CT Imaging Of Tumors Research Articles

Bi-DTPA, a potential CT-guided radiation sensitizer

The therapeutic effects and application of radiotherapy are restricted to some extent due to low radiosensitivity of tumor tissues and adverse effects by excess dosage. Current radiosensitizers are confronted with problems in clinical translation because of complicated manufacture technique and high cost. In this research, we have synthesized a radiosensitizer with advantages in low cost and mass production, which could be applied to CT imaging and enhanced radiotherapy in breast cancer, namely Bi-DTPA. It not only enhanced tumor CT imaging which resulted in better therapeutic accuracy, but also realized radiotherapy sensitization by producing massive ROS and inhibit tumor proliferation, providing a sound perspective in the clinical translation of the radiosensitizer.

Influence of Data Distribution on Federated Learning Performance in Tumor Segmentation.

To investigate the correlation between differences in data distributions and federated deep learning (Fed-DL) algorithm performance in tumor segmentation on CT and MR images. Two Fed-DL datasets were retrospectively collected (from November 2020 to December 2021): one dataset of liver tumor CT images (Federated Imaging in Liver Tumor Segmentation [or, FILTS]; three sites, 692 scans) and one publicly available dataset of brain tumor MR images (Federated Tumor Segmentation [or, FeTS]; 23 sites, 1251 scans). Scans from both datasets were grouped according to site, tumor type, tumor size, dataset size, and tumor intensity. To quantify differences in data distributions, the following four distance metrics were calculated: earth mover's distance (EMD), Bhattacharyya distance (BD), χ2 distance (CSD), and Kolmogorov-Smirnov distance (KSD). Both federated and centralized nnU-Net models were trained by using the same grouped datasets. Fed-DL model performance was evaluated by using the ratio of Dice coefficients, θ, between federated and centralized models trained and tested on the same 80:20 split datasets. The Dice coefficient ratio (θ) between federated and centralized models was strongly negatively correlated with the distances between data distributions, with correlation coefficients of -0.920 for EMD, -0.893 for BD, and -0.899 for CSD. However, KSD was weakly correlated with θ, with a correlation coefficient of -0.479. Performance of Fed-DL models in tumor segmentation on CT and MRI datasets was strongly negatively correlated with the distances between data distributions.Keywords: CT, Abdomen/GI, Liver, Comparative Studies, MR Imaging, Brain/Brain Stem, Convolutional Neural Network (CNN), Federated Deep Learning, Tumor Segmentation, Data Distribution Supplemental material is available for this article. © RSNA, 2023See also the commentary by Kwak and Bai in this issue.

PD22-07 A MULTICENTER STUDY PRESENTING AN EXPLAINABLE AI APPROACH TO AUTOMATIC SEGMENTATION OF RENAL TUMORS ON COMPUTED TOMOGRAPHY

You have accessJournal of UrologyCME1 Apr 2023PD22-07 A MULTICENTER STUDY PRESENTING AN EXPLAINABLE AI APPROACH TO AUTOMATIC SEGMENTATION OF RENAL TUMORS ON COMPUTED TOMOGRAPHY Annemarie Uhlig, Sophie Bachanek, Paul Romario Würzberg, Manuel Nietert, Lutz Trojan, Joachim Lotz, and Johannes Uhlig Annemarie UhligAnnemarie Uhlig More articles by this author , Sophie BachanekSophie Bachanek More articles by this author , Paul Romario WürzbergPaul Romario Würzberg More articles by this author , Manuel NietertManuel Nietert More articles by this author , Lutz TrojanLutz Trojan More articles by this author , Joachim LotzJoachim Lotz More articles by this author , and Johannes UhligJohannes Uhlig More articles by this author View All Author Informationhttps://doi.org/10.1097/JU.0000000000003295.07AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail Abstract INTRODUCTION AND OBJECTIVE: Automatic segmentation of renal tumors is a cornerstone for Artificial Intelligence (AI) based tumor diagnostics. This study presents an automatic segmentation algorithm and visualization method using CT-studies from renal tumor patients recruited in a multicenter setting. METHODS: Renal tumor patients diagnosed between 2018 and 2021 were retrospectively assessed for this study. Inclusion criteria were CT-imaging of renal tumors in corticomedullary (CM) or nephrogenic (NG) contrast media phase. Patients aged <18years, and those with cystic or infiltrative renal tumors (i.e. lymphomas) were excluded. Manual segmentation of the renal tumors was performed by a GU-radiologist on all axial CT slices. A convolutional neural network (UNET) was trained based on the radiologists´ manual segmentations. In an independent validation dataset, the accuracy of the UNETs predictions of renal tumor contours was compared to the references standard of manual segmentations and quantified using the DICE score. RESULTS: A total of n=394/ n=350 patients with renal tumors imaged in CM /NG phase, respectively, were included (median age 66y; 35% female; median tumor diameter 5.4cm). CT-studies from >20 radiological imaging centers were included with different imaging protocols and slice thickness. The UNET was trained on n=316 CM and n=294 NG contrast phase patients (n=7019 / n=6859 separate CT images). In the independent validation dataset (n=78 / n=56 patients with 1713 / 1298 CT images), the UNET achieved a DICE score of 0.88 and 0.90 for the corticomedullary and nephrogenic CM phase, respectively. The UNET predictions were visualized using a tile-based approach with color-coding and contour-lines that could be overlaid on CT-images to depict varying levels of prediction confidence. CONCLUSIONS: A UNET AI-approach yields a robust automatic delineation of renal tumors on CT-images acquired in clinical routine, irrespective of the contrast media phase. Using color-coding and contour-lines that could be overlaid on original CT-images provides an explainable approach to the UNETs predictions and might improve AI-acceptance in clinical practice. Source of Funding: This study received funding by the 2022 ESR Research Seed Grant. The 2022 ESR Research Seed Grants were kindly supported by an unrestricted, non-exclusive grant from GE Healthcare © 2023 by American Urological Association Education and Research, Inc.FiguresReferencesRelatedDetails Volume 209Issue Supplement 4April 2023Page: e667 Advertisement Copyright & Permissions© 2023 by American Urological Association Education and Research, Inc.MetricsAuthor Information Annemarie Uhlig More articles by this author Sophie Bachanek More articles by this author Paul Romario Würzberg More articles by this author Manuel Nietert More articles by this author Lutz Trojan More articles by this author Joachim Lotz More articles by this author Johannes Uhlig More articles by this author Expand All Advertisement PDF downloadLoading ...

Relevant Risk Factor and Follow-Up of Lung Nodules in Physical Examination with Low-Dose CT Screening.

We aimed to explore the risk factors of lung nodules and lung cancer in physical examination population with low-dose multi-slice spiral CT (LDCT) screening, to provide basis for lung cancer screening and follow-up management after CT examination. The general data, serum tumor markers and CT images of 2,274 patients underwent LDCT in the Physical Examination Center of the Fourth Hospital of Hebei Medical University, China in 2019 were retrospectively analyzed and followed up for three years. The detection rate of lung nodules was 48.42%. The detection rate of lung nodules was higher in females, those over 70, those with history of smoking, passive smoking, drinking, precious history of lung diseases and family history of malignant tumors, with statistically significant differences (P<0.05). The abnormal rate of serum tumor markers (CA199, CA125 female and CYFRA211) were higher than that in the non-nodule group, with statistically significant differences (P<0.05). Multivariate logistic regression analysis showed that gender, age, history of smoking, passive smoking, family history of malignant tumors and serum tumor markers (CYFRA211 and CA199) were independent risk factors for the occurrence of lung nodules. Gender female, age>35, history of smoking, passive smoking, history of drinking, history of past lung disease, family history of malignant tumors, abnormal CYFRA211 tumor markers were detected and low dose multi-slice spiral CT image showed ground-glass nodules are risk factors for lung nodules and lung cancer, which should be paid close attention to during physical examination and follow-up.

A dual branch and fine-grained enhancement network for pancreatic tumor segmentation in contrast enhanced CT images

Segmentation of pancreatic tumors in CT images is important for clinical diagnosis and treatment, but it faces challenges of small size, low contrast, and large position difference. To address these issues, the abnormal pancreas is first segmented based on a dual branch coding network (DB-Net) using a coarse-to-fine segmentation strategy. In the encoder part, one branch extracts the semantic features of the pancreas and its surroundings, and the other branch captures the complex pancreas through wide-channel convolution and few down-sampling operations. An aggregation layer is used to fuse the different feature maps obtained by the two branches, and a U-Net decoder is used to segment the abnormal pancreas in CT images with pancreatic tumors. DB-Net is further trained to obtain the accurate pancreatic segmentation. Then, pancreatic tumors are segmented in the pancreas based on the fine-grained enhancement network (FE-Net). The FE-Net integrates a contrast enhancement block with a reverse attention block to extract detailed features and excavate effective information from the feature maps of the encoder and decoder to segment pancreatic tumors. In order to segment the tumor more accurately, the pancreatic tumor is segmented in the cropped pancreas. Experiments on 116 contrast-enhanced abdominal CT volumes of pancreatic cancer and 42 contrast-enhanced abdominal CT volumes of normal pancreas verify the effectiveness of the proposed framework in pancreatic tumor segmentation by using the two-fold cross-validation strategy. Compared to state-of-the-art deep learning segmentation network, the proposed method can achieve better segmentation of pancreas and pancreatic tumors.

Liver tumor segmentation and classification using FLAS-UNet++ and an improved DenseNet.

The incidence of liver tumors is among the top three in China. The treatments of benign and malignant tumors are different. Accurate diagnosis plays an important role in guiding the treatment of tumors. The aim of this study is to solve the following: (1) blurred boundary between the liver tumor and other organs causes incorrect segmentation of liver tumor boundaries; (2) large difference in tumor size and the diversity in texture and grayscale are major challenges in liver tumor classification tasks. Firstly, the liver tumor is segmented from the original CT images by a tumor segmentation network, UNet++ with fusion loss and atrous spatial pyramid pooling (FLAS-UNet++). The proposed segmentation method can solve the problem of tumor edge segmentation error by learning the tumor edge information. Secondly they are adaptively cropped according to the tumor volume to reduce the over-fitting and over-sensitivity of the deep network. Thirdly an improved Dense Block is designed to pay more attention to the changes in grayscale and texture between benign and malignant tumors. Finally, the features extracted from the network combined with tumor volume, patient's sex and age, are sent to a classifier for diagnosis. Liver tumor segmentation results show that the dice, HD95 reached 71.9%, 12.1 mm, respectively. The classification results show that the accuracy, specificity, sensitivity and area under curve reached 82.4%, 79.8%, 84.4%, 87.5%, respectively. The segmentation and classification results are both better than other's methods and mainstream networks. In order to solve existing problems of liver tumor CT image classification methods, our method realizes the accurate segmentation and classification of liver tumors in CT images and has important clinical application value.

Auto-Diagnosis of Lung Cancer with the Proposed Feature Fusion-Based Hybrid Deep Model

Early detection of lung cancer increases the response rate to treatment. Therefore, the accuracy of diagnostic methods is of great importance. Reading the patient's medical images by radiologists can cause a severe time cost besides subjective result. In this context, Artificial Intelligence (AI) methods create an innovative field to reduce the workforce of radiologists and obtain objective results. AI methods play a vital role in improving the analysis of the dataset, extracting meaningful features, clustering, and classification. In our study, the data set contains healthy images besides CT images of malignant and benign tumors with lung cancer; AlexNet is trained using DenseNet 201, GoogleNet, MobileNetV2, and ResNet50 architectures. In addition, a hybrid model has been developed to classify lung CT images. The developed model constitutively used GoogleNet, MobileNetV2, and ResNet50 architectures. The feature maps obtained in these three architectures were combined and classified into different classifiers. Among the classifiers used in the study, the highest accuracy rate was achieved in the Ensemble Subspace KNN classifier. The accuracy value obtained in this classifier is 98.3%.

Into the future: a pilot study combining imaging with molecular profiling for preoperative prediction of resectability in ovarian cancer (068)

Objectives: We sought to determine the joint predictive value of tumor molecular subtype and CT imaging data in predicting surgical outcomes prior to primary debulking surgery (PDS) in Stage IIIC/IV high-grade serous ovarian cancer patients. Methods: Patients undergoing PDS for Stage IIIC/IV were included. Based on CT imaging, a continuous CT-score indicative of overall disease burden was defined based on six CT measurements of anatomic involvement: (1) diaphragm disease; (2) gastrohepatic/porta hepatis; (3) root of superior mesenteric artery; (4) presence of moderate to severe ascites; (5) intrahepatic lesion, and (6) diffuse peritoneal thickening > 4mm. Patients were then classified as CT-low or CT-high if their score was ≤2 or ≥3, respectively. Molecular subtypes were derived from mRNA profiling of chemo-naïve tumors; women were classified as having mesenchymal (MES) subtype or non-MES subtype tumors. Patients were classified as having no gross residual disease (RD0) or any gross residual disease less than 1 cm (RD1), or gross residual disease greater than 1 cm (RD2). Multivariate regression models were fit to predict RD. The four cohorts, (i) non-MES & CT-low, (ii) non-MES & CT-high, (iii) MES & CT-low, and (iv) MES & CT-high, were compared with Fisher’s exact tests to examine the association with RD and surgical complexity. Graphical Abstract View Large Image Figure Viewer Download Hi-res image Conclusions: Preoperative CT imaging combined with tumor molecular subtyping can identify a subset of women for whom successful primary surgery is unlikely. Preoperative tumor sampling may be useful in advanced OC to better triage these cases to alternative approaches. Objectives: We sought to determine the joint predictive value of tumor molecular subtype and CT imaging data in predicting surgical outcomes prior to primary debulking surgery (PDS) in Stage IIIC/IV high-grade serous ovarian cancer patients. Methods: Patients undergoing PDS for Stage IIIC/IV were included. Based on CT imaging, a continuous CT-score indicative of overall disease burden was defined based on six CT measurements of anatomic involvement: (1) diaphragm disease; (2) gastrohepatic/porta hepatis; (3) root of superior mesenteric artery; (4) presence of moderate to severe ascites; (5) intrahepatic lesion, and (6) diffuse peritoneal thickening > 4mm. Patients were then classified as CT-low or CT-high if their score was ≤2 or ≥3, respectively. Molecular subtypes were derived from mRNA profiling of chemo-naïve tumors; women were classified as having mesenchymal (MES) subtype or non-MES subtype tumors. Patients were classified as having no gross residual disease (RD0) or any gross residual disease less than 1 cm (RD1), or gross residual disease greater than 1 cm (RD2). Multivariate regression models were fit to predict RD. The four cohorts, (i) non-MES & CT-low, (ii) non-MES & CT-high, (iii) MES & CT-low, and (iv) MES & CT-high, were compared with Fisher’s exact tests to examine the association with RD and surgical complexity. Conclusions: Preoperative CT imaging combined with tumor molecular subtyping can identify a subset of women for whom successful primary surgery is unlikely. Preoperative tumor sampling may be useful in advanced OC to better triage these cases to alternative approaches.

Endogenous ROS-Mediated Covalent Immobilization of Gold Nanoparticles in Mitochondria: A "Sharp Sword" in Tumor Radiotherapy.

Mitochondria as one of the key subcellular organelles have been well recognized as a promising druggable target and are closely associated with energy supply and various cellular functions. Realizing high accumulation and prolonged retention of radiosensitizers in the cellular mitochondria of tumors is an effective way to improve radiotherapeutic efficacy. Herein, we develop mitochondria-targeting and protein sulfenic acid (PSA)-reactive gold nanoparticles (dAuNP-TPP) that are fabricated by incorporating triphenylphosphine and 1,3-cyclohexanedione onto the surface of AuNPs (∼20 nm) to improve CT imaging and radiotherapeutic efficacy of tumors. Taking advantage of the specific mitochondrial targeting and PSA-mediated on-site covalent immobilization, this nanosystem shows significantly enhanced accumulation and retention in mitochondria with approximately 5.22-fold higher enrichment than nonimmobilizable AuNP-TPP. More notably, the covalent immobilization of dAuNP-TPP in cellular mitochondria could induce a dramatic reduction of ATP, leading to serious mitochondrial disruption, combined with the radiosensitization effect, in consequence achieving efficient radiotherapy of breast tumors in vivo. This subcellular organelle-targeted and immobilizable strategy may offer a valuable and universal tool for efficient tumor treatment.

Study on the Correlation Between CT Features and Vascular Tumor Thrombus Together With Nerve Invasion in Surgically Resected Lung Adenocarcinoma.

BackgroundWe aimed to analyze the relationship between pulmonary adenocarcinoma patients with vascular tumor thrombus and nerve invasion and different CT features.MethodsThe preoperative CT scanning data of 86 patients with lung adenocarcinoma who underwent surgical resection in our hospital from January 2020 to January 2022 were analyzed in the form of retrospective analysis. The CT images of all patients were observed, and the relationship between them and vascular tumor thrombus and nerve invasion of lung adenocarcinoma was analyzed. At the same time, the sensitivity, specificity, and accuracy of enhanced CT and plain CT were compared to evaluate the diagnostic efficacy of both.ResultsThe results showed that the vascular tumor thrombus of lung adenocarcinoma was mainly related to the solid components and lobulated and calcified tumors in CT images, and the nerve invasion of lung adenocarcinoma was mainly related to the tumors with bronchial inflation sign in CT images (P < 0.05). The sensitivity, specificity, and accuracy of enhanced CT in the diagnosis of vascular tumor thrombus were 78.26%, 96.83%, and 91.86%, respectively, and the sensitivity, specificity, and accuracy in the diagnosis of nerve invasion were 75.00%, 98.72%, and 96.51%, respectively. The sensitivity, specificity, and accuracy of plain CT in the diagnosis of vascular tumor thrombus were 43.48%, 92.06%, and 79.07%, respectively, and the sensitivity, specificity, and accuracy in the diagnosis of nerve invasion were 25.00%, 94.87%, and 88.37%, respectively. The contrast showed that the sensitivity and accuracy of enhanced CT were higher than those of plain CT (P < 0.05), but the difference of specificity was not obvious (P > 0.05).ConclusionsSolid components and lobulated and calcified tumors in CT signs are closely related to vascular tumor thrombus of lung adenocarcinoma, while patients with bronchial inflation sign are related to nerve invasion.

Segmentation of renal tumors in CT images by 3D U-Net preserving rotational symmetry in axial slices

We present a method to improve accuracy of renal tumor segmentation in CT images by considering the structure of kidneys. We apply deep learning based on the following protocol. First, kidney regions are extracted from the CT images. Second, adjacent eight slices along axial direction are picked up as a patch. Third, the center of gravity of the kidney region in the patch is aligned to the center of the patch in sagittal and coronal direction. Fourth, we apply data augmentation with scaling and rotation around the center, which preserves the basic slice structure of kidneys. Finally, these patches are fed to a 3D U-Net for training. Compared with the conventional 3D U-Net, the proposed method improves the DICE score from 0.507 to 0.604.

Two-Stage Liver and Tumor Segmentation Algorithm Based on Convolutional Neural Network.

The liver is an essential metabolic organ of the human body, and malignant liver tumors seriously affect and threaten human life. The segmentation algorithm for liver and liver tumors is one of the essential branches of computer-aided diagnosis. This paper proposed a two-stage liver and tumor segmentation algorithm based on the convolutional neural network (CNN). In the present study, we used two stages to segment the liver and tumors: liver localization and tumor segmentation. In the liver localization stage, the network segments the liver region, adopts the encoding–decoding structure and long-distance feature fusion operation, and utilizes the shallow features’ spatial information to improve liver identification. In the tumor segmentation stage, based on the liver segmentation results of the first two steps, a CNN model was designed to accurately identify the liver tumors by using the 2D image features and 3D spatial features of the CT image slices. At the same time, we use the attention mechanism to improve the segmentation performance of small liver tumors. The proposed algorithm was tested on the public data set Liver Tumor Segmentation Challenge (LiTS). The Dice coefficient of liver segmentation was 0.967, and the Dice coefficient of tumor segmentation was 0.725. The proposed algorithm can accurately segment the liver and liver tumors in CT images. Compared with other state-of-the-art algorithms, the segmentation results of the proposed algorithm rank the highest in the Dice coefficient.

Automatic segmentation of organs at risk and tumors in CT images of lung cancer from partially labelled datasets with a semi-supervised conditional nnU-Net

Background and ObjectiveAccurately and reliably defining organs at risk (OARs) and tumors are the cornerstone of radiation therapy (RT) treatment planning for lung cancer. Almost all segmentation networks based on deep learning techniques rely on fully annotated data with strong supervision. However, existing public imaging datasets encountered in the RT domain frequently include singly labelled tumors or partially labelled organs because annotating full OARs and tumors in CT images is both rigorous and tedious. To utilize labelled data from different sources, we proposed a dual-path semi-supervised conditional nnU-Net for OARs and tumor segmentation that is trained on a union of partially labelled datasets. MethodsThe framework employs the nnU-Net as the base model and introduces a conditioning strategy by incorporating auxiliary information as an additional input layer into the decoder. The conditional nnU-Net efficiently leverages prior conditional information to classify the target class at the pixelwise level. Specifically, we employ the uncertainty-aware mean teacher (UA-MT) framework to assist in OARs segmentation, which can effectively leverage unlabelled data (images from a tumor labelled dataset) by encouraging consistent predictions of the same input under different perturbations. Furthermore, we individually design different combinations of loss functions to optimize the segmentation of OARs (Dice loss and cross-entropy loss) and tumors (Dice loss and focal loss) in a dual path. ResultsThe proposed method is evaluated on two publicly available datasets of the spinal cord, left and right lung, heart, esophagus, and lung tumor, in which satisfactory segmentation performance has been achieved in term of both the region-based Dice similarity coefficient (DSC) and the boundary-based Hausdorff distance (HD). ConclusionsThe proposed semi-supervised conditional nnU-Net breaks down the barriers between nonoverlapping labelled datasets and further alleviates the problem of “data hunger” and “data waste” in multi-class segmentation. The method has the potential to help radiologists with RT treatment planning in clinical practice.

CT Guided Diagnosis: Cascaded U-Net for 3D Segmentation of Liver and Tumor

Volumetric estimation of the liver tumor is the first step to identifying the livers critical disorder. The liver and its tumor ratio prerequisite measures to select the therapeutic procedure. 3D printing and virtual reality platform require a segmented liver entity mask to evaluate the pre and post-treatment analysis. A cascaded U-Net model is proposed for automatic segmentation of liver and tumor in CT images. LiTS CT data set utilized for this study. The images were pre-processed using the windowing technique for contrast enhancement. Two U-Net models were modified for liver and tumor segmentation, respectively and connected in a cascaded manner. U-Net decoder end was modified in comparison to the original U-Net. The probability map of the first U-Net fed to the second U-Net and the input image to segment out the liver tumor. Eight subject volumetric CT datasets were utilized to test the cascaded U-Net performance and achieved average Dice coefficient for liver and tumor 0.95 and 0.69, respectively. Liver tumor diagnosis and treatment accuracy depend upon the precision of segmentation algorithms. Designed model segmented liver almost accurately and tumor segmented with limited accuracy. A further modification is required for the tumor segmentation cause of the occurrence of false negative.

Esophageal Tumor Segmentation in CT Images Using a Dilated Dense Attention Unet (DDAUnet)

Manual or automatic delineation of the esophageal tumor in CT images is known to be very challenging. This is due to the low contrast between the tumor and adjacent tissues, the anatomical variation of the esophagus, as well as the occasional presence of foreign bodies (e.g. feeding tubes). Physicians therefore usually exploit additional knowledge such as endoscopic findings, clinical history, additional imaging modalities like PET scans. Achieving his additional information is time-consuming, while the results are error-prone and might lead to non-deterministic results. In this paper we aim to investigate if and to what extent a simplified clinical workflow based on CT alone, allows one to automatically segment the esophageal tumor with sufficient quality. For this purpose, we present a fully automatic end-to-end esophageal tumor segmentation method based on convolutional neural networks (CNNs). The proposed network, called Dilated Dense Attention Unet (DDAUnet), leverages spatial and channel attention gates in each dense block to selectively concentrate on determinant feature maps and regions. Dilated convolutional layers are used to manage GPU memory and increase the network receptive field. We collected a dataset of 792 scans from 288 distinct patients including varying anatomies with air pockets, feeding tubes and proximal tumors. Repeatability and reproducibility studies were conducted for three distinct splits of training and validation sets. The proposed network achieved a DSC value of 0.79 ± 0.20, a mean surface distance of 5.4 ± 20.2mm and 95% Hausdorff distance of 14.7 ± 25.0mm for 287 test scans, demonstrating promising results with a simplified clinical workflow based on CT alone. Our code is publicly available via <monospace xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><uri>https://github.com/yousefis/DenseUnet_Esophagus_Segmentation</uri></monospace> .

Construction of a Standard Dataset for Liver Tumors for Testing the Performance and Safety of Artificial Intelligence-Based Clinical Decision Support Systems

목적간 종양의 조영증강 컴퓨터단층촬영(이하 CT) 영상에 관한 인공지능 알고리즘의 성능과 안전성을 검증할 수 있는 표준 테스팅 데이터셋을 구축하고자 하였다.대상과 방법국내 4개 3차 의료기관의 복부 영상의학 전문가 4인이 모여 간 종양 진단 알고리즘의 성능과 안전성을 검증하기 위해 표준 데이터셋이 갖춰야 할 조건을 논의하였다. 각 기관마다 간세포암 75예, 전이암 75예, 그리고 양성 병변 30–50예씩 수집하여, 총 783명 환자의 CT 영상을 대상으로 하였다. 간세포암과 전이암의 경우 병리학적으로 확진된 경우만을 대상으로 하였다. 각 기관의 복부 영상의학 전문가들이 직접 환자의 임상정보를 추출하고 CT 영상에 관한 데이터 라벨링(labeling)을 수기로 시행하였다. CT 영상은 의료용 디지털 영상 및 통신(Digital Imaging and Communications in Medicine, DICOM) 파일로 저장하였다.결과복부 영상의학 전문가들이 수기 데이터 라벨링을 시행한 총 783 증례의 간 종양 조영증강 CT의 표준 데이터셋을 구축하였다. 알고리즘의 성능 및 안전성은 병변의 발견 여부 및 특성화의 정확도에 대해 민감도와 특이도를 계산하여 평가할 수 있다.결론본 연구에서 구축한 간 종양 조영증강 CT 영상의 표준 데이터셋은 임상의학 결정 지원 시스템을 위한 기계학습 기반 인공지능 알고리즘을 평가하는 데에 활용될 수 있다.

Cross-insitutional standardization of imaging protocols - A pilot study within the scope of the Comprehensive Cancer Center Lower Saxony.

Cross-institutional establishment of standardized protocols for CT and MR imaging of primary liver and pancreas tumors in an oncological center. This prospective dual-institution study was approved by the local IRBs. Minimum requirements (phases, sequences, slice thickness) for imaging of primary liver and pancreas tumors were defined and implemented at both sites. Between 06/19 and 08/19 in-house examinations were evaluated in terms of compliance with defined protocols and image quality. In addition, extramural examinations that were demonstrated at interdisciplinary tumor board meetings in the same study period were reviewed. Results were analyzed by means of descriptive statistics, and differences between centers, modalities and organs assessed (Fisher-exact Test, p < 0.05 deemed significant). 480 data sets (397 internal, 83 extramural) were included in this study and analyzed. Overall protocol compliance for in-house examinations was 93.5 % (371/397 datasets), without statistical significant difference between the two institutions (p = 0.0615). External studies met minimum requirements in 48.2 % (40/83 datasets). Regarding in-house imaging, significant differences were observed between CT of the liver and the pancreas (p < 0.05) and between CT and MRI of the pancreas (p < 0.05). As demonstrated in this pilot project, cross-institutional establishment of standardized imaging protocols is feasible with a compliance rate of more than 90 %. Standardized imaging protocols may serve as a quality indicator in oncological imaging, and over time, improve cross-institutional patient care. · Cross-institutional establishment of standardized imaging protocols is feasible with high compliance.. · Standards may serve as a quality indicator in oncological imaging.. · In perspective, cross-institutional patient care may be improved.. · Römermann I, Al-Bourini O, Seif Amir Hosseini A et al. Cross-insitutional standardization of imaging protocols - Apilot study within the scope of the Comprehensive Cancer Center Lower Saxony. Fortschr Röntgenstr 2021; 193: 410 - 416.

Abstract 4258: Molecular spectral CT imaging of Hsp70 positive solid tumors

Abstract Computed tomography (CT) imaging techniques play a major role in clinical imaging of malignant lesions. In recent years, this technique was expanded by the development of spectral CT techniques, offering the possibility to extract detailed information from spectral parameters of the imaged specimen. In particular, utilizing metal based contrast agent probes offer considerable advantages for tumor diagnosis and prognosis. Therefore, metal nanoparticle platforms enable effective payload delivery. In this work, we demonstrate the feasibility of anti plasmamembrane Heat shock protein 70 antibody conjugated spherical gold nanoparticles (cmHsp70.1-AuNP) for tumor specific in vivo spectral CT imaging, following intravenous application. The membrane associated form of Hsp70 (mHsp70) is exclusively presented on the plasma membrane of malignant cells of multiple tumor entities, but not on corresponding normal tissue cells, predestining this target epitope for tumor-selective in vivo targeting. Following incubation in vitro, the cmHsp70.1-AuNP were detectable in the cytoplasma of two cell lines being internalized via the endosomal-lysosomal pathway. In contrast, very low binding and internalization of isotype-labeled or blank NP were detectable. For characterization of the cell lines, cytoplasmatic and membrane-localized Hsp70 content was determined by flow cytometry, in-cell ELISA and Western Blot. Binding and internalization of the NP were investigated by Transition Electron Microscopy (TEM) and Mass Spectrometry (MS). In vivo, 24h after i.v. application in orthotopically and subcutaneously tumor bearing mice, the biodistribution as well as the intratumoral enrichment of AuNP were examined. AuNP were analyzed qualitatively and quantitatively by Silver enhancement technique of gold surfaces on histological slides, as well as by spectral CT analysis. Next to a minor enrichment in the tumor-associated Macrophages (due to camouflage by PEGylation), significant enrichment and explicit tumor delineation of cmHsp70.1-AuNP in tumors was detectable. Overall enrichment of the Isotype control mAb labelled AuNP was significantly lower. In contrast, the blank AuNP were found predominantly in tumor associated Macrophages near the main supplying blood vessels at the tumor borders.In summary, utilizing mHsp70 as target structure for guidance of Hsp70 functionalized Nanoparticles enables sufficient enrichment as well as uniform distribution of the cmHsp70.1-AuNP, adequate for spectral CT based tumor delineation. The study was supported by TaGoNaX DFG project № 336532926, STA1520/1-1 and the DFG grant SFB824/3 (TUM). Additional financial support was provided by RFBR 20-38-70039. Citation Format: Stefan Stangl, Maxim Shevtsov, Caroline Werner, Wolfgang Sievert, Gabriele Multhoff. Molecular spectral CT imaging of Hsp70 positive solid tumors [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 4258.

Research on the Practical Classification and Privacy Protection of CT Images of Parotid Tumors based on ResNet50 Model

Parotid gland disease is one of the main causes of facial paralysis, and parotid gland tumor is a great threat to the life of patients. The main diagnostic way of parotid diseases is imaging examination, so it is of great significance for the rapid classification of parotid image. In conclusion, 51 CT images of parotid malignant tumors and 101 CT images of parotid pleomorphic adenomas are selected as the research data set, and an intelligent and efficient machine learning algorithm is proposed for the practical classification of parotid images. At the same time, this paper also explores the privacy protection of medical images. Based on the advantages of deep learning, such as no feature engineering, strong adaptability and easy conversion, ResNet50 model in deep learning is selected as the basic network framework to achieve the purpose of rapid classification of parotid CT images. This is the first time that ResNet50 classification algorithm is applied to the practical classification of parotid tumor CT images. The results show that the accuracy of the test set converges to 90% when the model is iterated 1000 times, which also proves that this study has certain practical significance and application value for the auxiliary diagnosis of parotid gland tumor and other head and neck tumors. Simultaneously, this paper also explores the application of desensitization strategy in CT images of parotid tumors, which improves the performance of the model and also greatly protects the privacy of patients, and has a good application prospect in medical big data.

Desensitization and Cleaning Technology of CT Image of Parotid Tumor based on DICOM

CT image data are widely used in model training and data mining in the fields of medical image processing and medical big data analysis. Sensitive information on the data set needs to be desensitized before use. Traditional CT image data desensitization often wipe Remove most of the patient’s identification information. The data has not been cleaned, and the quality and analytical value of the data can’t be guaranteed after desensitization. In order to solve the privacy protection and data quality problems of DICOM-based image datasets during use, this paper uses symmetric encryption, numerical transformation and invalidation desensitization strategies to desensitize CT images of parotid tumor CT images based on DICOM. The desensitized encrypted data are decrypted to determine the true identity of the patient for data cleaning. Data cleaning is performed on the desensitized and decrypted data according to the formulated process. While meeting the desensitization standard to protect the privacy and safety of patients, the method of this paper greatly guarantees the quality, statistical characteristics and analytical value of the desensitized data set, and provide a feasible solution to the safe use and quality control of CT image data sets in the future.