Figure 1.
Example of lung CT images. From left to right: COVID-19, common pneumonia, and normal lung CT images.
Vitamin C or ascorbic acid has been proposed as an anticancer agent, as an intervention to reduce cancer incidence, and also as a medication to reduce cancer treatment-related side effects. While there has been significant basic science research that has evaluated the potential tumoricidal mechanisms of action, clinical studies have been underpowered, retrospective, or poorly designed. Current systematic reviews have suggested that the outcome data from vitamin C therapy is limited. There is an important need for prospective clinical trials and pharmacologic studies to fully evaluate the potential of vitamin C as an anticancer agent.
Citation: Annekathryn Goodman. Vitamin C and Cancer[J]. AIMS Medical Science, 2016, 3(1): 41-51. doi: 10.3934/medsci.2016.1.41
| [1] | Zhi-Jie Wang, Hong-Rui Sun . Normalized solutions for Kirchhoff equations with Choquard nonlinearity: mass Super-Critical Case. Communications in Analysis and Mechanics, 2025, 17(2): 317-340. doi: 10.3934/cam.2025013 |
| [2] | Yangyu Ni, Jijiang Sun, Jianhua Chen . Multiplicity and concentration of normalized solutions for a Kirchhoff type problem with $ L^2 $-subcritical nonlinearities. Communications in Analysis and Mechanics, 2024, 16(3): 633-654. doi: 10.3934/cam.2024029 |
| [3] | Xueqi Sun, Yongqiang Fu, Sihua Liang . Normalized solutions for pseudo-relativistic Schrödinger equations. Communications in Analysis and Mechanics, 2024, 16(1): 217-236. doi: 10.3934/cam.2024010 |
| [4] | Qi Li, Yuzhu Han, Bin Guo . A critical Kirchhoff problem with a logarithmic type perturbation in high dimension. Communications in Analysis and Mechanics, 2024, 16(3): 578-598. doi: 10.3934/cam.2024027 |
| [5] | Shengbing Deng, Qiaoran Wu . Existence of normalized solutions for the Schrödinger equation. Communications in Analysis and Mechanics, 2023, 15(3): 575-585. doi: 10.3934/cam.2023028 |
| [6] | Wenqian Lv . Ground states of a Kirchhoff equation with the potential on the lattice graphs. Communications in Analysis and Mechanics, 2023, 15(4): 792-810. doi: 10.3934/cam.2023038 |
| [7] | Yan-Fei Yang, Chun-Lei Tang . Positive and sign-changing solutions for Kirchhoff equations with indefinite potential. Communications in Analysis and Mechanics, 2025, 17(1): 159-187. doi: 10.3934/cam.2025008 |
| [8] | Henryk Żołądek . Normal forms, invariant manifolds and Lyapunov theorems. Communications in Analysis and Mechanics, 2023, 15(2): 300-341. doi: 10.3934/cam.2023016 |
| [9] | Panyu Deng, Jun Zheng, Guchuan Zhu . Well-posedness and stability for a nonlinear Euler-Bernoulli beam equation. Communications in Analysis and Mechanics, 2024, 16(1): 193-216. doi: 10.3934/cam.2024009 |
| [10] | Chang-Jian Wang, Jia-Yue Zhu . Global existence and uniform boundedness to a bi-attraction chemotaxis system with nonlinear indirect signal mechanisms. Communications in Analysis and Mechanics, 2023, 15(4): 743-762. doi: 10.3934/cam.2023036 |
Vitamin C or ascorbic acid has been proposed as an anticancer agent, as an intervention to reduce cancer incidence, and also as a medication to reduce cancer treatment-related side effects. While there has been significant basic science research that has evaluated the potential tumoricidal mechanisms of action, clinical studies have been underpowered, retrospective, or poorly designed. Current systematic reviews have suggested that the outcome data from vitamin C therapy is limited. There is an important need for prospective clinical trials and pharmacologic studies to fully evaluate the potential of vitamin C as an anticancer agent.
After the novel coronavirus pneumonia (Corona Virus Disease 2019, COVID-19) was first discovered in November 2019, it caused extremely high morbidity and mortality worldwide in a short period of time [1]. Radiological examination can be used for COVID-19 screening, that is, radiologists perform chest imaging (such as chest X-ray or computed tomography imaging) of the patient, and can look for visual indicators related to the COVID-19 virus infection [2]. Early studies have found that once COVID-19 patients are infected, there will be abnormalities in lung imaging [3,4,5]. For example, Huang et al. found in their research that most COVID-19 positive cases showed bilateral radiological abnormalities in CXR images. These abnormalities are most common with ground glass opacity, patchy shadows, warped spreading and lung consolidation. Therefore, researchers believe that lung imaging can be used as the main tool for COVID-19 screening in the hardest-hit areas [6].
In recent years, artificial intelligence (AI) technology represented by convolutional neural networks (CNN) has made great achievements in computer vision. As the CT radiography mode has been proved to have higher sensitivity and specificity than the RT-PCR detection method, many researchers have begun to use CT and CRX image data sets to train automatic classification AI algorithms. Ozturk et al. [7] proposed a deep model that uses DarkNet as a classifier and has an end-to-end architecture. However, the number of data sets needs to be increased to make the model more robust and accurate. Y. Pathak et al. [8] used transfer learning methods to classify patients infected with COVID-19. The author introduces the smooth loss function top-2 with cost-sensitive attributes to process noise and unbalanced data sets, and uses a 10-fold cross-validation method to prevent overfitting. Narin et al. [9] used three advanced deep learning models (ResNet50, Inception v3 and Inception-ResNetV2) to classify CRX images (50 COVID-19 images vs. 50 non-COVID-19 images). In June 2021, Adel Oulefki et al. proposed a method to automatically segment and measure COVID-19 lung infection area [10]. In November 2021, Chen Zhao et al. used chest CT imaging features for lung segmentation and COVID-19 automatic detection [11]. In December 2021, Juan Juan He et al. used the evolutionary countermeasure network with gradient punishment to realize the segmentation task of COVID-19 infection [12]. In December 2021, Nan Mu et al. applied progressive global perception and local polishing network to the segmentation task of pulmonary infection in COVID-19 CT image [13]. In 2021, Qimao proposed the intelligent immune clonal optimization algorithm for pulmonary nodule classification [14]. In February 2022, X. Liu et al. used scribble annotation to segment the infected area of COVID-19 CT image [15].
Main contributions:
(1) This research proposes a two-level deep network model, which mainly realizes the classification of lung CT images. The first level of the model uses convolutional neural networks to locate lung regions in lung CT images. The second level of the model uses the capsule network to classify and predict the segmented images.
(2) SegNet-Lung segmentation model is used to segment and preprocess the lung region of lung CT image (the segmentation model in this study is named SegNet-Lung model) in order to extract shape features such as color and texture from lung CT image. At the same time, the unlabeled lung CT images are predicted and labeled as the training/test data set in the classification processing stage.
(3) The CapsNet-COVID19 capsule network classification model is established to classify and reconstruct the segmented lung CT images, and the column diagram is used to visualize the prediction results. By stopping early and increasing the amount of training data, the over fitting of the model caused by excessive iterative training in the training process is avoided. The confusion matrix, accuracy, boundary loss, reconstruction loss and total loss are used to quantitatively evaluate the capsule network model.
(4) Our classification method is suitable for medical image classification with complex background, low recognition rate, blurred boundaries and large image noise. The accuracy of our method is 84.291% on the test set and 100% on the training set. We believe that this classification method is of great value for monitoring and controlling the growth of patients in COVID-19 infected areas.
All human lung CT image data used in this research are derived from the 2019 Novel Coronavirus Information Database (2019nCoVR) provided on the website of China National Bioinformatics Center [16]. The lung CT image data in this database is provided by the China Chest CT Image Investigation Association (CC-CCII), so 2019nCoVR has high professional and practical value. There are three main types of lung CT images in 2019nCoVR: The Novel coronavirus pneumonia data set caused by the SARS-CoV-2 virus, the common pneumonia data set, and the normal lung data set. The amount of data in 2019nCoVR is relatively large. In view of the limitations of the experimental hardware equipment in this research, we selected 29,278 images with clear and obvious lung texture characteristics from the information database as the data set of this research. An example of lung CT images in the dataset is shown in Figure 1.
The main features of lung CT images of COVID-19 patients are early unilateral or bilateral lungs with localized inflammatory infiltration, which is mainly manifested as patches, masses, segments or subpleural subsegmental ground glass shadow. In the late stage, consolidation is dominant in the lungs [17]. Common pneumonia is generally divided into lobar pneumonia and lobular pneumonia. The early stage of lobar pneumonia is mainly manifested as a ground-glass-like density shadow in a certain lung lobe with fuzzy edges. Lobular pneumonia is common in the middle and lower parts of the lungs. The local bronchial vascular bundles are thickened, and there are blurry patches of varying sizes. The texture of the lungs of normal people is very clear, and there is no thickening, disorder or abnormal lines. At the same time, there are no obvious high-density or low-density dense shadows and nodules in both lungs, and no pathological manifestations such as hyperplasia, cavitation, calcification and exudation. The CT images of common pneumonia and COVID-19 overlap to a certain extent, and it is often difficult to make an accurate judgment of these two types of pneumonia by the human eye alone. Therefore, it is necessary to increase the diagnostic accuracy of COVID-19 with the help of image analysis technology.
This research uses a deep network model to analyze lung CT images to achieve automatic segmentation and classification of a large number of lung CT images (three main types). First, use an image segmentation model similar to U-NET [18] to segment the lung area of the lung CT image (the segmentation model in this research is named SegNet-Lung segmentation model). This model can extract shape features such as color and texture from lung CT images. Then, a capsule network model [19] was established to classify and predict the segmented lung CT images (the classification model in this research is named CapsNet-COVID19 classification model). The specific process is as follows:
(1) Train the segmentation model SegNet-Lung segmentation model using the labeled lung CT image set. Then, the trained SegNet-Lung segmentation model is used to segment the lung CT images to be classified.
(2) Use a self-editing program to unify the format of the segmented lung CT images: (a) The edge of the lung region in the image predicted by SegNet-Lung segmentation model is outlined with red lines; (b) Extract the lung region of the outlined lung CT image and hide other irrelevant regions; (c) Convert the segmented lung region image format into the input format available for the CapsNet-COVID19 classification model.
(3) The above-mentioned image set is divided into train set, validation set and test set according to an appropriate ratio. Through training the CapsNet-COVID19 classification model, the prediction of image classification is finally realized.
The data sets of lung CT images used in this research include: COVID-19, common pneumonia (CP) and normal lungs (Normal). The numbers of them are 9420, 9858 and 10,000 respectively. Set the resolution of the image set to 128 × 128 and the format to.tif. If the lung CT images are directly classified, the image classification accuracy will be reduced because the images are interfered by the external environment when they are taken. Therefore, it is necessary to segment the lung area of the lung CT image. We have about 2532 images of COVID-19 lung CT masks that have been annotated by experts (as shown in Figures 3 and 4). Then, they are used as the train set and validation set of the SegNet-Lung segmentation model, and the unlabeled lung CT images to be classified are used as the test set. Among them, the training set and validation set (90%) are used to train the SegNet-Lung segmentation model to determine the network parameters of the segmentation model, and the test set (10%) is used to test and predict the recognition rate and generalization of the trained segmentation model (Table 1). The test set is used to test and predict the segmentation performance of the trained SegNet-Lung segmentation model. Finally, the trained SegNet-Lung segmentation model is used for segmentation prediction and labeling of unlabeled lung CT images. Therefore, the labeled lung CT images can be used as the training/test data set of CapsNet-COVID19 classification model in the next step.
| Training/Validation Data | Test Set | Total (Labeled Images) |
| 2279 (90%) | 253 (10%) | 2532 |
DownLoad:
CSV
The SegNet-Lung segmentation model used in this study is improved on the basis of the U-NET model. This model uses Nconvolution instead of MaxPool and adds a random dropout layer with a parameter of 0.5. At the same time, the Sigmoid function is used in the output convolutional layer. During model training, the batch_size parameter is set to 64, and the convolutional layer uses a 3 × 3 filter and activation function ELU, which can reduce the feature dimension of the subsequent pooling layer and avoid overfitting. The filter sizes used by the largest pooling layer and upsampling layer are 3 × 3 and 2 × 2. After the convolutional layers are merged, the network structure of the model will be flattened. The SegNet-Lung segmentation model structure is shown in Figure 5.
The structure of nconvolution of covseg net model is convolution + batchnormalization + ELU. Batch normalization (BN) is used in the convolution process, which can forcibly pull the more and more biased data distribution in model training back to the standard distribution through certain standardized means. In this way, the active input value can fall in the area where the nonlinear function is sensitive to the input, and the output of the network will not be very large. Therefore, using the structure of nconvolution instead of maxpool can speed up the training and convergence of the network, control the gradient explosion, prevent the gradient from disappearing and prevent over fitting. At the same time, the concept structure is introduced into covseg net model, which stacks the convolution layer of 1 × 1, 3 × 3 and 5 × 5 and the pooling layer of 3 × 3 together. On the one hand, it increases the width of the network, on the other hand, it increases the adaptability of the network to scale.
We use Focal-Tversky Unet model and nn-Unet model reported in isbi as comparison models (Table 2). Through comparison, it is found that in the focus segmentation task of COVID-19 CT image, the evaluation indexes such as Dice and Precision of COVSeg-NET model proposed in this study are better than the other two models. Especially for the situation that the number of samples in medical images is small and the problem space is small, COVSeg-NET model can achieve the goal of medical image segmentation by using a variety of ways of data set amplification and preprocessing.
| Method | Dice | Precision | Recall |
| COVSeg-NET | 55.5% | 83.3% | 44.5% |
| Focal-Tversky Unet [20] | 48.4% | 48.3% | 61.9% |
| nn-Unet [21] | 4.5% | 75.4% | 48.6% |
DownLoad:
CSV
The trained SegNet-Lung segmentation model is used to segment all lung CT images to be classified (the remaining 29,278 lung CT images without segmentation annotations), and the segmentation marks are automatically labeled. Then use a self-editing program to extract the lung area in the image. Some examples of image annotation are shown in Figure 6.
Next, the CapsNet-COVID19 classification model will be used to classify the pre-segmented lung CT images. CapsNet-COVID19 has made many adjustments and improvements to the CapsNet model.
In the CapsNet model [19], the capsule can contain the attributes and characteristics of the entity. The training goal of the capsule is to capture the possibility of features and their variants, not the characteristics of specific variants. Therefore, the capsule can be used not only to detect image features, but also to learn image variants. In this way, the same capsule can detect the same object category with different directions.
The input $ {u}_{i} $ of the capsule is a vector, apply a transformation matrix $ {W}_{ij} $ to get the output of the upper layer of the capsule. For an m × k matrix, the k-dimensional $ {u}_{i} $ is converted to the m-dimensional $ \hat{u}_{j \mid i} \quad((m \times k) \times(k \times 1) \Rightarrow m \times 1) $, and then the weight $ {c}_{ij} $ is used to calculate the weighted sum $ {s}_{ij} $:
|
$
$
|
(1) |
$ {c}_{ij} $ is the coupling coefficient trained by iterative dynamic routing, and $ {\sum }_{j}{c}_{ij} $ is designed as a sum. This model does not use the ReLU function, but uses a squashing function to scale the vector between 0 and 1:
| $ v_{j} = \frac{\left\|s_{j}\right\|^{2}}{1+\left\|s_{j}\right\|^{2}} \frac{s_{j}}{\left\|s_{j}\right\|} $ | (2) |
The squashing function sets the small vector as a zero vector and the long vector as a unit vector:
| $ v_{j} \approx\left\|s_{j}\right\| s_{j} \quad \text { for } s_{j} \text { is short } \\ v_{j} \approx \frac{s_{j}}{\left\|s_{j}\right\|} \text { for } s_{j} \text { is long } $ | (3) |
Dynamic routing is located between the capsule layers, that is, between PrimaryCaps and DigitCaps in the structure diagram. This algorithm describes the calculation method of the connection weight between the $ \ell $th layer and the $ (\ell+1) $th layer capsule, and calculates the correlation score $ {b}_{ij} $ according to the similarity:
| $ b_{i j} \leftarrow \hat{u}_{j \mid i} \cdot v_{j} $ | (4) |
The coupling coefficient $ {c}_{ij} $ is calculated by the softmax function of $ {b}_{ij} $:
| $ c_{i j} = \frac{\exp b_{i j}}{\sum_{k} \exp b_{i k}} $ | (5) |
After multiple iterations and updates, the obtained $ {b}_{ij} $ is more accurate:
| $ b_{i j} \leftarrow b_{i j}+\hat{u}_{j \mid i} \cdot v_{j} $ | (6) |
The 29,278 lung CT images processed by lung region segmentation can be used as the train/test data set of the CapsNet-COVID19 classification model (as shown in Figure 8 and Table 3). In order to train and test the CapsNet-COVID19 classification model, 20,000 images of the data set are used as the train set, and the remaining 9278 images are used as the test set and validationset, and the image size is set to 32 × 32.
| Train Set | Validation Set | Test Set |
| 20,000 (68.3%) | 3428 (11.7%) | 5850 (20%) |
DownLoad:
CSV
Since lung CT images are more difficult to identify and classify than ordinary images, the CapsNet-COVID19 classification model needs to be continuously modified during training to achieve better classification results. Tables 4 and 5 show the setting of the parameters of the CapsNet-COVID19 classification model.
| Name | Filter Size | Number of Filters | CapsNet-COVID19 |
| Conv1 layer | 9 × 9 | 256 | |
| Conv2 layer | 6 × 6 | 64 | |
| Conv2 layer dropout value | 0.7 | ||
| Caps1 layer | 5 × 5 | 16 | |
| Caps2 layer | 32 | ||
| Learning rate | 0.0001 | ||
| Number of routes | 2 |
DownLoad:
CSV
| Name | CapsNet-COVID19 |
| Input image pixel size | 32 × 32 |
| Transformation matrix $ {W}_{ij} $ | 16 × 32 |
| PrimaryCaps layer capsule dimension | 16-D |
| Number of PrimaryCaps layer capsules | 32 |
| DigitCaps layer capsule shape | 32 × 3 |
| Number of categories | 3 |
| Does the image to be classified have segmentation preprocessing | Segmentation preprocessing |
DownLoad:
CSV
The classification process of the CapsNet-COVID19 classification model is as follows:
(1) Input the segmented lung CT image training data set into the CapsNet-COVID19 classification model. The image set enters Conv1 first, passes through 256 9 × 9 kernel convolutions, and outputs feature maps of 256 channels (Figure 9);
(2) Then the data enters the PrimaryCaps layer. After the kernel is 5 × 5 convolution, the image size is reduced from 24 × 24 to 10 × 10. The output shape of the PrimaryCaps layer is 10 × 10 × 16 × 16;
(3) Then the data enters the DigiCaps layer, and the 16-dimensional capsule is converted into a 32-dimensional capsule using the 16 × 32 transformation matrix $ {W}_{ij} $. The output of the DigiCaps layer is 3 × 32, which means 3 types of 32-dimensional capsules;
(4) Finally, the data is input to the squashing nonlinear function, and the vector $ {v}_{j} $. is output. Take each vector $ {v}_{j} $. as a capsule of class j, where the length of $ {v}_{j} $. represents the probability of the image being recognized as each class. Finally, after the fully connected FC layer, a 32 × 32 image can be reconstructed. The reconstructed image is shown in Figure 10.
The hardware and software configuration used in the lung CT image classification experiment is as follows:
(1) Hard disk: 16 GB RAM, 1 TB + 256 GB;
(2) Graphics processor: NVIDIA Geforce GTX 1070 8 G;
(3) Central processing unit: Intel Core i7-6700HQ processor;
(4) Operating system: Ubuntu 16.04 operating system;
(5) Software: Python, Tensorflow, Cuda8.0, NumPy, docopt, Sklearn, Matplotlib.
The training of the CapsNet-COVID19 classification model on the GPU takes about 2 hours and 24 minutes. The confusion matrix can be used to evaluate the classification effect of the CapsNet-COVID19 classification model, as shown in Figure 11 and Table 6. Among them, True Positive (TP): predict the number of positive cases predicted as positive; False Negative (FN): predict the number of positive cases predicted as negative; False Positive (FP): predict the number of negative cases predicted as positive; True Negative (TN): predict the number of negative cases as negative cases. In a predictive classification model, it is usually hoped that the more accurate the predictive result, the better. That is, the larger the observed and predicted values of TP and TN, the better, on the contrary, the smaller the observed values of FP and FN, the better.
| COVID-19 | CP | Normal | |
| COVID-19 | 0.956 | 0.0313 | 0.013 |
| CP | 0.187 | 0.729 | 0.084 |
| Normal | 0.114 | 0.043 | 0.844 |
DownLoad:
CSV
In addition, the evaluation indicators of the CapsNet-COVID19 classification model also include accuracy, boundary loss, reconstruction loss, and total loss. Among them, the accuracy rate refers to the ratio of the correct recognition of lung CT image categories to the total number of lung CT images:
| $ \text { Accuracy } = \frac{\sum \text { correctly identified images }}{ { Total\; number \;of\; images }} $ | (7) |
Margin Loss: For each category $ c $ that appears in the lung CT image, the capsule uses a separate loss function $L_{c}$:
| $ L_{c} = T_{c} \max \left(0, m^{+}-\left\|v_{c}\right\|\right)^{2}+\lambda\left(1-T_{c}\right) \max \left(0, \left\|v_{c}\right\|-m^{-}\right)^{2} $ | (8) |
Among them, if there are $T_{c} = 1, \quad m^{+} = 0.9$, and $m^{-} = 0.1$ for the object of the class, the downward weighting of λ (the default is 0.5) prevents the initial learning and shrinks the activity vectors of all classes. The length of the instantiation output vector represents the probability of the existence of each capsule entity. Only when there is a lung CT image type in the input image, this type $ c $ has the longest vector output.
Reconstruction Loss: It is mainly used to promote the encoding of the input images by the DigitCaps layer:
| $ \text{R} = \operatorname{mean}\left((\text { decoded images }-\text { origin images })^{2}\right) $ | (9) |
Among them, R represents reconstruction loss, decoded images represents the encoded images, and origin images represents the original images.
Total Loss is the sum of boundary loss and reconstruction loss:
| $ \mathrm{F} = (\text { Margin Loss })+\lambda(\text { Reconstruction Loss }) $ | (10) |
Among them, F is the total loss, λ = 0.392, therefore, the boundary loss should always dominate the reconstruction loss. If the reconstruction loss is greater than the total loss, the model will try to accurately match the output image with the input image of the train data set, resulting in an overfitting of the model and the train data.
Due to the large noise interference of the lung CT image data set used in this research, the feature difference between lung CT image categories is not obvious enough, and the image recognition is low. Therefore, when the lung CT image classification model is initially trained, the model is prone to overfitting during training.
The reasons for the over fitting of the model are as follows: (1) the amount of data in the training set is not large enough, and the training set cannot evaluate the whole data distribution, resulting in the over fitting of the model; (2) The classification model is over iterative trained in the training process, which leads to the model fitting the noise in the data and the non representative characteristics in the samples, which also leads to the over fitting of the training model; (3) The noise interference of image data set is large, the feature difference between lung CT image categories is not obvious, and the image recognition degree is low, which makes the medical image more difficult to recognize than ordinary images, resulting in over fitting of the model.
In order to avoid or reduce the over fitting phenomenon of CapsNet-COVID19 classification model, the main methods adopted in this study are as follows: first, we widely collect open data sets to ensure that the number of images in the training set is as large as possible, so as to prevent the over fitting problem. Second, in the process of training CapsNet-COVID19 classification model, when the number of iterations is set to be large and the performance of the model in the validation set/test set begins to decline, early stop measures shall be implemented to prevent over fitting caused by continuous training. At the same time, save a copy of the network model corresponding to the current number of iterations and use it as the final/optimal network model.
As shown in Figure 12, during the training process of the CapsNet-COVID19 classification model, the accuracy of the train set and the test set continue to increase as the number of iterations increases. Among them, the red polyline represents the change of the accuracy of the train set, and the blue polyline represents the change of the accuracy of the test set. As the number of iterations increased, the accuracy of the train set approached 100%. In the later stage, the accuracy of the test set fluctuated between 80% and 98%, and the highest value reached 98% accuracy at 85k step. At the same time, as the number of iterations increases, the loss value gradually decreases and tends to stabilize. At the 130 k step position, an early stop was adopted to prevent over-fitting. It should be noted that the CapsNet-COVID19 classification model gives a separate boundary loss for each capsule of the characterization category, and the total loss is the sum of the losses of all categories. The overall evaluation index of the CapsNet-COVID19 classification model is: the accuracy of the train set is 100%, the accuracy of the test set is 84.291%, and the total loss value is 11.066%. It can be seen that the CapsNet-COVID19 classification model is a robust model (Table 7).
| Method | Loss | Train accuracy | Test accuracy |
| CapsNet-COVID19 | 11.066% | 100% | 84.291% |
DownLoad:
CSV
We compare the performance of the CapsNet-COVID19 classification model with several better lung CT image classification methods, as shown in Table 8. The CapsNet-COVID19 classification model has a clear advantage in accuracy over the other models mentioned in the table. The number of images that the CapsNet-COVID19 classification model can process at the same time also has greater advantages than other methods. Our model can meet the requirements of massive lung CT image analysis and improve the efficiency of image analysis.
| Method | Type | Number | Accuracy |
| CapsNet-COVID19 Model | Chest CT | COVID-19: 9420 CP: 9858 Normal: 10,000 |
84.291% (Test set) 100% (Train set) |
| 169-layer DenseNet [22] | Chest CT | COVID-19: 449 CP: 495 Normal: 425 |
83.3% |
| InceptionV3 [22] | Chest CT | COVID-19: 449 CP:495 Normal: 425 |
82.7% |
| ResNet50 [23] | Chest CT | COVID-19: 349 Non-COVID-19: 463 |
77.4% |
| M-Inception [24] | Chest CT | COVID-19: 44 Normal: 55 |
73.1% |
DownLoad:
CSV
Using CapsNet-COVID19 classification model can achieve classification prediction. Input the lung CT image into the CapsNet-COVID19 classification model, and then the type and Softmax value of the image can be predicted. As shown in Figure 13. The classification method of this research is suitable for complicated medical image classification problems such as complex background, low recognition degree, blurred boundary and large image noise.
In this study, a lung CT image classification method based on improved deep learning model is proposed. The main work and results are as follows: firstly, the SegNet-Lung segmentation model is used to segment and predict 29,278 lung CT images. Among them, SegNet-Lung segmentation model uses the structure of nconvolution (convolution2d + batchnormalization + ELU) instead of maxpool, which can speed up the training and convergence of the network, control the gradient explosion, prevent the gradient from disappearing and prevent over fitting. Secondly, the CapsNet-COVID19 classification model is used to classify and predict 29,278 segmented and preprocessed lung CT images with high accuracy. Among them, the CapsNet-COVID19 classification model improves the capsule shape of the primarycaps layer and digicaps layer in the capsule classification model. Thirdly, the evaluation method of classification model is proposed, including accuracy, boundary loss, reconstruction loss, total loss and other indicators, and the over fitting problem is solved by using early stop technology and increasing training data set. The accuracy of the test set of the classification model proposed in this study is higher than that of 169 layer densenet, inception V3, resnet50, m-inception and other methods, and once the model is trained, it can realize the rapid prediction of lung CT test images. The classification model improves the analysis efficiency of lung CT images and can meet the requirements of massive image analysis.
We declare that we do not have any commercial or associative interest that represents a conflict of interest in connection with the work submitted.
| [1] |
Padayatty SJ, Katz A, Wang Y, et al. (2003) Vitamin C as an antioxidant: evaluation of its role in disease prevention. J Am Coll Nutr 22: 18-35. doi: 10.1080/07315724.2003.10719272
|
| [2] | Peterkofsky B (1991) Ascorbate requirement for hydroxylation and secretion of procollagen: relationship to inhibition of collagen synthesis in scurvy. Am J Clin Nutr 54: 1135S-1140S. |
| [3] | McCormick WJ (1959) Cancer: a collagen disease, secondary to a nutritional deficiency. Arch Pediatr 76: 166-171. |
| [4] | Cameron E, Rotman D (1972) Ascorbic acid, cell proliferation, and cancer. Lancet 1: 542. |
| [5] |
Cameron E, Pauling L (1976) Supplemental ascorbate in the supportive treatment of cancer: prolongation of survival times in terminal human cancer. Proc Natl Acad Sci USA 73: 3685-3689. doi: 10.1073/pnas.73.10.3685
|
| [6] | Jacobs C, Hutton B, Ng T, et al. (2015) Is there a role for oral or intravenous ascorbate (vitamin C) in treating patients with cancer? A systematic review. The Oncologist 20: 210-223. |
| [7] |
Levine M, Espey MG, Chen Q (2009) Losing and finding a way at C: new promise for pharmacologic ascorbate in cancer treatment. Free Radic Biol Med 47: 27-29. doi: 10.1016/j.freeradbiomed.2009.04.001
|
| [8] |
Parrow NL, Leshin JA, Levine M (2013) Parenteral ascorbate as a cancer therapeutic: a reassessment based on pharmacokinetics. Antioxidants Redox Signal 19: 2141-2156. doi: 10.1089/ars.2013.5372
|
| [9] | Chen Q, Polireddy K, Chen P, et al. (2015) The unpaved journey of vitamin C in cancer treatment. Can J Physiol Pharmacol 27: 1-9. |
| [10] | Rebouche CJ (1991) Ascorbic acid and carnitine biosynthesis. Am J Clin Nutr 54: 1147S-1152S. |
| [11] | Eipper BA, Stoffers DA, Mains RE (1992) The biosynthesis of neuropeptides: peptide alpha-amidation. Annu Rev Neurosci 15: 57-85. |
| [12] |
Venturelli S, Sinnberg TW, Niessner H, et al. (2015) Molecular mechanisms of pharmacological doses of ascorbate on cancer cells. Wien Med Wochenschr 165: 251-257. doi: 10.1007/s10354-015-0356-7
|
| [13] | Ranzato E, Biffo S, Burlando B (2011) Selective ascorbate toxicity in malignant mesothelioma: a redox Trojan mechanism. Am J Respir Cell Mol Biol 44: 108-117. |
| [14] | Gonzalez MJ, Miranda-Massari JR (2014) New insights on vitamin C and cancer. Springer Brief Cancer Res. |
| [15] |
Oreopoulos DG, Lindeman RD, VanderJagt DJ, et al. (1993) Renal excretion of ascorbic acid: effect of age and sex. J Am Coll Nutr 12: 537-542. doi: 10.1080/07315724.1993.10718349
|
| [16] |
Stephenson CM, Levin RD, Spector T, et al. (2013) Phase I clinical trial to evaluate the safety, tolerability, and pharmacokinetics of high-dose intravenous ascorbic acid in patients with advanced cancer. Cancer Chemoth Pharmacol 72: 139-146. doi: 10.1007/s00280-013-2179-9
|
| [17] | Murata A, Morishige F, Yamaguchi H (1982) Prolongation of survival times of terminal cancer patients by administration of large doses of ascorbate. Int J Vitam Nutr Res Suppl 23:103-113. |
| [18] | Padayatty SJ, Sun H, Wang Y, et al. (2004) Vitamin C pharmacokinetics: implications for oral and intravenous use. Ann Intern Med 140: 533-537. |
| [19] |
Fritz H, Flower G, Weeks L, et al. (2014) Intravenous vitamin C and cancer: A systematic review. Integr Cancer Ther 13: 280-300. doi: 10.1177/1534735414534463
|
| [20] | McCarty MF, Contreras F (2014) Increasing superoxide production and the labile iron pool in tumor cells may sensitize them to extracellular ascorbate. Frontiers Oncology 4: 249. |
| [21] | Chen Q, Espey MG, Krishna MC, et al. (2005) Pharmacologic ascorbic acid concentrations selectively kill cancer cells: action as a pro-drug to deliver hydrogen peroxide to tissues. Proc Natl Acad Sci USA 102: 13604-13910. |
| [22] | Knowles HJ, Raval RR, Harris AL,et al. (2003) Effect of ascorbate on the activity of hypoxia-inducible factor in cancer cells. Cancer Res 63:1764-1768. |
| [23] |
Uetaki M, Tabata S, Nakasuka F, et al. (2015) Metabolomic alterations in human cancer cells by vitamin C-induced oxidative stress. Sci Report 5: 13896. doi: 10.1038/srep13896
|
| [24] |
Sinnberg T, Noor S, Venturelli S, et al. (2014) The ROS-induced cytotoxicity of ascorbate is attenuated by hypoxia and HIF-1alpha in the NCI60 cancer cell lines. J Cell Mol Med 18: 530-541. doi: 10.1111/jcmm.12207
|
| [25] |
Mastrangelo D, Massai L, Lo Cocco F, et al. (2015) Cytotoxic effects of high concentrations of sodium ascorbate on human myeloid cell lines. Ann Hematol 94: 1807-1816. doi: 10.1007/s00277-015-2464-2
|
| [26] | Pichler V, Göschi S, Schreiber-Brynzak E, et al. (2015) Influence of reducing agents on the cytotoxic activity of platinum (IV) complexes: induction of G2/M arrest, apoptosis and oxidative stress in A2780 and cisplatin resistant A2780cis cell lines. Metallomics 7: 1078-1090. |
| [27] |
Chtourou Y, Aouey B, Kebieche M, et al. (2015) Protective role of naringin against cisplatin induced oxidative stress, inflammatory response and apoptosis in rat striatum via suppressing ROS-mediated NF-kB and P53 signalling pathways. Chem Biol Interact 239: 76-86. doi: 10.1016/j.cbi.2015.06.036
|
| [28] | Gerhardt T, Jones R, Park J, et al. (2015) Effects of antioxidants and pro-oxidants on cytotoxicity of dihydroartemisinin to Molt-4 human leukemia cells. Anticancer Res 35: 1867-1871. |
| [29] |
Harada S, Ehara S, Ishi K, et al. (2015) Targeted concurrent chemotherapy, by using improved microcapsules that release carboplatin in response to radiation, improves detectability by computed tomography as well as antitumor activity while reducing adverse effect in vivo. Biomed Pharmacother 70: 196-205. doi: 10.1016/j.biopha.2015.01.006
|
| [30] | Cortés-Jofré M, Rueda JR, Corsini-Muñoz G, et al. (2012) Drugs for preventing lung cancer in healthy people. Cochrane Database Syst Rev 10: CD002141. |
| [31] | Luo J, Shen L, Zheng D (2014) Association between vitamin C intake and lung cancer: a dose-response meta-analysis. Sci Rep 4: 6161. |
| [32] |
Stratton J, Godwin M (2011) The effect of supplemental vitamins and minerals on the development of prostate cancer: a systematic review and meta-analysis. Fam Pract 28: 243-252. doi: 10.1093/fampra/cmq115
|
| [33] | Xu X, Yu E, Liu L, et al. (2013) Dietary intake of vitamins A, C, and E and the risk of colorectal adenoma: a meta-analysis of observational studies. Eur J Cancer Prev 22: 529-539. |
| [34] |
Papaioannou D, Cooper KL, Carroll C, et al. (2011) Antioxidants in the chemoprevention of colorectal cancer and colorectal adenomas in the general population: a systematic review and meta-analysis. Colorectal Dis 13: 1085-1099. doi: 10.1111/j.1463-1318.2010.02289.x
|
| [35] |
Fulan H, Changxing J, Baina WY, et al. (2011) Retinol, vitamins A, C, and E and breast cancer risk: a meta-analysis and meta-regression. Cancer Causes Control 22: 1383-1396. doi: 10.1007/s10552-011-9811-y
|
| [36] |
Fan H, Kou J, Han D, et al. (2015) Association between vitamin C intake and the risk of pancreatic cancer: a meta-analysis of observational studies. Sci Rep 5:13973. doi: 10.1038/srep13973
|
| [37] |
Bo Y, Lu Y, Zhao Y, et al (2015). Association between dietary vitamin C intake and risk of esophageal cancer: a dose-response meta-analysis. Int J Cancer. doi: 10.1002/ijc.29838. doi: 10.1002/ijc.29838
|
| [38] | Creagan ET, Moertel CG, O'Fallon JR, et al. (1979) Failure of high-dose vitamin C (ascorbic acid) therapy to benefit patients with advanced cancer. A controlled trial. N Engl J Med 301: 687-690. |
| [39] | Hoffer LJ (2001) Proof versus plausibility: rules of engagement for the struggle to evaluate alternative cancer therapies. CMAJ 164: 351-353. |
| [40] | Wilson MK, Baguley BC, Wall C, et al. (2014) Review of high-dose intravenous vitamin C as an anticancer agent. Asia Pac J Clin Oncol 10: 22-37. |
| [41] |
Itoh H, Ikeda S, Oohata Y, et al. (1988) Treatment of desmoid tumors in Gardner’s syndrome. Report of a case. Dis Colon Rectum 31: 459-461. doi: 10.1007/BF02552617
|
| [42] |
Berenson JR, Yellin O, Woytowitz D, et al. (2009) Bortezomib, ascorbic acid and melphalan (BAM) therapy for patients with newly diagnosed multiple myeloma: an effective and well-tolerated frontline regimen. Eur J Haematol 82: 433-439. doi: 10.1111/j.1600-0609.2009.01244.x
|
| [43] | Seo MS, Kim JK, Shim JY (2015) High-dose vitamin C promotes regression of multiple pulmonary metastases originating from hepatocellular carcinoma.Yonsei Med J 56: 1449-1452. |
| [44] |
Moertel CG, Fleming TR, Creagan ET, et al. (1985) High-dose vitamin C versus placebo in the treatment of patients with advanced cancer who have had no prior chemotherapy: a randomized double-blind comparison. N Engl J Med 312:137-141. doi: 10.1056/NEJM198501173120301
|
| [45] | Goel SAS, Mandal A, Singhal K, et al. (1999) Emerging role of ascorbic acid in the management of advanced breast carcinoma as a chemosensitizer. Asian J Surg 22: 333-336. |
| [46] | Li X, Sun WJ (2015) The clinical activity of arsenic trioxide, ascorbic acid, ifosfamide and prednisone combination therapy I patients with relapsed and refractory multiple myeloma. Onco Targets Ther 8: 775-781. |
| [47] | Ma Y, Chapman J, Levine M, et al. (2014) High-dose parenteral ascorbate enhanced chemosensitivity of ovarian cancer and reduced toxicity of chemotherapy. Sci Transl Med 6: 222-218. |
| [48] |
Harris HR, Orsini N, Wolk A (2014) Vitamin C and survival among women with breast cancer: a meta-analysis. Eur J Cancer 50: 1223-1231. doi: 10.1016/j.ejca.2014.02.013
|
| [49] |
Welsh JL, Wagner BA, van't Erve TJ, et al. (2013) Pharmacological ascorbate with gemcitabine for the control of metastatic and node-positive pancreatic cancer (PACMAN): results from a phase I clinical trial. Cancer Chemother Pharmacol 71: 765-775. doi: 10.1007/s00280-013-2070-8
|
| [50] | Cameron E, Campbell A (1974) The orthomolecular treatment of cancer II clinical trial of high-dose ascorbic acid supplements in advanced human cancer. Chem Biol Interact 9: 285-315. |
| [51] | Yeom CH, Jung GC, Song KJ (2007) Changes of terminal cancer patients’ health-related quality of life after high dose vitamin C administration. J Korean Med Sci 22: 7-11. |
| [52] | Vollbracht C, Schneider B, Leendert V, et al. (2011) Intravenous vitamin C administration improves quality of life in breast cancer patients during chemo-radiotherapy and aftercare: results of a retrospective, multicentre, epidemiological cohort study in Germany. In Vivo 25: 983-990. |
| [53] |
Günes-Bayir A, Kiziltan HS (2015) Palliative vitamin C application in patients with radiotherapy-resistant bone metastases: a retrospective study. Nutr Cancer 67: 921-925. doi: 10.1080/01635581.2015.1055366
|
| [54] | Carr AC, Vissers MCM, Cook JS (2014) The effect of intravenous vitamin C on cancer- and chemotherapy-related fatigue and quality of life. Frontiers Oncology 4: 283. |
| [55] |
Riordan NH, Riordan HD, Meng X, et al. (1995) Intravenous ascorbate as a tumor cytotoxic chemotherapeutic agent. Med Hypotheses 44: 207-213. doi: 10.1016/0306-9877(95)90137-X
|
| [56] | Molassiotis A, Fernadez-Ortega P, Pud D, et al. (2005) Use of complementary and alternative medicine in cancer patients: a European survey. Ann Oncol 16: 655-663. |
| [57] | Padayatty SJ, Sun AY, et al. (2010) Vitamin C: Intravenous use by complementary and alternative medicine practitioners and adverse effects. PLoS One 5: e11414. |
| [58] | Bjørklund G (2015) The Adjuvant Nutritional Intervention in Cancer (ANICA) Trial. Nutr Cancer 16:1-4. |
| [59] |
Lawton JM, Conway LT, Crosson JT, et al. (1985) Acute oxalate nephropathy after massive ascorbic acid administration. Arch Intern Med 145: 950-951. doi: 10.1001/archinte.1985.00360050220044
|
| [60] | McAllister CJ, Scowden EB, Dewberry FL, et al. (1984) Renal failure secondary to massive infusion of vitamin C. JAMA 252: 1684. |
| [61] | Rees DC, Kelsey H, Richards JD (1993) Acute haemolysis induced by high dose ascorbic acid in glucose-6- phosphate dehydrogenase deficiency. BMJ 306: 841-842. |
| [62] | Hoffer LJ, Levine M, Assouline S, et al. (2008) Phase I clinical trial of I.V. ascorbic acid in advanced malignancy. Ann Oncol 19: 1969-1974. |
| [63] |
Monti DA, Mitchell E, Bazzan AJ, et al. (2012) Phase I evaluation of intravenous ascorbic acid in combination with gemcitabine and erlotinib in patients with metastatic pancreatic cancer. PLoS One 7: e29794. doi: 10.1371/journal.pone.0029794
|
| [64] |
Nielsen TK, Højgaard M, Andersen JT, et al (2015) Elimination of ascorbic acid after high-dose infusion in prostate cancer patients: a pharmacokinetic evaluation. Basic Clin Pharmacol Toxicol 116: 343-348. doi: 10.1111/bcpt.12323
|
| [65] |
Sterne JA, Gavaghan D, Egger M (2000) Publication and related bias in meta-analysis: power of statistical tests and prevalence in the literature. J Clin Epidemiol 53: 1119-1129. doi: 10.1016/S0895-4356(00)00242-0
|
| 1. | Jingyao Liu, Qinghe Feng, Yu Miao, Wei He, Weili Shi, Zhengang Jiang, COVID-19 disease identification network based on weakly supervised feature selection, 2023, 20, 1551-0018, 9327, 10.3934/mbe.2023409 | |
| 2. | Zhan He, Chunju Zhang, Shu Wang, Jianwei Huang, Xiaoyun Zheng, Weijie Jiang, Jiachen Bo, Yucheng Yang, A Grad-CAM and capsule network hybrid method for remote sensing image scene classification, 2024, 18, 2095-0195, 538, 10.1007/s11707-022-1079-x | |
| 3. | Wafa Gtifa, Marwa Fradi, Anis Sakly, Mohsen Machhout, Refining COVID‐19 Lesion Segmentation in Lung CT Scans Using Swarm Intelligence and Evolutionary Algorithms, 2025, 21, 1478-5951, 10.1002/rcs.70044 | |
| 4. | A. V. P. Sarvari, K. Sridevi, Optimized Residual Attention Based Generalized Adversarial Network for COVID‐19 Classification Using Chest CT Images, 2025, 41, 0824-7935, 10.1111/coin.70031 |
Annekathryn Goodman. Vitamin C and Cancer[J]. AIMS Medical Science, 2016, 3(1): 41-51. doi: 10.3934/medsci.2016.1.41
| Training/Validation Data | Test Set | Total (Labeled Images) |
| 2279 (90%) | 253 (10%) | 2532 |
DownLoad:
CSV
| Method | Dice | Precision | Recall |
| COVSeg-NET | 55.5% | 83.3% | 44.5% |
| Focal-Tversky Unet [20] | 48.4% | 48.3% | 61.9% |
| nn-Unet [21] | 4.5% | 75.4% | 48.6% |
DownLoad:
CSV
| Train Set | Validation Set | Test Set |
| 20,000 (68.3%) | 3428 (11.7%) | 5850 (20%) |
DownLoad:
CSV
| Name | Filter Size | Number of Filters | CapsNet-COVID19 |
| Conv1 layer | 9 × 9 | 256 | |
| Conv2 layer | 6 × 6 | 64 | |
| Conv2 layer dropout value | 0.7 | ||
| Caps1 layer | 5 × 5 | 16 | |
| Caps2 layer | 32 | ||
| Learning rate | 0.0001 | ||
| Number of routes | 2 |
DownLoad:
CSV
| Name | CapsNet-COVID19 |
| Input image pixel size | 32 × 32 |
| Transformation matrix $ {W}_{ij} $ | 16 × 32 |
| PrimaryCaps layer capsule dimension | 16-D |
| Number of PrimaryCaps layer capsules | 32 |
| DigitCaps layer capsule shape | 32 × 3 |
| Number of categories | 3 |
| Does the image to be classified have segmentation preprocessing | Segmentation preprocessing |
DownLoad:
CSV
| COVID-19 | CP | Normal | |
| COVID-19 | 0.956 | 0.0313 | 0.013 |
| CP | 0.187 | 0.729 | 0.084 |
| Normal | 0.114 | 0.043 | 0.844 |
DownLoad:
CSV
| Method | Loss | Train accuracy | Test accuracy |
| CapsNet-COVID19 | 11.066% | 100% | 84.291% |
DownLoad:
CSV
| Method | Type | Number | Accuracy |
| CapsNet-COVID19 Model | Chest CT | COVID-19: 9420 CP: 9858 Normal: 10,000 |
84.291% (Test set) 100% (Train set) |
| 169-layer DenseNet [22] | Chest CT | COVID-19: 449 CP: 495 Normal: 425 |
83.3% |
| InceptionV3 [22] | Chest CT | COVID-19: 449 CP:495 Normal: 425 |
82.7% |
| ResNet50 [23] | Chest CT | COVID-19: 349 Non-COVID-19: 463 |
77.4% |
| M-Inception [24] | Chest CT | COVID-19: 44 Normal: 55 |
73.1% |
DownLoad:
CSV
| Training/Validation Data | Test Set | Total (Labeled Images) |
| 2279 (90%) | 253 (10%) | 2532 |
| Method | Dice | Precision | Recall |
| COVSeg-NET | 55.5% | 83.3% | 44.5% |
| Focal-Tversky Unet [20] | 48.4% | 48.3% | 61.9% |
| nn-Unet [21] | 4.5% | 75.4% | 48.6% |
| Train Set | Validation Set | Test Set |
| 20,000 (68.3%) | 3428 (11.7%) | 5850 (20%) |
| Name | Filter Size | Number of Filters | CapsNet-COVID19 |
| Conv1 layer | 9 × 9 | 256 | |
| Conv2 layer | 6 × 6 | 64 | |
| Conv2 layer dropout value | 0.7 | ||
| Caps1 layer | 5 × 5 | 16 | |
| Caps2 layer | 32 | ||
| Learning rate | 0.0001 | ||
| Number of routes | 2 |
| Name | CapsNet-COVID19 |
| Input image pixel size | 32 × 32 |
| Transformation matrix $ {W}_{ij} $ | 16 × 32 |
| PrimaryCaps layer capsule dimension | 16-D |
| Number of PrimaryCaps layer capsules | 32 |
| DigitCaps layer capsule shape | 32 × 3 |
| Number of categories | 3 |
| Does the image to be classified have segmentation preprocessing | Segmentation preprocessing |
| COVID-19 | CP | Normal | |
| COVID-19 | 0.956 | 0.0313 | 0.013 |
| CP | 0.187 | 0.729 | 0.084 |
| Normal | 0.114 | 0.043 | 0.844 |
| Method | Loss | Train accuracy | Test accuracy |
| CapsNet-COVID19 | 11.066% | 100% | 84.291% |
| Method | Type | Number | Accuracy |
| CapsNet-COVID19 Model | Chest CT | COVID-19: 9420 CP: 9858 Normal: 10,000 |
84.291% (Test set) 100% (Train set) |
| 169-layer DenseNet [22] | Chest CT | COVID-19: 449 CP: 495 Normal: 425 |
83.3% |
| InceptionV3 [22] | Chest CT | COVID-19: 449 CP:495 Normal: 425 |
82.7% |
| ResNet50 [23] | Chest CT | COVID-19: 349 Non-COVID-19: 463 |
77.4% |
| M-Inception [24] | Chest CT | COVID-19: 44 Normal: 55 |
73.1% |
