodules: sorting the wheat from the chaff

Pulmonary nodules are a common finding on CT scans of the chest. In the United Kingdom, management should follow British Thoracic Society Guidelines, which were published in 2015. This review covers key aspects of nodule management also looks at new and emerging evidence since then.

A predictive nomogram for two-year growth of CT-indeterminate small pulmonary nodules

OBJECTIVES

Lung cancer is the most common cancer and the leading cause of cancer-related death worldwide. The optimal management of computed tomography (CT)-indeterminate pulmonary nodules is important. To optimize individualized follow-up strategies, we developed a radiomics nomogram for predicting 2-year growth in case of indeterminate small pulmonary nodules.

METHODS

A total of 215 histopathology-confirmed small pulmonary nodules (21 benign and 194 malignant) in 205 patients with ultra-high-resolution CT (U-HRCT) were divided into growth and nongrowth nodules and were randomly allocated to the primary (n = 151) or validation (n = 64) group. The least absolute shrinkage and selection operator (LASSO) method was used for radiomics feature selection and radiomics signature determination. Multivariable logistic regression analysis was used to develop a radiomics nomogram that integrated the radiomics signature with significant clinical parameters (sex and nodule type). The area under the curve (AUC) was applied to assess the predictive performance of the radiomics nomogram. The net benefit of the radiomics nomogram was assessed using a clinical decision curve.

RESULTS

The radiomics signature and nomogram yielded AUCs of 0.892 (95% confidence interval [CI]: 0.843-0.940) and 0.911 (95% CI: 0.867-0.955), respectively, in the primary group and 0.826 (95% CI: 0.727-0.926) and 0.843 (95% CI: 0.749-0.937), respectively, in the validation group. The clinical usefulness of the nomogram was demonstrated by decision curve analysis.

CONCLUSIONS

A radiomics nomogram was developed by integrating the radiomics signature with clinical parameters and was easily used for the individualized prediction of two-year growth in case of CT-indeterminate small pulmonary nodules.

KEY POINTS

• A radiomics nomogram was developed for predicting the two-year growth of CT-indeterminate small pulmonary nodules. • The nomogram integrated a CT-based radiomics signature with clinical parameters and was valuable in developing an individualized follow-up strategy for patients with indeterminate small pulmonary nodules.

The promises and challenges of early non-small cell lung cancer detection: patient perceptions, low-dose CT screening, bronchoscopy and biomarkers

Lung cancer survival statistics are sobering with survival ranking among the poorest of all cancers despite the addition of targeted therapies and immunotherapies. However, improvements in tools for early detection hold promise. The Nederlands–Leuvens Longkanker Screenings Onderzoek (NELSON) trial recently corroborated the findings from the previous National Lung Screening Trial low‐dose Computerised Tomography (NLST) screening trial in reducing lung cancer mortality. Biomarker research and development is increasing at pace as the molecular life histories of lung cancers become further unravelled. Low‐dose CT screening (LDCT) is effective but targets only those at the highest risk and is burdensome on healthcare. An optimally designed CT screening programme at best will only detect a low proportion of overall lung cancers as only those at very high‐risk meet screening criteria. Biomarkers that help risk stratify suitable patients for LDCT screening, and those that assist in determining which LDCT detected nodules are likely to represent malignant disease are needed. Some biomarkers have been proposed as standalone lung cancer diagnosis tools. Bronchoscopy technology is improving, with better capacity to identify and obtain samples from early lung cancers. Clinicians need to be aware of each early lung cancer detection method’s inherent limitations. We anticipate that the future of early lung cancer diagnosis will involve a synergistic, multimodal approach, combining several early detection methods.

Assessing the Accuracy of a Deep Learning Method to Risk Stratify Indeterminate Pulmonary Nodules

Rationale: The management of indeterminate pulmonary nodules (IPNs) remains challenging, resulting in invasive procedures and delays in diagnosis and treatment. Strategies to decrease the rate of unnecessary invasive procedures and optimize surveillance regimens are needed.Objectives: To develop and validate a deep learning method to improve the management of IPNs.Methods: A Lung Cancer Prediction Convolutional Neural Network model was trained using computed tomography images of IPNs from the National Lung Screening Trial, internally validated, and externally tested on cohorts from two academic institutions.Measurements and Main Results: The areas under the receiver operating characteristic curve in the external validation cohorts were 83.5% (95% confidence interval [CI], 75.4-90.7%) and 91.9% (95% CI, 88.7-94.7%), compared with 78.1% (95% CI, 68.7-86.4%) and 81.9 (95% CI, 76.1-87.1%), respectively, for a commonly used clinical risk model for incidental nodules. Using 5% and 65% malignancy thresholds defining low- and high-risk categories, the overall net reclassifications in the validation cohorts for cancers and benign nodules compared with the Mayo model were 0.34 (Vanderbilt) and 0.30 (Oxford) as a rule-in test, and 0.33 (Vanderbilt) and 0.58 (Oxford) as a rule-out test. Compared with traditional risk prediction models, the Lung Cancer Prediction Convolutional Neural Network was associated with improved accuracy in predicting the likelihood of disease at each threshold of management and in our external validation cohorts.Conclusions: This study demonstrates that this deep learning algorithm can correctly reclassify IPNs into low- or high-risk categories in more than a third of cancers and benign nodules when compared with conventional risk models, potentially reducing the number of unnecessary invasive procedures and delays in diagnosis.

External validation of a convolutional neural network artificial intelligence tool to predict malignancy in pulmonary nodules

BACKGROUND

Estimation of the risk of malignancy in pulmonary nodules detected by CT is central in clinical management. The use of artificial intelligence (AI) offers an opportunity to improve risk prediction. Here we compare the performance of an AI algorithm, the lung cancer prediction convolutional neural network (LCP-CNN), with that of the Brock University model, recommended in UK guidelines.

METHODS

A dataset of incidentally detected pulmonary nodules measuring 5-15 mm was collected retrospectively from three UK hospitals for use in a validation study. Ground truth diagnosis for each nodule was based on histology (required for any cancer), resolution, stability or (for pulmonary lymph nodes only) expert opinion. There were 1397 nodules in 1187 patients, of which 234 nodules in 229 (19.3%) patients were cancer. Model discrimination and performance statistics at predefined score thresholds were compared between the Brock model and the LCP-CNN.

RESULTS

The area under the curve for LCP-CNN was 89.6% (95% CI 87.6 to 91.5), compared with 86.8% (95% CI 84.3 to 89.1) for the Brock model (p≤0.005). Using the LCP-CNN, we found that 24.5% of nodules scored below the lowest cancer nodule score, compared with 10.9% using the Brock score. Using the predefined thresholds, we found that the LCP-CNN gave one false negative (0.4% of cancers), whereas the Brock model gave six (2.5%), while specificity statistics were similar between the two models.

CONCLUSION

The LCP-CNN score has better discrimination and allows a larger proportion of benign nodules to be identified without missing cancers than the Brock model. This has the potential to substantially reduce the proportion of surveillance CT scans required and thus save significant resources.

Volumetric assessment of solid pulmonary nodules on ultralow-dose CT: a phantom study

Background

To reduce the radiation exposure from chest computed tomography (CT), ultralow-dose CT (ULDCT) protocols performed at sub-millisievert levels were previously tested for the evaluation of pulmonary nodules (PNs). The purpose of our study was to investigate the effect of ULDCT and iterative image reconstruction on volumetric measurements of solid PNs.

Methods

CT datasets of an anthropomorphic chest phantom containing solid microspheres were obtained with a third-generation dual-source CT at standard dose, 1/8th, 1/20th and 1/70th of standard dose [CT volume dose index (CTDI_vol): 0.03-2.03 mGy]. Semi-automated volumetric measurements were performed on CT datasets reconstructed with filtered back projection (FBP) and advanced modelled iterative reconstruction (ADMIRE), at strength level 3 and 5. Absolute percentage error (APE) evaluated measurement accuracy related to the effective volume. Scan repetition differences were evaluated using Bland-Altman analysis. Two-way analysis of variance (ANOVA) assessed influence of different scan parameters on APE. Proportional differences (PDs) tested the effect of dose settings and reconstruction algorithms on volumetric measurements, as compared to the standard protocol (standard dose-FBP).

Results

Bland-Altman analysis revealed small mean interscan differences of APE with narrow limits of agreement (-0.1%±4.3% to -0.3%±3.8%). Dose settings (P<0.001), reconstruction algorithms (P<0.001), nodule diameters (P<0.001) and nodule density (P=0.011) had statistically significant influence on APE. Post-hoc Bonferroni tests showed slightly higher APE when scanning with 1/70th of standard dose [mean difference: 3.4%, 95% confidence interval (CI): 2.5-4.3%; P<0.001], and for image reconstruction with ADMIRE5 (mean difference: 1.8%, 95% CI: 1.0-2.5%; P<0.001). No significant differences for scanning with 1/20th of standard dose (P=0.42), and image reconstruction with ADMIRE3 (P=0.19) were found. Scanning with 1/70th of standard dose and image reconstruction with FBP showed the widest range of PDs (-16.8% to 23.4%) compared to standard dose-FBP.

Conclusions

Our phantom study showed no significant difference between nodule volume measurements on standard dose CT (CTDI_vol: 2 mGy) and ULDCT with 1/20th of standard dose (CTDI_vol: 0.10 mGy).

Detection of Pulmonary Nodule Growth with Chest Tomosynthesis: A Human Observer Study Using Simulated Nodules

RATIONALE AND OBJECTIVES

Chest tomosynthesis has been suggested as a suitable alternative to CT for follow-up of pulmonary nodules. The aim of the present study was to investigate the possibility of detecting pulmonary nodule growth using chest tomosynthesis.

MATERIALS AND METHODS

Simulated nodules with volumes of approximately 100 mm³ and 300 mm³ as well as additional versions with increasing volumes were created. The nodules were inserted into images from pairs of chest tomosynthesis examinations, simulating cases where the nodule had either remained stable in size or increased in size between the two imaging occasions. Nodule volume growths ranging from 11% to 252% were included. A simulated dose reduction was applied to a subset of the cases. Cases differing in terms of nodule size, dose level, and nodule position relative to the plane of image reconstruction were included. Observers rated their confidence that the nodules were stable in size or not. The rating data for the nodules that were stable in size was compared to the rating data for the nodules simulated to have increased in size using ROC analysis.

RESULTS

Area under the curve values ranging from 0.65 to 1 were found. The lowest area under the curve values were found when there was a mismatch in nodule position relative to the reconstructed image plane between the two examinations. Nodule size and dose level affected the results.

CONCLUSION

The study indicates that chest tomosynthesis can be used to detect pulmonary nodule growth. Nodule size, dose level, and mismatch in position relative to the image reconstruction plane in the baseline and follow-up examination may affect the precision.

Management of incidental lung nodules <8 mm in diameter

Due to the increase of incidentally detected pulmonary nodules and the information obtained from several screening programs, updated guidelines with new recommendations for the management of small pulmonary nodules have been proposed. These international guidelines coincide in proposing periodic follow-up for small nodules, less than 8 mm of diameter. Fleischner and British Thoracic Society guidelines are the most recent and popular guidelines for incidental pulmonary nodules management. They have specific recommendations according to nodule characteristics (density and size) and cancer risk of the patient. Both guidelines separate recommendations for solid and subsolid nodules. Predictive risk models have been developed to improve the nodule management. In certain cases follow up may not be the best option. We discuss the scenarios and options to achieve a histologic diagnosis of these tiny pulmonary nodules.

Lung cancer screening with low-dose CT in Europe: strength and weakness of diverse independent screening trials

A North American trial reported a significant reduction of lung cancer mortality and overall mortality as a result of annual screening using low-dose computed tomography (LDCT). European trials prospectively tested a variety of possible screening strategies. The main topics of current discussion regarding the optimal screening strategy are pre-test selection of the high-risk population, interval length of LDCT rounds, definition of positive finding, and post-test apportioning of lung cancer risk based on LDCT findings. Despite the current lack of statistical evidence regarding mortality reduction, the European independent diverse strategies offer a multi-perspective view on screening complexity, with remarkable indications for improvements in cost-effectiveness and harm-benefit balance. The UKLS trial reported the advantage of a comprehensive and simple risk model for selection of patients with 5% risk of lung cancer in 5 years. Subjective risk prediction by biological sampling is under investigation. The MILD trial reported equal efficiency for biennial and annual screening rounds, with a significant reduction in the total number of LDCT examinations. The NELSON trial introduced volumetric quantification of nodules at baseline and volume-doubling time (VDT) for assessment of progression. Post-test risk refinement based on LDCT findings (qualitative or quantitative) is under investigation. Smoking cessation remains the most appropriate strategy for mortality reduction, and it must therefore remain an integral component of any lung cancer screening programme.

The UK Lung Cancer Screening Trial: a pilot randomised controlled trial of low-dose computed tomography screening for the early detection of lung cancer

BACKGROUND

Lung cancer kills more people than any other cancer in the UK (5-year survival < 13%). Early diagnosis can save lives. The USA-based National Lung Cancer Screening Trial reported a 20% relative reduction in lung cancer mortality and 6.7% all-cause mortality in low-dose computed tomography (LDCT)-screened subjects.

OBJECTIVES

To (1) analyse LDCT lung cancer screening in a high-risk UK population, determine optimum recruitment, screening, reading and care pathway strategies; and (2) assess the psychological consequences and the health-economic implications of screening.

DESIGN

A pilot randomised controlled trial comparing intervention with usual care. A population-based risk questionnaire identified individuals who were at high risk of developing lung cancer (≥ 5% over 5 years).

SETTING

Thoracic centres with expertise in lung cancer imaging, respiratory medicine, pathology and surgery: Liverpool Heart & Chest Hospital, Merseyside, and Papworth Hospital, Cambridgeshire.

PARTICIPANTS

Individuals aged 50-75 years, at high risk of lung cancer, in the primary care trusts adjacent to the centres.

INTERVENTIONS

A thoracic LDCT scan. Follow-up computed tomography (CT) scans as per protocol. Referral to multidisciplinary team clinics was determined by nodule size criteria.

MAIN OUTCOME MEASURES

Population-based recruitment based on risk stratification; management of the trial through web-based database; optimal characteristics of CT scan readers (radiologists vs. radiographers); characterisation of CT-detected nodules utilising volumetric analysis; prevalence of lung cancer at baseline; sociodemographic factors affecting participation; psychosocial measures (cancer distress, anxiety, depression, decision satisfaction); and cost-effectiveness modelling.

RESULTS

A total of 247,354 individuals were approached to take part in the trial; 30.7% responded positively to the screening invitation. Recruitment of participants resulted in 2028 in the CT arm and 2027 in the control arm. A total of 1994 participants underwent CT scanning: 42 participants (2.1%) were diagnosed with lung cancer; 36 out of 42 (85.7%) of the screen-detected cancers were identified as stage 1 or 2, and 35 (83.3%) underwent surgical resection as their primary treatment. Lung cancer was more common in the lowest socioeconomic group. Short-term adverse psychosocial consequences were observed in participants who were randomised to the intervention arm and in those who had a major lung abnormality detected, but these differences were modest and temporary. Rollout of screening as a service or design of a full trial would need to address issues of outreach. The health-economic analysis suggests that the intervention could be cost-effective but this needs to be confirmed using data on actual lung cancer mortality.

CONCLUSIONS

The UK Lung Cancer Screening (UKLS) pilot was successfully undertaken with 4055 randomised individuals. The data from the UKLS provide evidence that adds to existing data to suggest that lung cancer screening in the UK could potentially be implemented in the 60-75 years age group, selected via the Liverpool Lung Project risk model version 2 and using CT volumetry-based management protocols.

FUTURE WORK

The UKLS data will be pooled with the NELSON (Nederlands Leuvens Longkanker Screenings Onderzoek: Dutch-Belgian Randomised Lung Cancer Screening Trial) and other European Union trials in 2017 which will provide European mortality and cost-effectiveness data. For now, there is a clear need for mortality results from other trials and further research to identify optimal methods of implementation and delivery. Strategies for increasing uptake and providing support for underserved groups will be key to implementation.

TRIAL REGISTRATION

Current Controlled Trials ISRCTN78513845.

FUNDING

This project was funded by the National Institute for Health Research (NIHR) Health Technology Assessment programme and will be published in full in Health Technology Assessment; Vol. 20, No. 40. See the NIHR Journals Library website for further project information.

The IASLC Lung Cancer Staging Project: Proposals for Coding T Categories for Subsolid Nodules and Assessment of Tumor Size in Part-Solid Tumors in the Forthcoming Eighth Edition of the TNM Classification of Lung Cancer

This article proposes codes for the primary tumor categories of adenocarcinoma in situ (AIS) and minimally invasive adenocarcinoma (MIA) and a uniform way to measure tumor size in part-solid tumors for the eighth edition of the tumor, node, and metastasis classification of lung cancer. In 2011, new entities of AIS, MIA, and lepidic predominant adenocarcinoma were defined, and they were later incorporated into the 2015 World Health Organization classification of lung cancer. To fit these entities into the T component of the staging system, the Tis category is proposed for AIS, with Tis (AIS) specified if it is to be distinguished from squamous cell carcinoma in situ (SCIS), which is to be designated Tis (SCIS). We also propose that MIA be classified as T1mi. Furthermore, the use of the invasive size for T descriptor size follows a recommendation made in three editions of the Union for International Cancer Control tumor, node, and metastasis supplement since 2003. For tumor size, the greatest dimension should be reported both clinically and pathologically. In nonmucinous lung adenocarcinomas, the computed tomography (CT) findings of ground glass versus solid opacities tend to correspond respectively to lepidic versus invasive patterns seen pathologically. However, this correlation is not absolute; so when CT features suggest nonmucinous AIS, MIA, and lepidic predominant adenocarcinoma, the suspected diagnosis and clinical staging should be regarded as a preliminary assessment that is subject to revision after pathologic evaluation of resected specimens. The ability to predict invasive versus noninvasive size on the basis of solid versus ground glass components is not applicable to mucinous AIS, MIA, or invasive mucinous adenocarcinomas because they generally show solid nodules or consolidation on CT.

Multiple pulmonary nodules and unexplained fever: when the pulmonologist fails

We describe herein a difficult case of persistent and refractory fever, associated with multiple lung nodules, progressive respiratory failure and general deterioration. Our patient was carefully investigated for the possible causes of his symptoms, using current and advanced diagnostic procedures, either serological or by imaging. The confirmatory diagnosis of anaplastic T-cell lymphoma, was obtained only after an invasive procedure (with severe pneumothorax), although it was too late. This suggests that also very rare diseases should be considered in the presence of unexplained signs/symptoms, and that in such cases, aggressive diagnostic procedures should be applied as early as possible.

Early detection of lung cancer with low-dose computed tomography: an update on recently presented data

SUMMARY Trials on lung cancer screening with computed tomography have been conducted worldwide for over a decade, but more recently a wealth of new data has emerged on this subject. This review will discuss the results of the American NLST, published in 2011, and evaluate the initial results that have been produced from the European randomized controlled trials. Additionally, the review will outline some of the many outstanding questions that persist relating to lung cancer screening with computed tomography that may potentially be answered in the near future as ongoing studies are concluded.