Search

Your search keyword '"Pomykala, Kelsey L."' showing total 155 results
Start Over Author "Pomykala, Kelsey L."
155 results on '"Pomykala, Kelsey L."'

2. 'A net for everyone': fully personalized and unsupervised neural networks trained with longitudinal data from a single patient

Author

Strack, Christian, Pomykala, Kelsey L., Schlemmer, Heinz-Peter, Egger, Jan, and Kleesiek, Jens

Subjects
Computer Science - Machine Learning, Computer Science - Artificial Intelligence, Computer Science - Computer Vision and Pattern Recognition
Abstract

With the rise in importance of personalized medicine, we trained personalized neural networks to detect tumor progression in longitudinal datasets. The model was evaluated on two datasets with a total of 64 scans from 32 patients diagnosed with glioblastoma multiforme (GBM). Contrast-enhanced T1w sequences of brain magnetic resonance imaging (MRI) images were used in this study. For each patient, we trained their own neural network using just two images from different timepoints. Our approach uses a Wasserstein-GAN (generative adversarial network), an unsupervised network architecture, to map the differences between the two images. Using this map, the change in tumor volume can be evaluated. Due to the combination of data augmentation and the network architecture, co-registration of the two images is not needed. Furthermore, we do not rely on any additional training data, (manual) annotations or pre-training neural networks. The model received an AUC-score of 0.87 for tumor change. We also introduced a modified RANO criteria, for which an accuracy of 66% can be achieved. We show that using data from just one patient can be used to train deep neural networks to monitor tumor change.

Published
2022

3. FakeNews: GAN-based generation of realistic 3D volumetric data -- A systematic review and taxonomy

Author

Ferreira, André, Li, Jianning, Pomykala, Kelsey L., Kleesiek, Jens, Alves, Victor, and Egger, Jan

Subjects
Computer Science - Computer Vision and Pattern Recognition, Computer Science - Databases, Computer Science - Machine Learning
Abstract

With the massive proliferation of data-driven algorithms, such as deep learning-based approaches, the availability of high-quality data is of great interest. Volumetric data is very important in medicine, as it ranges from disease diagnoses to therapy monitoring. When the dataset is sufficient, models can be trained to help doctors with these tasks. Unfortunately, there are scenarios where large amounts of data is unavailable. For example, rare diseases and privacy issues can lead to restricted data availability. In non-medical fields, the high cost of obtaining enough high-quality data can also be a concern. A solution to these problems can be the generation of realistic synthetic data using Generative Adversarial Networks (GANs). The existence of these mechanisms is a good asset, especially in healthcare, as the data must be of good quality, realistic, and without privacy issues. Therefore, most of the publications on volumetric GANs are within the medical domain. In this review, we provide a summary of works that generate realistic volumetric synthetic data using GANs. We therefore outline GAN-based methods in these areas with common architectures, loss functions and evaluation metrics, including their advantages and disadvantages. We present a novel taxonomy, evaluations, challenges, and research opportunities to provide a holistic overview of the current state of volumetric GANs., Comment: 88 pages

Published
2022
Full Text
View/download PDF

4. k-strip: A novel segmentation algorithm in k-space for the application of skull stripping

Author

Rempe, Moritz, Mentzel, Florian, Pomykala, Kelsey L., Haubold, Johannes, Nensa, Felix, Kröninger, Kevin, Egger, Jan, and Kleesiek, Jens

Subjects
Electrical Engineering and Systems Science - Image and Video Processing, Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning
Abstract

Objectives: Present a novel deep learning-based skull stripping algorithm for magnetic resonance imaging (MRI) that works directly in the information rich k-space. Materials and Methods: Using two datasets from different institutions with a total of 36,900 MRI slices, we trained a deep learning-based model to work directly with the complex raw k-space data. Skull stripping performed by HD-BET (Brain Extraction Tool) in the image domain were used as the ground truth. Results: Both datasets were very similar to the ground truth (DICE scores of 92\%-98\% and Hausdorff distances of under 5.5 mm). Results on slices above the eye-region reach DICE scores of up to 99\%, while the accuracy drops in regions around the eyes and below, with partially blurred output. The output of k-strip often smoothed edges at the demarcation to the skull. Binary masks are created with an appropriate threshold. Conclusion: With this proof-of-concept study, we were able to show the feasibility of working in the k-space frequency domain, preserving phase information, with consistent results. Future research should be dedicated to discovering additional ways the k-space can be used for innovative image analysis and further workflows., Comment: 11 pages, 6 figures, 2 tables

Published
2022

6. AI-based Aortic Vessel Tree Segmentation for Cardiovascular Diseases Treatment: Status Quo

Author

Jin, Yuan, Pepe, Antonio, Li, Jianning, Gsaxner, Christina, Zhao, Fen-hua, Pomykala, Kelsey L., Kleesiek, Jens, Frangi, Alejandro F., and Egger, Jan

Subjects
Electrical Engineering and Systems Science - Image and Video Processing, Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning, Physics - Medical Physics
Abstract

The aortic vessel tree is composed of the aorta and its branching arteries, and plays a key role in supplying the whole body with blood. Aortic diseases, like aneurysms or dissections, can lead to an aortic rupture, whose treatment with open surgery is highly risky. Therefore, patients commonly undergo drug treatment under constant monitoring, which requires regular inspections of the vessels through imaging. The standard imaging modality for diagnosis and monitoring is computed tomography (CT), which can provide a detailed picture of the aorta and its branching vessels if completed with a contrast agent, called CT angiography (CTA). Optimally, the whole aortic vessel tree geometry from consecutive CTAs is overlaid and compared. This allows not only detection of changes in the aorta, but also of its branches, caused by the primary pathology or newly developed. When performed manually, this reconstruction requires slice by slice contouring, which could easily take a whole day for a single aortic vessel tree, and is therefore not feasible in clinical practice. Automatic or semi-automatic vessel tree segmentation algorithms, however, can complete this task in a fraction of the manual execution time and run in parallel to the clinical routine of the clinicians. In this paper, we systematically review computing techniques for the automatic and semi-automatic segmentation of the aortic vessel tree. The review concludes with an in-depth discussion on how close these state-of-the-art approaches are to an application in clinical practice and how active this research field is, taking into account the number of publications, datasets and challenges.

Published
2021

11. Simultaneous multislice diffusion-weighted imaging versus standard diffusion-weighted imaging in whole-body PET/MRI

Author

Furtado, Felipe S., Mercaldo, Nathaniel D., Vahle, Thomas, Benkert, Thomas, Bradley, William R., Ratanaprasatporn, Lisa, Seethamraju, Ravi Teja, Harisinghani, Mukesh G., Lee, Susanna, Suarez-Weiss, Krista, Umutlu, Lale, Catana, Ciprian, Pomykala, Kelsey L., Domachevsky, Liran, Bernstine, Hanna, Groshar, David, Rosen, Bruse R., and Catalano, Onofrio Antonio

Published
2023
Full Text
View/download PDF

13. Medical Deep Learning -- A systematic Meta-Review

Author

Egger, Jan, Gsaxner, Christina, Pepe, Antonio, Pomykala, Kelsey L., Jonske, Frederic, Kurz, Manuel, Li, Jianning, and Kleesiek, Jens

Subjects
Electrical Engineering and Systems Science - Image and Video Processing, Computer Science - Computer Vision and Pattern Recognition, Computer Science - Machine Learning
Abstract

Deep learning (DL) has remarkably impacted several different scientific disciplines over the last few years. E.g., in image processing and analysis, DL algorithms were able to outperform other cutting-edge methods. Additionally, DL has delivered state-of-the-art results in tasks like autonomous driving, outclassing previous attempts. There are even instances where DL outperformed humans, for example with object recognition and gaming. DL is also showing vast potential in the medical domain. With the collection of large quantities of patient records and data, and a trend towards personalized treatments, there is a great need for automated and reliable processing and analysis of health information. Patient data is not only collected in clinical centers, like hospitals and private practices, but also by mobile healthcare apps or online websites. The abundance of collected patient data and the recent growth in the DL field has resulted in a large increase in research efforts. In Q2/2020, the search engine PubMed returned already over 11,000 results for the search term 'deep learning', and around 90% of these publications are from the last three years. However, even though PubMed represents the largest search engine in the medical field, it does not cover all medical-related publications. Hence, a complete overview of the field of 'medical deep learning' is almost impossible to obtain and acquiring a full overview of medical sub-fields is becoming increasingly more difficult. Nevertheless, several review and survey articles about medical DL have been published within the last few years. They focus, in general, on specific medical scenarios, like the analysis of medical images containing specific pathologies. With these surveys as a foundation, the aim of this article is to provide the first high-level, systematic meta-review of medical DL surveys., Comment: 22 pages, 7 figures, 7 tables, 159 references. Computer Methods and Programs in Biomedicine (CMPB), Elsevier, May 2022

Published
2020
Full Text
View/download PDF

14. Total-Body 68Ga-PSMA-11 PET/CT for Bone Metastasis Detection in Prostate Cancer Patients: Potential Impact on Bone Scan Guidelines

Author

Pomykala, Kelsey L, Czernin, Johannes, Grogan, Tristan R, Armstrong, Wesley R, Williams, John, and Calais, Jeremie

Subjects
Biomedical and Clinical Sciences, Clinical Sciences, Oncology and Carcinogenesis, Aging, Cancer, Prostate Cancer, Biomedical Imaging, Urologic Diseases, 4.2 Evaluation of markers and technologies, Adult, Aged, Aged, 80 and over, Bone Neoplasms, Bone and Bones, Edetic Acid, Gallium Isotopes, Gallium Radioisotopes, Humans, Image Processing, Computer-Assisted, Male, Middle Aged, Oligopeptides, Positron Emission Tomography Computed Tomography, Practice Guidelines as Topic, Prostatic Neoplasms, Retrospective Studies, Whole-Body Irradiation, PSMA, PET/CT, prostate cancer, bone metastasis, field of view, guidelines, Nuclear Medicine & Medical Imaging, Clinical sciences
Abstract

Our purpose was to determine the relationship between serum prostate-specific antigen (PSA) level categories (20 ng/mL) and the incidence of bone metastases detected by total-body 68Ga-prostate-specific membrane antigen (PSMA)-11 PET/CT and to assess if expanding the 68Ga-PSMA-11 PET/CT imaging field to include the vertex and lower extremities (total-body acquisition) affects bone metastasis detection rates and patient management. Methods: This was a retrospective analysis of 388 prostate cancer patients enrolled in 5 prospective studies (NCT02940262, NCT03368547, NCT03042312, NCT04050215, and NCT03515577). All underwent 68Ga-PSMA-11 PET/CT scans acquired from vertex to toes for primary staging (n = 93/388, 24%), biochemical recurrence (BCR) localization (n = 225/388, 58%), or restaging metastatic disease (M1) before or during systemic therapy (n = 70/388, 18%) between September 2017 and May 2018. Results: In total, 321 of 388 patients (83%) had a positive 68Ga-PSMA-11 study. PSMA-positive bone lesions were found in 105 of 388 (27%) patients, with an incidence that was positively associated with serum PSA level (

Published
2020

19. Perceived Misinterpretation Rates in Oncologic 18F-FDG PET/CT Studies: A Survey of Referring Physicians

Author

Karantanis, Dimitrios, Kalkanis, Dimitrios, Czernin, Johannes, Herrmann, Ken, Pomykala, Kelsey L, Bogsrud, Trond V, Subramaniam, Rathan M, Lowe, Val J, and Allen-Auerbach, Martin S

Subjects
Biomedical and Clinical Sciences, Clinical Sciences, Oncology and Carcinogenesis, Biomedical Imaging, Clinical Trials and Supportive Activities, Clinical Research, Cancer, Diagnostic Errors, Fluorodeoxyglucose F18, Health Care Surveys, Humans, Internet, Medical Oncology, Multimodal Imaging, Neoplasms, Positron-Emission Tomography, Radiopharmaceuticals, Referral and Consultation, Surveys and Questionnaires, overinterpretation, false-positive results, false-negative results, accuracy, pitfall, Nuclear Medicine & Medical Imaging, Clinical sciences
Abstract

UnlabelledBecause only pathologic examination can confirm the presence or absence of malignant disease in cancer patients, a certain rate of misinterpretation in any kind of imaging study is inevitable. For the accuracy of interpretation to be improved, determination of the nature, causes, and magnitude of this problem is needed. This study was designed to collect pertinent information from physicians referring patients for oncologic (18)F-FDG PET/CT.MethodsA total of 662 referring physicians completed an 11-question survey focused on their experience with the interpretation of oncologic (18)F-FDG PET/CT studies. The participants were oncologists (36.1%; n = 239), hematologists (14.5%; n = 96), radiation oncologists (7.4%; n = 49), surgeons (33.8%; n = 224), and other physicians (8.2%; n = 54). Questions were aimed at determining the frequency, nature, and causes of scan misinterpretations as well as potential solutions to reduce the frequency of misinterpretations.ResultsPerceived misinterpretation rates ranged from 5% to 20%, according to most (59.3%) of the participants; 20.8% of respondents reported rates of less than 5%. Overinterpretation rather than underinterpretation was more frequently encountered (68.9% vs. 8.7%, respectively). Limited availability of a patient's history and limited experience of interpreters were the major contributors to this phenomenon, according to 46.8% and 26.7% of the participants, respectively. The actions most commonly suggested to reduce misinterpretation rates (multiple suggestions were possible) were the institution of multidisciplinary meetings (59.8%), the provision of adequate history when ordering an examination (37.4%), and a discussion with imaging specialists when receiving the results of the examination (38.4%).ConclusionOverinterpretation rather than underinterpretation of oncologic (18)F-FDG PET/CT studies prevails in clinical practice, according to referring physicians. Closer collaboration of imaging specialists with referring physicians through more multidisciplinary meetings, improved communication, and targeted training of interpreting physicians are actions suggested to reduce the rates of misinterpretation of oncologic (18)F-FDG PET/CT studies.

Published
2014

21. Comparison of visual and semiquantitative analysis of 18F-FDOPA-PET/CT for recurrence detection in glioblastoma patients

Author

Herrmann, Ken, Czernin, Johannes, Cloughesy, Timothy, Lai, Albert, Pomykala, Kelsey L, Benz, Matthias R, Buck, Andreas K, Phelps, Michael E, and Chen, Wei

Subjects
Biomedical and Clinical Sciences, Clinical Sciences, Oncology and Carcinogenesis, Biomedical Imaging, Rare Diseases, Neurosciences, Cancer, Brain Cancer, Brain Disorders, 4.1 Discovery and preclinical testing of markers and technologies, Detection, screening and diagnosis, 4.2 Evaluation of markers and technologies, Adult, Aged, Aged, 80 and over, Dihydroxyphenylalanine, Female, Fluorodeoxyglucose F18, Follow-Up Studies, Glioblastoma, Humans, Male, Middle Aged, Neoplasm Recurrence, Local, Positron-Emission Tomography, Prognosis, Radiopharmaceuticals, Survival Rate, Tomography, X-Ray Computed, Young Adult, glioblastoma, F-18-FDOPA, recurrence detection, 18F-FDOPA, Oncology & Carcinogenesis, Oncology and carcinogenesis
Abstract

BackgroundAmino acid transport imaging with 18F-FDOPA PET is increasingly used for detection of glioblastoma recurrence. However, a standardized image interpretation for 18F-FDOPA brain PET studies has not yet been established. This study compares visual and semiquantitative analysis parameters for detection of tumor recurrence and correlates them with progression-free survival (PFS).MethodsOne-hundred ten patients (72 male:38 female) with suspected tumor recurrence who underwent 18F-FDOPA PET imaging were studied. PET scans were analyzed visually (5-point scale) and semiquantitatively (lesion-to-striatum- and lesion- to-normal-brain-tissue ratios using both SUV(mean) and SUV(max)). Accuracies for recurrence detection were calculated using histopathology and clinical follow-up for validation. Receiving operator characteristic and Kaplan-Meier survival analysis were performed to derive imaging-based prediction of PFS and overall survival (OS).ResultsAccuracies for detection of glioblastoma recurrence were similar for visual (82%) and semiquantitative (range, 77%-82%) analysis. Both visual and semiquantitative indices were significant predictors of PFS, with mean lesion-to normal brain tissue ratios providing the best discriminator (mean survival, 39.4 vs 9.3 months; P < .001). None of the investigated parameters was predictive for OS.ConclusionsBoth visual and semiquantitative indices detected glioblastoma recurrence with high accuracy and were predictive for PFS. Lesion-to-normal-tissue ratios were the best discriminators of PFS; however, none of the investigated parameters predicted OS. These retrospectively established analysis parameters need to be confirmed prospectively.

Published
2014

22. Oncologic Staging with68Ga-FAPI PET/CT Demonstrates a Lower Rate of Nonspecific Lymph Node Findings Than18F-FDG PET/CT

Author

Demmert, Tristan T., primary, Pomykala, Kelsey L., additional, Lanzafame, Helena, additional, Pabst, Kim M., additional, Lueckerath, Katharina, additional, Siveke, Jens, additional, Umutlu, Lale, additional, Hautzel, Hubertus, additional, Hamacher, Rainer, additional, Herrmann, Ken, additional, and Fendler, Wolfgang P., additional

Published
2023
Full Text
View/download PDF

23. Histology-Based Prediction of Therapy Response to Neoadjuvant Chemotherapy for Esophageal and Esophagogastric Junction Adenocarcinomas Using Deep Learning

Author

Hörst, Fabian, primary, Ting, Saskia, additional, Liffers, Sven-Thorsten, additional, Pomykala, Kelsey L., additional, Steiger, Katja, additional, Albertsmeier, Markus, additional, Angele, Martin K., additional, Lorenzen, Sylvie, additional, Quante, Michael, additional, Weichert, Wilko, additional, Egger, Jan, additional, Siveke, Jens T., additional, and Kleesiek, Jens, additional

Published
2023
Full Text
View/download PDF

24. Non-exposed endoscopic wall-inversion surgery with one-step nucleic acid amplification for early gastrointestinal tumors: Personal experience and literature review

Author

Crafa, Francesco, primary, Vanella, Serafino, additional, Morante, Aristide, additional, Catalano, Onofrio A, additional, Pomykala, Kelsey L, additional, Baiamonte, Mario, additional, Godas, Maria, additional, Antunes, Alexandra, additional, Costa Pereira, Joaquim, additional, and Giaccaglia, Valentina, additional

Published
2023
Full Text
View/download PDF

25. Voxel-Based Dosimetry Predicts Hepatotoxicity in Hepatocellular Carcinoma Patients Undergoing Radioembolization with90Y Glass Microspheres

Author

Watanabe, Masao, primary, Grafe, Hong, additional, Theysohn, Jens, additional, Schaarschmidt, Benedikt, additional, Ludwig, Johannes, additional, Jochheim, Leonie, additional, Jeschke, Matthias, additional, Schmidt, Hartmut, additional, Fendler, Wolfgang P., additional, Moraitis, Alexandros, additional, Herrmann, Ken, additional, Pomykala, Kelsey L., additional, and Weber, Manuel, additional

Published
2023
Full Text
View/download PDF

30. Novel ⁶⁸Ga-FAPI PET/CT Offers Oncologic Staging Without COVID-19 Vaccine-Related Pitfalls

Author

Demmert, Tristan T., Maric, Ines, Pomykala, Kelsey L., Lueckerath, Katharina, Siveke, Jens, Schaarschmidt, Benedikt M., Hamacher, Rainer, Herrmann, Ken, and Fendler, Wolfgang P.

Subjects
Medizin
Abstract

In the setting of ongoing coronavirus disease 2019 vaccination, vaccine-related tracer uptake in locoregional lymph nodes has become a well-known issue in tumor staging by 18F-FDG PET/CT. 68Ga-fibroblast-activation protein inhibitor (FAPI) PET/CT is a new oncologic imaging tool that may overcome this limitation. Methods: We assessed postvaccine head-to-head and same-day 18F-FDG and 68Ga-FAPI-46 PET/CT findings in a series of 11 patients from a large, prospective imaging registry. All patients with documented tracer uptake in locoregional lymph nodes on PET/CT or PET/MRI, after vaccination within 6 wk, were eligible for investigation. Result: Significant visual lymph node uptake adjacent to the injection site was noted in 11 of 11 (100%) patients with 18F-FDG PET/CT, versus 0 of 11 (0%) with 68Ga-FAPI PET/CT. 18F-FDG detected 73% and 68Ga-FAPI PET/CT 94% of all tumor lesions. Conclusion: In this case-series study, 68Ga-FAPI showed its potential to avoid 18F-FDG PET/CT postvaccination pitfalls and presented superior tumor localization.

Published
2023

31. Evaluation of thresholding methods for the quantification of [⁶⁸Ga]Ga-PSMA-11 PET molecular tumor volume and their effect on survival prediction in patients with advanced prostate cancer undergoing [¹⁷⁷Lu]Lu-PSMA-617 radioligand therapy

Author

Kim, Moon, Seifert, Robert, Fragemann, Jana, Kersting, David, Murray, Jacob, Jonske, Frederic, Pomykala, Kelsey L., Egger, Jan, Fendler, Wolfgang, Herrmann, Ken, and Kleesiek, Jens

Subjects
Medizin
Abstract

Purpose: The aim of this study was to systematically evaluate the effect of thresholding algorithms used in computer vision for the quantification of prostate-specific membrane antigen positron emission tomography (PET) derived tumor volume (PSMA-TV) in patients with advanced prostate cancer. The results were validated with respect to the prognostication of overall survival in patients with advanced-stage prostate cancer. Materials and methods: A total of 78 patients who underwent [¹⁷⁷Lu]Lu-PSMA-617 radionuclide therapy from January 2018 to December 2020 were retrospectively included in this study. [⁶⁸Ga]Ga-PSMA-11 PET images, acquired prior to radionuclide therapy, were used for the analysis of thresholding algorithms. All PET images were first analyzed semi-automatically using a pre-evaluated, proprietary software solution as the baseline method. Subsequently, five histogram-based thresholding methods and two local adaptive thresholding methods that are well established in computer vision were applied to quantify molecular tumor volume. The resulting whole-body molecular tumor volumes were validated with respect to the prognostication of overall patient survival as well as their statistical correlation to the baseline methods and their performance on standardized phantom scans. Results: The whole-body PSMA-TVs, quantified using different thresholding methods, demonstrate a high positive correlation with the baseline methods. We observed the highest correlation with generalized histogram thresholding (GHT) (Pearson r (r), p value (p): r = 0.977, p < 0.001) and Sauvola thresholding (r = 0.974, p < 0.001) and the lowest correlation with Multiotsu (r = 0.877, p < 0.001) and Yen thresholding methods (r = 0.878, p < 0.001). The median survival time of all patients was 9.87 months (95% CI [9.3 to 10.13]). Stratification by median whole-body PSMA-TV resulted in a median survival time from 11.8 to 13.5 months for the patient group with lower tumor burden and 6.5 to 6.6 months for the patient group with higher tumor burden. The patient group with lower tumor burden had significantly higher probability of survival (p < 0.00625) in eight out of nine thresholding methods (Fig. 2); those methods were SUVmax50 (p = 0.0038), SUV ≥3 (p = 0.0034), Multiotsu (p = 0.0015), Yen (p = 0.0015), Niblack (p = 0.001), Sauvola (p = 0.0001), Otsu (p = 0.0053), and Li thresholding (p = 0.0053). Conclusion: Thresholding methods commonly used in computer vision are promising tools for the semiautomatic quantification of whole-body PSMA-TV in [⁶⁸Ga]Ga-PSMA-11-PET. The proposed algorithm-driven thresholding strategy is less arbitrary and less prone to biases than thresholding with predefined values, potentially improving the application of whole-body PSMA-TV as an imaging biomarker.

Published
2023

34. Simultaneous multislice diffusion-weighted imaging versus standard diffusion-weighted imaging in whole-body PET/MRI

Author

Furtado, Felipe S., primary, Mercaldo, Nathaniel D., additional, Vahle, Thomas, additional, Benkert, Thomas, additional, Bradley, William R., additional, Ratanaprasatporn, Lisa, additional, Seethamraju, Ravi Teja, additional, Harisinghani, Mukesh G., additional, Lee, Susanna, additional, Suarez-Weiss, Krista, additional, Umutlu, Lale, additional, Catana, Ciprian, additional, Pomykala, Kelsey L., additional, Domachevsky, Liran, additional, Bernstine, Hanna, additional, Groshar, David, additional, Rosen, Bruse R., additional, and Catalano, Onofrio Antonio, additional

Published
2022
Full Text
View/download PDF

39. Virtual Prostate Biopsy with Prostate-specific Membrane Antigen and Magnetic Resonance Imaging : Closer to Reality in a Subgroup of Prostate Cancer Patients?

Author

Pomykala, Kelsey L, Herrmann, Ken, Emmett, Louise, Lalumera, Elisabetta, Fanti, Stefano, Pomykala, Kelsey L, Herrmann, Ken, Emmett, Louise, Lalumera, Elisabetta, and Fanti, Stefano

Subjects
Urology, Medizin, Prostate-specific membrane antigen (PSMA), Prostate cancer, diagnosis, PET, virtual biopsy
Abstract

CA Pomykala

Published
2022

40. GAN-based generation of realistic 3D data: A systematic review and taxonomy

Author

Ferreira, André, Li, Jianning, Pomykala, Kelsey L., Kleesiek, Jens, Alves, Victor, and Egger, Jan

Subjects
FOS: Computer and information sciences, Computer Science - Machine Learning, Computer Science - Databases, Computer Vision and Pattern Recognition (cs.CV), Computer Science - Computer Vision and Pattern Recognition, Databases (cs.DB), Machine Learning (cs.LG)
Abstract

Data has become the most valuable resource in today's world. With the massive proliferation of data-driven algorithms, such as deep learning-based approaches, the availability of data is of great interest. In this context, high-quality training, validation and testing datasets are particularly needed. Volumetric data is a very important resource in medicine, as it ranges from disease diagnoses to therapy monitoring. When the dataset is sufficient, models can be trained to help doctors with these tasks. Unfortunately, there are scenarios and applications where large amounts of data is unavailable. For example, in the medical field, rare diseases and privacy issues can lead to restricted data availability. In non-medical fields, the high cost of obtaining a sufficient amount of high-quality data can also be a concern. A solution to these problems can be the generation of synthetic data to perform data augmentation in combination with other more traditional methods of data augmentation. Therefore, most of the publications on 3D Generative Adversarial Networks (GANs) are within the medical domain. The existence of mechanisms to generate realistic synthetic data is a good asset to overcome this challenge, especially in healthcare, as the data must be of good quality and close to reality, i.e. realistic, and without privacy issues. In this review, we provide a summary of works that generate realistic 3D synthetic data using GANs. We therefore outline GAN-based methods in these areas with common architectures, advantages and disadvantages. We present a novel taxonomy, evaluations, challenges and research opportunities to provide a holistic overview of the current state of GANs in medicine and other fields., 52 pages

Published
2022

43. Evaluation of thresholding methods for the quantification of [68Ga]Ga-PSMA-11 PET molecular tumor volume and their effect on survival prediction in patients with advanced prostate cancer undergoing [177Lu]Lu-PSMA-617 radioligand therapy

Author

Kim, Moon, Seifert, Robert, Fragemann, Jana, Kersting, David, Murray, Jacob, Jonske, Frederic, Pomykala, Kelsey L., Egger, Jan, Fendler, Wolfgang P., Herrmann, Ken, and Kleesiek, Jens

Subjects
CANCER patients, PROSTATE cancer patients, MOLECULAR volume, OVERALL survival, HOLMIUM, STATISTICAL correlation, ANDROGEN receptors
Abstract

Purpose: The aim of this study was to systematically evaluate the effect of thresholding algorithms used in computer vision for the quantification of prostate-specific membrane antigen positron emission tomography (PET) derived tumor volume (PSMA-TV) in patients with advanced prostate cancer. The results were validated with respect to the prognostication of overall survival in patients with advanced-stage prostate cancer. Materials and methods: A total of 78 patients who underwent [177Lu]Lu-PSMA-617 radionuclide therapy from January 2018 to December 2020 were retrospectively included in this study. [68Ga]Ga-PSMA-11 PET images, acquired prior to radionuclide therapy, were used for the analysis of thresholding algorithms. All PET images were first analyzed semi-automatically using a pre-evaluated, proprietary software solution as the baseline method. Subsequently, five histogram-based thresholding methods and two local adaptive thresholding methods that are well established in computer vision were applied to quantify molecular tumor volume. The resulting whole-body molecular tumor volumes were validated with respect to the prognostication of overall patient survival as well as their statistical correlation to the baseline methods and their performance on standardized phantom scans. Results: The whole-body PSMA-TVs, quantified using different thresholding methods, demonstrate a high positive correlation with the baseline methods. We observed the highest correlation with generalized histogram thresholding (GHT) (Pearson r (r), p value (p): r = 0.977, p < 0.001) and Sauvola thresholding (r = 0.974, p < 0.001) and the lowest correlation with Multiotsu (r = 0.877, p < 0.001) and Yen thresholding methods (r = 0.878, p < 0.001). The median survival time of all patients was 9.87 months (95% CI [9.3 to 10.13]). Stratification by median whole-body PSMA-TV resulted in a median survival time from 11.8 to 13.5 months for the patient group with lower tumor burden and 6.5 to 6.6 months for the patient group with higher tumor burden. The patient group with lower tumor burden had significantly higher probability of survival (p < 0.00625) in eight out of nine thresholding methods (Fig. 2); those methods were SUVmax50 (p = 0.0038), SUV ≥3 (p = 0.0034), Multiotsu (p = 0.0015), Yen (p = 0.0015), Niblack (p = 0.001), Sauvola (p = 0.0001), Otsu (p = 0.0053), and Li thresholding (p = 0.0053). Conclusion: Thresholding methods commonly used in computer vision are promising tools for the semiautomatic quantification of whole-body PSMA-TV in [68Ga]Ga-PSMA-11-PET. The proposed algorithm-driven thresholding strategy is less arbitrary and less prone to biases than thresholding with predefined values, potentially improving the application of whole-body PSMA-TV as an imaging biomarker. [ABSTRACT FROM AUTHOR]

Published
2023
Full Text
View/download PDF

45. Role of one-step nucleic acid amplification in colorectal cancer lymph node metastases detection

Author

Crafa, Francesco, Vanella, Serafino, Catalano, Onofrio A., Pomykala, Kelsey L., and Baiamonte, Mario

Subjects
Medizin
Abstract

Current histopathological staging procedures in colorectal cancer (CRC) depend on midline division of the lymph nodes (LNs) with one section of hematoxylin and eosin staining. Cancer cells outside this transection line may be missed, which could lead to understaging of Union for International Cancer Control Stage II high-risk patients. The one-step nucleic acid amplification (OSNA) assay has emerged as a rapid molecular diagnostic tool for LN metastases detection. It is a molecular technique that can analyze the entire LN tissue using a reversetranscriptase loop-mediated isothermal amplification reaction to detect tumorspecific cytokeratin 19 mRNA. Our findings suggest that the OSNA assay has a high diagnostic accuracy in detecting metastatic LNs in CRC and a high negative predictive value. OSNA is a standardized, observer-independent technique, which may lead to more accurate staging. It has been suggested that in stage II CRC, the upstaging can reach 25% and these patients can access postoperative adjuvant chemotherapy. Moreover, intraoperative OSNA sentinel node evaluation may allow early CRC to be treated with organ-preserving surgery, while in more advanced-stage disease, a tailored lymphadenectomy can be performed considering the presence of aberrant lymphatic drainage and skip metastases.

Published
2022

Catalog

Books, media, physical & digital resources
See catalog results