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

Fang, Jiansheng; Wang, Jingwen; Li, Anwei; Yan, Yuguang; Hou, Yonghe; Song, Chao; Liu, Hongbo; Liu, Jiang

doi:10.1007/978-3-031-16431-6_46

Cited by 6 publications

(4 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Compared with STM ( 13 ), our SiamModel obtained better performance with the AUC of 0.858 (95% CI 0.786–0.921) vs. 0.823 (95% CI 0.731–0.898) in the validation set and 0.862 (95% CI 0.789–0.927) vs. 0.806 (95% CI 0.693–0.902) in the external test set, which indicated the superiority of our proposed weighted smooth-l1 loss for SSN growth prediction.…”

Section: Resultsmentioning

confidence: 99%

“…In order to analyze the changes in diameter, volume, and mass of SSNs over consecutive years, we followed the method in Fang et al., to pair the same nodules between different CT scans ( 13 ). Our data organization approach aimed to ascertain the diameter, volume, and mass change of SSNs in consecutive CT scans ( Figure 3 ).…”

Section: Methodsmentioning

confidence: 99%

“…Following selection ( 13 ), the deep learning-based model was identified (called SiamModel, Figure 4 ), where FG t , FL t , and FL t−1 represent global feature embedding of T t VOI, local feature embedding of T t VOI, and local feature embedding of T t−1 VOI, respectively. A learnable embedding FL t−1 was provided if T t−1 VOI was unavailable, which occurred when there were only two CT scans.…”

Section: Methodsmentioning

confidence: 99%

“…For a given subject, the 3D VOI pairs (T t−1 and T t ) taken from CT scans at sequential time points were fed into the Siamese encoder for extracting feature embedding. After fusing the features using the spatial-temporal mixer (STM) module ( 13 ), the fully connected layer was used to predict the growth probability. It was worth mentioning that the global information of VOIs was changeless on T t−1 and T t .…”

Section: Methodsmentioning

confidence: 99%

See 3 more Smart Citations

Deep learning-based growth prediction for sub-solid pulmonary nodules on CT images

Liao

Yan

et al. 2022

Front. Oncol.

View full text Add to dashboard Cite

BackgroundEstimating the growth of pulmonary sub-solid nodules (SSNs) is crucial to the successful management of them during follow-up periods. The purpose of this study is to (1) investigate the measurement sensitivity of diameter, volume, and mass of SSNs for identifying growth and (2) seek to establish a deep learning-based model to predict the growth of SSNs.MethodsA total of 2,523 patients underwent at least 2-year examination records retrospectively collected with sub-solid nodules. A total of 2,358 patients with 3,120 SSNs from the NLST dataset were randomly divided into training and validation sets. Patients from the Yibicom Health Management Center and Guangdong Provincial People’s Hospital were collected as an external test set (165 patients with 213 SSN). Trained models based on LUNA16 and Lndb19 datasets were employed to automatically obtain the diameter, volume, and mass of SSNs. Then, the increase rate in measurements between cancer and non-cancer groups was studied to evaluate the most appropriate way to identify growth-associated lung cancer. Further, according to the selected measurement, all SSNs were classified into two groups: growth and non-growth. Based on the data, the deep learning-based model (SiamModel) and radiomics model were developed and verified.ResultsThe double time of diameter, volume, and mass were 711 vs. 963 days (P = 0.20), 552 vs. 621 days (P = 0.04) and 488 vs. 623 days (P< 0.001) in the cancer and non-cancer groups, respectively. Our proposed SiamModel performed better than the radiomics model in both the NLST validation set and external test set, with an AUC of 0.858 (95% CI 0.786–0.921) and 0.760 (95% CI 0.646–0.857) in the validation set and 0.862 (95% CI 0.789–0.927) and 0.681 (95% CI 0.506–0.841) in the external test set, respectively. Furthermore, our SiamModel could use the data from first-time CT to predict the growth of SSNs, with an AUC of 0.855 (95% CI 0.793–0.908) in the NLST validation set and 0.821 (95% CI 0.725–0.904) in the external test set.ConclusionMass increase rate can reflect more sensitively the growth of SSNs associated with lung cancer than diameter and volume increase rates. A deep learning-based model has a great potential to predict the growth of SSNs.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

Section: Methodsmentioning

confidence: 99%

See 2 more Smart Citations