Lung Nodule Malignancy Classification Based ON NLSTx Data

Veasey, Benjamin; Farhangi, M. Mehdi; Frigui, Hichem; Broadhead, Justin; Dahle, Michael; Pezeshk, Aria; Seow, Albert; Amini, Amir A.

doi:10.1109/isbi45749.2020.9098486

Cited by 11 publications

(16 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The approach provides flexibility by allowing various number of inputs to be processed concurrently while also reducing the number of overall weight parameters since they are shared across twin branches. As we and others have previously shown [12]- [14], 2-D CNN-based systems can effectively encode 3-D information from multiple 2-D slices. This is important because it permits pre-trained 2-D networks with substantially fewer parameters to be applied to 3-D image data with comparable results.…”

Section: Figmentioning

confidence: 66%

See 1 more Smart Citation

Lung Nodule Malignancy Prediction From Longitudinal CT Scans With Siamese Convolutional Attention Networks

Veasey

Broadhead

Dahle

et al. 2020

IEEE Open J. Eng. Med. Biol.

Self Cite

View full text Add to dashboard Cite

We propose a convolutional attentionbased network that allows for use of pre-trained 2-D convolutional feature extractors and is extendable to multitime-point classification in a Siamese structure. Methods: Our proposed framework is evaluated for single-and multitime-point classification to explore the value that temporal information, such as nodule growth, adds to malignancy prediction. Results: Our results show that the proposed method outperforms a comparable 3-D network with less than half the parameters on single-time-point classification and further achieves performance gains on multi-time-point classification. Conclusions: Attention-based, Siamese 2-D pre-trained CNNs lead to fast training times and are effective for malignancy prediction from single-time-point or multiple-time-point imaging data.

show abstract

Section: Figmentioning

confidence: 66%

“…A description of all networks used in this paper is given in Table III for quick referencing. [12] that use recurrent modules to encode slice-wise features. Each CRN network uses different weight initializations or a different number of inputs.…”

Section: Table II Network Architecturesmentioning

confidence: 99%

Lung Nodule Malignancy Prediction From Longitudinal CT Scans With Siamese Convolutional Attention Networks

Veasey

Broadhead

Dahle

et al. 2020

IEEE Open J. Eng. Med. Biol.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Image analysis for lung nodules: Lung cancer is an important disease and attracts a great deal of interest in the field of radiology. Many applications for detection [2], [3], [4], [19], segmentation [20], and characterization [5], [6], [7], [8] of lung nodules, which are candidate lesions for lung cancer, have been developed so far. Most of the studies on characterization of lung nodules are on classification of their malignancy [7], [8], and there are still few studies on the prediction of imaging features such as morphology and internal characteristics of lung nodules [5].…”

Section: Related Workmentioning

confidence: 99%

“…Based on this background, many researchers have tackled pulmonary nodule detection from CT images [2], [3], [4] and characterization [5], [6], [7], [8]. An issue of these tasks is the intensive annotation cost to build training data.…”

Section: Introductionmentioning

confidence: 99%

“…Thus, the manual annotation is cost intensive and not scalable. Since it is difficult to construct a large training dataset which has many kinds of labels solely by manual annotation, researchers focus on nodule characterization of a single category, such as malignancy [7], [8] and opacity [6]. Although a previous study [5] that can classify multiple characteristics of lung nodules exists, it does not cover the characteristics required for qualitative assessment of lung nodules.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Characterization of Pulmonary Nodules in Computed Tomography Images Based on Pseudo-Labeling Using Radiology Reports

Momoki

Ichinose

Shigeto

et al. 2022

IEEE Trans. Circuits Syst. Video Technol.

View full text Add to dashboard Cite

A computer-aided diagnosis (CAD) system that characterizes nodules in medical images can help radiologists determine its malignancy. Preparing large volumes of labeled data for CAD systems, however, requires advanced medical knowledge. This makes it extremely difficult to develop such systems, despite their growing demand. In this paper, we propose a new training method to build an image classifier for characterization of nodules utilizing pseudo-labels, i.e., image labels automatically retrieved from radiology reports. A radiology report is a type of record in which radiologists present a summary of lesion characteristics and diagnosis. Labeling radiology reports is much easier than labeling radiology images, and can be done without high expertise. Using several thousand labeled reports, we constructed a hierarchical attention network-based text classifier to assign pseudo-labels of the characteristics of pulmonary nodules with high accuracy (macro F1-score of 0.941). Experimental results show that the image classifier trained with the pseudo-labels can achieve almost the same performance as the one trained with the labels annotated by radiologists: AUC 0.848 for the model trained with the pseudo-labels on 3,000 computed tomography (CT) images and 0.847 for the model trained with the manual labels on 800 CT images.

show abstract

Siamese Encoder-based Spatial-Temporal Mixer for Growth Trend Prediction of Lung Nodules on CT Scans

Fang

Wang²,

Li³

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Lung Nodule Malignancy Classification Based ON NLSTx Data

Cited by 11 publications

References 11 publications

Lung Nodule Malignancy Prediction From Longitudinal CT Scans With Siamese Convolutional Attention Networks

Lung Nodule Malignancy Prediction From Longitudinal CT Scans With Siamese Convolutional Attention Networks

Characterization of Pulmonary Nodules in Computed Tomography Images Based on Pseudo-Labeling Using Radiology Reports

Siamese Encoder-based Spatial-Temporal Mixer for Growth Trend Prediction of Lung Nodules on CT Scans

Contact Info

Product

Resources

About