The effect of pulmonary vessel suppression on computerized detection of nodules in chest CT scans

[Segmentation of Mass in Mammogram Using Gaze Search Patterns]

PURPOSE

It is very difficult for a radiologist to correctly detect small lesions and lesions hidden on dense breast tissue on a mammogram. Therefore, recently, computer-aided detection (CAD) systems have been widely used to assist radiologists in interpreting images. Thus, in this study, we aimed to segment mass on the mammogram with high accuracy by using focus images obtained from an eye-tracking device.

METHODS

We obtained focus images for two mammography expert radiologists and 19 mammography technologists on 8 abnormal and 8 normal mammograms published by the DDSM. Next, the auto-encoder, Pix2Pix, and UNIT learned the relationship between the actual mammogram and the focus image, and generated the focus image for the unknown mammogram. Finally, we segmented regions of mass on mammogram using the U-Net for each focus image generated by the auto-encoder, Pix2Pix, and UNIT.

RESULTS

The dice coefficient in the UNIT was 0.64±0.14. The dice coefficient in the UNIT was higher than that in the auto-encoder and Pix2Pix, and there was a statistically significant difference (p<0.05). The dice coefficient of the proposed method, which combines the focus images generated by the UNIT and the original mammogram, was 0.66±0.15, which is equivalent to the method using the original mammogram.

CONCLUSION

In the future, it will be necessary to increase the number of cases and further improve the segmentation.

BiRPN-YOLOvX: A weighted bidirectional recursive feature pyramid algorithm for lung nodule detection

BACKGROUND

Lung cancer has the second highest cancer mortality rate in the world today. Although lung cancer screening using CT images is a common way for early lung cancer detection, accurately detecting lung nodules remains a challenged issue in clinical practice.

OBJECTIVE

This study aims to develop a new weighted bidirectional recursive pyramid algorithm to address the problems of small size of lung nodules, large proportion of background region, and complex lung structures in lung nodule detection of CT images.

METHODS

First, the weighted bidirectional recursive feature pyramid network (BiPRN) is proposed, which can increase the ability of network model to extract feature information and achieve multi-scale fusion information. Second, a CBAM_CSPDarknet53 structure is developed to incorporate an attention mechanism as a feature extraction module, which can aggregate both spatial information and channel information of the feature map. Third, the weighted BiRPN and CBAM_CSPDarknet53 are applied to the YOLOvX model for lung nodule detection experiments, named BiRPN-YOLOvX, where YOLOvX represents different versions of YOLO. To verify the effectiveness of our weighted BiRPN and CBAM_ CSPDarknet53 algorithm, they are fused with different models of YOLOv3, YOLOv4 and YOLOv5, and extensive experiments are carried out using the publicly available lung nodule datasets LUNA16 and LIDC-IDRI. The training set of LUNA16 contains 949 images, and the validation and testing sets each contain 118 images. There are 1987, 248 and 248 images in LIDC-IDRI's training, validation and testing sets, respectively.

RESULTS

The sensitivity of lung nodule detection using BiRPN-YOLOv5 reaches 98.7% on LUNA16 and 96.2% on LIDC-IDRI, respectively.

CONCLUSION

This study demonstrates that the proposed new method has potential to help improve the sensitivity of lung nodule detection in future clinical practice.

PUNDIT: Pulmonary nodule detection with image category transformation

BACKGROUND

Convolutional neural networks (CNNs) have achieved great success in pulmonary nodules detection, which plays an important role in lung cancer screening.

PURPOSE

In this paper, we proposed a novel strategy for pulmonary nodule detection by learning it from a harder task, which was to transform nodule images into normal images. We named this strategy as pulmonary nodule detection with image category transformation (PUNDIT).

METHODS

There were two steps for nodules detection, nodule candidate detection and false positive (FP) reduction. In nodule candidate detection step, a segmentation-based framework was built for detection. We designed an image category transformation (ICT) task to translate nodule images into pixel-to-pixel normal images and share the information of detection and transformation tasks by multitask learning. As for references of transformation tasks, we proposed background consistency losses into standard cycle-consistent adversarial networks, which can solve the problem of background uncontrolled changing. A three-dimensional network was used in FP reduction step.

RESULTS

PUNDIT was evaluated in two datasets, cancer screening dataset (CSD) with 1186 nodules for cross-validation and (CTD) with 3668 nodules for external test. Results were mainly evaluated by competition performance metric (CPM), the average sensitivity at seven predefined FP rates. The CPM was improved from 0.906 to 0.931 in CSD, and from 0.835 to 0.848 in CTD.

CONCLUSIONS

Experimental results showed that PUNDIT can improve the performance of pulmonary nodules detection effectively.

Lung Nodule Detectability of Artificial Intelligence-assisted CT Image Reading in Lung Cancer Screening

BACKGROUND

Artificial intelligence (AI)-based automatic lung nodule detection system improves the detection rate of nodules. It is important to evaluate the clinical value of AI system by comparing AI-assisted nodule detection with actu-al radiology reports.

OBJECTIVE

To compare the detection rate of lung nodules between the actual radiology reports and AI-assisted reading in lung cancer CT screening.

METHODS

Participants in chest CT screening from November to December 2019 were retrospectively included. In the real-world radiologist observation, 14 residents and 15 radiologists participated to finalize radiology reports. In AI-assisted reading, one resident and one radiologist reevaluated all subjects with the assistance of an AI system to lo-cate and measure the detected lung nodules. A reading panel determined the type and number of detected lung nodules between these two methods.

RESULTS

In 860 participants (57±7 years), the reading panel confirmed 250 patients with >1 solid nodule, while radiolo-gists observed 131, lower than 247 by AI-assisted reading (p<0.001). The panel confirmed 111 patients with >1 non-solid nodule, whereas radiologist observation identified 28, lower than 110 by AI-assisted reading (p<0.001). The accuracy and sensitivity of radiologist observation for solid nodules were 86.2% and 52.4%, lower than 99.1% and 98.8% by AI-assisted reading, respectively. These metrics were 90.4% and 25.2% for non-solid nodules, lower than 98.8% and 99.1% by AI-assisted reading, respectively.

CONCLUSION

Comparing with the actual radiology reports, AI-assisted reading greatly improves the accuracy and sensi-tivity of nodule detection in chest CT, which benefits lung nodule detection, especially for non-solid nodules.

Comparison of performances of conventional and deep learning-based methods in segmentation of lung vessels and registration of chest radiographs

Conventional machine learning-based methods have been effective in assisting physicians in making accurate decisions and utilized in computer-aided diagnosis for more than 30 years. Recently, deep learning-based methods, and convolutional neural networks in particular, have rapidly become preferred options in medical image analysis because of their state-of-the-art performance. However, the performances of conventional and deep learning-based methods cannot be compared reliably because of their evaluations on different datasets. Hence, we developed both conventional and deep learning-based methods for lung vessel segmentation and chest radiograph registration, and subsequently compared their performances on the same datasets. The results strongly indicated the superiority of deep learning-based methods over their conventional counterparts.