A Novel Deep Learning Model to Distinguish Malignant Versus Benign Solid Lung Nodules

Wang, Shuwen; Zhou, Leilei; Li, Xiaoran; Tang, Jie; Wu, Jing; Xia, Yin; Chen, Yu‐Chen; Lu, Lingeng

doi:10.12659/msm.936830

Cited by 7 publications

(5 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The size of the pulmonary nodule was an important indicator of malignancy risk, and only nodules with a diameter larger than 6 mm have a slightly elevated risk (≥ 1%) of malignancy based on Lung CT Screening Reporting and Data System (Lung-RADS) [20]. In our research, the minimum size of malignant nodules in our study was larger than benign nodules in both training and validation data set (12mm Verus 10mm, and 12mm Verus 11mm), which were smaller than in previous reports [13]. What's more, our benign nodules contained many uncommon categories including granulomatous in ammation, hamartomas, and others, which are di cult to differentiate from lung cancer by means of clinical characteristics.…”

Section: Discussioncontrasting

confidence: 54%

“…Lately, with the development of arti cial intelligence, deep learning has developed rapidly and has applied to medical research. Among them, CNNs whose main tasks were feature extraction and classi cation, have achieved success in the detection and differentiation of pulmonary nodules [13][14][15]21]. In this research, we established E cientNet-B0 and Res2Net models as 2D and 3D modules, respectively.…”

Section: Discussionmentioning

confidence: 99%

“…E cientNet-B0 and Res2Net models are the most common 2D and 3D modules which applied in medical elds, such as the E cientNet-B0 model, which performed well in lung ultrasound classi cation [6], colposcopy diagnosis for precancerous cervical lesions [7], and diabetic retinopathy diagnosis[8] and the Res2Net model, which is widely applied in the diagnosis of COVID-19 [9], classi cation of breast cancer [10], and pulmonary nodules detection [11,12]. According to a newly published study, a novel deep learning model (Inception V3 model) based three-dimensional (3D) images showed good differentiation accuracy and speci city to classify benign vs malignant lung nodules and achieved 98.9% accuracy [13]. Ohno et al used CNN to improve the differentiation e cacy of benign from malignant pulmonary nodules [14].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Comparison of EfficientNet-B0 and Res2Net Models for Distinguishing Malignant and Benign Pulmonary Nodules

Shao,

Xu,

Chen

et al. 2024

Preprint

View full text Add to dashboard Cite

To develop and validate two models for convolutional neural networks (CNNs), namely EfficientNet-B0 and Res2Net, constructed from non-contrast CT images for discriminate malignant and benign pulmonary nodules.We recruited 3579 cases of solitary pulmonary nodules, among which 335 cases were benign and 3244 cases were malignant. The EfficientNet-B0 model and Res2Net model were constructed as two-dimensional(2D), and three- dimensional (3D) models, respectively. Furthermore, 4 clinical factors (sex, age, smoking status, and tumor marker) and all clinical factors were combined with Res2Net model to build Res2Net-4F model, and Res2Net-20F model, respectively. The receiver operating characteristic (ROC) curves were utilized to evaluate the diagnostic efficiency and discriminative capability of these models, and ROC curves of these models were compared with Delong test.The diagnostic accuracy of Res2Net, Res2Net-4F, and Res2Net-20F [areas under ROC curves (AUC) = 0.9301, AUC = 0.9811, and AUC = 0.9357, respectively] were higher than that of the EfficientNet B0 (AUC = 0.8801) in the training data set. The results were confirmed by the validation data set (AUC = 0.8282 for the Res2Net; AUC = 0.8299 for Res2Net-4F; AUC = 0.8468 for Res2Net-20F; AUC = 0.7737 for the EfficientNet B0). There was a significant difference between Res2Net model and EfficientNet-B0 model in discriminating malignant and benign pulmonary nodules in both the training data set and validation set (Delong test, both p < 0.05). We developed two novel deep learning models to distinguish malignant and benign pulmonary nodules, and the Res2Net model showed better differentiation accuracy and sensitivity than EfficientNet-B0 model.

show abstract

Section: Discussioncontrasting

confidence: 54%

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Comparison of EfficientNet-B0 and Res2Net Models for Distinguishing Malignant and Benign Pulmonary Nodules

Shao,

Xu,

Chen

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Wang et al created a model based on retrospective non-contrast thin-layer CT scans to distinguish between benign and malignant solid lung nodules. The model’s accuracy was 98.9%, which was higher than other models’ accuracies ( 68 ). Schwyzer et al used ANNs to assess the detection of lung cancer based on varying doses of PET ( 69 ).…”

Section: Machine Learningmentioning

confidence: 55%

Developing technologies and areas of interest in lung cancer screening adjuncts

Park,

Pham,

Yalamanchili

et al. 2024

J Thorac Dis

View full text Add to dashboard Cite

Lung cancer remains the leading cause of cancer mortality. Screening guidelines have been implemented in the past decade to aid in earlier detection of at-risk groups. Nevertheless, computed tomography (CT) scans, the principal screening modality in use today, are still low yield, with 3.6% of lung cancer confirmed amongst 39.1% of lesions detected over a 3-year period. They also carry relatively high false positive rates, between 9% and 27%, which can bear unnecessary financial and emotional costs to patients. As such, research efforts have been dedicated to the development of lung cancer screening adjuncts to improve detection reliability. We herein review several emerging technologies in this specific arena and their efficacy. These include plasma markers (microDNA, DNA methylation, and tumor-associated antibodies), breath/sputum biomarkers [volatile organic compounds (VOCs) and exhaled breath condensate (EBC)], proteomics, metabolomics, and machine learning, such as radiomics technology. We find that, across the board, they offer promising results in terms of non-invasive diagnostics, genetic sequencing for higher-risk individuals, and accessibility for a diverse cohort of patients. While these screening adjuncts are unlikely to completely replace the current standard of care at the moment, continued research into these technologies is crucial to improve and personalize the identification, treatment, and outcome of lung cancer patients in the near future.

show abstract

“…ANN diagnosis achieves growing acceptance in the field of radiology [ 23 ]. Taking into account the aging population and the growing number of cancer patients, the future role of ANNs in cancer diagnosis seems inevitable [ 1 ].…”

Section: Discussionmentioning

confidence: 99%

Improving Renal Tumor Diagnosis with Computed Tomography and Artificial Neural Networks

Glembin¹,

Obuchowski

Klaudel

et al. 2023

Med Sci Monit

View full text Add to dashboard Cite

Background:Renal cell carcinoma is one of the most common cancers in Europe, with a total incidence rate of 18.4 cases per 100 000 population. There is currently significant overdiagnosis (11% to 30.9%) at times of planned surgery based on radiological studies. The purpose of this study was to create an artificial neural network (ANN) solution based on computed tomography (CT) images as an additional tool to improve the differentiation of malignant and benign renal tumors and to aid active surveillance. Material/Methods:A retrospective study based on CT images was conducted. Axial CT images of 357 renal tumor cases were collected. There were 265 (74.2%) cases histologically proven to be malignant, while 34 (9.5%) cases were benign. Radiologists diagnosed 58 (16.3%) cases as angiomyolipoma (AML), based on characteristic appearance, not confirmed histopathologically. For ANN training, the arterial CT phase images were used. A total of 7207 arterial-phase images were collected, then cropped and added to the database with the associated diagnosis. For the test dataset (ANN validation), 38 cases (10 benign, 28 malignant) were chosen by subgroup randomization to correspond to statistical tumor type distribution. The VGG-16 ANN architecture was used in this study. Results:Trained ANN correctly classified 23 out of 28 malignant tumors and 8 out of 10 benign tumors. Accuracy was 81.6% (95% confidence interval, 65.7-92.3%), sensitivity was 82.1% (63.1-93.9%), specificity was 80.0% (44.4-97.5%), and F1 score was 86.8% (74.7-94.5%). Conclusions:The created ANN achieved promising accuracy in differentiating benign vs malignant renal tumors.

show abstract

A Novel Deep Learning Model to Distinguish Malignant Versus Benign Solid Lung Nodules

Cited by 7 publications

References 50 publications

Comparison of EfficientNet-B0 and Res2Net Models for Distinguishing Malignant and Benign Pulmonary Nodules

Comparison of EfficientNet-B0 and Res2Net Models for Distinguishing Malignant and Benign Pulmonary Nodules

Developing technologies and areas of interest in lung cancer screening adjuncts

Improving Renal Tumor Diagnosis with Computed Tomography and Artificial Neural Networks

Contact Info

Product

Resources

About