Robust deep learning from incomplete annotation for accurate lung nodule detection.

Deep learning plays a significant role in the detection of pulmonary nodules in low-dose computed tomography (LDCT) scans, contributing to the diagnosis and treatment of lung cancer. Nevertheless, its effectiveness often relies on the availability of extensive, meticulously annotated dataset. In this paper, we explore the utilization of an incompletely annotated dataset for pulmonary nodules detection and introduce the FULFIL (Forecasting Uncompleted Labels For Inexpensive Lung nodule detection) algorithm as an innovative approach. By instructing annotators to label only the nodules they are most confident about, without requiring complete coverage, we can substantially reduce annotation costs. Nevertheless, this approach results in an incompletely annotated dataset, which presents challenges when training deep learning models. Within the FULFIL algorithm, we employ Graph Convolution Network (GCN) to discover the relationships between annotated and unannotated nodules for self-adaptively completing the annotation. Meanwhile, a teacher-student framework is employed for self-adaptive learning using the completed annotation dataset. Furthermore, we have designed a Dual-Views loss to leverage different data perspectives, aiding the model in acquiring robust features and enhancing generalization. We carried out experiments using the LUng Nodule Analysis (LUNA) dataset, achieving a sensitivity of 0.574 at a False positives per scan (FPs/scan) of 0.125 with only 10% instance-level annotations for nodules. This performance outperformed comparative methods by 7.00%. Experimental comparisons were conducted to evaluate the performance of our model and human experts on test dataset. The results demonstrate that our model can achieve a comparable level of performance to that of human experts. The comprehensive experimental results demonstrate that FULFIL can effectively leverage an incomplete pulmonary nodule dataset to develop a robust deep learning model, making it a promising tool for assisting in lung nodule detection.

Comparison of laser guidance and freehand hook-wire for CT-guided preoperative localization of pulmonary nodules.

PURPOSE: In VATS surgery, precise preoperative localization is particularly crucial when dealing with small-diameter pulmonary nodules located deep within the lung parenchyma. The purpose of this study was to compare the efficacy and safety of laser guidance and freehand hook-wire for CT-guided preoperative localization of pulmonary nodules. METHODS: This retrospective study was conducted on 164 patients who received either laser guidance or freehand hook-wire localization prior to Uni-port VATS from September 1st, 2022 to September 30th, 2023 at The First Affiliated Hospital of Soochow University. Patients were divided into laser guidance group and freehand group based on which technology was used. Preoperative localization data from all patients were compiled. The localization success and complication rates associated with the two groups were compared. The risk factors for common complications were analyzed. RESULTS: The average time of the localization duration in the laser guidance group was shorter than the freehand group (p<0.001), and the average CT scan times in the laser guidance group was less than that in the freehand group (p<0.001). The hook-wire was closer to the nodule in the laser guidance group (p<0.001). After the localization of pulmonary nodules, a CT scan showed 14 cases of minor pneumothorax (22.58%) in the laser guidance group and 21 cases (20.59%) in the freehand group, indicating no statistical difference between the two groups (p=0.763). CT scans in the laser guidance group showed pulmonary minor hemorrhage in 8 cases (12.90%) and 6 cases (5.88%) in the freehand group, indicating no statistically significant difference between the two groups (p=0.119). Three patients (4.84%) in the laser guidance group and six patients (5.88%) in the freehand group had hook-wire dislodgement, showing no statistical difference between the two groups (p=0.776). CONCLUSION: The laser guidance localization method possessed a greater precision and less localization duration and CT scan times compared to the freehand method. However, laser guidance group and freehand group do not differ in the appearance of complications such as pulmonary hemorrhage, pneumothorax and hook-wire dislodgement.

Management of incidental pulmonary nodules in CT examination.

Lung cancer is the leading cause of cancer-related death in Denmark and the world. The increase in CT examinations has led to an increase in detection of pulmonary nodules divided into solid and subsolid (including ground glass and part solid). Risk factors for malignancy include age, smoking, female gender, and specific ethnicities. Nodule traits like size, spiculation, upper-lobe location, and emphysema correlate with higher malignancy risk. Managing these potentially malignant nodules relies on evidence-based guidelines and risk stratification. These risk stratification models can standardize the approach for the management of incidental pulmonary findings, as argued in this review.

Development of a modified 3D region proposal network for lung nodule detection in computed tomography scans: a secondary analysis of lung nodule datasets.

BACKGROUND: Low-dose computed tomography (LDCT) has been shown useful in early lung cancer detection. This study aimed to develop a novel deep learning model for detecting pulmonary nodules on chest LDCT images. METHODS: In this secondary analysis, three lung nodule datasets, including Lung Nodule Analysis 2016 (LUNA16), Lung Nodule Received Operation (LNOP), and Lung Nodule in Health Examination (LNHE), were used to train and test deep learning models. The 3D region proposal network (RPN) was modified via a series of pruning experiments for better predictive performance. The performance of each modified deep leaning model was evaluated based on sensitivity and competition performance metric (CPM). Furthermore, the performance of the modified 3D RPN trained on three datasets was evaluated by 10-fold cross validation. Temporal validation was conducted to assess the reliability of the modified 3D RPN for detecting lung nodules. RESULTS: The results of pruning experiments indicated that the modified 3D RPN composed of the Cross Stage Partial Network (CSPNet) approach to Residual Network (ResNet) Xt (CSP-ResNeXt) module, feature pyramid network (FPN), nearest anchor method, and post-processing masking, had the optimal predictive performance with a CPM of 92.2%. The modified 3D RPN trained on the LUNA16 dataset had the highest CPM (90.1%), followed by the LNOP dataset (CPM: 74.1%) and the LNHE dataset (CPM: 70.2%). When the modified 3D RPN trained and tested on the same datasets, the sensitivities were 94.6%, 84.8%, and 79.7% for LUNA16, LNOP, and LNHE, respectively. The temporal validation analysis revealed that the modified 3D RPN tested on LNOP test set achieved a CPM of 71.6% and a sensitivity of 85.7%, and the modified 3D RPN tested on LNHE test set had a CPM of 71.7% and a sensitivity of 83.5%. CONCLUSION: A modified 3D RPN for detecting lung nodules on LDCT scans was designed and validated, which may serve as a computer-aided diagnosis system to facilitate lung nodule detection and lung cancer diagnosis.

A modified 3D RPN for detecting lung nodules on CT images that exhibited greater sensitivity and CPM than did several previously reported CAD detection models was established.

Incidence and risk factors for pulmonary hemorrhage after percutaneous CT-guided pulmonary nodule biopsy: an observational study.

To evaluate the current incidence of pulmonary hemorrhage and the potential factors contributing to its increased risk after percutaneous CT-guided pulmonary nodule biopsy and to summarize the technical recommendations for its treatment. In this observational study, patient data were collected from ten medical centers from April 2021 to April 2022. The incidence of pulmonary hemorrhage was as follows: grade 0, 36.1% (214/593); grade 1, 36.8% (218/593); grade 2, 18.9% (112/593); grade 3, 3.5% (21/593); and grade 4, 4.7% (28/593). High-grade hemorrhage (HGH) occurred in 27.2% (161/593) of the patients. The use of preoperative breathing exercises (PBE, p =0.000), semiautomatic cutting needles (SCN, p = 0.004), immediate contrast enhancement (ICE, p =0.021), and the coaxial technique (CoT, p = 0.000) were found to be protective factors for HGH. A greater length of puncture (p =0.021), the presence of hilar nodules (p = 0.001), the presence of intermediate nodules (p = 0.026), a main pulmonary artery diameter (mPAD) larger than 29 mm (p = 0.015), and a small nodule size (p = 0.014) were risk factors for high-grade hemorrhage. The area under the curve (AUC) was 0.783. These findings contribute to a deeper understanding of the risks associated with percutaneous CT-guided pulmonary nodule biopsy and provide valuable insights for developing strategies to minimize pulmonary hemorrhage.

Development of automatic generation system for lung nodule finding descriptions.

Worldwide, lung cancer is the leading cause of cancer-related deaths. To manage lung nodules, radiologists observe computed tomography images, review various imaging findings, and record these in radiology reports. The report contents should be of high quality and uniform regardless of the radiologist. Here, we propose an artificial intelligence system that automatically generates descriptions related to lung nodules in computed tomography images. Our system consists of an image recognition method for extracting contents-namely, bronchopulmonary segments and nodule characteristics from images-and a natural language processing method to generate fluent descriptions. To verify our system's clinical usefulness, we conducted an experiment in which two radiologists created nodule descriptions of findings using our system. Through our system, the similarity of the described contents between the two radiologists (p = 0.001) and the comprehensiveness of the contents (p = 0.025) improved, while the accuracy did not significantly deteriorate (p = 0.484).

Association of patient and health care organization factors with incidental nodule guidelines adherence: A multi-system observational study.

BACKGROUND: Health care organizations are increasingly developing systems to ensure patients with pulmonary nodules receive guideline-adherent care. Our goal was to determine patient and organization factors that are associated with radiologist adherence as well as clinician and patient concordance to 2005 Fleischner Society guidelines for incidental pulmonary nodule follow-up. MATERIALS: Trained researchers abstracted data from the electronic health record from two Veterans Affairs health care systems for patients with incidental pulmonary nodules as identified by interpreting radiologists from 2008 to 2016. METHODS: We classified radiology reports and patient follow-up into two categories. Radiologist-Fleischner Adherence was the agreement between the radiologist's recommendation in the computed tomography report and the 2005 Fleischner Society guidelines. Clinician/Patient-Fleischner Concordance was agreement between patient follow-up and the guidelines. We calculated multivariable-adjusted predicted probabilities for factors associated with Radiologist-Fleischner Adherence and Clinician/Patient-Fleischner Concordance. RESULTS: Among 3150 patients, 69% of radiologist recommendations were adherent to 2005 Fleischner guidelines, 4% were more aggressive, and 27% recommended less aggressive follow-up. Overall, only 48% of patients underwent follow-up concordant with 2005 Fleischner Society guidelines, 37% had less aggressive follow-up, and 15% had more aggressive follow-up. Radiologist-Fleischner Adherence was associated with Clinician/Patient-Fleischner Concordance with evidence for effect modification by health care system. CONCLUSION: Clinicians and patients seem to follow radiologists' recommendations but often do not obtain concordant follow-up, likely due to downstream differential processes in each health care system. Health care organizations need to develop comprehensive and rigorous tools to ensure high levels of appropriate follow-up for patients with pulmonary nodules.

Decreased levels of sex hormones in females with solitary pulmonary nodules were risk factors for malignancy.

OBJECTIVE: The purpose of this research was to detect the relationship between the levels of sex hormones in females with solitary pulmonary nodules (SPNs) and their potential malignancies. METHODS: A total of 187 consecutive patients with pathologically confirmed SPNs by chest CT were enrolled in our study. They were divided into two groups based on the pathologic findings of SPNs after surgery: benign and malignant SPNs. Progesterone (P), estradiol (E2), and testosterone (T) levels in the two groups were measured. Meanwhile, we used binary logistic regression analysis to analyze the risk factors for SPNs. RESULTS: Of these 187 patients, 73 had benign SPNs, while 114 had malignant SPNs. We found that the levels of progesterone (P), estradiol (E2), and testosterone (T) were decreased significantly in patients with malignant SPNs compared to patients with benign SPNs (all P < 0.05). Multivariate logistic regression analysis revealed that second-hand smoke, burr sign, lobulation sign, pleural traction sign, vascular convergence sign, vacuole sign, and ≥ 1 cm nodules were independent risk factors for malignant pulmonary nodules in females. CONCLUSIONS: Decreased levels of sex hormones in females were associated with malignant pulmonary nodules, suggesting that they can contribute to the diagnosis of lung cancer.

Intelligent diagnosis of lung nodule images based on machine learning in the context of lung teaching.

The vast majority of intelligent diagnosis models have widespread problems, which seriously affect the medical staff judgment of patients' injuries. So depending on the situation, you need to use different algorithms, The study suggests a model for intelligent diagnosis of lung nodule images based on machine learning, and a support vector machine-based machine learning algorithm is selected. In order to improve the diagnostic accuracy of intelligent diagnosis of lung nodule images as well as the diagnostic model of lung nodule images. The objectives are broken down into algorithm determination and model construction, and the proposed optimized model is solved using machine learning techniques in order to achieve the original algorithm selected for intelligent diagnosis of lung nodule photos. The validation findings demonstrated that dimensionality reduction of the features produced 17 × 1120 and 17 × 2980 non-node matrices with 1216 nodes and 3407 non-nodes in 17 features. The support vector machine classification method has more benefits in terms of accuracy, sensitivity, and specificity when compared to other classification methods. Since there were some anomalies among both benign and malignant tumors and no discernible difference between them, the distribution of median values revealed that the data was symmetrical in terms of texture and gray scale. Non-small nodules can be identified from benign nodules, but more training is needed to separate them from the other 2 types. Pulmonary nodules are a common disease. MN are distinct from the other 2 types, non-small nodules and benign small nodules, which require further training to differentiate. This has great practical value in teaching practice. Therefore, building a machine learning-based intelligent diagnostic model for pulmonary nodules is of significant importance in helping to solve medical imaging diagnostic problems.

S-Net: an S-shaped network for nodule detection in 3D CT images.

Objective. Accurate and automatic detection of pulmonary nodules is critical for early lung cancer diagnosis, and promising progress has been achieved in developing effective deep models for nodule detection. However, most existing nodule detection methods merely focus on integrating elaborately designed feature extraction modules into the backbone of the detection network to extract rich nodule features while ignore disadvantages of the structure of detection network itself. This study aims to address these disadvantages and develop a deep learning-based algorithm for pulmonary nodule detection to improve the accuracy of early lung cancer diagnosis.Approach. In this paper, an S-shaped network called S-Net is developed with the U-shaped network as backbone, where an information fusion branch is used to propagate lower-level details and positional information critical for nodule detection to higher-level feature maps, head shared scale adaptive detection strategy is utilized to capture information from different scales for better detecting nodules with different shapes and sizes and the feature decoupling detection head is used to allow the classification and regression branches to focus on the information required for their respective tasks. A hybrid loss function is utilized to fully exploit the interplay between the classification and regression branches.Main results. The proposed S-Net network with ResSENet and other three U-shaped backbones from SANet, OSAF-YOLOv3 and MSANet (R+SC+ECA) models achieve average CPM scores of 0.914, 0.915, 0.917 and 0.923 on the LUNA16 dataset, which are significantly higher than those achieved with other existing state-of-the-art models.Significance. The experimental results demonstrate that our proposed method effectively improves nodule detection performance, which implies potential applications of the proposed method in clinical practice.

Hybrid operating room with ceiling mounted imaging system assisted pre-operative and intra-operative lung nodule localization for thoracoscopic resections: a 5-year case series.

BACKGROUND: Video-assisted thoracoscopic (VATS) lung resections are increasingly popular and localization techniques are necessary to aid resection. We describe our experience with hybrid operating room (OR) cone-beam computed tomography (CT) assisted pre-operative and intra-operative lesion localization of lung nodules for VATS wedge resections, including our novel workflow using the hybrid OR cone-beam CT to re-evaluate patients who have undergone pre-operative localization for those who are unsuitable for intra-operative localization. METHODS: Retrospective analysis of all consecutive patients with small (≤ 20 mm), deep (≥ 10 mm distance from pleura) and/or predominantly ground-glass nodules selected for lesion localization in the Interventional Radiology suite followed by re-evaluation with cone-beam CT in the hybrid OR (pre-operative), or in the hybrid OR alone (intra-operative), prior to intentional VATS wedge performed by a single surgeon at our centre from January 2017 to December 2021. RESULTS: 30 patients with 36 nodules underwent localization. All nodules were successfully resected with a VATS wedge resection, although 10% of localizations had hookwire or coil dislodgement. The median effective radiation dose in the pre-operative group was 10.4 mSV including a median additional radiation exposure of 0.9 mSV in the hybrid OR for reconfirmation of hookwire or coil position prior to surgery (p = 0.87). The median effective radiation dose in the intra-operative group was 3.2 mSV with a higher mean rank than the intra-operative group, suggesting a higher radiation dose (p = 0.01). CONCLUSIONS: We demonstrate that our multidisciplinary approach utilizing the hybrid OR is safe and effective. Intra-operative localization is associated with lower radiation doses. Routine use of cone-beam CT to confirm the position of the physical marker prior to surgery in the hybrid OR helps mitigate consequences of localization failure with only a modest increase in radiation exposure.

Presurgical computed tomography-guided localization of lung ground glass nodules: comparing hook-wire and indocyanine green.

BACKGROUND: Presurgical computed tomography (CT)-guided localization is frequently employed to reduce the thoracotomy conversion rate, while increasing the rate of successful sublobar resection of ground glass nodules (GGNs) via video-assisted thoracoscopic surgery (VATS). In this study, we compared the clinical efficacies of presurgical CT-guided hook-wire and indocyanine green (IG)-based localization of GGNs. METHODS: Between January 2018 and December 2021, we recruited 86 patients who underwent CT-guided hook-wire or IG-based GGN localization before VATS resection in our hospital, and compared the clinical efficiency and safety of both techniques. RESULTS: A total of 38 patients with 39 GGNs were included in the hook-wire group, whereas 48 patients with 50 GGNs were included in the IG group. There were no significant disparities in the baseline data between the two groups of patients. According to our investigation, the technical success rates of CT-based hook-wire- and IG-based localization procedures were 97.4% and 100%, respectively (P = 1.000). Moreover, the significantly longer localization duration (15.3 ± 6.3 min vs. 11.2 ± 5.3 min, P = 0.002) and higher visual analog scale (4.5 ± 0.6 vs. 3.0 ± 0.5, P = 0.001) were observed in the hook-wire patients, than in the IG patients. Occurrence of pneumothorax was significantly higher in hook-wire patients (27.3% vs. 6.3%, P = 0.048). Lung hemorrhage seemed higher in hook-wire patients (28.9% vs. 12.5%, P = 0.057) but did not reach statistical significance. Lastly, the technical success rates of VATS sublobar resection were 97.4% and 100% in hook-wire and IG patients, respectively (P = 1.000). CONCLUSIONS: Both hook-wire- and IG-based localization methods can effectively identified GGNs before VATS resection. Furthermore, IG-based localization resulted in fewer complications, lower pain scores, and a shorter duration of localization.

Dual-layer detector spectral CT-based machine learning models in the differential diagnosis of solitary pulmonary nodules.

The benign and malignant status of solitary pulmonary nodules (SPNs) is a key determinant of treatment decisions. The main objective of this study was to validate the efficacy of machine learning (ML) models featured with dual-layer detector spectral computed tomography (DLCT) parameters in identifying the benign and malignant status of SPNs. 250 patients with pathologically confirmed SPN were included in this study. 8 quantitative and 16 derived parameters were obtained based on the regions of interest of the lesions on the patients' DLCT chest enhancement images. 6 ML models were constructed from 10 parameters selected after combining the patients' clinical parameters, including gender, age, and smoking history. The logistic regression model showed the best diagnostic performance with an area under the receiver operating characteristic curve (AUC) of 0.812, accuracy of 0.813, sensitivity of 0.750 and specificity of 0.791 on the test set. The results suggest that the ML models based on DLCT parameters are superior to the traditional CT parameter models in identifying the benign and malignant nature of SPNs, and have greater potential for application.

Effective lung nodule detection using deep CNN with dual attention mechanisms.

Novel methods are required to enhance lung cancer detection, which has overtaken other cancer-related causes of death as the major cause of cancer-related mortality. Radiologists have long-standing methods for locating lung nodules in patients with lung cancer, such as computed tomography (CT) scans. Radiologists must manually review a significant amount of CT scan pictures, which makes the process time-consuming and prone to human error. Computer-aided diagnosis (CAD) systems have been created to help radiologists with their evaluations in order to overcome these difficulties. These systems make use of cutting-edge deep learning architectures. These CAD systems are designed to improve lung nodule diagnosis efficiency and accuracy. In this study, a bespoke convolutional neural network (CNN) with a dual attention mechanism was created, which was especially crafted to concentrate on the most important elements in images of lung nodules. The CNN model extracts informative features from the images, while the attention module incorporates both channel attention and spatial attention mechanisms to selectively highlight significant features. After the attention module, global average pooling is applied to summarize the spatial information. To evaluate the performance of the proposed model, extensive experiments were conducted using benchmark dataset of lung nodules. The results of these experiments demonstrated that our model surpasses recent models and achieves state-of-the-art accuracy in lung nodule detection and classification tasks.

Evaluation of an artificial intelligence U-net algorithm for pulmonary nodule tracking on chest computed tomography images.

OBJECTIVES: To apply image registration in the follow up of lung nodules and verify the feasibility of automatic tracking of lung nodules using an artificial intelligence (AI) method. METHODS: For this retrospective, observational study, patients with pulmonary nodules 5-30 mm in diameter on computed tomography (CT) and who had at least six months follow-up were identified. Two radiologists defined a 'correct' cuboid circumscribing each nodule which was used to judge the success/failure of nodule tracking. An AI algorithm was applied in which a U-net type neural network model was trained to predict the deformation vector field between two examinations. When the estimated position was within a defined cuboid, the AI algorithm was judged a success. RESULTS: In total, 49 lung nodules in 40 patients, with a total of 368 follow-up CT examinations were examined. The success rate for each time evaluation was 94% (345/368) and for 'nodule-by-nodule evaluation' was 78% (38/49). Reasons for a decrease in success rate were related to small nodules and those that decreased in size. CONCLUSION: Automatic tracking of lung nodules is highly feasible.

Safety and feasibility comparison between three different CT-guided localization techniques under systemic approach algorithm.

INTRODUCTION: In the era of lung cancer screening, more and more sub-centimeter indeterminate lung lesions are being identified. It is difficult to approach these lesions and obtain tissue to confirm diagnosis. CT-guided navigation followed by surgical resection is the best way to overcome this difficulty. The aim of this study is to compare the safety and feasibility of wire and dye-tattoo CT-guided localization techniques. MATERIALS AND METHODS: From September 2019 to August 2021, 418 patients who presented with single lung lesion and received single CT-guided localization were included in this study. Procedure details, navigation results, and related complications were compared. RESULTS: For patients who received wire localization, majority (98.3 %) had perihilar lesions. In addition, 68 (57.1 %) patients received tangential approach because of lesions were blocked by bony or vital structure, abutting major fissure, or previous approach failure. The characteristics of lesion location was quite different than dye-tattooing technique (p = 0.033). As regards persistence of the target lesion localization, the interval between localization and surgery using ICG tattooing was 829.0 ± 552.9 min; much longer than the other two navigation techniques (p < 0.0001). As regards safety, patients who received wire localization had a higher rate of pneumothorax (p = 0.042) and pulmonary hemorrhage (p < 0.001) than the dye-tattooing techniques. DISCUSSION: CT-guided navigation techniques are safe and feasible. Wire localization is suitable for centrally located lesions but the wire needs to be fixed properly and symptomatic pneumothorax monitored for. Dye-tattooing is more suitable for peripheral lesions, while ICG localization persists longer than other techniques.