Management of Incidental Lung Nodules

Outcomes in Incidentally Versus Screening Detected Stage I Lung Cancer Surgery Patients

INTRODUCTION

Although the importance of lung cancer screening for early diagnosis is established, because of poor enrollment, incidental findings still play a role in diagnosis of patients who qualify. Nevertheless, analysis of this incidental cohort is lacking. We present a retrospective analysis comparing patients with thoracic surgery with incidental versus screening detected stage I lung cancer.

METHODS

Thoracic surgery cases at Mount Sinai Hospital from March, 1, 2012, to June, 30, 2022, were queried for patients eligible for lung cancer screening and a stage I diagnosis. The basis of lung nodule detection (incidental versus screening detected) was identified. We compared demographic variables, comorbidities, tumor staging, procedure details, and postoperative outcomes between the cohorts.

RESULTS

Of the patients eligible for screening with lung cancer resection and stage I diagnosis at Mount Sinai, 153 were identified incidentally and 67 through screening. The patients in the incidental cohort were older (p = 0.005), more likely to have quit smoking (p = 0.04), and had a greater number of comorbidities (p = 0.0002). There was no statistically significant difference between the groups with regard to pack-year smoking history, lung cancer histological type, location or size of tumor, and surgical approach, length of surgery or stay, number of postoperative outcomes, and survival.

CONCLUSIONS

In stage I lung cancers, no significant differences were identified between incidentally and screening detected lung nodules with regard to tumor characteristics, surgical approach, and postoperative outcomes. Imaging conducted for other reasons should be considered as a valid and important diagnostic tool, similar to traditional low-dose computed tomography, in patients who qualify for screening.

Development of a combined radiomics and CT feature-based model for differentiating malignant from benign subcentimeter solid pulmonary nodules

BACKGROUND

We aimed to develop a combined model based on radiomics and computed tomography (CT) imaging features for use in differential diagnosis of benign and malignant subcentimeter (≤ 10 mm) solid pulmonary nodules (SSPNs).

METHODS

A total of 324 patients with SSPNs were analyzed retrospectively between May 2016 and June 2022. Malignant nodules (n = 158) were confirmed by pathology, and benign nodules (n = 166) were confirmed by follow-up or pathology. SSPNs were divided into training (n = 226) and testing (n = 98) cohorts. A total of 2107 radiomics features were extracted from contrast-enhanced CT. The clinical and CT characteristics retained after univariate and multivariable logistic regression analyses were used to develop the clinical model. The combined model was established by associating radiomics features with CT imaging features using logistic regression. The performance of each model was evaluated using the area under the receiver-operating characteristic curve (AUC).

RESULTS

Six CT imaging features were independent predictors of SSPNs, and four radiomics features were selected after a dimensionality reduction. The combined model constructed by the logistic regression method had the best performance in differentiating malignant from benign SSPNs, with an AUC of 0.942 (95% confidence interval 0.918-0.966) in the training group and an AUC of 0.930 (0.902-0.957) in the testing group. The decision curve analysis showed that the combined model had clinical application value.

CONCLUSIONS

The combined model incorporating radiomics and CT imaging features had excellent discriminative ability and can potentially aid radiologists in diagnosing malignant from benign SSPNs.

RELEVANCE STATEMENT

The model combined radiomics features and clinical features achieved good efficiency in predicting malignant from benign SSPNs, having the potential to assist in early diagnosis of lung cancer and improving follow-up strategies in clinical work.

KEY POINTS

• We developed a pulmonary nodule diagnostic model including radiomics and CT features. • The model yielded the best performance in differentiating malignant from benign nodules. • The combined model had clinical application value and excellent discriminative ability. • The model can assist radiologists in diagnosing malignant from benign pulmonary nodules.

Natural Language Processing for the Identification of Incidental Lung Nodules in Computed Tomography Reports: A Quality Control Tool

PURPOSE

To evaluate the diagnostic performance of a natural language processing (NLP) model in detecting incidental lung nodules (ILNs) in unstructured chest computed tomography (CT) reports.

METHODS

All unstructured consecutive reports of chest CT scans performed at a tertiary hospital between 2020 and 2021 were retrospectively reviewed (n = 21,542) to train the NLP tool. Internal validation was performed using reference readings by two radiologists of both CT scans and reports, using a different external cohort of 300 chest CT scans. Second, external validation was performed in a cohort of all random unstructured chest CT reports from 57 different hospitals conducted in May 2022. A review by the same thoracic radiologists was used as the gold standard. The sensitivity, specificity, and accuracy were calculated.

RESULTS

Of 21,542 CT reports, 484 mentioned at least one ILN (mean age, 71 ± 17.6 [standard deviation] years; women, 52%) and were included in the training set. In the internal validation (n = 300), the NLP tool detected ILN with a sensitivity of 100.0% (95% CI, 97.6 to 100.0), a specificity of 95.9% (95% CI, 91.3 to 98.5), and an accuracy of 98.0% (95% CI, 95.7 to 99.3). In the external validation (n = 977), the NLP tool yielded a sensitivity of 98.4% (95% CI, 94.5 to 99.8), a specificity of 98.6% (95% CI, 97.5 to 99.3), and an accuracy of 98.6% (95% CI, 97.6 to 99.2). Twelve months after the initial reports, 8 (8.60%) patients had a final diagnosis of lung cancer, among which 2 (2.15%) would have been lost to follow-up without the NLP tool.

CONCLUSION

NLP can be used to identify ILNs in unstructured reports with high accuracy, allowing a timely recall of patients and a potential diagnosis of early-stage lung cancer that might have been lost to follow-up.

A combined non-enhanced CT radiomics and clinical variable machine learning model for differentiating benign and malignant sub-centimeter pulmonary solid nodules

BACKGROUND

Radiomics has been used to predict pulmonary nodule (PN) malignancy. However, most of the studies focused on pulmonary ground-glass nodules. The use of computed tomography (CT) radiomics in pulmonary solid nodules, particularly sub-centimeter solid nodules, is rare.

PURPOSE

This study aims to develop a radiomics model based on non-enhanced CT images that can distinguish between benign and malignant sub-centimeter pulmonary solid nodules (SPSNs, <1 cm).

METHODS

The clinical and CT data of 180 SPSNs confirmed by pathology were analyzed retrospectively. All SPSNs were divided into two groups: training set (n = 144) and testing set (n = 36). From non-enhanced chest CT images, over 1000 radiomics features were extracted. Radiomics feature selection was performed using the analysis of variance and principal component analysis. The selected radiomics features were fed into a support vector machine (SVM) to develop a radiomics model. The clinical and CT characteristics were used to develop a clinical model. Associating non-enhanced CT radiomics features with clinical factors were used to develop a combined model using SVM. The performance was evaluated using the area under the receiver-operating characteristic curve (AUC).

RESULTS

The radiomics model performed well in distinguishing between benign and malignant SPSNs, with an AUC of 0.913 (95% confidence interval [CI], 0.862-0.954) in the training set and an AUC of 0.877 (95% CI, 0.817-0.924) in the testing set. The combined model outperformed the clinical and radiomics models with an AUC of 0.940 (95% CI, 0.906-0.969) in the training set and an AUC of 0.903 (95% CI, 0.857-0.944) in the testing set.

CONCLUSIONS

Radiomics features based on non-enhanced CT images can be used to differentiate SPSNs. The combined model, which included radiomics and clinical factors, had the best discrimination power between benign and malignant SPSNs.

Artificial Intelligence Algorithm-Based Feature Extraction of Computed Tomography Images and Analysis of Benign and Malignant Pulmonary Nodules

This study was aimed to explore the effect of CT image feature extraction of pulmonary nodules based on an artificial intelligence algorithm and the image performance of benign and malignant pulmonary nodules. In this study, the CT images of pulmonary nodules were collected as the research object, and the lung nodule feature extraction model based on expectation maximization (EM) was used to extract the image features. The Dice similarity coefficient, accuracy, benign and malignant nodule edges, internal signs, and adjacent structures were compared and analyzed to obtain the extraction effect of this feature extraction model and the image performance of benign and malignant pulmonary nodules. The results showed that the detection sensitivity of pulmonary nodules in this model was 0.955, and the pulmonary nodules and blood vessels were well preserved in the image. The probability of burr sign detection in the malignant group was 73.09% and that in the benign group was 8.41%. The difference was statistically significant (P < 0.05). The probability of malignant component leaf sign (69.96%) was higher than that of a benign component leaf sign (0), and the difference was statistically significant (P < 0.05). The probability of cavitation signs in the malignant group (59.19%) was higher than that in the benign group (3.74%), and the probability of blood vessel collection signs in the malignant group (74.89%) was higher than that in the benign group (11.21%), with statistical significance (P < 0.05). The probability of the pleural traction sign in the malignant group was 17.49% higher than that in the benign group (4.67%), and the difference was statistically significant (P < 0.05). In summary, the feature extraction effect of CT images based on the EM algorithm was ideal. Imaging findings, such as the burr sign, lobulation sign, vacuole sign, vascular bundle sign, and pleural traction sign, can be used as indicators to distinguish benign and malignant nodules.

Preoperative computed tomography-guided localization for lung nodules: localization needle versus coil

PURPOSE

To compare the clinical efficacy of computed tomography (CT)-guided localization needle and coil insertion as approaches to preoperative lung nodule (LN) localization.

MATERIAL AND METHODS

Between January 2018 and December 2019, 52 patients awaiting video-assisted thoracoscopic surgery (VATS) resection underwent CT-guided coil insertion to facilitate LN localization. Additionally, 41 patients underwent CT-guided localization needle insertion between January and June 2021.

RESULTS

In total, 62 and 54 LNs were localized in 52 and 41 patients in the coil and localization needle groups, respectively, with respective technical localization success rates of 96.8% and 100% (p = .498). The localization needle group exhibited a significantly shorter duration of localization relative to the coil group (p < .001), whereas comparable rates of pneumothorax (p = .918) and hemorrhage (p = .712) were evident in these groups. VATS-guided LN resection procedures achieved 100% technical success rates in both groups, and there were no significant differences between groups with respect to the type of resection (p = .113) or the mean duration of VATS (p = .778).

CONCLUSION

Coil- and localization needle-based approaches can be successfully used for LN localization prior to VATS resection, with localization needle insertion being associated with a shorter duration of localization.

Seasonal prevalence and characteristics of low-dose CT detected lung nodules in a general Dutch population

We investigated whether presence and characteristics of lung nodules in the general population using low-dose computed tomography (LDCT) varied by season. Imaging in Lifelines (ImaLife) study participants who underwent chest LDCT-scanning between October 2018 and October 2019 were included in this sub-study. Hay fever season (summer) was defined as 1st April to 30th September and Influenza season (winter) as 1st October to 31st March. All lung nodules with volume of ≥ 30 mm³ (approximately 3 mm in diameter) were registered. In total, 2496 lung nodules were found in 1312 (38%) of the 3456 included participants (nodules per participant ranging from 1 to 21, median 1). In summer, 711 (54%) participants had 1 or more lung nodule(s) compared to 601 (46%) participants in winter (p = 0.002). Of the spherical, perifissural and left-upper-lobe nodules, relatively more were detected in winter, whereas of the polygonal-, irregular-shaped and centrally-calcified nodules, relatively more were detected in summer. Various seasonal diseases with inflammation as underlying pathophysiology may influence presence and characteristics of lung nodules. Further investigation into underlying pathophysiology using short-term LDCT follow-up could help optimize the management strategy for CT-detected lung nodules in clinical practice.

Human-recognizable CT image features of subsolid lung nodules associated with diagnosis and classification by convolutional neural networks

OBJECTIVES

The interpretability of convolutional neural networks (CNNs) for classifying subsolid nodules (SSNs) is insufficient for clinicians. Our purpose was to develop CNN models to classify SSNs on CT images and to investigate image features associated with the CNN classification.

METHODS

CT images containing SSNs with a diameter of ≤ 3 cm were retrospectively collected. We trained and validated CNNs by a 5-fold cross-validation method for classifying SSNs into three categories (benign and preinvasive lesions [PL], minimally invasive adenocarcinoma [MIA], and invasive adenocarcinoma [IA]) that were histologically confirmed or followed up for 6.4 years. The mechanism of CNNs on human-recognizable CT image features was investigated and visualized by gradient-weighted class activation map (Grad-CAM), separated activation channels and areas, and DeepDream algorithm.

RESULTS

The accuracy was 93% for classifying 586 SSNs from 569 patients into three categories (346 benign and PL, 144 MIA, and 96 IA in 5-fold cross-validation). The Grad-CAM successfully located the entire region of image features that determined the final classification. Activated areas in the benign and PL group were primarily smooth margins (p < 0.001) and ground-glass components (p = 0.033), whereas in the IA group, the activated areas were mainly part-solid (p < 0.001) and solid components (p < 0.001), lobulated shapes (p < 0.001), and air bronchograms (p < 0.001). However, the activated areas for MIA were variable. The DeepDream algorithm showed the image features in a human-recognizable pattern that the CNN learned from a training dataset.

CONCLUSION

This study provides medical evidence to interpret the mechanism of CNNs that helps support the clinical application of artificial intelligence.

KEY POINTS

• CNN achieved high accuracy (93%) in classifying subsolid nodules on CT images into three categories: benign and preinvasive lesions, MIA, and IA. • The gradient-weighted class activation map (Grad-CAM) located the entire region of image features that determined the final classification, and the visualization of the separated activated areas was consistent with radiologists' expertise for diagnosing subsolid nodules. • DeepDream showed the image features that CNN learned from a training dataset in a human-recognizable pattern.

Computed tomography-guided coil localization for sub-fissural lung nodules

This study was designed to explore the safety, feasibility, and clinical efficacy of preoperative computed tomography (CT)-guided coil localization of sub-fissural lung nodules (LNs). A total of 105 LN patients underwent CT-guided coil localization followed by video-assisted thoracoscopic surgery (VATS)-guided wedge resection at our hospital from January 2016 to December 2019. Of these patients, 4 had sub-fissural LNs and were therefore subjected to trans-fissural coil localization procedures. We analyzed data pertaining to the coil localization and VATS procedures in these patients. A total of 4 coils were used to localize 4 LNs in 4 separate patients. One of these patients suffered from parenchymal hemorrhage around the needle path, while one other patient exhibited asymptomatic pneumothorax following coil localization. A thoracoscope was able to successfully visualize the coil tails in all of these patients. There were no instances of coils having been dislodged, and wedge resection was conducted with a 100% technical success rate in these patients. These 4 LNs were subsequently diagnosed as adenocarcinomas in situ (n = 3) and benign nodules (n = 1). CT-guided coil localization can be used to safely and easily localize sub-fissural LNs in patients scheduled to undergo VATS.

Implementation of a Standardized Template for Reporting of Incidental Pulmonary Nodules: Feasibility, Acceptability, and Outcomes

OBJECTIVE

Incidental pulmonary nodules (IPNs) are common. Up to 70% are not followed up according to current guidelines. Follow-up recommendations are based on the characteristics of the patient and the IPN. However, many IPNs are incompletely characterized in CT reports. Structured radiology reports have been shown to reduce missing information. We sought to improve IPN reporting by assessing the feasibility, acceptability, and effectiveness of a structured dictation template to increase the presence of six key nodule descriptors.

METHODS

We performed a mixed methods, pre- and postimplementation assessment. A template was developed with a multidisciplinary group based on Fleischner Society guidelines. A standardized checklist was used to determine the presence of documented descriptors pre- and postimplementation for sequential radiology reports of patients with an IPN present (n = 400 pre-implementation and n = 400 postimplementation) on a CT performed at the Minneapolis Veterans Affairs Health Care System. We conducted qualitative interviews with radiologists (n = 4) and members of the lung nodule tracking team (n = 2) to elicit their experiences of the template implementation process.

RESULTS

The proportion of radiology reports including all six elements increased from 12% to 47% (P < .001). Postimplementation, the template was used in 40% of interpretations involving lung nodules, 67% of follow-up scans, and 8% of initial identifications. Response to the template was overall positive.

DISCUSSION

Use of a dictation template seems to be effective in increasing compliance with full IPN documentation, streamlining the follow-up process. Low utilization rates of the template for initial nodule identification is a limitation, which may be combated through clearer communication and advances in technology.

Diagnostic Accuracy of Maximum Intensity Projection in Diagnosis of Malignant Pulmonary Nodules

Introduction Pulmonary nodules are frequently encountered during chest imaging, and its evaluation is usually done by chest radiograph and computed tomography (CT) scan of chest. High resolution of multidetector CT (MDCT) has improved the nodule detection. Post processing techniques such as maximum intensity projection (MIP) can further improve the sensitivity of MDCT for nodule detection. Failure to diagnose malignancy in pulmonary nodules can delay the treatment. Therefore, the aim of this study was to determine the diagnostic accuracy of MIP in the diagnosis of malignant pulmonary nodules taking histopathology findings as gold standard. Materials and methods A retrospective cross-sectional study was conducted at Dow Institute of Radiology, Dow University of Health Sciences, from 1 December 2018 till 30 June 2019. Both male and female patients aged 18 years and above who underwent CT scan of chest with suspicion of pulmonary nodules were included. Patients already diagnosed with malignant pulmonary nodules and presenting for follow-up were excluded. Contrast-enhanced CT chest was performed on a multi-slice scanner. MIP reconstruction and evaluation was performed on the workstation. Sensitivity, specificity, positive predictive value, negative predictive value, and diagnostic accuracy of MIP were calculated taking histopathology findings as gold standard. Results A total of 202 patients were included in this study. The mean age of the patients was 55.87 ± 13.08 years. A total of 103 patients (51.0%) were males and 99 patients (49.0%) were females. There were 131 (64.9%) nodules with smooth margins and 71 (35.1%) nodules with irregular margins. The mean size of nodule was 3.1 ± 0.7 cm. Sensitivity, specificity, positive predictive value, negative predictive value, and diagnostic accuracy of MIP in diagnosing malignant pulmonary nodules were found to be 85.82%, 82.35%, 90.55%, 74.67%, and 84.65%, respectively, taking histopathology findings as gold standard. The nodules >3 cm in size had a higher sensitivity for diagnosing malignant pulmonary nodules. Smooth margin nodule had high sensitivity, specificity, and diagnostic accuracy for diagnosing malignant pulmonary nodules. Conclusion MIP images have high sensitivity, specificity, and diagnostic accuracy in the diagnosis of malignant pulmonary nodules. The utilization of MIP images can aid in the detection of malignant pulmonary nodules and help in formulating early treatment strategies for the patients. Other post processing techniques such as volume rendering and computer-aided detection can help in further improving patient care.

Application of Radiomics in Predicting the Malignancy of Pulmonary Nodules in Different Sizes

OBJECTIVE. The purpose of this study was to investigate the utility of radiomics for predicting the malignancy of pulmonary nodules (PNs) of different sizes using unenhanced, thin-section CT images. MATERIALS AND METHODS. Patients with a single PN (n = 373) who underwent a preoperative chest CT were recruited retrospectively at Beijing Friendship Hospital from March 2015 to March 2018. Of the 373 PNs studied, 192 were benign and 181 were malignant. The lesions were classified into three groups (T1a, T1b, or T1c according to the 8th edition of the TNM staging system for lung cancer) on the basis of lesion diameters: T1a (diameter, 0-1 cm), T1b (1 cm < diameter ≤ 2 cm) and T1c (2 cm < diameter ≤ 3 cm). A total of 1160 radiomic features were extracted from PN segmentation on unenhanced CT images. We developed three radiomic models to predict PN malignancy in each group on the basis of the extracted radiomic features. Fivefold cross-validation was used to estimate AUC, accuracy, sensitivity, and specificity for indicating the performance of prediction models. RESULTS. The AUC, accuracy, sensitivity, and specificity for predicting PN malignancy in each group were 0.84, 0.77, 0.89, and 0.74 with the T1a model; 0.78, 0.73, 0.74, and 0.71 with the T1b model, and 0.79, 0.76, 0.77, and 0.73 with the T1c model, respectively. The most contributive radiomic features for predicting PN malignancy for groups T1a, T1b, and T1c were LoG_X_Uniformity, Intensity_Minimum, and Shape_SI9, respectively. CONCLUSION. Radiomic features based on unenhanced CT images can be used to predict the malignancy of pulmonary nodules. The radiomic T1a model showed superior prediction performance to the T1b and T1c models, and the best performance in terms of AUC and sensitivity was found for predicting the malignancy of T1a PN.

Changes in quantitative parameters of pulmonary nonsolid nodule induced by lung inflation according to paired inspiratory and expiratory computed tomography imaging

OBJECTIVE

To evaluate quantitative parameters of nonsolid nodules on paired inspiratory and expiratory computed tomography (CT) and to examine whether these parameters are sensitive to lung inflation reflected by lung volume.

METHODS

Thirty-three patients with 41 nonsolid nodules were included in this prospective study. Paired inspiratory and low-dose respiratory plain chest CT were performed. The volume and density of nonsolid nodule(s), both lungs, the right and left lung, and five lobes, were analyzed in inspiratory and expiratory CT scans. The ratio of expiratory to inspiratory parameters was calculated and labeled as parameter_(E-I)/I. To standardize the changes in nonsolid nodule quantitative parameters, the ratio of nonsolid nodule parameter to lung parameter was also calculated. Quantitative parameters were compared between inspiratory and expiratory CT.

RESULTS

Nonsolid nodule volumes on expiratory CT were reduced by 19.8% ± 12.9%, while the density was increased by 11.4% ± 8.8%. The volume of nonsolid nodules was significantly greater on inspiratory compared with expiratory CT (p < 0.001). The density of nonsolid nodules was significantly greater on expiratory than inspiratory CT (p < 0.001). The volume_(E-I)/I was significantly greater than density_(E-I)/I both in nonsolid nodules and lung. The volume_(E-I)/I and density_(E-I)/I of nonsolid nodules were independent of size. The density_(E-I)/I of nonsolid nodule was greater in the lower lobe than that in the upper lobe (p = 0.002).

CONCLUSION

Volume changes in nonsolid nodules were more sensitive than density changes in expiratory phase. The density of lower lobe nodules was more susceptible to respiration. Expiratory scanning is not recommended for quantification of nonsolid nodules and/or follow-up.

KEY POINTS

• The nonsolid nodule volume on expiratory CT was reduced by 19.8% ± 12.9%. • The nonsolid nodule density on expiratory CT was increased by 11.4% ± 8.8%. • The volume _(E-I)/I and density _(E-I)/I of nonsolid nodules were independent of size.