Long-term follow-up of persistent pulmonary pure ground-glass nodules with deep learning-assisted nodule segmentation

Deep learning for malignancy risk estimation of incidental sub-centimeter pulmonary nodules on CT images

OBJECTIVES

To establish deep learning models for malignancy risk estimation of sub-centimeter pulmonary nodules incidentally detected by chest CT and managed in clinical settings.

MATERIALS AND METHODS

Four deep learning models were trained using CT images of sub-centimeter pulmonary nodules from West China Hospital, internally tested, and externally validated on three cohorts. The four models respectively learned 3D deep features from the baseline whole lung region, baseline image patch where the nodule located, baseline nodule box, and baseline plus follow-up nodule boxes. All regions of interest were automatically segmented except that the nodule boxes were additionally manually checked. The performance of models was compared with each other and that of three respiratory clinicians.

RESULTS

There were 1822 nodules (981 malignant) in the training set, 806 (416 malignant) in the testing set, and 357 (253 malignant) totally in the external sets. The area under the curve (AUC) in the testing set was 0.754, 0.855, 0.928, and 0.942, respectively, for models derived from baseline whole lung, image patch, nodule box, and the baseline plus follow-up nodule boxes. When baseline models externally validated (follow-up images not available), the nodule-box model outperformed the other two with AUC being 0.808, 0.848, and 0.939 respectively in the three external datasets. The resident, junior, and senior clinicians achieved an accuracy of 67.0%, 82.5%, and 90.0%, respectively, in the testing set. The follow-up model performed comparably to the senior clinician.

CONCLUSION

The deep learning algorithms solely mining nodule information can efficiently predict malignancy of incidental sub-centimeter pulmonary nodules.

CLINICAL RELEVANCE STATEMENT

The established models may be valuable for supporting clinicians in routine clinical practice, potentially reducing the number of unnecessary examinations and also delays in diagnosis.

KEY POINTS

• According to different regions of interest, four deep learning models were developed and compared to evaluate the malignancy of sub-centimeter pulmonary nodules by CT images. • The models derived from baseline nodule box or baseline plus follow-up nodule boxes demonstrated sufficient diagnostic accuracy (86.4% and 90.4% in the testing set), outperforming the respiratory resident (67.0%) and junior clinician (82.5%). • The proposed deep learning methods may aid clinicians in optimizing follow-up recommendations for sub-centimeter pulmonary nodules and may lead to fewer unnecessary diagnostic interventions.

A case of lung metastasis from gastric cancer presenting as ground-glass opacity dominant nodules

BACKGROUND

Most metastatic lung tumors present as solid nodules on chest computed tomography (CT). In contrast, ground-glass opacity on chest computed tomography usually suggests low-grade malignant lesions such as adenocarcinoma in situ or atypical adenomatous hyperplasia of the lung.

CASE PRESENTATION

A 75-year-old woman with a history of gastric cancer surgery approximately 5 years prior was referred to the Department of Thoracic Surgery at our hospital because of two newly appearing pulmonary ground-glass opacity-dominant nodules on chest computed tomography. She had two ground-glass opacities in the right lower lobe, one in the S6 segment was 12 mm and the other in the S10 segment was 8 mm. On chest computed tomography 15 months prior to referral, the lesion in the S6 segment was 8 mm, and the lesion in the S10 segment was 2 mm. She was suspected to have primary lung cancer and underwent wide-wedge resection of the nodule in the S6 segment. In the resected specimen, polygonal tumor cells infiltrated the alveolar septa, with some tumor cells exhibiting signet ring cell morphology. Based on morphological similarities to the tumor cells of previous gastric cancers and the results of immunostaining, the patient was diagnosed with lung metastases of gastric cancer.

CONCLUSIONS

Pulmonary nodules in patients with a history of cancer in other organs, even if ground-glass opacity is predominant, should also be considered for the possibility of metastatic pulmonary tumors if they are growing rapidly.

Artificial intelligence in lung cancer screening: Detection, classification, prediction, and prognosis

BACKGROUND

The exceptional capabilities of artificial intelligence (AI) in extracting image information and processing complex models have led to its recognition across various medical fields. With the continuous evolution of AI technologies based on deep learning, particularly the advent of convolutional neural networks (CNNs), AI presents an expanded horizon of applications in lung cancer screening, including lung segmentation, nodule detection, false-positive reduction, nodule classification, and prognosis.

METHODOLOGY

This review initially analyzes the current status of AI technologies. It then explores the applications of AI in lung cancer screening, including lung segmentation, nodule detection, and classification, and assesses the potential of AI in enhancing the sensitivity of nodule detection and reducing false-positive rates. Finally, it addresses the challenges and future directions of AI in lung cancer screening.

RESULTS

AI holds substantial prospects in lung cancer screening. It demonstrates significant potential in improving nodule detection sensitivity, reducing false-positive rates, and classifying nodules, while also showing value in predicting nodule growth and pathological/genetic typing.

CONCLUSIONS

AI offers a promising supportive approach to lung cancer screening, presenting considerable potential in enhancing nodule detection sensitivity, reducing false-positive rates, and classifying nodules. However, the universality and interpretability of AI results need further enhancement. Future research should focus on the large-scale validation of new deep learning-based algorithms and multi-center studies to improve the efficacy of AI in lung cancer screening.

[The contribution of artificial intelligence (AI) subsequent to the processing of thoracic imaging]

The contribution of artificial intelligence (AI) to medical imaging is currently the object of widespread experimentation. The development of deep learning (DL) methods, particularly convolution neural networks (CNNs), has led to performance gains often superior to those achieved by conventional methods such as machine learning. Radiomics is an approach aimed at extracting quantitative data not accessible to the human eye from images expressing a disease. The data subsequently feed machine learning models and produce diagnostic or prognostic probabilities. As for the multiple applications of AI methods in thoracic imaging, they are undergoing evaluation. Chest radiography is a practically ideal field for the development of DL algorithms able to automatically interpret X-rays. Current algorithms can detect up to 14 different abnormalities present either in isolation or in combination. Chest CT is another area offering numerous AI applications. Various algorithms have been specifically formed and validated for the detection and characterization of pulmonary nodules and pulmonary embolism, as well as segmentation and quantitative analysis of the extent of diffuse lung diseases (emphysema, infectious pneumonias, interstitial lung disease). In addition, the analysis of medical images can be associated with clinical, biological, and functional data (multi-omics analysis), the objective being to construct predictive approaches regarding disease prognosis and response to treatment.

Two-stage minimally invasive pulmonary resections with intraoperative localization technique for bilateral multiple primary lung cancers: A case report

Multiple primary lung cancers (MPLCs) are becoming more and more common and these patients can benefit from minimally invasive surgery. Here, we report a case of a patient diagnosed with synchronous MPLCs who underwent bilateral thoracoscopic pulmonary resections in a two-stage strategy, and achieved a good surgical outcome and high quality of life. A 66-year-old female was found to have one major ground-glass nodule (GGN) in the right upper lobe and eight minor GGNs in the left upper and lower lobes. The patient underwent right upper lobe resection and systematic mediastinal lymph node dissection via single-utility port thoracoscopic surgery in September 2018. Pathology was lepidic predominant adenocarcinoma pT1bN0M0, IA2. Regular high-resolution computed tomography examination during 36 months after right upper lobectomy showed gradually increasing diameter and solid component of multiple GGNs in left lung. The patient underwent thoracoscopic multiple pulmonary resections using an intraoperative localization technique in a hybrid operating room in October 2021 and all eight nodules in the left lung were resected. Two segmentectomies and four wedge resections were performed, and the pathological results of the eight nodules included four adenocarcinomas, three adenocarcinomas in situ, and one alveolar epithelial hyperplasia. The two operations were successful with no intra- or postoperative 90-day complications. During more than 20 months of follow-up after the second operation, the patient had well recovered pulmonary function and physical status with a Karnofsky performance status score of 90 and no local recurrence or metastasis. A two-stage surgical strategy for synchronous MPLCs is therefore feasible. The surgical strategy, timing of intervention, and extent of pulmonary resection should be individually designed according to the location and characteristics of each nodule. Intraoperative localization of small GGNs is very important to ensure that all nodules are completely and accurately resected during the operation.

Individualised prediction of progression of solitary sub-solid pulmonary nodules based on CT semantic and clinical features: a 3-year follow-up study

AIM

To develop and validate a progressive prediction model for estimating the time to progression (TTP) of sub-solid pulmonary nodules (SSNs).

MATERIALS AND METHODS

A total of 126 cases who met inclusion and exclusion criteria were included in the study. The primary endpoint of the study was TTP of SSNs. Baseline characteristics were assessed in terms of clinical and CT semantic features. Kaplan-Meier analysis and Cox regression analysis were performed to determine the relationship between SSNs TTP and factors from the entire data set. The nomogram was constructed based on the result of multivariate analysis and internal validation was performed using the bootstrapping. The nomogram's performance was assessed with the C-index, calibration curves, and decision curve analysis.

RESULTS

The median follow-up time of the population was 42.5 (21.5) months. On Kaplan-Meier analysis, patients with higher or positive values of the indices had higher cumulative progression rates (p<0.05). Multivariate Cox regression models identified diameter, consolidation tumour ratio (CTR), morphology, and vasodilation sign (VDS) as independent risk factors of TTP. These predictors were included in the final model to estimate individual probabilities of progression in the 3 years, which performed well in the discrimination (the C-index was 0.901 [95%CI: 0.830-0.981] and 0.875 [95%CI: 0.805-0.942] in the training and internally validation sets).

CONCLUSION

The radiological semantic features nomogram is a promising and favourable prognostic biomarker for predicting progression and may aid in clinical risk stratification and decision-making for SSNs.

Developing a multi-institutional nomogram for assessing lung cancer risk in patients with 5-30 mm pulmonary nodules: a retrospective analysis

Background

The diagnosis of benign and malignant solitary pulmonary nodules based on personal experience has several limitations. Therefore, this study aims to establish a nomogram for the diagnosis of benign and malignant solitary pulmonary nodules using clinical information and computed tomography (CT) results.

Methods

Retrospectively, we collected clinical and CT characteristics of 1,160 patients with pulmonary nodules in Guang'an People's Hospital and the hospital affiliated with North Sichuan Medical College between 2019 and 2021. Among these patients, data from 773 patients with pulmonary nodules were used as the training set. We used the least absolute shrinkage and selection operator (LASSO) to optimize clinical and imaging features and performed a multivariate logistic regression to identify features with independent predictive ability to develop the nomogram model. The area under the receiver operating characteristic curve (AUC), C-index, decision curve analysis, and calibration plot were used to evaluate the performance of the nomogram model in terms of predictive ability, discrimination, calibration, and clinical utility. Finally, data from 387 patients with pulmonary nodules were utilized for validation.

Results

In the training set, the predictors for the nomogram were gender, density of the nodule, nodule diameter, lobulation, calcification, vacuole, vascular convergence, bronchiole, and pleural traction, selected through LASSO and logistic regression analysis. The resulting model had a C-index of 0.842 (95% CI [0.812-0.872]) and AUCs of 0.842 (95% CI [0.812-0.872]). In the validation set, the C-index was 0.856 (95% CI [0.811-0.901]), and the AUCs were 0.844 (95% CI [0.797-0.891]). Results from the calibration curve and clinical decision curve analyses indicate that the nomogram has a high fit and clinical benefit in both the training and validation sets.

Conclusion

The establishment of a nomogram for predicting the benign or malignant diagnosis of solitary pulmonary nodules by this study has shown good efficacy. Such a nomogram may help to guide the diagnosis, follow-up, and treatment of patients.

Radiomics Signatures of Carotid Plaque on Computed Tomography Angiography : An Approach to Identify Symptomatic Plaques

PURPOSE

To develop and validate a combined model incorporating conventional clinical and imaging characteristics and radiomics signatures based on head and neck computed tomography angiography (CTA) to assess plaque vulnerability.

METHODS

We retrospectively analyzed 167 patients with carotid atherosclerosis who underwent head and neck CTA and brain magnetic resonance imaging (MRI) within 1 month. Clinical risk factors and conventional plaque characteristics were evaluated, and radiomic features were extracted from the carotid plaques. The conventional, radiomics and combined models were developed using fivefold cross-validation. Model performance was evaluated using receiver operating characteristic (ROC), calibration, and decision curve analyses.

RESULTS

Patients were divided into symptomatic (n = 70) and asymptomatic (n = 97) groups based on MRI results. Homocysteine (odds ratio, OR 1.057; 95% confidence interval, CI 1.001-1.116), plaque ulceration (OR 6.106; 95% CI 1.933-19.287), and carotid rim sign (OR 3.285; 95% CI 1.203-8.969) were independently associated with symptomatic status and were used to construct the conventional model and s radiomic features were retained to establish the radiomics model. Radiomics scores incorporated with conventional characteristics were used to establish the combined model. The area under the ROC curve (AUC) of the combined model was 0.832, which outperformed the conventional (AUC = 0.767) and radiomics (AUC = 0.797) models. Calibration and decision curves analysis showed that the combined model was clinically useful.

CONCLUSION

Radiomics signatures of carotid plaque on CTA can well predict plaque vulnerability, which may provide additional value to identify high-risk patients and improve outcomes.

Tumor blood vessel in 3D reconstruction CT imaging as an risk indicator for growth of pulmonary nodule with ground-glass opacity

OBJECTIVE

Despite the vital role of blood perfusion in tumor progression, in patients with persistent pulmonary nodule with ground-glass opacity (GGO) is still unclear. This study aims to investigate the relationship between tumor blood vessel and the growth of persistent malignant pulmonary nodules with ground-glass opacity (GGO).

METHODS

We collected 116 cases with persistent malignant pulmonary nodules, including 62 patients as stable versus 54 patients in the growth group, from 2017 to 2021. Three statistical methods of logistic regression model, Kaplan-Meier analysis regression analysis were used to explore the potential risk factors for growth of malignant pulmonary nodules with GGO.

RESULTS

Multivariate variables logistic regression analysis and Kaplan-Meier analysis identified that tumor blood vessel diameter (p = 0.013) was an significant risk factor in the growth of nodules and Cut-off value of tumor blood vessel diameter was 0.9 mm with its specificity 82.3% and sensitivity 66.7%.While in subgroup analysis, for the GGO CTR < 0.5[C(the maximum diameter of consolidation in tumor)/T(the maximum diameter of the whole tumor including GGO) ratio], tumor blood vessel diameter (p = 0.027) was important during the growing processes of nodules.

CONCLUSIONS

The tumor blood vessel diameter of GGO lesion was closely associated with the growth of malignant pulmonary nodules. The results of this study would provide evidence for effective follow-up strategies for pulmonary nodule screening.

Development of a novel combined nomogram integrating deep-learning-assisted CT texture and clinical-radiological features to predict the invasiveness of clinical stage IA part-solid lung adenocarcinoma: a multicentre study

AIM

To develop a novel combined nomogram based on deep-learning-assisted computed tomography (CT) texture (DL-TA) and clinical-radiological features for the preoperative prediction of invasiveness in patients with clinical stage IA lung adenocarcinoma manifesting as part-solid nodules (PSNs).

MATERIALS AND METHODS

This study was conducted from January 2015 to October 2021 at three centres: 355 patients with 355 PSN lung adenocarcinomas who underwent surgical resection were included and classified into the training (n=222) and validation (n=133) cohorts. PSN segmentation on CT images was performed automatically with a commercial deep-learning algorithm, and CT texture features were extracted. The least absolute shrinkage and selection operator was used for feature selection and transformed into a DL-TA score. The combined nomogram that incorporated the DL-TA score and identified clinical-radiological features was developed for the prediction of pathological invasiveness of the PSNs and validated in terms of discrimination and calibration.

RESULTS

The present study generated a combined nomogram for predicting the invasiveness of PSNs that included age, consolidation-to-tumour ratio, smoking status, and DL-TA score, with a C-index of 0.851 (95% confidence interval: 0.826-0.877) for the training cohort and 0.854 (95% confidence interval: 0.817-0.891) for the validation cohort, indicating good discrimination. Furthermore, the model had a Brier score of 0.153 for the training cohort and 0.135 for the validation cohort, indicating good calibration.

CONCLUSION

The developed combined nomogram consisting of the DL-TA score and clinical-radiological features and has the potential to predict the individual risk for the invasiveness of stage IA PSN lung adenocarcinomas.

Radiomics-based Machine Learning Methods for Volume Doubling Time Prediction of Pulmonary Ground-glass Nodules With Baseline Chest Computed Tomography

PURPOSE

Reliable prediction of volume doubling time (VDT) is essential for the personalized management of pulmonary ground-glass nodules (GGNs). We aimed to determine the optimal VDT prediction method by comparing different machine learning methods only based on the baseline chest computed tomography (CT) images.

MATERIALS AND METHODS

Seven classical machine learning methods were evaluated in terms of their stability and performance for VDT prediction. The VDT, calculated by the preoperative and baseline CT, was divided into 2 groups with a cutoff value of 400 days. A total of 90 GGNs from 3 hospitals constituted the training set, and 86 GGNs from the fourth hospital served as the external validation set. The training set was used for feature selection and model training, and the validation set was used to evaluate the predictive performance of the model independently.

RESULTS

The eXtreme Gradient Boosting showed the highest predictive performance (accuracy: 0.890±0.128 and area under the ROC curve (AUC): 0.896±0.134), followed by the neural network (NNet) (accuracy: 0.865±0.103 and AUC: 0.886±0.097). While regarding stability, the NNet showed the highest robustness against data perturbation (relative SDs [%] of mean AUC: 10.9%). Therefore, the NNet was chosen as the final model, achieving high accuracy of 0.756 in the external validation set.

CONCLUSION

The NNet is a promising machine learning method to predict the VDT of GGNs, which would assist in the personalized follow-up and treatment strategies for GGNs reducing unnecessary follow-up and radiation dose.

Baseline whole-lung CT features deriving from deep learning and radiomics: prediction of benign and malignant pulmonary ground-glass nodules

Objective

To develop and validate the model for predicting benign and malignant ground-glass nodules (GGNs) based on the whole-lung baseline CT features deriving from deep learning and radiomics.

Methods

This retrospective study included 385 GGNs from 3 hospitals, confirmed by pathology. We used 239 GGNs from Hospital 1 as the training and internal validation set; 115 and 31 GGNs from Hospital 2 and Hospital 3 as the external test sets 1 and 2, respectively. An additional 32 stable GGNs from Hospital 3 with more than five years of follow-up were used as the external test set 3. We evaluated clinical and morphological features of GGNs at baseline chest CT and extracted the whole-lung radiomics features simultaneously. Besides, baseline whole-lung CT image features are further assisted and extracted using the convolutional neural network. We used the back-propagation neural network to construct five prediction models based on different collocations of the features used for training. The area under the receiver operator characteristic curve (AUC) was used to compare the prediction performance among the five models. The Delong test was used to compare the differences in AUC between models pairwise.

Results

The model integrated clinical-morphological features, whole-lung radiomic features, and whole-lung image features (CMRI) performed best among the five models, and achieved the highest AUC in the internal validation set, external test set 1, and external test set 2, which were 0.886 (95% CI: 0.841-0.921), 0.830 (95%CI: 0.749-0.893) and 0.879 (95%CI: 0.712-0.968), respectively. In the above three sets, the differences in AUC between the CMRI model and other models were significant (all P < 0.05). Moreover, the accuracy of the CMRI model in the external test set 3 was 96.88%.

Conclusion

The baseline whole-lung CT features were feasible to predict the benign and malignant of GGNs, which is helpful for more refined management of GGNs.

Managing Persistent Subsolid Nodules in Lung Cancer: Education, Decision Making, and Impact of Interval Growth Patterns

With the popularization of lung cancer screening, many persistent subsolid nodules (SSNs) have been identified clinically, especially in Asian non-smokers. However, many studies have found that SSNs exhibit heterogeneous growth trends during long-term follow ups. This article adopted a narrative approach to extensively review the available literature on the topic to explore the definitions, rationale, and clinical application of different interval growths of subsolid pulmonary nodule management and follow-up strategies. The development of SSN growth thresholds with different growth patterns could support clinical decision making with follow-up guidelines to reduce over- and delayed diagnoses. In conclusion, using different SSN growth thresholds could optimize the follow-up management and clinical decision making of SSNs in lung cancer screening programs. This could further reduce the lung cancer mortality rate and potential harm from overdiagnosis and over management.

Software-based quantitative CT analysis to predict the growth trend of persistent nonsolid pulmonary nodules: a retrospective study

PURPOSE

Persistent nonsolid nodules (NSNs) usually exhibit an indolent course and may remain stable for several years; however, some NSNs grow quickly and require surgical excision. Therefore, identifying quantitative features capable of early discrimination between growing and nongrowing NSNs is becoming a crucial aspect of radiological analysis. The main purpose of this study was to evaluate the performance of an open-source software (ImageJ) to predict the future growth of NSNs detected in a Caucasian (Italian) population.

MATERIAL AND METHODS

We retrospectively selected 60 NSNs with an axial diameter of 6-30 mm scanned with the same acquisition-reconstruction parameters and the same computed tomography (CT) scanner. Software-based analysis was performed on thin-section CT images using ImageJ. For each NSNs, several quantitative features were extracted from the baseline CT images. The relationships of NSN growth with quantitative CT features and other categorical variables were analyzed using univariate and multivariable logistic regression analyses.

RESULTS

In multivariable analysis, only the skewness and linear mass density (LMD) were significantly associated with NSN growth, and the skewness was the strongest predictor of growth. In receiver operating characteristic curve analyses, the optimal cutoff values of skewness and LMD were 0.90 and 19.16 mg/mm, respectively. The two predictive models that included the skewness, with or without LMD, exhibited an excellent power for predicting NSN growth.

CONCLUSION

According to our results, NSNs with a skewness value > 0.90, specifically those with a LMD > 19.16 mg/mm, should require closer follow-up due to their higher growth potential, and higher risk of becoming an active cancer.

Neural network-based model for evaluating inert nodules and volume doubling time in T1 lung adenocarcinoma: a nested case-control study

Objective

The purpose of this study is to establish model for assessing inert nodules predicting nodule volume-doubling.

Methods

A total of 201 patients with T1 lung adenocarcinoma were analysed retrospectively pulmonary nodule information was predicted by an AI pulmonary nodule auxiliary diagnosis system. The nodules were classified into two groups: inert nodules (volume-doubling time (VDT)>600 days n=152) noninert nodules (VDT<600 days n=49). Then taking the clinical imaging features obtained at the first examination as predictive variables the inert nodule judgement model <sn</sn>>(INM) volume-doubling time estimation model (VDTM) were constructed based on a deep learning-based neural network. The performance of the INM was evaluated by the area under the curve (AUC) obtained from receiver operating characteristic (ROC) analysis the performance of the VDTM was evaluated by R²(determination coefficient).

Results

The accuracy of the INM in the training and testing cohorts was 81.13% and 77.50%, respectively. The AUC of the INM in the training and testing cohorts was 0.7707 (95% CI 0.6779-0.8636) and 0.7700 (95% CI 0.5988-0.9412), respectively. The INM was effective in identifying inert pulmonary nodules; additionally, the R2 of the VDTM in the training cohort was 0.8008, and that in the testing cohort was 0.6268. The VDTM showed moderate performance in estimating the VDT, which can provide some reference during a patients' first examination and consultation.

Conclusion

The INM and the VDTM based on deep learning can help radiologists and clinicians distinguish among inert nodules and predict the nodule volume-doubling time to accurately treat patients with pulmonary nodules.

Management of pulmonary nodules in women with pregnant intention: A review with perspective

The process for the management of pulmonary nodules in women with pregnant intention remains a challenge. There was a certain proportion of targeted female patients with high-risk lung cancer, and anxiety for suspicious lung cancer in early stage also exists. A comprehensive review of hereditary of lung cancer, effects of sexual hormone on lung cancer, natural history of pulmonary nodules, and computed tomography imaging with radiation exposure based on PubMed search was completed. The heredity of lung cancer and effects of sexual hormone on lung cancer are not the decisive factors, and the natural history of pulmonary nodules and the radiation exposure of imaging should be the main concerns. The management of incidental pulmonary nodules in young women with pregnant intention is an intricate and indecisive problem we have to encounter. The balance between the natural history of pulmonary nodules and the radiation exposure of imaging should be weighed.

Deep learning-powered 3D segmentation derives factors associated with lymphovascular invasion and prognosis in clinical T1 stage non-small cell lung cancer

Background

Lymphovascular invasion (LVI) is an invasive biologic behavior that affects the treatment and prognosis of patients with early-stage lung cancer. This study aimed to identify LVI diagnostic and prognostic biomarkers using deep learning-powered 3D segmentation with artificial intelligence (AI) technology.

Methods

Between January 2016 and October 2021, we enrolled patients with clinical T1 stage non-small cell lung cancer (NSCLC). We used commercially available AI software (Dr. Wise system, Deep-wise Corporation, China) to extract quantitative AI features of pulmonary nodules automatically. Dimensionality reduction was achieved through least absolute shrinkage and selection operator regression; subsequently, the AI score was calculated.Then, the univariate and multivariate analysis was further performed on the AI score and patient baseline parameters.

Results

Among 175 enrolled patients, 22 tested positive for LVI at pathology review. Based on the multivariate logistic regression results, we incorporated the AI score, carcinoembryonic antigen, spiculation, and pleural indentation into the nomogram for predicting LVI. The nomogram showed good discrimination (C-index = 0.915 [95% confidence interval: 0.89-0.94]); moreover, calibration of the nomogram revealed good predictive ability (Brier score = 0.072). Kaplan-Meier analysis revealed that relapse-free survival and overall survival were significantly higher among patients with a low-risk AI score and without LVI than those among patients with a high-risk AI score (p = 0.008 and p = 0.002, respectively) and with LVI (p = 0.013 and p = 0.008, respectively).

Conclusions

Our findings indicate that a high-risk AI score is a diagnostic biomarker for LVI in patients with clinical T1 stage NSCLC; accordingly, it can serve as a prognostic biomarker for these patients.

Long-term follow-up of persistent pulmonary subsolid nodules: Natural course of pure, heterogeneous, and real part-solid ground-glass nodules

BACKGROUND

Previous studies have suggested the applicability of three classifications of subsolid nodules (SSNs). However, few studies have unraveled the natural history of the three types of SSNs.

METHODS

A retrospective study from two medical centers between November 2007 and November 2017 was conducted to explore the long-term follow-up results of three different types of SSNs, which were divided into pure ground-glass nodules (pGGNs), heterogeneous ground-glass nodules (hGGNs), and real part-solid nodules (rPSNs).

RESULTS

A total of 306 consecutive patients, including 361 SSNs with long-term follow-up, were reviewed. The median growth times of pGGNs, hGGNs, and rPSNs were 7.7, 6.0, and 2.0 years, respectively. For pGGNs, the median period of development into rPSNs was 4.6 years, while that of hGGNs was 1.8 years, and the time from pGGNs to hGGNs was 3.1 years (p < 0.05). In SSNs with an initial lung window consolidation tumor ratio (LW-CTR) >0.5 and mediastinum window (MW)-CTR >0.2, all cases with growth were identified within 5 years. Meanwhile, in SSNs whose LW-CTR and MW-CTR were 0, it took over 5 years to detect nodular growth. Pathologically, 90.6% of initial SSNs with LW-CTR >0 were invasive carcinomas (invasive adenocarcinoma and micro-invasive adenocarcinoma). Among patients with rPSNs in the initial state, 100.0% of the final pathological results were invasive carcinoma. Cox regression showed that age (p = 0.038), initial maximal diameter (p < 0.001), and LW-CTR (p = 0.002) were independent risk factors for SSN growth.

CONCLUSIONS

pGGNs, hGGNs, and rPSNs have significantly different natural histories. Age, initial nodule diameter, and LW-CTR are important risk factors for SSN growth.

Natural progression of persistent pure ground-glass nodules 10 mm or smaller: long-term observation and risk factor assessment

PURPOSE

Semi-automatic segmentation was used to investigate the natural progression of pure ground-glass nodules (pGGNs) of 5-10 mm in long-term follow-up and to analyze independent risk factors for subsequent growth.

MATERIALS AND METHODS

A total of 154 pGGNs of 5-10 mm from 132 patients with 698 follow-up CT scans were retrospectively identified. Subsequently, enrolled pGGNs were semiautomatically segmented on initial and follow-up CT to obtain diameter, density and volume, thus calculating mass, volume doubling time (VDT), and mass doubling time (MDT). Kaplan‒Meier analysis and multivariate Cox proportional risk regression were performed to explore independent predictors of pGGN growth. We analyzed growth differences among different pathological results of pGGNs confirmed by surgery. The prognosis was analyzed using the total diameter or solid size of the nodules on the last preoperative CT.

RESULTS

Among the 85 (55.2%) pGGNs with growth, 5.9%, 51.8%, and 80.0% showed growth within 1, 3, and 5 years, respectively. The median VDT and MDT were 1206.4 (range 349.8-5134.4) days and 1161.3 (range 339.4-6630.4) days, respectively. The multivariate Cox risk regression analysis showed that mean CT attenuation (m-CTA) [hazard ratio (HR) = 2.098, p = 0.010] and roundness index (HR = 1.892, p = 0.021) were independent risk factors for pGGN growth. In total, 67.6% of surgically resected and growing pGGNs were invasive non-mucinous adenocarcinoma (IA), including 2 cases of endpoint events, showing a PSN with solid components of 5.6 mm and a solid nodule with a diameter of 19.9 mm.

CONCLUSIONS

pGGNs of 5-10 mm showed an indolent clinical course. Follow-up CT imaging of pGGNs in the latter half of the first two years should be a rational management strategy. Small pGGNs with a larger overall m-CTA and roundness index on baseline CT are more likely to grow.

Correlation between lung cancer probability and number of pulmonary nodules in baseline computed tomography lung cancer screening: A retrospective study based on the Chinese population

Background

Screening for lung cancer with LDCT detects a large number of nodules. However, it is unclear whether nodule number influences lung cancer probability. This study aimed to acquire deeply insight into the distribution characteristics of nodule number in the Chinese population and to reveal the association between the nodule number and the probability of lung cancer (LC).

Methods

10,167 asymptomatic participants who underwent LDCT LC screening were collected. Noncalcified nodules larger than 4 mm were included. The nodule number per participant was determined. We defined five categories according to the number of nodules (based on nodule type and size): one, two, three, four, and more than four nodules. We stratified the nodules as groups A, B, and C and participants as Amax, Bmax, and Cmax groups, and explored the association between nodule number and the probability of LC on nodule and participant levels.

Results

97 participants were confirmed to have LC. The probabilities of LC were 49/1719, 22/689, 11/327, 6/166, and 9/175 in participants with one, two, three, four, and more than four nodules (p>0.05), respectively. In the Bmax group, the probability of LC was significantly higher in participants with one nodule than those with >4 nodules (p<0.05), and the probability of LC showed a negative linear trend with increasing nodule numbers (p<0.05). Based on the nodule-level analyses, in Group B, LC probability was significantly higher when participants had a solitary nodule than when they had >4 nodules (p<0.05).

Conclusion

LC probability does not significantly change with the number of nodules. However, when stratified by the nodule size, the effect of nodule number on LC probability was nodule-size dependent, and greater attention and active follow-up are required for solitary nodules especially SNs/solid component of PSNs measuring 6-15 mm or NSNs measuring 8-15 mm. Assessing the nodule number in conjunction with nodule size in baseline LDCT LC screening is considered beneficial.

The long-term course of subsolid nodules and predictors of interval growth on chest CT: a systematic review and meta-analysis

OBJECTIVES

To calculate the pooled incidence of interval growth after long-term follow-up and identify predictors of interval growth in subsolid nodules (SSNs) on chest CT.

METHODS

A search of MEDLINE (PubMed), Cochrane Library, Web of Science Core Collection, and Embase was performed on November 08, 2021, for relevant studies. Patient information, CT scanner, and SSN follow-up information were extracted from each included study. A random-effects model was applied along with subgroup and meta-regression analyses. Study quality was assessed by the Newcastle-Ottawa scale, and publication bias was assessed by Egger's test.

RESULTS

Of the 6802 retrieved articles, 16 articles were included and analyzed, providing a total of 2898 available SSNs. The pooled incidence of growth in the 2898 SSNs was 22% (95% confidence interval [CI], 15-29%). The pooled incidence of growth in the subgroup analysis of pure ground-glass nodules was 26% (95% CI: 12-39%). The incidence of SSN growth after 2 or more years of stability was only 5% (95% CI: 3-7%). An initially large SSN size was found to be the most frequent risk factor affecting the incidence of SSN growth and the time of growth.

CONCLUSIONS

The pooled incidence of SSN growth was as high as 22%, with a 26% incidence reported for pure ground-glass nodules. Although the incidence of growth was only 5% after 2 or more years of stability, long-term follow-up is needed in certain cases. Moreover, the initial size of the SSN was the most frequent risk factor for growth.

KEY POINTS

• Based on a meta-analysis of 2898 available subsolid nodules in the literature, the pooled incidence of growth was 22% for all subsolid nodules and 26% for pure ground-glass nodules. • After 2 or more years of stability on follow-up CT, the pooled incidence of subsolid nodule growth was only 5%. • Given the incidence of subsolid nodule growth, management of these lesions with long-term follow-up is preferred.

A pre-treatment CT-based weighted radiomic approach combined with clinical characteristics to predict durable clinical benefits of immunotherapy in advanced lung cancer

OBJECTIVES

To develop a pre-treatment CT-based predictive model to anticipate inoperable lung cancer patients' progression-free survival (PFS) to immunotherapy.

METHODS

This single-center retrospective study developed and cross-validated a radiomic model in 185 patients and tested it in 48 patients. The binary endpoint is the durable clinical benefit (DCB, PFS ≥ 6 months) and non-DCB (NDCB, PFS < 6 months). Radiomic features were extracted from multiple intrapulmonary lesions and weighted by an attention-based multiple-instance learning model. Aggregated features were then selected through L2-regularized ridge regression. Five machine-learning classifiers were conducted to build predictive models using radiomic and clinical features alone and then together. Lastly, the predictive value of the model with the best performance was validated by Kaplan-Meier survival analysis.

RESULTS

The predictive models based on the weighted radiomic approach showed superior performance across all classifiers (AUCs: 0.75-0.82) compared with the largest lesion approach (AUCs: 0.70-0.78) and the average sum approach (AUCs: 0.64-0.80). Among them, the logistic regression model yielded the most balanced performance (AUC = 0.87 [95%CI 0.84-0.89], 0.75 [0.68-0.82], 0.80 [0.68-0.92] in the training, validation, and test cohort respectively). The addition of five clinical characteristics significantly enhanced the performance of radiomic-only model (train: AUC 0.91 [0.89-0.93], p = .042; validation: AUC 0.86 [0.80-0.91], p = .011; test: AUC 0.86 [0.76-0.96], p = .026). Kaplan-Meier analysis of the radiomic-based predictive models showed a clear stratification between classifier-predicted DCB versus NDCB for PFS (HR = 2.40-2.95, p < 0.05).

CONCLUSIONS

The adoption of weighted radiomic features from multiple intrapulmonary lesions has the potential to predict long-term PFS benefits for patients who are candidates for PD-1/PD-L1 immunotherapies.

KEY POINTS

• Weighted radiomic-based model derived from multiple intrapulmonary lesions on pre-treatment CT images has the potential to predict durable clinical benefits of immunotherapy in lung cancer. • Early line immunotherapy is associated with longer progression-free survival in advanced lung cancer.

The natural growth history of persistent pulmonary subsolid nodules: Radiology, genetics, and clinical management

The high detection rate of pulmonary subsolid nodules (SSN) is an increasingly crucial clinical issue due to the increased number of screening tests and the growing popularity of low-dose computed tomography (LDCT). The persistence of SSN strongly suggests the possibility of malignancy. Guidelines have been published over the past few years and guide the optimal management of SSNs, but many remain controversial and confusing for clinicians. Therefore, in-depth research on the natural growth history of persistent pulmonary SSN can help provide evidence-based medical recommendations for nodule management. In this review, we briefly describe the differential diagnosis, growth patterns and rates, genetic characteristics, and factors that influence the growth of persistent SSN. With the advancement of radiomics and artificial intelligence (AI) technology, individualized evaluation of SSN becomes possible. These technologies together with liquid biopsy, will promote the transformation of current diagnosis and follow-up strategies and provide significant progress in the precise management of subsolid nodules in the early stage of lung cancer.

CT and CEA-based machine learning model for predicting malignant pulmonary nodules

Computed tomography (CT), an efficient radiological technology, is used to detect lung cancer in the clinic. Carcinoembryonic antigen (CEA), a common tumor biomarker, is applied in the detection of various tumors. To highlight the advantages of two-dimensional techniques and assist clinicians in optimizing lung cancer diagnostic schemes, we established a favorable model combining CT and CEA. In the study, univariate analysis was performed to screen independent predictors in a training cohort of 271 patients with malignant pulmonary nodules (MPNs) and 92 with benign pulmonary nodules (BPNs). Six machine learning-based models involving five CT predictors (mediastinal lymph node enlargement, lobulation, vascular notch sign, spiculation, and nodule number) and lnCEA were constructed and validated in an independent cohort of 129 participants (92 MPNs and 37 BPNs) by SPSS Modeler. A nomogram and the Delong test were generated by R software. Finally, the model established by logistic regression had highest diagnostic efficiency (area under the curve [AUC] = 0.912). Moreover, the diagnostic ability of the logistic model in the validation cohort (AUC = 0.882, 80.4% sensitivity, 75.7% specificity) was higher than that of the Peking University model (AUC = 0.712, 68.5% sensitivity, 70.3% specificity) and the Mayo model (AUC = 0.745, 62.0% sensitivity, 75.7% specificity). Interestingly, for the participants with intermediate (10-30 mm) and CEA-negative nodule, the model reached an AUC of 0.835 (72.3% sensitivity, 83.3% specificity). The AUC for the early lung cancer was as high as 0.822 with 67.3% sensitivity and 78.9% specificity. As a conclusion, this promising model presents a new diagnostic strategy for the clinic to distinguish MPNs from BPNs.

Artificial Intelligence (AI) for Lung Nodules, From the AJR Special Series on AI Applications

Interest in artificial intelligence (AI) applications for lung nodules continues to grow among radiologists, particularly with the expanding eligibility criteria and clinical utilization of lung cancer screening CT. AI has been heavily investigated for detecting and characterizing lung nodules and for guiding prognostic assessment. AI tools have also been used for image postprocessing (e.g., rib suppression on radiography or vessel suppression on CT) and for noninterpretive aspects of reporting and workflow, including management of nodule follow-up. Despite growing interest in and rapid development of AI tools and FDA approval of AI tools for pulmonary nodule evaluation, integration into clinical practice has been limited. Challenges to clinical adoption have included concerns about generalizability, regulatory issues, technical hurdles in implementation, and human skepticism. Further validation of AI tools for clinical use and demonstration of benefit in terms of patient-oriented outcomes also are needed. This article provides an overview of potential applications of AI tools in the imaging evaluation of lung nodules and discusses the challenges faced by practices interested in clinical implementation of such tools.

Measuring pure ground-glass nodules on computed tomography: assessing agreement between a commercially available deep learning algorithm and radiologists’ readings

Background

Deep learning algorithms (DLAs) could enable automatic measurements of solid portions of mixed ground-glass nodules (mGGNs) in agreement with the invasive component sizes measured during pathologic examinations. However, the measurement of pure ground-glass nodules (pGGNs) based on DLAs has rarely been reported in the literature.

Purpose

To evaluate the use of a commercially available DLA for the automatic measurement of pGGNs on computed tomography (CT).

Material and Methods

In this retrospective study, we included 68 patients with 81 pGGNs. The maximum diameter of the nodules was manually measured by senior radiologists and automatically segmented and measured by the DLA. Agreement between the measurements by the radiologist and DLA was assessed using Bland–Altman plots, and correlations were analyzed using Pearson correlation. Finally, we evaluated the association between the radiologist and DLA measurements and the invasiveness of lung adenocarcinoma in patients with pGGNs on preoperative CT.

Results

The radiologist and DLA measurements exhibited good agreement with a Bland–Altman bias of 3.0%, which were clinically acceptable. The correlation between both sets of maximum diameters was also strong, with a Pearson correlation coefficient of 0.968 ( P < 0.001). In addition, both sets of maximum diameters were larger in the invasive adenocarcinoma group than in the non-invasive adenocarcinoma group ( P < 0.001).

Conclusion

Automatic pGGNs measurements by the DLA were comparable with those measured manually and were closely associated with the invasiveness of lung adenocarcinoma.

Assessing the inflammatory severity of the terminal ileum in Crohn disease using radiomics based on MRI

BACKGROUND

Evaluating inflammatory severity using imaging is essential for Crohn's disease, but it is limited by potential interobserver variation and subjectivity. We compared the efficiency of magnetic resonance index of activity (MaRIA) collected by radiologists and a radiomics model in assessing the inflammatory severity of terminal ileum (TI).

METHODS

121 patients were collected from two centers. Patients were divided into ulcerative group and mucosal remission group based on the TI Crohn's disease Endoscopic Severity Index. The consistency of bowel wall thickness (BWT), relative contrast enhancement (RCE), edema, ulcer, MaRIA and features of the region of interest between radiologists were described by weighted Kappa test and intraclass correlation coefficient (ICC), and developed receiver operating curve of MaRIA. The radiomics model was established using reproducible features of logistic regression based on arterial staging of T1WI sequences. Delong test was used to compare radiomics with MaRIA.

RESULTS

The consistency between radiologists were moderate in BWT (ICC = 0.638), fair in edema (κ = 0.541), RCE (ICC = 0.461), MaRIA (ICC = 0.579) and poor in ulcer (κ = 0.271). Radiomics model was developed by 6 reproducible features (ICC = 0.93-0.96) and equivalent to MaRIA which evaluated by the senior radiologist (0.872 vs 0.883 in training group, 0.824 vs 0.783 in validation group, P = 0.847, 0.471), both of which were significantly higher than MaRIA evaluated by junior radiologist (AUC: 0.621 in training group, 0.557 in validation group, all, P < 0.05).

CONCLUSION

The evaluation of inflammatory severity could be performed by radiomics objectively and reproducibly, and was comparable to MaRIA evaluated by the senior radiologist. Radiomics may be an important method to assist junior radiologists to assess the severity of inflammation objectively and accurately.

Performance Analysis of State-of-the-Art CNN Architectures for LUNA16

The convolutional neural network (CNN) has become a powerful tool in machine learning (ML) that is used to solve complex problems such as image recognition, natural language processing, and video analysis. Notably, the idea of exploring convolutional neural network architecture has gained substantial attention as well as popularity. This study focuses on the intrinsic various CNN architectures: LeNet, AlexNet, VGG16, ResNet-50, and Inception-V1, which have been scrutinized and compared with each other for the detection of lung cancer using publicly available LUNA16 datasets. Furthermore, multiple performance optimizers: root mean square propagation (RMSProp), adaptive moment estimation (Adam), and stochastic gradient descent (SGD), were applied for this comparative study. The performances of the three CNN architectures were measured for accuracy, specificity, sensitivity, positive predictive value, false omission rate, negative predictive value, and F1 score. The experimental results showed that the CNN AlexNet architecture with the SGD optimizer achieved the highest validation accuracy for CT lung cancer with an accuracy of 97.42%, misclassification rate of 2.58%, 97.58% sensitivity, 97.25% specificity, 97.58% positive predictive value, 97.25% negative predictive value, false omission rate of 2.75%, and F1 score of 97.58%. AlexNet with the SGD optimizer was the best and outperformed compared to the other state-of-the-art CNN architectures.

Radiomics for differentiating minimally invasive adenocarcinoma from precursor lesions in pure ground-glass opacities on chest computed tomography

OBJECTIVE

To explore the correlation between radiomic features and the pathology of pure ground-glass opacities (pGGOs), we established a radiomics model for predicting the pathological subtypes of minimally invasive adenocarcinoma (MIA) and precursor lesions.

METHODS

CT images of 1521 patients with lung adenocarcinoma or precursor lesions appearing as pGGOs on CT in our hospital (The Third Affiliated Hospital of Sun Yat-sen University) from January 2015 to March 2021 were analyzed retrospectively and selected based on inclusion and exclusion criteria. pGGOs were divided into an atypical adenomatous hyperplasia (AAH)/adenocarcinoma in situ (AIS) group and an MIA group. Radiomic features were extracted from the original and preprocessed images of the region of interest. ANOVA and least absolute shrinkage and selection operator feature selection algorithm were used for feature selection. Logistic regression algorithm was used to construct radiomics prediction model. Receiver operating characteristic curves were used to evaluate the classification efficiency.

RESULTS

129 pGGOs were included. 2107 radiomic features were extracted from each region of interest. 18 radiomic features were eventually selected for model construction. The area under the curve of the radiomics model was 0.884 [95% confidence interval (CI), 0.818-0.949] in the training set and 0.872 (95% CI, 0.756-0.988) in the test set, with a sensitivity of 72.73%, specificity of 88.24% and accuracy of 79.47%. The decision curve indicated that the model had a high net benefit rate.

CONCLUSION

The prediction model for pathological subtypes of MIA and precursor lesions in pGGOs demonstrated a high diagnostic accuracy.

ADVANCES IN KNOWLEDGE

We focused on lesions appearing as pGGOs on CT and revealed the differences in radiomic features between MIA and precursor lesions. We constructed a radiomics prediction model and improved the diagnostic accuracy for the pathology of MIA and precursor lesions.

Invasive Prediction of Ground Glass Nodule Based on Clinical Characteristics and Radiomics Feature

Objective: To explore the diagnostic value of CT radiographic images and radiomics features for invasive classification of lung adenocarcinoma manifesting as ground-glass nodules (GGNs) in computer tomography (CT).

Methods: A total of 312 GGNs were enrolled in this retrospective study. All GGNs were randomly divided into training set (n = 219) and test set (n = 93). Univariate and multivariate logistic regressions were used to establish a clinical model, while the minimum redundancy maximum relevance (mRMR) and least absolute shrinkage and selection operator (LASSO) algorithm were used to select the radiomics features and construct the radiomics model. A combined model was finally built by combining these two models. The performance of these models was assessed in both training and test set. A combined nomogram was developed based on the combined model and evaluated with its calibration curves and C-index.

Results: Diameter [odds ratio (OR), 1.159; p ＜ 0.001], lobulation (OR, 2.953; p = 0.002), and vascular changes (OR, 3.431; p ＜ 0.001) were retained as independent predictors of the invasive adenocarcinoma (IAC) group. Eleven radiomics features were selected by mRMR and LASSO method to established radiomics model. The clinical model and radiomics mode showed good predictive ability in both training set and test set. When two models were combined, the diagnostic area under the curve (AUC) value was higher than the single clinical or radiomics model (training set: 0.86 vs. 0.83 vs. 0.82; test set: 0.80 vs. 0.78 vs. 0.79). The constructed combined nomogram could effectively quantify the risk degree of 3 image features and Rad score with a C-index of 0.855 (95%: 0.805∼0.905).

Conclusion: Radiographic and radiomics features show high accuracy in the invasive diagnosis of GGNs, and their combined analysis can improve the diagnostic efficacy of IAC manifesting as GGNs. The nomogram, serving as a noninvasive and accurate predictive tool, can help judge the invasiveness of GGNs prior to surgery and assist clinicians in creating personalized treatment strategies.

Analysis of Tracheobronchial Diverticula Based on Semantic Segmentation of CT Images via the Dual-Channel Attention Network

Tracheobronchial diverticula (TD) is a common cystic lesion that can be easily neglected; hence accurate and rapid identification is critical for later diagnosis. There is a strong need to automate this diagnostic process because traditional manual observations are time-consuming and laborious. However, most studies have only focused on the case report or listed the relationship between the disease and other physiological indicators, but a few have adopted advanced technologies such as deep learning for automated identification and diagnosis. To fill this gap, this study interpreted TD recognition as semantic segmentation and proposed a novel attention-based network for TD semantic segmentation. Since the area of TD lesion is small and similar to surrounding organs, we designed the atrous spatial pyramid pooling (ASPP) and attention mechanisms, which can efficiently complete the segmentation of TD with robust results. The proposed attention model can selectively gather features from different branches according to the amount of information they contain. Besides, to the best of our knowledge, no public research data is available yet. For efficient network training, we constructed a data set containing 218 TD and related ground truth (GT). We evaluated different models based on the proposed data set, among which the highest MIOU can reach 0.92. The experiments show that our model can outperform state-of-the-art methods, indicating that the deep learning method has great potential for TD recognition.

Identifying Cancer Subtypes Using a Residual Graph Convolution Model on a Sample Similarity Network

Cancer subtype classification helps us to understand the pathogenesis of cancer and develop new cancer drugs, treatment from which patients would benefit most. Most previous studies detect cancer subtypes by extracting features from individual samples, ignoring their associations with others. We believe that the interactions of cancer samples can help identify cancer subtypes. This work proposes a cancer subtype classification method based on a residual graph convolutional network and a sample similarity network. First, we constructed a sample similarity network regarding cancer gene co-expression patterns. Then, the gene expression profiles of cancer samples as initial features and the sample similarity network were passed into a two-layer graph convolutional network (GCN) model. We introduced the initial features to the GCN model to avoid over-smoothing during the training process. Finally, the classification of cancer subtypes was obtained through a softmax activation function. Our model was applied to breast invasive carcinoma (BRCA), glioblastoma multiforme (GBM) and lung cancer (LUNG) datasets. The accuracy values of our model reached 82.58%, 85.13% and 79.18% for BRCA, GBM and LUNG, respectively, which outperformed the existing methods. The survival analysis of our results proves the significant clinical features of the cancer subtypes identified by our model. Moreover, we can leverage our model to detect the essential genes enriched in gene ontology (GO) terms and the biological pathways related to a cancer subtype.

Volumetric segmentation of ground glass nodule based on 3D attentional cascaded residual U-net and conditional radom field

BACKGROUND

Ground glass nodule (GGN) segmentation is one of the important and challenging tasks in diagnosing early-stage lung adenocarcinomas. Manually delineation of 3D GGN in computed tomography (CT) image is a subjective, laborious and tedious task, which presents poor repeatability.

PURPOSE

To reduce the annotation burden and improve the segmentation performance, this study proposes a 3D deep learning-based volumetric segmentation model to segment the GGN in CT images.

METHODS

A total of 379 GGNs were retrospectively collected from the public database, Shanghai Pulmonary Hospital (SHPH) and Fudan University Shanghai Cancer Center (FUSCC). First, a series of image pre-processing techniques involving image resampling, intensity normalization, 3D nodule patch cropping, and data augmentation, were adopted to generate the input images for the deep learning model by using the CT scans. Then, a 3D attentional cascaded residual network (ACRU-Net) was proposed to develop the deep learning-based segmentation model by using residual network and atrous spatial pyramid pooling module. To improve the model performance, a voxel-based conditional random field (CRF) method was used to optimize the segmentation results. Finally, a balanced cross-entropy and Dice combined loss function was applied to train and build the segmentation model.

RESULTS

Testing on SHPH and FUSCC datasets, the proposed method generates the Dice coefficients of 0.721±0.167 and 0.733±0.100, respectively, which are higher than that of 3D residual U-Net and ACRU-Net without CRF optimization.

CONCLUSIONS

The results demonstrated that combining 3D ACRU-Net and CRF effectively improved the GGN segmentation performance. The proposed segmentation model may provide a potential tool to help the radiologist in the segmentation and diagnosis of 3D GGN. This article is protected by copyright. All rights reserved.

Convolutional Neural Network-Based Diagnostic Model for a Solid, Indeterminate Solitary Pulmonary Nodule or Mass on Computed Tomography

PURPOSE

To establish a non-invasive diagnostic model based on convolutional neural networks (CNNs) to distinguish benign from malignant lesions manifesting as a solid, indeterminate solitary pulmonary nodule (SPN) or mass (SPM) on computed tomography (CT).

METHOD

A total of 459 patients with solid indeterminate SPNs/SPMs on CT were ultimately included in this retrospective study and assigned to the train (n=366), validation (n=46), and test (n=47) sets. Histopathologic analysis was available for each patient. An end-to-end CNN model was proposed to predict the natural history of solid indeterminate SPN/SPMs on CT. Receiver operating characteristic curves were plotted to evaluate the predictive performance of the proposed CNN model. The accuracy, sensitivity, and specificity of diagnoses by radiologists alone were compared with those of diagnoses by radiologists by using the CNN model to assess its clinical utility.

RESULTS

For the CNN model, the AUC was 91% (95% confidence interval [CI]: 0.83-0.99) in the test set. The diagnostic accuracy of radiologists with the CNN model was significantly higher than that without the model (89 vs. 66%, P<0.01; 87 vs. 61%, P<0.01; 85 vs. 66%, P=0.03, in the train, validation, and test sets, respectively). In addition, while there was a slight increase in sensitivity, the specificity improved significantly by an average of 42% (the corresponding improvements in the three sets ranged from 43, 33, and 42% to 82, 78, and 84%, respectively; P<0.01 for all).

CONCLUSION

The CNN model could be a valuable tool in non-invasively differentiating benign from malignant lesions manifesting as solid, indeterminate SPNs/SPMs on CT.

Consecutive Serial Non-Contrast CT Scan-Based Deep Learning Model Facilitates the Prediction of Tumor Invasiveness of Ground-Glass Nodules

Introduction

Tumors are continuously evolving biological systems which can be monitored by medical imaging. Previous studies only focus on single timepoint images, whether the performance could be further improved by using serial noncontrast CT imaging obtained during nodule follow-up management remains unclear. In this study, we evaluated DL model for predicting tumor invasiveness of GGNs through analyzing time series CT images.

Methods

A total of 168 pathologically confirmed GGN cases (48 noninvasive lesions and 120 invasive lesions) were retrospectively collected and randomly assigned to the development dataset (n = 123) and independent testing dataset (n = 45). All patients underwent consecutive noncontrast CT examinations, and the baseline CT and 3-month follow-up CT images were collected. The gross region of interest (ROI) patches containing only tumor region and the full ROI patches including both tumor and peritumor regions were cropped from CT images. A baseline model was built on the image features and demographic features. Four DL models were proposed: two single-DL model using gross ROI (model 1) or full ROI patches (model 3) from baseline CT images, and two serial-DL models using gross ROI (model 2) or full ROI patches (model 4) from consecutive CT images (baseline scan and 3-month follow-up scan). In addition, a combined model integrating serial full ROI patches and clinical information was also constructed. The performance of these predictive models was assessed with respect to discrimination and clinical usefulness.

Results

The area under the curve (AUC) of the baseline model, models 1, 2, 3, and 4 were 0.562 [(95% confidence interval (C)], 0.406~0.710), 0.693 (95% CI, 0.538-0.822), 0.787 (95% CI, 0.639-0.895), 0.727 (95% CI, 0.573-0.849), and 0.811 (95% CI, 0.667-0.912) in the independent testing dataset, respectively. The results indicated that the peritumor region had potential to contribute to tumor invasiveness prediction, and the model performance was further improved by integrating imaging scans at multiple timepoints. Furthermore, the combined model showed best discrimination ability, with AUC, sensitivity, specificity, and accuracy achieving 0.831 (95% CI, 0.690-0.926), 86.7%, 73.3%, and 82.2%, respectively.

Conclusion

The DL model integrating full ROIs from serial CT images shows improved predictive performance in differentiating noninvasive from invasive GGNs than the model using only baseline CT images, which could benefit the clinical management of GGNs.

High-resolution CT image analysis based on 3D convolutional neural network can enhance the classification performance of radiologists in classifying pulmonary non-solid nodules

OBJECTIVE

To investigate whether 3D convolutional neural network (CNN) is able to enhance the classification performance of radiologists in classifying pulmonary non-solid nodules (NSNs).

MATERIALS AND METHODS

Data of patients with solitary NSNs and diagnosed as adenocarcinoma in situ (AIS), minimally invasive adenocarcinoma (MIA), or invasive adenocarcinoma (IAC) in pathological after surgical resection were analyzed retrospectively. Ultimately, 532 patients in our institution were included in the study: 427 cases (144 AIS, 167 MIA, 116 IAC) were assigned to training dataset and 105 cases (36 AIS, 41 MIA and 28 IAC) were assigned to validation dataset. For external validation, 177 patients (60 AIS, 69 MIA and 48 IAC) from another hospital were assigned to testing dataset. The clinical and morphological characteristics of NSNs were established as radiologists' model. The trained classification model based on 3D CNN was used to identify NSNs types automatically. The evaluation and comparison on classification performance of the two models and CNN + radiologists' model were performed via receiver operating curve (ROC) analysis and integrated discrimination improvement (IDI) index. The Akaike information criterion (AIC) was calculated to find the best-fit model.

RESULTS

In external testing dataset, radiologists' model showed inferior classification performance than CNN model both in discriminating AIS from MIA-IAC and AIS-MIA from IAC (the area under the ROC curve (Az value), 0.693 vs 0.820, P = 0.011; 0.746 vs 0.833, P = 0.026, respectively). However, combining CNN significantly enhanced the classification performance of radiologists and exhibited higher Az values than CNN model alone (Az values, 0.893 vs 0.820, P < 0.001; 0.906 vs 0.833, P < 0.001, respectively). The IDI index further confirmed CNN's contribution to radiologists in classifying NSNs (IDI = 25.8 % (18.3-46.1 %), P < 0.001; IDI = 30.1 % (26.1-45.2 %), P < 0.001, respectively). The CNN + radiologists' model also provided the best fit over radiologists' model and CNN model alone (AIC value 63.3 % vs. 29.5 %, 49.5 %, P < 0.001; 69.2 % vs. 34.9 %, 53.6 %, P < 0.001, respectively).

CONCLUSION

CNN successfully classified NSNs based on CT images and its classification performance were superior to radiologists' model. But the classification performance of radiologists can be significantly enhanced when combined with CNN in classifying NSNs.

Invasive adenocarcinoma manifesting as pure ground glass nodule with different size: radiological characteristics differ while prognosis remains the same

BACKGROUND

Invasive adenocarcinoma (IA) manifesting as pure ground-glass nodule is rare and not been well studied. Meanwhile, tumor size is considered as a predictor of invasiveness in lung adenocarcinoma. The present study aimed to investigate the radiological and pathological characteristics as well as prognosis of IA manifesting as pure ground-glass nodule with different sizes.

METHODS

Patients with solitary pure ground glass nodule (GGN) who underwent resection and were pathologically diagnosed as IA between July 2013 and July 2015 were included. Nodules were divided into four groups according to size: A, B, C, and D, corresponding to "≤1 cm," "1-2 cm," "2-3 cm," and ">3 cm," respectively. The correlations and differences in radiological and pathological characteristics as well as prognosis among these groups were analyzed.

RESULTS

The amounts of nodules in groups A, B, C, and D are 17, 148, 78, and 30, respectively. The average diameter of these 273 nodules is 1.9 (1.5-2.4) cm. A large tumor is likely to have low computed tomography (CT) value (P<0.001), irregular shape (P=0.001), spiculation appearance (P<0.001) and exhibit pleural indentation (P<0.001) and air bronchogram (P<0.001). The proportion of lepidic predominant adenocarcinoma (LPA) (n=239, 87.5%) is much higher than that of other subtypes (n=34, 12.5%). Currently, there is no case with lymphatic, pleural, or vessel invasion and lymph node involvement, and none died of recurrence or metastasis within 5 years after resection.

CONCLUSIONS

For IA manifesting as pure ground-glass nodule, size is correlated to invasiveness, and large tumors tend to have lower CT value, an irregular shape, lobulation and spiculation appearance and exhibit pleural indentation and air bronchogram. Nevertheless, the prognosis is excellent with 100% 5-year disease-free survival regardless of the size and pathological subtype.

Lung Nodule Segmentation with a Region-Based Fast Marching Method

When dealing with computed tomography volume data, the accurate segmentation of lung nodules is of great importance to lung cancer analysis and diagnosis, being a vital part of computer-aided diagnosis systems. However, due to the variety of lung nodules and the similarity of visual characteristics for nodules and their surroundings, robust segmentation of nodules becomes a challenging problem. A segmentation algorithm based on the fast marching method is proposed that separates the image into regions with similar features, which are then merged by combining regions growing with k-means. An evaluation was performed with two distinct methods (objective and subjective) that were applied on two different datasets, containing simulation data generated for this study and real patient data, respectively. The objective experimental results show that the proposed technique can accurately segment nodules, especially in solid cases, given the mean Dice scores of 0.933 and 0.901 for round and irregular nodules. For non-solid and cavitary nodules the performance dropped—0.799 and 0.614 mean Dice scores, respectively. The proposed method was compared to active contour models and to two modern deep learning networks. It reached better overall accuracy than active contour models, having comparable results to DBResNet but lesser accuracy than 3D-UNet. The results show promise for the proposed method in computer-aided diagnosis applications.

Determining the invasiveness of ground-glass nodules using a 3D multi-task network

OBJECTIVES

The aim of this study was to determine the invasiveness of ground-glass nodules (GGNs) using a 3D multi-task deep learning network.

METHODS

We propose a novel architecture based on 3D multi-task learning to determine the invasiveness of GGNs. In total, 770 patients with 909 GGNs who underwent lung CT scans were enrolled. The patients were divided into the training (n = 626) and test sets (n = 144). In the test set, invasiveness was classified using deep learning into three categories: atypical adenomatous hyperplasia (AAH) and adenocarcinoma in situ (AIS), minimally invasive adenocarcinoma (MIA), and invasive pulmonary adenocarcinoma (IA). Furthermore, binary classifications (AAH/AIS/MIA vs. IA) were made by two thoracic radiologists and compared with the deep learning results.

RESULTS

In the three-category classification task, the sensitivity, specificity, and accuracy were 65.41%, 82.21%, and 64.9%, respectively. In the binary classification task, the sensitivity, specificity, accuracy, and area under the ROC curve (AUC) values were 69.57%, 95.24%, 87.42%, and 0.89, respectively. In the visual assessment of GGN invasiveness of binary classification by the two thoracic radiologists, the sensitivity, specificity, and accuracy of the senior and junior radiologists were 58.93%, 90.51%, and 81.35% and 76.79%, 55.47%, and 61.66%, respectively.

CONCLUSIONS

The proposed multi-task deep learning model achieved good classification results in determining the invasiveness of GGNs. This model may help to select patients with invasive lesions who need surgery and the proper surgical methods.

KEY POINTS

• The proposed multi-task model has achieved good classification results for the invasiveness of GGNs. • The proposed network includes a classification and segmentation branch to learn global and regional features, respectively. • The multi-task model could assist doctors in selecting patients with invasive lesions who need surgery and choosing appropriate surgical methods.

Natural history of pathologically confirmed pulmonary subsolid nodules with deep learning-assisted nodule segmentation

OBJECTIVE

To explore the natural history of pulmonary subsolid nodules (SSNs) with different pathological types by deep learning-assisted nodule segmentation.

METHODS

Between June 2012 and June 2019, 95 resected SSNs with preoperative long-term follow-up were enrolled in this retrospective study. SSN detection and segmentation were performed on preoperative follow-up CTs using the deep learning-based Dr. Wise system. SSNs were categorized into invasive adenocarcinoma (IAC, n = 47) and non-IAC (n = 48) groups; according to the interval change during the preoperative follow-up, SSNs were divided into growth (n = 68), nongrowth (n = 22), and new emergence (n = 5) groups. We analyzed the cumulative percentages and pattern of SSN growth and identified significant factors for IAC diagnosis and SSN growth.

RESULTS

The mean preoperative follow-up was 42.1 ± 17.0 months. More SSNs showed growth or new emergence in the IAC than in the non-IAC group (89.4% vs. 64.6%, p = 0.009). Volume doubling time was non-significantly shorter for IACs than for non-IACs (1436.0 ± 1188.2 vs. 2087.5 ± 1799.7 days, p = 0.077). Median mass doubling time was significantly shorter for IACs than for non-IACs (821.7 vs. 1944.1 days, p = 0.001). Lobulated sign (p = 0.002) and SSN mass (p = 0.004) were significant factors for differentiating IACs. IACs showed significantly higher cumulative growth percentages than non-IACs in the first 70 months of follow-up. The growth pattern of SSNs may conform to the exponential model. The initial volume (p = 0.042) was a predictor for SSN growth.

CONCLUSIONS

IACs appearing as SSNs showed an indolent course. The mean growth rate was larger for IACs than for non-IACs. SSNs with larger initial volume are more likely to grow.

KEY POINTS

• Invasive adenocarcinomas (IACs) appearing as subsolid nodules (SSNs), with a mean volume doubling time (VDT) of 1436.0 ± 1188.2 days and median mass doubling time (MDT) of 821.7 days, showed an indolent course. • The VDT was shorter for IACs than for non-IACs (1436.0 ± 1188.2 vs. 2087.5 ± 1799.7 days), but the difference was not significant (p = 0.077). The median MDT was significantly shorter for IACs than for non-IACs (821.7 vs. 1944.1 days, p = 0.001). • SSNs with lobulated sign and larger mass (> 390.5 mg) may very likely be IACs. SSNs with larger initial volume are more likely to grow.

Dynamic evaluation of lung involvement during coronavirus disease-2019 (COVID-19) with quantitative lung CT

Purpose

To identify and quantify lung changes associated with coronavirus disease-2019 (COVID-19) with quantitative lung CT during the disease.

Methods

This retrospective study reviewed COVID-19 patients who underwent multiple chest CT scans during their disease course. Quantitative lung CT was used to determine the nature and volume of lung involvement. A semi-quantitative scoring system was also used to evaluate lung lesions.

Results

This study included eighteen cases (4 cases in mild type, 10 cases in moderate type, 4 cases in severe type, and without critical type cases) with confirmed COVID-19. Patients had a mean hospitalized period of 24.1 ± 7.1 days (range: 14–38 days) and underwent an average CT scans of 3.9 ± 1.6 (range: 2–8). The total volumes of lung abnormalities reached a peak of 8.8 ± 4.1 days (range: 2–14 days). The ground-glass opacity (GGO) volume percentage was higher than the consolidative opacity (CO) volume percentage on the first CT examination (Z = 2.229, P = 0.026), and there was no significant difference between the GGO volume percentage and that of CO at the peak stage (Z = - 0.628, P = 0.53). The volume percentage of lung involvement identified by AI demonstrated a strong correlation with the total CT scores at each stage (r = 0.873, P = 0.0001).

Conclusions

Quantitative lung CT can automatically identify the nature of lung involvement and quantify the dynamic changes of lung lesions on CT during COVID-19. For patients who recovered from COVID-19, GGO was the predominant imaging feature on the initial CT scan, while GGO and CO were the main appearances at peak stage.

Electronic supplementary material

The online version of this article (10.1007/s10140-020-01856-4) contains supplementary material, which is available to authorized users.

Application of artificial intelligence in clinical non-small cell lung cancer

Lung cancer is the most common cause of cancer death in the world. Early diagnosis, screening and precise individualized treatment can significantly reduce the death rate of lung cancer. Artificial intelligence (AI) has been shown to be able to help clinicians make more accurate judgments and decisions in many ways. It has been involved in the screening of lung cancer, the judgment of benign and malignant degree of pulmonary nodules, the classification of histological cancer, the differentiation of histological subtypes, the identification of genomics, the judgment of the effectiveness of treatment and even the prognosis. AI has shown that it can be an excellent assistant for clinicians. This paper reviews the application of AI in the field of non-small cell lung cancer and describes the relevant progress. Although most of the studies to evaluate the clinical application of AI in non-small cell lung cancer have not been repeatable and generalizable, the research results highlight the efforts to promote the clinical application of AI technology and influence the future treatment direction.

Recent developments and advances in secondary prevention of lung cancer

Lung cancer prevention may include primary prevention strategies, such as corrections of working conditions and life style - primarily smoking cessation - as well as secondary prevention strategies, aiming at early detection that allows better survival rates and limited resections. This review summarizes recent developments and advances in secondary prevention, focusing on recent technological tools for an effective early diagnosis.

Clinical, Conventional CT and Radiomic Feature-Based Machine Learning Models for Predicting ALK Rearrangement Status in Lung Adenocarcinoma Patients

Objectives: To predict the anaplastic lymphoma kinase (ALK) mutations in lung adenocarcinoma patients non-invasively with machine learning models that combine clinical, conventional CT and radiomic features.

Methods: This retrospective study included 335 lung adenocarcinoma patients who were randomly divided into a primary cohort (268 patients; 90 ALK-rearranged; and 178 ALK wild-type) and a test cohort (67 patients; 22 ALK-rearranged; and 45 ALK wild-type). One thousand two hundred and eighteen quantitative radiomic features were extracted from the semi-automatically delineated volume of interest (VOI) of the entire tumor using both the original and the pre-processed non-enhanced CT images. Twelve conventional CT features and seven clinical features were also collected. Normalized features were selected using a sequential of the F-test-based method, the density-based spatial clustering of applications with noise (DBSCAN) method, and the recursive feature elimination (RFE) method. Selected features were then used to build three predictive models (radiomic, radiological, and integrated models) for the ALK-rearranged phenotype by a soft voting classifier. Models were evaluated in the test cohort using the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, and specificity, and the performances of three models were compared using the DeLong test.

Results: Our results showed that the addition of clinical information and conventional CT features significantly enhanced the validation performance of the radiomic model in the primary cohort (AUC = 0.83–0.88, P = 0.01), but not in the test cohort (AUC = 0.80–0.88, P = 0.29). The majority of radiomic features associated with ALK mutations reflected information around and within the high-intensity voxels of lesions. The presence of the cavity and left lower lobe location were new imaging phenotypic patterns in association with ALK-rearranged tumors. Current smoking was strongly correlated with non-ALK-mutated lung adenocarcinoma.

Conclusions: Our study demonstrates that radiomics-derived machine learning models can potentially serve as a non-invasive tool to identify ALK mutation of lung adenocarcinoma.

What's New on Quantitative CT Analysis as a Tool to Predict Growth in Persistent Pulmonary Subsolid Nodules? A Literature Review

Pulmonary subsolid nodules (SSNs) are observed not infrequently on thin-section chest computed tomography (CT) images. SSNs persisting after a follow-up period of three to six months have a high likelihood of being pre-malignant or malignant lesions. Malignant SSNs usually represent the histologic spectrum of pulmonary adenocarcinomas, and pulmonary adenocarcinomas presenting as SSNs exhibit quite heterogeneous behavior. In fact, while most lesions show an indolent course and may grow very slowly or remain stable for many years, others may exhibit significant growth in a relatively short time. Therefore, it is not yet clear which persistent SSNs should be surgically removed and for how many years stable SSNs should be monitored. In order to solve these two open issues, the use of quantitative analysis has been proposed to define the "tailored" management of persistent SSNs. The main purpose of this review was to summarize recent results about quantitative CT analysis as a diagnostic tool for predicting the behavior of persistent SSNs. Thus, a literature search was conducted in PubMed/MEDLINE, Scopus, and Web of Science databases to find original articles published from January 2014 to October 2019. The results of the selected studies are presented and compared in a narrative way.