Wenqing Sun

Redox Regulation of Pancreatic Cancer Cell Growth: Role of Glutathione Peroxidase in the Suppression of the Malignant Phenotype

Pancreatic cancer has low levels of antioxidant enzymes including manganese superoxide dismutase (MnSOD), which converts superoxide radical (O(2)(*-)) into hydrogen peroxide (H(2)O(2)), and glutathione peroxidase (GPx), which converts H(2)O(2) into water. Recent studies have demonstrated that overexpression of MnSOD has a tumor-suppressive effect in pancreatic cancer. However, GPx overexpression has been shown to reverse the tumor cell growth inhibition caused by MnSOD overexpression in other types of cancer. Our aims were to determine if overexpression of GPx alters in vitro pancreatic cancer cell behavior and if delivering the GPx gene directly to tumor xenografts alters growth and survival. In vitro, AdGPx slowed tumor growth by 39% and AdMnSOD slowed tumor growth by 35%. AdGPx also decreased plating efficiency and growth in soft agar. The combination of AdGPx and AdMnSOD had the greatest effect on tumor cell growth suppression with a 71% reduction in cell growth compared to controls. In vivo, either AdGPx or AdMnSOD alone slowed tumor growth by 51% and 54%, respectively, while the combination of AdGPx and AdMnSOD potentiated tumor growth suppression by 81% of controls and increased animal survival. GPx may be a tumor suppressor gene in pancreatic cancer. Delivery of the GPx gene alone or in combination with the MnSOD gene may prove beneficial for treatment of pancreatic cancer.

Optimal Co-Segmentation of Tumor in PET-CT Images With Context Information

Positron emission tomography (PET)-computed tomography (CT) images have been widely used in clinical practice for radiotherapy treatment planning of the radiotherapy. Many existing segmentation approaches only work for a single imaging modality, which suffer from the low spatial resolution in PET or low contrast in CT. In this work, we propose a novel method for the co-segmentation of the tumor in both PET and CT images, which makes use of advantages from each modality: the functionality information from PET and the anatomical structure information from CT. The approach formulates the segmentation problem as a minimization problem of a Markov random field model, which encodes the information from both modalities. The optimization is solved using a graph-cut based method. Two sub-graphs are constructed for the segmentation of the PET and the CT images, respectively. To achieve consistent results in two modalities, an adaptive context cost is enforced by adding context arcs between the two sub-graphs. An optimal solution can be obtained by solving a single maximum flow problem, which leads to simultaneous segmentation of the tumor volumes in both modalities. The proposed algorithm was validated in robust delineation of lung tumors on 23 PET-CT datasets and two head-and-neck cancer subjects. Both qualitative and quantitative results show significant improvement compared to the graph cut methods solely using PET or CT.

Modulation of Reactive Oxygen Species in Pancreatic Cancer

Purpose: The aim of the present study was to compare the effects of the three different forms of the antioxidant enzyme superoxide dismutase [i.e., manganese superoxide dismutase (MnSOD), copper zinc superoxide dismutase (CuZnSOD), and extracellular superoxide dismutase (EcSOD)] on the malignant phenotype of human pancreatic cancer. Experimental Design: Human pancreatic cancer cell lines were infected with adenoviral vectors containing the cDNAs for three different forms of the antioxidant enzyme SOD. Intratumoral injections of the adenoviral vectors were used in nude mice with human tumor xenografts. Results: Increases in immunoreactive protein and enzymatic activity were seen after infections with the AdMnSOD, AdCuZnSOD, or AdEcSOD constructs. Increased SOD activity decreased superoxide levels and increased hydrogen peroxide levels. Increasing SOD levels correlated with increased doubling time. Cell growth and plating efficiency decreased with increasing amounts of the adenoviral constructs, with the AdCuZnSOD vector having the greatest effect in decreasing in vitro tumor growth. In contrast, inhibiting endogenous SOD with small interfering RNA increased superoxide levels and promoted tumor growth. Of the three SODs, tumors grew the slowest and survival was increased the greatest in nude mice injected with the AdEcSOD construct. Conclusions: Scavenging plasma membrane–generated superoxide may prove beneficial for suppression of pancreatic cancer growth.

Enhancing the Antitumor Activity of Adriamycin and Ionizing Radiation

Overexpression of manganese superoxide dismutase (MnSOD), when combined with certain chemicals that inhibit peroxide removal, increases cancer cell cytotoxicity. Elevating MnSOD levels in cells enhances the conversion of superoxide (O(2)(*-)) to hydrogen peroxide (H(2)O(2)), combined with inhibiting the removal of H(2)O(2), further increases H(2)O(2) levels, leading to increased cytotoxicity. We hypothesized that increasing endogenous O(2)(*-) production in cells that were pretreated with adenoviral MnSOD (AdMnSOD) plus 1,3-bis(2-chloroethyl)-1-nitrosourea (BCNU) would lead to an increased level of intracellular H(2)O(2) accumulation and increased cell killing. The cytotoxic effects of Adriamycin or radiation, agents known to produce O(2)(*-), were determined in MDA-MB-231 breast cancer cells pretreated with AdMnSOD plus BCNU both in vitro and in vivo. In vitro, AdMnSOD plus BCNU sensitized cells to the cytotoxicity of Adriamycin or radiation. In vivo, AdMnSOD, BCNU, and Adriamycin or ionizing radiation inhibited tumor growth and prolonged survival. The results suggest that agents that produce O(2)(*-) in combination with AdMnSOD plus BCNU may represent a powerful new antitumor regimen against breast cancer.

Prediction of near-term risk of developing breast cancer using computerized features from bilateral mammograms

Asymmetry of bilateral mammographic tissue density and patterns is a potentially strong indicator of having or developing breast abnormalities or early cancers. The purpose of this study is to design and test the global asymmetry features from bilateral mammograms to predict the near-term risk of women developing detectable high risk breast lesions or cancer in the next sequential screening mammography examination. The image dataset includes mammograms acquired from 90 women who underwent routine screening examinations, all interpreted as negative and not recalled by the radiologists during the original screening procedures. A computerized breast cancer risk analysis scheme using four image processing modules, including image preprocessing, suspicious region segmentation, image feature extraction, and classification was designed to detect and compute image feature asymmetry between the left and right breasts imaged on the mammograms. The highest computed area under curve (AUC) is 0.754±0.024 when applying the new computerized aided diagnosis (CAD) scheme to our testing dataset. The positive predictive value and the negative predictive value were 0.58 and 0.80, respectively.

Enhancing deep convolutional neural network scheme for breast cancer diagnosis with unlabeled data

In this study we developed a graph based semi-supervised learning (SSL) scheme using deep convolutional neural network (CNN) for breast cancer diagnosis. CNN usually needs a large amount of labeled data for training and fine tuning the parameters, and our proposed scheme only requires a small portion of labeled data in training set. Four modules were included in the diagnosis system: data weighing, feature selection, dividing co-training data labeling, and CNN. 3158 region of interests (ROIs) with each containing a mass extracted from 1874 pairs of mammogram images were used for this study. Among them 100 ROIs were treated as labeled data while the rest were treated as unlabeled. The area under the curve (AUC) observed in our study was 0.8818, and the accuracy of CNN is 0.8243 using the mixed labeled and unlabeled data.

Computer aided lung cancer diagnosis with deep learning algorithms

Deep learning is considered as a popular and powerful method in pattern recognition and classification. However, there are not many deep structured applications used in medical imaging diagnosis area, because large dataset is not always available for medical images. In this study we tested the feasibility of using deep learning algorithms for lung cancer diagnosis with the cases from Lung Image Database Consortium (LIDC) database. The nodules on each computed tomography (CT) slice were segmented according to marks provided by the radiologists. After down sampling and rotating we acquired 174412 samples with 52 by 52 pixel each and the corresponding truth files. Three deep learning algorithms were designed and implemented, including Convolutional Neural Network (CNN), Deep Belief Networks (DBNs), Stacked Denoising Autoencoder (SDAE). To compare the performance of deep learning algorithms with traditional computer aided diagnosis (CADx) system, we designed a scheme with 28 image features and support vector machine. The accuracies of CNN, DBNs, and SDAE are 0.7976, 0.8119, and 0.7929, respectively; the accuracy of our designed traditional CADx is 0.7940, which is slightly lower than CNN and DBNs. We also noticed that the mislabeled nodules using DBNs are 4% larger than using traditional CADx, this might be resulting from down sampling process lost some size information of the nodules.

Dosimetric impacts of applicator displacements and applicator reconstruction-uncertainties on 3D image-guided brachytherapy for cervical cancer

Purpose To quantify the dosimetric impact of applicator displacements and applicator reconstruction-uncertainties through simulated planning studies of virtual applicator shifts. Material and methods Twenty randomly selected high-dose-rate (HDR) titanium tandem-and-ovoid (T&O) plans were retrospectively studied. MRI-guided, conformal brachytherapy (MRIG-CBT) plans were retrospectively generated. To simulate T&O displacement, the whole T&O set was virtually shifted on treatment planning system in the cranial (+) and the caudal (–) direction after each dose calculation. Each shifted plan was compared to an unshifted plan. To simulate T&O reconstruction-uncertainties, each tandem and ovoid was separately shifted along its axis before performing the dose calculation. After the dose calculation, the calculated isodose lines and T&O were moved back to unshifted T&O position. Shifted and shifted-back plan were compared. Results Regarding the dosimetric impact of the simulated T&O displacements, rectal D2cc values were observed as being the most sensitive to change due to T&O displacement among all dosimetric metrics regardless of point A (p < 0.013) or MRIG-CBT plans (p < 0.0277). To avoid more than 10% change, ± 1.5 mm T&O displacements were accommodated for both point A and MRIG-CBT plans. The dosimetric impact of T&O displacements on sigmoid (p < 0.0005), bladder (p < 0.0001), HR-CTV (p < 0.0036), and point A (p < 0.0015) were significantly larger in the MRIG-CBT plans than point A plans. Regarding the dosimetric impact of T&O reconstruction-uncertainties, less than ± 3.0 mm reconstruction-uncertainties were also required in order to avoid more than 10% dosimetric change in either the point A or MRIG-CBT plans. Conclusions The dosimetric impact of simulated T&O displacements was significantly larger in the MRIG-CBT plans than in the point A plans. Either ± 3 mm T&O displacement or a ± 4.5 mm T&O reconstruction-uncertainty could cause greater than 10% dosimetric change for both point A plans and MRIG-CBT plans.

Automatic feature learning using multichannel ROI based on deep structured algorithms for computerized lung cancer diagnosis

This study aimed to analyze the ability of extracting automatically generated features using deep structured algorithms in lung nodule CT image diagnosis, and compare its performance with traditional computer aided diagnosis (CADx) systems using hand-crafted features. All of the 1018 cases were acquired from Lung Image Database Consortium (LIDC) public lung cancer database. The nodules were segmented according to four radiologists markings, and 13,668 samples were generated by rotating every slice of nodule images. Three multichannel ROI based deep structured algorithms were designed and implemented in this study: convolutional neural network (CNN), deep belief network (DBN), and stacked denoising autoencoder (SDAE). For the comparison purpose, we also implemented a CADx system using hand-crafted features including density features, texture features and morphological features. The performance of every scheme was evaluated by using a 10-fold cross-validation method and an assessment index of the area under the receiver operating characteristic curve (AUC). The observed highest area under the curve (AUC) was 0.899±0.018 achieved by CNN, which was significantly higher than traditional CADx with the AUC=0.848±0.026. The results from DBN was also slightly higher than CADx, while SDAE was slightly lower. By visualizing the automatic generated features, we found some meaningful detectors like curvy stroke detectors from deep structured schemes. The study results showed the deep structured algorithms with automatically generated features can achieve desirable performance in lung nodule diagnosis. With well-tuned parameters and large enough dataset, the deep learning algorithms can have better performance than current popular CADx. We believe the deep learning algorithms with similar data preprocessing procedure can be used in other medical image analysis areas as well.

Efficacy of nelfinavir as monotherapy in refractory adenoid cystic carcinoma: Results of a phase II clinical trial

Adenoid cystic carcinomas (ACCs) are malignant salivary gland tumors noteworthy for high rates of late failure with limited salvage therapy options. We have previously shown increased Akt signaling is common in ACC and the human immunodeficiency virus (HIV) protease inhibitor nelfinavir (NFV) inhibits in vitro tumor growth by suppressing Akt signaling. This phase II trial was conducted to determine progression‐free survival in response to NFV in patients with recurrent/endstage ACC who have failed standard therapies.