Dual energy CT image prediction on primary tumor of lung cancer for nodal metastasis using deep learning

Wang, Youwei; Chen, Chii-Jen; Huang, Hsu-Cheng; Wang, Teh-Chen; Chen, Hsin‐Ming; Shih, Jin‐Yuan; Chen, Jin‐Shing; Huang, Yu‐Sen; Chang, Yeun‐Chung; Chang, Ruey-Feng

doi:10.1016/j.compmedimag.2021.101935

Cited by 20 publications

(8 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Conventional CT examination has become the preferred imaging method for the diagnosis and di erential diagnosis of lung cancer at this stage, but it mainly relies on the morphological changes of the lesions for diagnosis, resulting in high falsenegative and false-positive rates in the diagnosis of lymph nodes metastasis. Spectral CT imaging technology is an excellent crystallization of the continuous optimization and development of imaging technology, which can improve more imaging information than conventional CT [12][13][14].…”

Section: Discussionmentioning

confidence: 99%

[Retracted] Correlation with Spectral CT Imaging Parameters and Occult Lymph Nodes Metastases in Sufferers with Isolated Lung Adenocarcinoma

Liu

Nie

2022

Contrast Media & Molecular Imaging

View full text Add to dashboard Cite

For investigating the correlation with spectral CT imaging parameters and occult lymph nodes metastasis in sufferers with isolated lung adenocarcinoma. The clinic cases data of 352 sufferers with isolated lung adenocarcinoma from January 2019 to January 2022 were assembled. In line with whether the sufferers had occult lymph nodes metastasis, they were taken as a part in the metastasis group (n = 172) and the nonmetastasis group. All sufferers were scanned by spectral CT with a dual-phase contrast-enhanced method, and the recording of spectral CT imaging parameters in arteriovenous phase, iodine concentration (IC), water concentration (WC), the slope rate of the spectral HU curve (λHU), the normalized iodine concentration(NIC), the normalized water concentration(NWC), the normalized effective atomic number (Neff-Z)], and receiver operating characteristic (ROC) were employed to analyze the spectral CT imaging parameters of the arteriovenous phase. Evaluation of occult lymph nodes metastases in sufferers with isolated lung adenocarcinoma. The IC, NIC, λHU, and Neff-Z in the arteriovenous phase spectral CT imaging parameters of the metastasis group were obviously smaller than that of the nonmetastasis group, and the discrepancies were statistically obvious ( P < 0.05 ). The results of ROC curve analysis manifested that the area under the curve (AUC) of λHU, IC, NIC, and Neff-Z in the CT parameters of the arterial phase were 0. 840 (95%CI : 0. 796–0.883), 0.763 (95% CI : 0.708–0.818), 0.918 (95% CI : 0.888–0.948), 0.778 (95% CI : 0.731–0.826). The AUCs of λHU, IC, NIC, and Neff-Z in the venous phase spectral CT parameters were 0.909 (95% CI : 0.877–0.941), 0.837 (95% CI : 0.792–0.881), and 0.980 (95% CI : 0.968–0.968), respectively. 0.993), 0.792 (95% CI : 0.742∼0.842). Spectral CT imaging parameters have a certain value in evaluating occult lymph nodes metastasis in sufferers with isolated lung adenocarcinoma, which is helpful for doctors to judge the lymph nodes metastasis in sufferers with this disease before surgery.

show abstract

Section: Discussionmentioning

confidence: 99%

[Retracted] Correlation with Spectral CT Imaging Parameters and Occult Lymph Nodes Metastases in Sufferers with Isolated Lung Adenocarcinoma

Liu

Nie

2022

Contrast Media & Molecular Imaging

View full text Add to dashboard Cite

show abstract

“…This makes artificial neural networks to be a valuable toolbox for prediction. Indeed, deep learning models have been successfully applied in health care to predict clinical events, disease classification, and electronic health record data augmentation [23][24][25][26][27]. However, the use of deep learning to detect disease and complications in elderly patients is limited.…”

Section: Discussionmentioning

confidence: 99%

Prediction of postoperative infection in elderly using deep learning-based analysis: an observational cohort study

Wang

et al. 2023

Aging Clin Exp Res

View full text Add to dashboard Cite

Elderly patients are susceptible to postoperative infections with increased mortality. Analyzing with a deep learning model, the perioperative factors that could predict and/or contribute to postoperative infections may improve the outcome in elderly. This was an observational cohort study with 2014 elderly patients who had elective surgery from 28 hospitals in China from April to June 2014. We aimed to develop and validate deep learning-based predictive models for postoperative infections in the elderly. 1510 patients were randomly assigned to be training dataset for establishing deep learning-based models, and 504 patients were used to validate the effectiveness of these models. The conventional model predicted postoperative infections was 0.728 (95% CI 0.688–0.768) with the sensitivity of 66.2% (95% CI 58.2–73.6) and specificity of 66.8% (95% CI 64.6–68.9). The deep learning model including risk factors relevant to baseline clinical characteristics predicted postoperative infections was 0.641 (95% CI 0.545–0.737), and sensitivity and specificity were 34.2% (95% CI 19.6–51.4) and 88.8% (95% CI 85.6–91.6), respectively. Including risk factors relevant to baseline variables and surgery, the deep learning model predicted postoperative infections was 0.763 (95% CI 0.681–0.844) with the sensitivity of 63.2% (95% CI 46–78.2) and specificity of 80.5% (95% CI 76.6–84). Our feasibility study indicated that a deep learning model including risk factors for the prediction of postoperative infections can be achieved in elderly. Further study is needed to assess whether this model can be used to guide clinical practice to improve surgical outcomes in elderly.

show abstract

“…Since lung GGNs are histologically special, with small lesions and great affection by respiration, studies on GGN imaging at home and abroad are mainly focused on the characteristics of lesions in unenhanced CT scanning and follow-up according to GGN growth to confirm the doubling time [ 3 – 6 ]. There have been relatively few studies in China on the dynamic enhanced CT of early lung adenocarcinoma with GGNs.…”

Section: Discussionmentioning

confidence: 99%

“…The separation of components of tissues helps produce two groups of pairwise base matter density map. The quantitative measurement of the concentrations of base matter provides quantitative information for the correct diagnosis of the lesion [ 3 – 6 ]. Iodine is the main component of the contrast agent of enhanced CT, whose concentrations are measured in GGN lesions after contrast enhancement of iodine-based images.…”

Section: Discussionmentioning

confidence: 99%

Application Value of Spectral CT Imaging in Quantitative Analysis of Early Lung Adenocarcinoma

Yu²,

Luo

et al. 2022

Journal of Oncology

View full text Add to dashboard Cite

Objective. To investigate the clinical value of gemstone energy spectral CT imaging for the quantitative analysis of early lung adenocarcinoma. Methods. 76 cases of pulmonary ground-glass nodules pathologically confirmed as precancerous lesion and early lung cancer (including pure ground-glass nodules in 46 cases and mixed ground-glass nodules in 30 cases) underwent chest CT scan first and then underwent contrast-enhanced gemstone energy spectral CT to get arterial phase images, venous phase images, and delayed phase images. All the lesions were set the region of interest (ROI). Cases of the pure ground-glass nodule (pGGN) were measured at the maximum level of lesions, cases of the mixed ground-glass nodule (mGGN) were measured in two areas of ground-glass and solid components, CT value and iodine concentrations of lesions in three-phase scanning were separately measured, and at the same time, iodine concentrations of the thoracic aorta were also measured. The normalized iodine concentrations (NICs) were calculated, that is, the ratio of iodine concentrations of lesions and the thoracic aorta. CT values of lesions were also measured at each stage of 70 keV. All the quantitative data were expressed by the mean ± standard deviation, and paired t-test was used for pairwise comparison. Results. In 76 cases, in the spectral CT imaging mode, the NIC value of solid components of the GGN was 0.33 ± 0.16 in the arterial phase (AP), 0.52 ± 0.25 in the venous phase (VP), and 0.58 ± 0.34 in the delayed phase (DP). There were significant differences of P values of NICs between each two phases in both solid component cases and the ground-glass component cases in AP/VP, VP/DP, and AP/DP ( P < 0.05 ); there were no statistically significant P values of CT values between each two phases in three-period enhanced CT in both the solid component cases and the ground-glass component cases in AP/VP, VP/DP, and AP/DP ( P > 0.05 ). Conclusion. Gemstone energy spectral CT with quantitative imaging can dynamically reflect the enhancement features of the pulmonary GGN.

show abstract

Dual energy CT image prediction on primary tumor of lung cancer for nodal metastasis using deep learning

Cited by 20 publications

References 26 publications

[Retracted] Correlation with Spectral CT Imaging Parameters and Occult Lymph Nodes Metastases in Sufferers with Isolated Lung Adenocarcinoma

[Retracted] Correlation with Spectral CT Imaging Parameters and Occult Lymph Nodes Metastases in Sufferers with Isolated Lung Adenocarcinoma

Prediction of postoperative infection in elderly using deep learning-based analysis: an observational cohort study

Application Value of Spectral CT Imaging in Quantitative Analysis of Early Lung Adenocarcinoma

Contact Info

Product

Resources

About