Subsolid Nodules Research Articles

Cross-Scale Integration of Nano-Sized Extracellular Vesicle-Based Biomarker and Radiomics Features for Predicting Suspected Sub-Solid Pulmonary Nodules.

Sub-solid nodules (SSN) are common radiographic findings. Due to possibility of malignancy, further evaluation is urgentlyneeded for prevention and management of lung cancer (LC). This current study enrolled patients with SSN, including LC, benign nodules (BN), and healthy individuals as a control, to discover small extracellular vesicles (sEVs) differentially expressed miRNAs (DEMs) as biomarker by next-generation sequencing (NGS) and validation by RT-qPCR. Through cross-scale integration of validated small-molecule and macro-imaging, the prediction model was developed by logistic algorithms and further interpreted into an easy-to-use Nomogram by Cox-proportional hazards modeling. Present study has discovered various sEVs DEMs and sEVs-miR-424-5p that were selected and validated as novel potential biomarkers for cancerous nodule, namely LC. Furthermore, the 10 radiomics signs and 4 clinical features of SSN were merged with sEVs-miR-424-5p and proceeded in multivariate logistic regression analysis to develop the cross-scale integrated modeling, which yielded a significantly higher area under the curve (AUC). Finally, visualization of an easy-to-use nomogram was invented to potentially predict suspected SSN. sEVs-miR-424-5p could be a novel biomarker for distinguishing SSN from LC and BN populations. Its association with cross-scale fusion of radiomics-clinical features will provide great potential to be an errorless prediction of malignant SSN.

Clinical Value and Pathologic Basis of Cystic Airspace Within Subsolid Nodules Confirmed as Lung Adenocarcinomas by Surgery

Purpose. To investigate the clinical value and pathologic basis of cystic airspace within lung adenocarcinomas manifesting as subsolid nodules.Patients and Methods. A retrospective study was conducted on a total of 541 surgically confirmed lung adenocarcinomas manifesting as subsolid nodules in computed tomography images, including 87 cases with cystic airspace and 454 cases without cystic airspace. The pathologic characteristics of the cases with and without cystic airspace were compared. The investigation of the pathologic structure of cystic airspace was attempted on the postoperative paraffin sections.Results. There was a significant difference in the containing of cystic airspace between preinvasive and invasive adenocarcinomas (10.5 vs 26.6%; P < .001). Multivariate analysis indicated that cystic airspace is an independent predictor of invasive adenocarcinomas (odds ratio, 3.220; 95% confidence interval, 1.822-5.687). Nodules containing multiple cystic airspaces are more likely to be invasive adenocarcinomas than nodules with a single cystic airspace (47.1 vs 72.2%; P < .05). On paraffin sections, the walls of the cystic airspace seemed to be mainly composed of atypical hyperplasia and/or tumor cells on the surface and the remaining smooth muscle cells and stroma below, which is similar to the structure of bronchi.Conclusions. Cystic airspace may be a reliable predictor of invasive adenocarcinomas, the classification method based on the number of cystic airspaces might be suitable for the computed tomography-based typing of cystic airspace within subsolid nodules. Cystic airspace may derive from the destroyed and enlarged bronchi owing to the growth or infiltration of atypical hyperplasia and/or tumor cells.

The Value of PETRA in Pulmonary Nodules of &lt;3 cm Among Patients With Lung Cancer

ObjectiveThis study aimed to evaluate the visibility of different subgroups of lung nodules of <3 cm using the pointwise encoding time reduction with radial acquisition (PETRA) sequence on 3T magnetic resonance imaging (MRI) in comparison with that obtained using low-dose computed tomography (LDCT).MethodsThe appropriate detection rate was calculated for each of the different subgroups of lung nodules of <3 cm. The mean diameter of each detected nodule was determined. The detection rates and diameters of the lung nodules detected by MRI with the PETRA sequence were compared with those detected by computed tomography (CT). The sensitivity of detection for the different subgroups of pulmonary nodules was determined based on the location, size, type of nodules and morphologic characteristics. Agreement of nodule characteristics between CT and MRI were assessed by intraclass correlation coefficient (ICC) and Kappa test.ResultsThe CT scans detected 256 lung nodules, comprising 99 solid nodules (SNs) and 157 subsolid nodules with a mean nodule diameter of 8.3 mm. For the SNs, the MRI detected 30/47 nodules of <6 mm in diameter and 52/52 nodules of ≥6 mm in diameter. For the subsolid nodules, the MRI detected 30/51 nodules of <6 mm in diameter and 102/106 nodules of ≥6 mm in diameter. The PETRA sequence returned a high detection rate (84%). The detection rates of SN, ground glass nodules, and PSN were 82%, 72%, and 94%, respectively. For nodules with a diameter of >6 mm, the sensitivity of the PETRA sequence reached 97%, with a higher rate for nodules located in the upper lung fields than those in the middle and lower lung fields. Strong agreement was found between the CT and PETRA results (correlation coefficients = 0.97).ConclusionThe PETRA technique had high sensitivity for different type of nodule detection and enabled accurate assessment of their diameter and morphologic characteristics. It may be an effective alternative to CT as a tool for screening and follow up pulmonary nodules.

Pearls and Pitfalls in Lung Cancer Imaging

Imaging plays an essential role in the diagnosis and staging of pulmonary malignancy. Familiarity of less common manifestations of lung cancer including subsolid nodule, consolidation, and cyst associated lung cancer is important to avoid delayed diagnosis and improve patient outcomes. In this article, we review the staging of multifocal lung cancer, PET negative lung cancers (carcinoid and indolent lung adenocarcinoma), and false positive lymph nodes on PET due to infectious and inflammatory etiologies. Knowledge of these potential pitfalls and pearls in lung cancer imaging and correlation with patients' clinical history are essential to prevent misinterpretation.

Managing screening-detected subsolid nodules-the Asian perspective.

The broad application of low-dose computed tomography (CT) screening has resulted in the detection of many small pulmonary nodules. In Asia, a large number of these detected nodules with a radiological ground glass pattern are reported as lung adenocarcinomas or premalignant lesions, especially among female non-smokers. In this review article, we discuss controversial issues and conditions involving these subsolid pulmonary nodules that we often face in Asia, including a lack or insufficiency of current guidelines; the roles of preoperative biopsy and imaging; the location of lesions; appropriate selection of localization techniques; the roles of dissection and sampling of frozen sections and lymph nodes; multifocal lesions; and the roles of non-surgical treatment modalities. For these complex issues, we have tried to present up-to-date evidence and our own opinions regarding the management of subsolid nodules. It is our hope that this article helps surgeons and physicians to manage the complex issues involving ground glass nodules (GGNs) in a balanced manner in their daily practice and provokes further discussion towards better guidelines and/or algorithms.

Managing of screening-detected sub-solid nodules-a European perspective.

Since the National Lung Screening Trial in 2011 showed a 20% reduction in lung cancer mortality using annual low-dose computed tomography (LDCT), several randomised controlled trials and studies have been started in Europe. These include the Italian lung study (ITALUNG), the Dutch-Belgian lung cancer screening trial (NELSON), the UK lung cancer screening trial (UKLS), the Detection and screening of early lung cancer with novel imaging technology (DANTE), the Danish lung cancer screening trial (DLCST), the German lung cancer screening intervention trial (LUSI), the Multicentric Italian lung detection trial (MILD) and the CT screening for lung cancer study (COSMOS). As a result of the increasing number of screening trials and the growing utilization of LDCT, the high detection of subsolid nodules is an increasingly important clinical problem. In the last few years, several guidelines have been published and providing guidance on the optimal management of subsolid nodules, but many controversies still exist. Follow-up imaging plays an important role in clinical assessment and subsequent management of this particular type of lung nodules, since they can be transient inflammatory lesions, and if persistent they can be both benign lesions or lung cancers of variable clinical behaviour. However, the vast majority of subsolid nodules retain an indolent course over many years. The aim of this review is to present a European perspective in management of screening detected subsolid nodules.

Differentiating focal interstitial fibrosis from adenocarcinoma in persistent pulmonary subsolid nodules (&gt; 5 mm and &lt; 20 mm): the role of coronal thin-section CT images

To investigate thin-section computed tomography (CT) features of pulmonary subsolid nodules (SSNs) with sizes between 5 and 20 mm to determine predictive factors for differentiating focal interstitial fibrosis (FIF) from adenocarcinoma. From January 2017 to December 2018, 169 patients who had persistent SSNs 5-20 mm in size and underwent preoperative nodule localization were enrolled. Patient characteristics and thin-section CT features of the SSNs were reviewed and compared between the FIF and adenocarcinoma groups. Univariable and multivariable analyses were used to identify predictive factors of malignancy. Receiver operating characteristic (ROC) curve analysis was used to quantify the performance of these factors. Among the 169 enrolled SSNs, 103 nodules (60.9%) presented as pure ground-glass opacities (GGOs), and 40 (23.7%) were FIFs. Between the FIF and adenocarcinoma groups, there were significant differences (p< 0.05) in nodule border, shape, thickness, and coronal/axial (C/A) ratio. Multivariable analysis demonstrated that a well-defined border, a nodule thickness >4.2, and a C/A ratio >0.62 were significant independent predictors of malignancy. The performance of a model that incorporated these three predictors in discriminating FIF from adenocarcinoma achieved a high area under the ROC curve (AUC, 0.979) and specificity (97.5%). For evaluating persistent SSNs 5-20 mm in size, the combination of a well-defined border, a nodule thickness > 4.2, and a C/A ratio > 0.62 is strongly correlated with malignancy. High accuracy and specificity can be achieved by using this predictive model. • Thin-section coronal images play an important role in differentiating FIF from adenocarcinoma. • The combination of a well-defined border, nodule thickness>4.2 mm, and C/A ratio >0.62 is associated with malignancy. • This predictive model may be helpful for managing persistent SSNs between 5 and 20 mm in size.

Clinical T1aN0M0 lung cancer: differences in clinicopathological patterns and oncological outcomes based on the findings on high-resolution computed tomography

To elucidate the clinicopathological characteristics and oncological outcomes of clinical T1aN0M0 (c-T1N0M0) lung cancer based on the newest 8th TNM classification. A total of 257 patients with c-T1aN0M0 lung cancer were retrospectively included in this study. According to the solid component size manifesting on the high-resolution computed tomography (HRCT), all lesions were classified as the pure ground-glass nodule (pure-GGN) with a diameter > 3 cm (n = 19), part-solid (n = 174), and pure-solid (n = 64) groups. We evaluated the prognostic impact of clinicopathologic variables including radiological presentations by establishing Cox proportional hazards model. When we evaluated the prognostic impact based on the radiological subtypes, the 5-year recurrence-free survival (RFS) and overall survival (OS) were significantly different among pure-GGN, part-solid, and pure-solid groups (RFS: 100% versus 95.4% versus 76.6%, p < 0.0001; OS: 100% versus 98.9% versus 87.5%, p < 0.0001). Cox regression analysis revealed the preoperative carcinoembryonic antigen (CEA) level and consolidation tumor ratio (CTR) were independently significant prognosticators related to RFS and OS. Furthermore, a receiver operating characteristic (ROC) verified the CTR (area under ROC [AUC] 0.784, 95%CI 0.697-0.869) was equipped with good performance to predict the postoperative recurrence with a cutoff point at 0.5. Lung cancer with higher CTR tended to be associated with lower survival in the c-T1aN0M0 stage. For the c-T1aN0M0 lung cancer, pulmonary nodules manifested as the pure-GGN and part-solid subtypes had an excellent prognosis and may be considered as the "early-stage" cancer, whereas those with pure-solid appearance were associated with the high risk of recurrence despite the sub-centimeter size. • Radiological subtypes could further stratify the risk of lung cancer in cT1a. • Sub-solid nodule has a favorable survival in c-T1a lung cancer, whereas pure-solid nodule is not always "early-stage" lung cancer and is relatively prone to postoperative recurrence despite the sub-centimeter size. • The preoperative CEA level and CTR are valuable prognosticators to predict the recurrence in c-T1a lung cancer.

Biopsy of Subsolid Nodules Suspicious for Adenocarcinoma: Counterpoint—Biopsy Has Limited Utility in the Diagnostic Evaluation of Subsolid Nodules

This article does not include an abstract. Please see the accompanying Point by Terrance T. Healey and Jo-Anne O. Shepard.

Biopsy of Subsolid Nodules Suspicious for Adenocarcinoma: Point-CT-Guided Biopsy of Subsolid Nodules Is a Safe and Effective Means to Establish a Definitive Preoperative Diagnosis.

This article does not include an abstract. Please see the accompanying Counterpoint by Brett M. Elicker and Kimberly G. Kallianos.

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

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

Artificial intelligence-based vessel suppression for detection of sub-solid nodules in lung cancer screening computed tomography.

Lung cancer screening (LCS) with low-dose computed tomography (LDCT) helps early lung cancer detection, commonly presenting as small pulmonary nodules. Artificial intelligence (AI)-based vessel suppression (AI-VS) and automatic detection (AI-AD) algorithm can improve detection of subsolid nodules (SSNs) on LDCT. We assessed the impact of AI-VS and AI-AD in detection and classification of SSNs [ground-glass nodules (GGNs) and part-solid nodules (PSNs)], on LDCT performed for LCS. Following regulatory approval, 123 LDCT examinations with sub-solid pulmonary nodules (average diameter ≥6 mm) were processed to generate three image series for each examination-unprocessed, AI-VS, and AI-AD series with annotated lung nodules. Two thoracic radiologists in consensus formed the standard of reference (SOR) for this study. Two other thoracic radiologists (R1 and R2; 5 and 10 years of experience in thoracic CT image interpretation) independently assessed the unprocessed images alone, then together with AI-VS series, and finally with AI-AD for detecting all ≥6 mm GGN and PSN. We performed receiver operator characteristics (ROC) and Cohen's Kappa analyses for statistical analyses. On unprocessed images, R1 and R2 detected 232/310 nodules (R1: 114 GGN, 118 PSN) and 255/310 nodules (R2: 122 GGN, 133 PSN), respectively (P>0.05). On AI-VS images, they detected 249/310 nodules (119 GGN, 130 PSN) and 277/310 nodules (128 GGN, 149 PSN), respectively (P≥0.12). When compared to the SOR, accuracy (AUC) for detection of PSN on the AI-VS images (AUC 0.80-0.81) was greater than on the unprocessed images (AUC 0.70-0.76). AI-VS images enabled detection of solid components in five nodules deemed as GGN on the unprocessed images. Accuracy of AI-AD was lower than both the radiologists (AUC 0.60-0.72). AI-VS improved the detection and classification of SSN into GGN and PSN on LDCT of the chest for the two radiologist (R1 and R2) readers.

Is flexible bronchoscopy necessary in the preoperative workup of patients with peripheral cT1N0 subsolid lung cancer? -a prospective multi-center cohort study.

BackgroundNecessity of flexible bronchoscopy (FB) examination as a routine preoperative work-up for peripheral clinical T1N0 subsolid lung cancer was unknown.MethodsThis was a prospective, multi-center clinical trial (NCT03591445). Patients with peripheral GGO nodules (GGNs) who were candidates for surgical resection were enrolled. FB examination was performed preoperatively. Surgical plan could be changed if any aberrant histologic and anatomic findings were detected by FB examination. Primary endpoint was the rate that surgical plan was changed by positive FB findings. Secondary endpoints were rate of positive FB findings and rate of procedural complications.ResultsSix hundred and fifteen patients with peripheral subsolid nodules detected by thoracic CT were enrolled. There were 187 (30.4%) male and 428 (69.6%) female patients, mean age was 54.85±10.41 y (range, 26–78). 262 (42.6%) patients had pure GGNs and 353 (57.4%) patients had part-solid nodules. Mean size of nodules was 13.87±6.37 mm (range, 5–30). FB examinations confirmed one (0.16%) adenocarcinoma, seven (1.14%) bronchial variations, one (0.16%) segmental bronchostenosis, one (0.16%) segmental bronchial occlusion and one (0.16%) bronchial inflammation. No complications of FB examinations occurred. 568 (92.35%) thoracoscopic and 47 (7.65%) open surgeries were performed. No established surgical plan was changed by positive FB findings. Final pathologies revealed 26 (4.2%) adenocarcinoma in situ (AIS), 240 (39%) minimal invasive adenocarcinomas (MIAs), 343 (55.8%) invasive adenocarcinomas (IADs), one (0.2%) adenosquamous cell carcinoma, one (0.2%) squamous cell carcinoma, two (0.3%) atypical adenoid hyperplasia and two (0.3%) inflammations.ConclusionsFB examination was unnecessary in the preoperative assessment of peripheral clinical T1N0 subsolid lung cancer.

8P Distinct tumor bacterial microbiome in lung adenocarcinomas manifested as radiological subsolid nodules

8P Distinct tumor bacterial microbiome in lung adenocarcinomas manifested as radiological subsolid nodules

Lung Cancer Screening: Evidence, Risks, and Opportunities forImplementation.

Lung cancer is the most common cause of cancer death worldwide. Several trials with different screening approaches have recognized the role of lung cancer screening with low-dose CT for reducing lung cancer mortality. The efficacy of lung cancer screening depends on many factors and implementation is still pending in most European countries. This review aims to portray current evidence on lung cancer screening with a focus on the potential for opportunities for implementation strategies. Pillars of lung cancer screening practice will be discussed according to the most updated literature (PubMed search until November 16, 2020). The NELSON trial showed reduction of lung cancer mortality, thus confirming previous results of independent European studies, notably by volume of lung nodules. Heterogeneity in patient recruitment could influence screening efficacy, hence the importance of risk models and community-based screening. Recruitment strategies develop and adapt continuously to address the specific needs of the heterogeneous population of potential participants, the most updated evidence comes from the UK. The future of lung cancer screening is a tailored approach with personalized continuous stratification of risk, aimed at reducing costs and risks. · Secondary prevention of lung cancer by low-dose computed tomography showed a reduction of lung cancer mortality.. · Semi-automated volume measurement and use of volume doubling time should be the reference method for optimization of risks, namely controlling measurement variability and the false-positive rate.. · A conservative approach with surveillance of subsolid nodules can be one of the strategies to reduce the risk of overdiagnosis and overtreatment.. · The goal of a tailored approach with personalized risk stratification aims to reduce costs and risks. A longer interval between rounds is one option for participants at lower risk.. · Tringali G, Milanese G, Ledda RE et al. Lung Cancer Screening: Evidence, Risks, and Opportunities for Implementation. Fortschr Röntgenstr 2021; 193: 1153 - 1161.

Distinct tumor bacterial microbiome in lung adenocarcinomas manifested as radiological subsolid nodules

ObjectivesIncreasing evidence indicates that microbiota dysbiosis in the human body may play vital roles in carcinogenesis. However, the relationship between microbiome and lung cancer remains unclear. In this study, we aimed to characterize the microbiome in early stage of lung adenocarcinoma (LUAD), which presented as subsolid nodules (SSN) or solid nodules (SN). Materials and MethodsWe performed 16S rRNA sequencing of 35 pairs (10 SSN and 25 SN) of LUAD tumor tissues and paired adjacent normal tissues. Machine learning was used to identify microbial signatures and construct predictive models. ResultsSSN has higher microbiome richness and diversity compared with SN (richness p = 0.017, Shannon index p = 0.17), and the microbiome composition of SSN is distinct from that of SN (Bray-Curtis p = 0.013, unweighted unifrac p = 0.001). Phylum Chloroflexi (p = 0.009), Gemmatimonadetes (p = 0.018) and genus including Cloacibacterium (p = 0.003), Subdoligranulum (p = 0.002), and Mycobacterium (p = 0.034) were significantly increased in SSN. Tumor and normal tissues had similar richness and diversity, as well as overall microbiome composition. Probiotics with anti-cancer potential, like Lactobacillus, showed elevated levels in normal tissues (p = 0.018). A random forest model with 20 genera-based biomarkers achieved high accuracy for LUAD prediction (area under curve, AUC = 0.879). Meanwhile, a five genera-based signature can accurately discriminate SSN between SN (AUC = 0.950). Cross-validation of these two models also showed high predictive performance (LUAD AUC = 0.813, SSN AUC = 0.933). ConclusionsThis study demonstrates, for the first time, the tumor bacterial microbiome composition of LUAD manifested as SSN is distinct from that presented as SN, which adds new knowledge to SSN in the perspective of microbiome. Furthermore, microbiome signatures showed good performance to predict LUAD or SSN.

Interobserver variability in Lung CT Screening Reporting and Data System categorisation in subsolid nodule-enriched lung cancer screening CTs.

To assess interobserver agreement in Lung CT Screening Reporting and Data System (Lung-RADS) categorisation in subsolid nodule-enriched low-dose screening CTs. A retrospective review of low-dose screening CT reports from 2013 to 2017 using keyword searches for subsolid nodules identified 54 baseline CT scans. With an additional 108 negative screening CT scans, a total of 162 CT scans were categorised according to the Lung-RADS by two fellowship-trained thoracic radiologists in consensus. We randomly selected 20, 20, 10, and 10 scans from categories 1/2, 3, 4A, and 4B CT scans, respectively, to ensure balanced category representation. Five radiologists classified the 60 CT scans into Lung-RADS categories. The frequencies of concordance and minor and major discordance were calculated, with major discordance defined as at least 6 months of management discrepancy. We used Cohen's κ statistics to analyse reader agreement. An average of 60.3% (181 of 300) of all cases and 45.0% (90 of 200) of positive screens were correctly categorised. The minor and major discordance rates were 12.3% and 27.3% overall and 18.5% and 36.5% in positive screens, respectively. The concordance rate was significantly higher among experienced thoracic radiologists. Overall, the interobserver agreement was moderate (mean κ, 0.45; 95% confidence interval: 0.40-0.51). The proportion of part-solid risk-dominant nodules was significantly higher in cases with low rates of accurate categorisation. This retrospective study observed variable accuracy and moderate interobserver agreement in radiologist categorisation of subsolid nodules in screening CTs. This inconsistency may affect management recommendations for lung cancer screening. • Diagnostic performance for Lung-RADS categorisation is variable among radiologists with fair to moderate interobserver agreement in subsolid nodule-enriched CT scans. • Experienced thoracic radiologists showed more accurate and consistent Lung-RADS categorisation than radiology residents. • The relative abundance of part-solid nodules was a potential factor related to increased disagreement in Lung-RADS categorisation.

ES10.06 Putting It Together: Current Best Practices in the Evaluation and Follow-Up of the Indeterminate Pulmonary Nodule

ES10.06 Putting It Together: Current Best Practices in the Evaluation and Follow-Up of the Indeterminate Pulmonary Nodule

Solid Components in the Mediastinal Window of Computed Tomography Define a Distinct Subtype of Subsolid Nodules in Clinical Stage I Lung Cancers

BackgroundWe aimed to validate the clinicopathologic characteristics and prognostic value of the presence of solid components in the mediastinal window of computed tomography scan in clinical stage I pulmonary subsolid nodules (SSNs). MethodsWe retrospectively evaluated patients with pulmonary SSNs resected between 2011 and 2016. We classified SSNs into heterogeneous ground-glass nodules (HGGNs) (solid component detected only in lung window) and part-solid nodules (PSNs) (solid component detected both in lung/mediastinal windows). ResultsA total of 487 patients (216 PSNs) were included. PSNs were associated with higher frequencies of micropapillary or solid pathologic patterns (18.1% vs. 3.3%; P < .001), epidermal growth factor receptor gene mutation (39.4% vs. 32.8%), and other types of gene mutations (2.3% vs. 1.1%; P = .043). Logistic regression analysis revealed that male sex (odds ratio [OR], 2.58; 95% confidence interval [CI], 1.20-5.57; P = .016) and higher consolidation tumor ratio (CTR) (OR, 110.04; 95% CI, 8.56-1414.39; P < .001) remained independent for invasive adenocarcinomas with poor differentiation. Receiver operating characteristic analyses revealed that solid component size in the mediastinal window (area under the curve [AUC], 0.731; 95% CI, 0.653-0.808; P < .0001) showed a better predictive ability to poor differentiation compared with solid component size in the lung window and CTR. The 5-year recurrence-free survival (RFS) rate of PSNs was worse than that of HGGNs (94.6% vs. 99.1%; P = .019). Multivariate Cox regression revealed that positive lymph node status (hazard ratio, 22.99; 95% CI, 4.52-116.86; P < .001) indicated worse RFS for PSNs. ConclusionSSNs with solid components in mediastinal window demonstrated clinicopathologic and prognostic features different from those without in clinical stage I lung cancer. Solid components in mediastinal window was a strong predictor of poor differentiation.

A deep learning-based lung nodule density classification and segmentation method and its effectiveness under different CT reconstruction algorithms

Objective: To evaluate the diagnostic value of the lung nodule classification and segmentation algorithm based on deep learning among different CT reconstruction algorithms. Methods: Chest CT of 363 patients from June 2019 to September 2019 in Radiology Department of Tianjin Medical University Chu Hsien-I Memorial Hospital were retrospectively collected in this study, each of which consisted of images by three different reconstruction methods (lung reconstruction, mediastinal reconstruction, bone reconstruction).These collected data were used as testing set and a total of 4 185 Chest CTs including the public data set and the constructed private data set were used as the training set. A model combines 3D deep convolutional neural network and recurrent neural network under a multi-task joint learning algorithm for lung nodule classification and segmentation were constructed. The well-trained method was tested on 363 test cases using two metrics, i.e., the accuracy of the density classification and the Dice coefficient of nodule segmentation. The performances under three reconstruction methods were statistically analyzed according to the variance analysis among three different reconstruction methods. Results: The average classification accuracies of the nodule under three reconstruction methods were 98.67%±5.70%, 98.38%±6.61% and 97.89%±7.32%. Specifically, the accuracies of the solid nodules under three reconstruction methods were 98.79%±5.58%, 98.49%±6.89% and 97.90%±7.41% and the accuracies of the sub-solid nodules were 97.57%±10.19%, 98.52%±7.77% and 98.52%±7.77%. There was no significant difference in the classification accuracy of pulmonary nodules under three different reconstruction algorithms (all P>0.05). The average Dice coefficients of nodule segmentation was 79.87%±5.78%, 79.02%±6.04% and 79.31%±5.95%. There was no significant difference in the average Dice coefficients of nodule segmentation under three different reconstruction algorithms (all P>0.05). Conclusion: Deep learning algorithm which combined with 3D convolutional neural network and recurrent neural network has demonstrated relatively stable in classification and segmentation of lung nodules under different CT reconstruction method.