Background:
Fully automated and volumetric segmentation of critical tumors may play a crucial role in diagnosis and surgical planning. One of the most challenging tumor segmentation tasks is localization of Pancreatic Ductal Adenocarcinoma (PDAC). Exclusive application of conventional methods does not appear promising. Deep learning approaches have achieved great success in the computer aided diagnosis, especially in biomedical image segmentation.
Methods:
In this retrospective study, multi slice CT scans of 157 cases with pathologically proven adenocarcinoma of pancreas were enrolled. CT scans were acquired prior to obtaining tissue sample by percutaneous core needle biopsy or fine needle aspiration using endosonography. We introduced a framework based on convolutional neural network (CNN) for segmentation of PDAC mass and surrounding vessels in CT images by incorporating powerful classic features, as well. Segmentation of PDAC mass in the obtained slices was subsequently performed using 2D attention U-Net and Texture Attention U-Net (TAU-Net). TAU-Net was introduced by fusion of dense Scale-Invariant Feature Transform (SIFT) and LBP descriptors into the attention U-Net. Then, an ensemble model was used to cumulate the advantages of both networks using a 3D-CNN. Due to insufficient sample size for vessel segmentation, we used the above-mentioned pre-trained networks and fin-tuned them.
Results:
Experimental results show that the proposed method improves the Dice score for PDAC mass segmentation in portal-venous phase by 7.52%. This algorithm showed a significant different (p = 0.007) compared to state-of-the-art methods. Fully automated segmentation of surrounding vessels around the PDAC are carried out using small dataset. The results are shown that pertained networks on PDAC segmentation can be effective in vessel segmentation.
Conclusion:
Three dimensional visualization of the tumor and surrounding vessels can facilitate decision making, treatment response assessment and surgical planning in PDAC.
Combination of imaging and treatment strategies in one nanosystem has received a great attention for cancer detection and treatment. The aim of this study was to design a nanosystem consisting of dual modality computed tomography / magnetic resonance imaging (CT / MRI) with the ability to load the 5-fluorouracil (5FU) as an anticancer drug.
In this study, a new nanosystem (AMP / 5FU) based on gold nanoparticles (Au) and manganese oxide (MnO) coated with polyethylene glycol (PEG) for imaging and 5FU delivery was prepared in the Faculty of Chemistry, University of Tabriz. Size and shape of nanoparticles using TEM, hydrodynamic size with DLS device, magnetic property using VSM device, cell and blood compatibility with hemolysis test, drug loading and release feature, X-ray attenuation using CT scan and r1 relaxivity of nanoparticles were examined by MRI.
Spherical nanoparticles had paramagnetic properties, an average size of 20 nm and a hydrodynamic size of 78 nm. 5FU non-loaded nanoparticles showed high compatibility for A549 cells for all concentrations, while drug-loaded nanoparticles showed up to 89% toxicity for cancer cells. The results of hemolysis test showed that the nanoparticles have blood compatibility. The drug loading rate of 5FU was 90% and its release rate showed a pH dependence of about 73%. AMP / 5FU nanoparticles showed a considerable X-ray attenuation for CT scan as well as appropriate r1 relaxivity of MRI (r1=4.36 mM−1s−1).
AMP/5FU NPs can be considered as a high potential candidate for bimodal CT/MRI and 5FU anticancer drug delivery.