The Experts below are selected from a list of 1212 Experts worldwide ranked by ideXlab platform
Martin W Huellner - One of the best experts on this subject based on the ideXlab platform.
-
artificial intelligence for detecting small fdg positive lung nodules in digital pet ct impact of image reconstructions on diagnostic performance
2020Co-Authors: Moritz Schwyzer, Katharina Martini, Dominik C. Benz, Daniela A. Ferraro, Ken Kudura, Philipp A Kaufmann, Irene A. Burger, Valerie Treyer, Gustav K Von Schulthess, Martin W HuellnerAbstract:To evaluate the diagnostic performance of a deep learning algorithm for automated detection of small 18F-FDG-avid pulmonary nodules in PET scans, and to assess whether novel block sequential regularized expectation maximization (BSREM) reconstruction affects detection accuracy as compared to ordered subset expectation maximization (OSEM) reconstruction. Fifty-seven patients with 92 18F-FDG-avid pulmonary nodules (all ≤ 2 cm) undergoing PET/CT for oncological (re-)staging were retrospectively included and a total of 8824 PET images of the lungs were extracted using OSEM and BSREM reconstruction. Per-slice and per-nodule sensitivity of a deep learning algorithm was assessed, with an expert readout by a radiologist/Nuclear Medicine Physician serving as standard of reference. Receiver-operator characteristic (ROC) curve of OSEM and BSREM were assessed and the areas under the ROC curve (AUC) were compared. A maximum standardized uptake value (SUVmax)–based sensitivity analysis and a size-based sensitivity analysis with subgroups defined by nodule size was performed. The AUC of the deep learning algorithm for nodule detection using OSEM reconstruction was 0.796 (CI 95%; 0.772–0.869), and 0.848 (CI 95%; 0.828–0.869) using BSREM reconstruction. The AUC was significantly higher for BSREM compared to OSEM (p = 0.001). On a per-slice analysis, sensitivity and specificity were 66.7% and 79.0% for OSEM, and 69.2% and 84.5% for BSREM. On a per-nodule analysis, the overall sensitivity of OSEM was 81.5% compared to 87.0% for BSREM. Our results suggest that machine learning algorithms may aid detection of small 18F-FDG-avid pulmonary nodules in clinical PET/CT. AI performed significantly better on images with BSREM than OSEM. • The diagnostic value of deep learning for detecting small lung nodules (≤ 2 cm) in PET images using BSREM and OSEM reconstruction was assessed. • BSREM yields higher SUV max of small pulmonary nodules as compared to OSEM reconstruction. • The use of BSREM translates into a higher detectability of small pulmonary nodules in PET images as assessed with artificial intelligence.
-
Artificial intelligence for detecting small FDG-positive lung nodules in digital PET/CT: impact of image reconstructions on diagnostic performance
2019Co-Authors: Moritz Schwyzer, Katharina Martini, Dominik C. Benz, Daniela A. Ferraro, Ken Kudura, Philipp A Kaufmann, Gustav K. Von Schulthess, Irene A. Burger, Valerie Treyer, Martin W HuellnerAbstract:ObjectivesTo evaluate the diagnostic performance of a deep learning algorithm for automated detection of small ^18F-FDG-avid pulmonary nodules in PET scans, and to assess whether novel block sequential regularized expectation maximization (BSREM) reconstruction affects detection accuracy as compared to ordered subset expectation maximization (OSEM) reconstruction.MethodsFifty-seven patients with 92 ^18F-FDG-avid pulmonary nodules (all ≤ 2 cm) undergoing PET/CT for oncological (re-)staging were retrospectively included and a total of 8824 PET images of the lungs were extracted using OSEM and BSREM reconstruction. Per-slice and per-nodule sensitivity of a deep learning algorithm was assessed, with an expert readout by a radiologist/Nuclear Medicine Physician serving as standard of reference. Receiver-operator characteristic (ROC) curve of OSEM and BSREM were assessed and the areas under the ROC curve (AUC) were compared. A maximum standardized uptake value (SUV_max)–based sensitivity analysis and a size-based sensitivity analysis with subgroups defined by nodule size was performed.ResultsThe AUC of the deep learning algorithm for nodule detection using OSEM reconstruction was 0.796 (CI 95%; 0.772–0.869), and 0.848 (CI 95%; 0.828–0.869) using BSREM reconstruction. The AUC was significantly higher for BSREM compared to OSEM ( p = 0.001). On a per-slice analysis, sensitivity and specificity were 66.7% and 79.0% for OSEM, and 69.2% and 84.5% for BSREM. On a per-nodule analysis, the overall sensitivity of OSEM was 81.5% compared to 87.0% for BSREM.ConclusionsOur results suggest that machine learning algorithms may aid detection of small ^18F-FDG-avid pulmonary nodules in clinical PET/CT. AI performed significantly better on images with BSREM than OSEM.Key Points• The diagnostic value of deep learning for detecting small lung nodules (≤ 2 cm) in PET images using BSREM and OSEM reconstruction was assessed. • BSREM yields higher SUV _ max of small pulmonary nodules as compared to OSEM reconstruction. • The use of BSREM translates into a higher detectability of small pulmonary nodules in PET images as assessed with artificial intelligence .
Michael Lassmann - One of the best experts on this subject based on the ideXlab platform.
-
eanm position paper on article 56 of the council directive 2013 59 euratom basic safety standards for Nuclear Medicine therapy
2021Co-Authors: Mark Konijnenberg, Roland Hustinx, Frederik A Verburg, Ken Herrmann, Carsten Kobe, Cecilia Hindorf, Michael LassmannAbstract:The EC Directive 2013/59/Euratom states in article 56 that exposures of target volumes in Nuclear Medicine treatments shall be individually planned and their delivery appropriately verified. The Directive also mentions that medical physics experts should always be appropriately involved in those treatments. Although it is obvious that, in Nuclear Medicine practice, every Nuclear Medicine Physician and physicist should follow national rules and legislation, the EANM considered it necessary to provide guidance on how to interpret the Directive statements for Nuclear Medicine treatments. For this purpose, the EANM proposes to distinguish three levels in compliance to the optimization principle in the directive, inspired by the indication of levels in prescribing, recording and reporting of absorbed doses after radiotherapy defined by the International Commission on Radiation Units and Measurements (ICRU): Most Nuclear Medicine treatments currently applied in Europe are standardized. The minimum requirement for those treatments is ICRU level 1 (“activity-based prescription and patient-averaged dosimetry”), which is defined by administering the activity within 10% of the intended activity, typically according to the package insert or to the respective EANM guidelines, followed by verification of the therapy delivery, if applicable.Non-standardized treatments are essentially those in developmental phase or approved radiopharmaceuticals being used off-label with significantly (> 25% more than in the label) higher activities. These treatments should comply with ICRU level 2 (“activity-based prescription and patient-specific dosimetry”), which implies recording and reporting of the absorbed dose to organs at risk and optionally the absorbed dose to treatment regions.The EANM strongly encourages to foster research that eventually leads to treatment planning according to ICRU level 3 (“dosimetry-guided patient-specific prescription and verification”), whenever possible and relevant. Evidence for superiority of therapy prescription on basis of patient-specific dosimetry has not been obtained. However, the authors believe that a better understanding of therapy dosimetry, i.e. how much and where the energy is delivered, and radiobiology, i.e. radiation-related processes in tissues, are keys to the long-term improvement of our treatments. (Less)
-
EANM position paper on article 56 of the Council Directive 2013/59/Euratom (basic safety standards) for Nuclear Medicine therapy
2020Co-Authors: Mark Konijnenberg, Roland Hustinx, Ken Herrmann, Carsten Kobe, Cecilia Hindorf, Frederik Verburg, Michael LassmannAbstract:The EC Directive 2013/59/Euratom states in article 56 that exposures of target volumes in Nuclear Medicine treatments shall be individually planned and their delivery appropriately verified. The Directive also mentions that medical physics experts should always be appropriately involved in those treatments. Although it is obvious that, in Nuclear Medicine practice, every Nuclear Medicine Physician and physicist should follow national rules and legislation, the EANM considered it necessary to provide guidance on how to interpret the Directive statements for Nuclear Medicine treatments. For this purpose, the EANM proposes to distinguish three levels in compliance to the optimization principle in the directive, inspired by the indication of levels in prescribing, recording and reporting of absorbed doses after radiotherapy defined by the International Commission on Radiation Units and Measurements (ICRU): Most Nuclear Medicine treatments currently applied in Europe are standardized. The minimum requirement for those treatments is ICRU level 1 (“activity-based prescription and patient-averaged dosimetry”), which is defined by administering the activity within 10% of the intended activity, typically according to the package insert or to the respective EANM guidelines, followed by verification of the therapy delivery, if applicable. Non-standardized treatments are essentially those in developmental phase or approved radiopharmaceuticals being used off-label with significantly (> 25% more than in the label) higher activities. These treatments should comply with ICRU level 2 (“activity-based prescription and patient-specific dosimetry”), which implies recording and reporting of the absorbed dose to organs at risk and optionally the absorbed dose to treatment regions. The EANM strongly encourages to foster research that eventually leads to treatment planning according to ICRU level 3 (“dosimetry-guided patient-specific prescription and verification”), whenever possible and relevant. Evidence for superiority of therapy prescription on basis of patient-specific dosimetry has not been obtained. However, the authors believe that a better understanding of therapy dosimetry, i.e. how much and where the energy is delivered, and radiobiology, i.e. radiation-related processes in tissues, are keys to the long-term improvement of our treatments.
-
eanm procedure guideline for radio immunotherapy for b cell lymphoma with 90 y radiolabelled ibritumomab tiuxetan zevalin
2007Co-Authors: Jan Tennvall, Manfred Fischer, Angelika Bischof Delaloye, Emilio Bombardieri, Lisa Bodei, Francesco Giammarile, Michael Lassmann, Wim J G Oyen, Boudewijn BransAbstract:Background In January 2004, EMEA approved Y-90-radiolabelled ibritumomab tiuxetan, Zevalin, in Europe for the treatment of adult patients with rituximab-relapsed or -refractory CD20+ follicular B-cell non-Hodgkin's lymphoma. The number of European Nuclear Medicine departments using Zevalin is continuously increasing, since the therapy is often considered successful. The Therapy, Oncology and Dosimetry Committees have worked together in order to define some EANM guidelines on the use of Zevalin, paying particular attention to the problems related to Nuclear Medicine. Purpose The purpose of this guideline is to assist the Nuclear Medicine Physician in treating and managing patients who may be candidates for radio-immunotherapy. The guideline also stresses the need for close collaboration with the Physician(s) treating the patient for the underlying disease.
Moritz Schwyzer - One of the best experts on this subject based on the ideXlab platform.
-
artificial intelligence for detecting small fdg positive lung nodules in digital pet ct impact of image reconstructions on diagnostic performance
2020Co-Authors: Moritz Schwyzer, Katharina Martini, Dominik C. Benz, Daniela A. Ferraro, Ken Kudura, Philipp A Kaufmann, Irene A. Burger, Valerie Treyer, Gustav K Von Schulthess, Martin W HuellnerAbstract:To evaluate the diagnostic performance of a deep learning algorithm for automated detection of small 18F-FDG-avid pulmonary nodules in PET scans, and to assess whether novel block sequential regularized expectation maximization (BSREM) reconstruction affects detection accuracy as compared to ordered subset expectation maximization (OSEM) reconstruction. Fifty-seven patients with 92 18F-FDG-avid pulmonary nodules (all ≤ 2 cm) undergoing PET/CT for oncological (re-)staging were retrospectively included and a total of 8824 PET images of the lungs were extracted using OSEM and BSREM reconstruction. Per-slice and per-nodule sensitivity of a deep learning algorithm was assessed, with an expert readout by a radiologist/Nuclear Medicine Physician serving as standard of reference. Receiver-operator characteristic (ROC) curve of OSEM and BSREM were assessed and the areas under the ROC curve (AUC) were compared. A maximum standardized uptake value (SUVmax)–based sensitivity analysis and a size-based sensitivity analysis with subgroups defined by nodule size was performed. The AUC of the deep learning algorithm for nodule detection using OSEM reconstruction was 0.796 (CI 95%; 0.772–0.869), and 0.848 (CI 95%; 0.828–0.869) using BSREM reconstruction. The AUC was significantly higher for BSREM compared to OSEM (p = 0.001). On a per-slice analysis, sensitivity and specificity were 66.7% and 79.0% for OSEM, and 69.2% and 84.5% for BSREM. On a per-nodule analysis, the overall sensitivity of OSEM was 81.5% compared to 87.0% for BSREM. Our results suggest that machine learning algorithms may aid detection of small 18F-FDG-avid pulmonary nodules in clinical PET/CT. AI performed significantly better on images with BSREM than OSEM. • The diagnostic value of deep learning for detecting small lung nodules (≤ 2 cm) in PET images using BSREM and OSEM reconstruction was assessed. • BSREM yields higher SUV max of small pulmonary nodules as compared to OSEM reconstruction. • The use of BSREM translates into a higher detectability of small pulmonary nodules in PET images as assessed with artificial intelligence.
-
Artificial intelligence for detecting small FDG-positive lung nodules in digital PET/CT: impact of image reconstructions on diagnostic performance
2019Co-Authors: Moritz Schwyzer, Katharina Martini, Dominik C. Benz, Daniela A. Ferraro, Ken Kudura, Philipp A Kaufmann, Gustav K. Von Schulthess, Irene A. Burger, Valerie Treyer, Martin W HuellnerAbstract:ObjectivesTo evaluate the diagnostic performance of a deep learning algorithm for automated detection of small ^18F-FDG-avid pulmonary nodules in PET scans, and to assess whether novel block sequential regularized expectation maximization (BSREM) reconstruction affects detection accuracy as compared to ordered subset expectation maximization (OSEM) reconstruction.MethodsFifty-seven patients with 92 ^18F-FDG-avid pulmonary nodules (all ≤ 2 cm) undergoing PET/CT for oncological (re-)staging were retrospectively included and a total of 8824 PET images of the lungs were extracted using OSEM and BSREM reconstruction. Per-slice and per-nodule sensitivity of a deep learning algorithm was assessed, with an expert readout by a radiologist/Nuclear Medicine Physician serving as standard of reference. Receiver-operator characteristic (ROC) curve of OSEM and BSREM were assessed and the areas under the ROC curve (AUC) were compared. A maximum standardized uptake value (SUV_max)–based sensitivity analysis and a size-based sensitivity analysis with subgroups defined by nodule size was performed.ResultsThe AUC of the deep learning algorithm for nodule detection using OSEM reconstruction was 0.796 (CI 95%; 0.772–0.869), and 0.848 (CI 95%; 0.828–0.869) using BSREM reconstruction. The AUC was significantly higher for BSREM compared to OSEM ( p = 0.001). On a per-slice analysis, sensitivity and specificity were 66.7% and 79.0% for OSEM, and 69.2% and 84.5% for BSREM. On a per-nodule analysis, the overall sensitivity of OSEM was 81.5% compared to 87.0% for BSREM.ConclusionsOur results suggest that machine learning algorithms may aid detection of small ^18F-FDG-avid pulmonary nodules in clinical PET/CT. AI performed significantly better on images with BSREM than OSEM.Key Points• The diagnostic value of deep learning for detecting small lung nodules (≤ 2 cm) in PET images using BSREM and OSEM reconstruction was assessed. • BSREM yields higher SUV _ max of small pulmonary nodules as compared to OSEM reconstruction. • The use of BSREM translates into a higher detectability of small pulmonary nodules in PET images as assessed with artificial intelligence .
Dominik C. Benz - One of the best experts on this subject based on the ideXlab platform.
-
artificial intelligence for detecting small fdg positive lung nodules in digital pet ct impact of image reconstructions on diagnostic performance
2020Co-Authors: Moritz Schwyzer, Katharina Martini, Dominik C. Benz, Daniela A. Ferraro, Ken Kudura, Philipp A Kaufmann, Irene A. Burger, Valerie Treyer, Gustav K Von Schulthess, Martin W HuellnerAbstract:To evaluate the diagnostic performance of a deep learning algorithm for automated detection of small 18F-FDG-avid pulmonary nodules in PET scans, and to assess whether novel block sequential regularized expectation maximization (BSREM) reconstruction affects detection accuracy as compared to ordered subset expectation maximization (OSEM) reconstruction. Fifty-seven patients with 92 18F-FDG-avid pulmonary nodules (all ≤ 2 cm) undergoing PET/CT for oncological (re-)staging were retrospectively included and a total of 8824 PET images of the lungs were extracted using OSEM and BSREM reconstruction. Per-slice and per-nodule sensitivity of a deep learning algorithm was assessed, with an expert readout by a radiologist/Nuclear Medicine Physician serving as standard of reference. Receiver-operator characteristic (ROC) curve of OSEM and BSREM were assessed and the areas under the ROC curve (AUC) were compared. A maximum standardized uptake value (SUVmax)–based sensitivity analysis and a size-based sensitivity analysis with subgroups defined by nodule size was performed. The AUC of the deep learning algorithm for nodule detection using OSEM reconstruction was 0.796 (CI 95%; 0.772–0.869), and 0.848 (CI 95%; 0.828–0.869) using BSREM reconstruction. The AUC was significantly higher for BSREM compared to OSEM (p = 0.001). On a per-slice analysis, sensitivity and specificity were 66.7% and 79.0% for OSEM, and 69.2% and 84.5% for BSREM. On a per-nodule analysis, the overall sensitivity of OSEM was 81.5% compared to 87.0% for BSREM. Our results suggest that machine learning algorithms may aid detection of small 18F-FDG-avid pulmonary nodules in clinical PET/CT. AI performed significantly better on images with BSREM than OSEM. • The diagnostic value of deep learning for detecting small lung nodules (≤ 2 cm) in PET images using BSREM and OSEM reconstruction was assessed. • BSREM yields higher SUV max of small pulmonary nodules as compared to OSEM reconstruction. • The use of BSREM translates into a higher detectability of small pulmonary nodules in PET images as assessed with artificial intelligence.
-
Artificial intelligence for detecting small FDG-positive lung nodules in digital PET/CT: impact of image reconstructions on diagnostic performance
2019Co-Authors: Moritz Schwyzer, Katharina Martini, Dominik C. Benz, Daniela A. Ferraro, Ken Kudura, Philipp A Kaufmann, Gustav K. Von Schulthess, Irene A. Burger, Valerie Treyer, Martin W HuellnerAbstract:ObjectivesTo evaluate the diagnostic performance of a deep learning algorithm for automated detection of small ^18F-FDG-avid pulmonary nodules in PET scans, and to assess whether novel block sequential regularized expectation maximization (BSREM) reconstruction affects detection accuracy as compared to ordered subset expectation maximization (OSEM) reconstruction.MethodsFifty-seven patients with 92 ^18F-FDG-avid pulmonary nodules (all ≤ 2 cm) undergoing PET/CT for oncological (re-)staging were retrospectively included and a total of 8824 PET images of the lungs were extracted using OSEM and BSREM reconstruction. Per-slice and per-nodule sensitivity of a deep learning algorithm was assessed, with an expert readout by a radiologist/Nuclear Medicine Physician serving as standard of reference. Receiver-operator characteristic (ROC) curve of OSEM and BSREM were assessed and the areas under the ROC curve (AUC) were compared. A maximum standardized uptake value (SUV_max)–based sensitivity analysis and a size-based sensitivity analysis with subgroups defined by nodule size was performed.ResultsThe AUC of the deep learning algorithm for nodule detection using OSEM reconstruction was 0.796 (CI 95%; 0.772–0.869), and 0.848 (CI 95%; 0.828–0.869) using BSREM reconstruction. The AUC was significantly higher for BSREM compared to OSEM ( p = 0.001). On a per-slice analysis, sensitivity and specificity were 66.7% and 79.0% for OSEM, and 69.2% and 84.5% for BSREM. On a per-nodule analysis, the overall sensitivity of OSEM was 81.5% compared to 87.0% for BSREM.ConclusionsOur results suggest that machine learning algorithms may aid detection of small ^18F-FDG-avid pulmonary nodules in clinical PET/CT. AI performed significantly better on images with BSREM than OSEM.Key Points• The diagnostic value of deep learning for detecting small lung nodules (≤ 2 cm) in PET images using BSREM and OSEM reconstruction was assessed. • BSREM yields higher SUV _ max of small pulmonary nodules as compared to OSEM reconstruction. • The use of BSREM translates into a higher detectability of small pulmonary nodules in PET images as assessed with artificial intelligence .
Daniela A. Ferraro - One of the best experts on this subject based on the ideXlab platform.
-
artificial intelligence for detecting small fdg positive lung nodules in digital pet ct impact of image reconstructions on diagnostic performance
2020Co-Authors: Moritz Schwyzer, Katharina Martini, Dominik C. Benz, Daniela A. Ferraro, Ken Kudura, Philipp A Kaufmann, Irene A. Burger, Valerie Treyer, Gustav K Von Schulthess, Martin W HuellnerAbstract:To evaluate the diagnostic performance of a deep learning algorithm for automated detection of small 18F-FDG-avid pulmonary nodules in PET scans, and to assess whether novel block sequential regularized expectation maximization (BSREM) reconstruction affects detection accuracy as compared to ordered subset expectation maximization (OSEM) reconstruction. Fifty-seven patients with 92 18F-FDG-avid pulmonary nodules (all ≤ 2 cm) undergoing PET/CT for oncological (re-)staging were retrospectively included and a total of 8824 PET images of the lungs were extracted using OSEM and BSREM reconstruction. Per-slice and per-nodule sensitivity of a deep learning algorithm was assessed, with an expert readout by a radiologist/Nuclear Medicine Physician serving as standard of reference. Receiver-operator characteristic (ROC) curve of OSEM and BSREM were assessed and the areas under the ROC curve (AUC) were compared. A maximum standardized uptake value (SUVmax)–based sensitivity analysis and a size-based sensitivity analysis with subgroups defined by nodule size was performed. The AUC of the deep learning algorithm for nodule detection using OSEM reconstruction was 0.796 (CI 95%; 0.772–0.869), and 0.848 (CI 95%; 0.828–0.869) using BSREM reconstruction. The AUC was significantly higher for BSREM compared to OSEM (p = 0.001). On a per-slice analysis, sensitivity and specificity were 66.7% and 79.0% for OSEM, and 69.2% and 84.5% for BSREM. On a per-nodule analysis, the overall sensitivity of OSEM was 81.5% compared to 87.0% for BSREM. Our results suggest that machine learning algorithms may aid detection of small 18F-FDG-avid pulmonary nodules in clinical PET/CT. AI performed significantly better on images with BSREM than OSEM. • The diagnostic value of deep learning for detecting small lung nodules (≤ 2 cm) in PET images using BSREM and OSEM reconstruction was assessed. • BSREM yields higher SUV max of small pulmonary nodules as compared to OSEM reconstruction. • The use of BSREM translates into a higher detectability of small pulmonary nodules in PET images as assessed with artificial intelligence.
-
Artificial intelligence for detecting small FDG-positive lung nodules in digital PET/CT: impact of image reconstructions on diagnostic performance
2019Co-Authors: Moritz Schwyzer, Katharina Martini, Dominik C. Benz, Daniela A. Ferraro, Ken Kudura, Philipp A Kaufmann, Gustav K. Von Schulthess, Irene A. Burger, Valerie Treyer, Martin W HuellnerAbstract:ObjectivesTo evaluate the diagnostic performance of a deep learning algorithm for automated detection of small ^18F-FDG-avid pulmonary nodules in PET scans, and to assess whether novel block sequential regularized expectation maximization (BSREM) reconstruction affects detection accuracy as compared to ordered subset expectation maximization (OSEM) reconstruction.MethodsFifty-seven patients with 92 ^18F-FDG-avid pulmonary nodules (all ≤ 2 cm) undergoing PET/CT for oncological (re-)staging were retrospectively included and a total of 8824 PET images of the lungs were extracted using OSEM and BSREM reconstruction. Per-slice and per-nodule sensitivity of a deep learning algorithm was assessed, with an expert readout by a radiologist/Nuclear Medicine Physician serving as standard of reference. Receiver-operator characteristic (ROC) curve of OSEM and BSREM were assessed and the areas under the ROC curve (AUC) were compared. A maximum standardized uptake value (SUV_max)–based sensitivity analysis and a size-based sensitivity analysis with subgroups defined by nodule size was performed.ResultsThe AUC of the deep learning algorithm for nodule detection using OSEM reconstruction was 0.796 (CI 95%; 0.772–0.869), and 0.848 (CI 95%; 0.828–0.869) using BSREM reconstruction. The AUC was significantly higher for BSREM compared to OSEM ( p = 0.001). On a per-slice analysis, sensitivity and specificity were 66.7% and 79.0% for OSEM, and 69.2% and 84.5% for BSREM. On a per-nodule analysis, the overall sensitivity of OSEM was 81.5% compared to 87.0% for BSREM.ConclusionsOur results suggest that machine learning algorithms may aid detection of small ^18F-FDG-avid pulmonary nodules in clinical PET/CT. AI performed significantly better on images with BSREM than OSEM.Key Points• The diagnostic value of deep learning for detecting small lung nodules (≤ 2 cm) in PET images using BSREM and OSEM reconstruction was assessed. • BSREM yields higher SUV _ max of small pulmonary nodules as compared to OSEM reconstruction. • The use of BSREM translates into a higher detectability of small pulmonary nodules in PET images as assessed with artificial intelligence .
