1
|
Peters S, Hertel T. ESTRO congress and AI: (No) more questions! Radiother Oncol 2024; 199:110428. [PMID: 39029589 DOI: 10.1016/j.radonc.2024.110428] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 06/26/2024] [Accepted: 07/01/2024] [Indexed: 07/21/2024]
Affiliation(s)
- Samuel Peters
- Department of Radiation Oncology, Kantonsspital St. Gallen, St. Gallen, Switzerland.
| | - Tanja Hertel
- Department of Radiation Oncology, Kantonsspital St. Gallen, St. Gallen, Switzerland
| |
Collapse
|
2
|
Wahid KA, Kaffey ZY, Farris DP, Humbert-Vidan L, Moreno AC, Rasmussen M, Ren J, Naser MA, Netherton TJ, Korreman S, Balakrishnan G, Fuller CD, Fuentes D, Dohopolski MJ. Artificial intelligence uncertainty quantification in radiotherapy applications - A scoping review. Radiother Oncol 2024; 201:110542. [PMID: 39299574 DOI: 10.1016/j.radonc.2024.110542] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Grants] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 05/11/2024] [Revised: 08/18/2024] [Accepted: 09/09/2024] [Indexed: 09/22/2024]
Abstract
BACKGROUND/PURPOSE The use of artificial intelligence (AI) in radiotherapy (RT) is expanding rapidly. However, there exists a notable lack of clinician trust in AI models, underscoring the need for effective uncertainty quantification (UQ) methods. The purpose of this study was to scope existing literature related to UQ in RT, identify areas of improvement, and determine future directions. METHODS We followed the PRISMA-ScR scoping review reporting guidelines. We utilized the population (human cancer patients), concept (utilization of AI UQ), context (radiotherapy applications) framework to structure our search and screening process. We conducted a systematic search spanning seven databases, supplemented by manual curation, up to January 2024. Our search yielded a total of 8980 articles for initial review. Manuscript screening and data extraction was performed in Covidence. Data extraction categories included general study characteristics, RT characteristics, AI characteristics, and UQ characteristics. RESULTS We identified 56 articles published from 2015 to 2024. 10 domains of RT applications were represented; most studies evaluated auto-contouring (50 %), followed by image-synthesis (13 %), and multiple applications simultaneously (11 %). 12 disease sites were represented, with head and neck cancer being the most common disease site independent of application space (32 %). Imaging data was used in 91 % of studies, while only 13 % incorporated RT dose information. Most studies focused on failure detection as the main application of UQ (60 %), with Monte Carlo dropout being the most commonly implemented UQ method (32 %) followed by ensembling (16 %). 55 % of studies did not share code or datasets. CONCLUSION Our review revealed a lack of diversity in UQ for RT applications beyond auto-contouring. Moreover, we identified a clear need to study additional UQ methods, such as conformal prediction. Our results may incentivize the development of guidelines for reporting and implementation of UQ in RT.
Collapse
Affiliation(s)
- Kareem A Wahid
- Department of Imaging Physics, The University of Texas MD Anderson Cancer Center, Houston, TX, USA; Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
| | - Zaphanlene Y Kaffey
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
| | - David P Farris
- Research Medical Library, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
| | - Laia Humbert-Vidan
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
| | - Amy C Moreno
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
| | | | - Jintao Ren
- Department of Oncology, Aarhus University Hospital, Denmark
| | - Mohamed A Naser
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
| | - Tucker J Netherton
- Department of Radiation Physics, University of Texas MD Anderson Cancer Center, Houston, TX, USA
| | - Stine Korreman
- Department of Oncology, Aarhus University Hospital, Denmark
| | | | - Clifton D Fuller
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
| | - David Fuentes
- Department of Imaging Physics, The University of Texas MD Anderson Cancer Center, Houston, TX, USA.
| | - Michael J Dohopolski
- Department of Radiation Oncology, The University of Texas Southwestern Medical Center, Dallas, TX, USA.
| |
Collapse
|
3
|
Chlap P, Min H, Dowling J, Field M, Cloak K, Leong T, Lee M, Chu J, Tan J, Tran P, Kron T, Sidhom M, Wiltshire K, Keats S, Kneebone A, Haworth A, Ebert MA, Vinod SK, Holloway L. Uncertainty estimation using a 3D probabilistic U-Net for segmentation with small radiotherapy clinical trial datasets. Comput Med Imaging Graph 2024; 116:102403. [PMID: 38878632 DOI: 10.1016/j.compmedimag.2024.102403] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [MESH Headings] [Track Full Text] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/02/2023] [Revised: 03/17/2024] [Accepted: 05/21/2024] [Indexed: 09/02/2024]
Abstract
BACKGROUND AND OBJECTIVES Bio-medical image segmentation models typically attempt to predict one segmentation that resembles a ground-truth structure as closely as possible. However, as medical images are not perfect representations of anatomy, obtaining this ground truth is not possible. A surrogate commonly used is to have multiple expert observers define the same structure for a dataset. When multiple observers define the same structure on the same image there can be significant differences depending on the structure, image quality/modality and the region being defined. It is often desirable to estimate this type of aleatoric uncertainty in a segmentation model to help understand the region in which the true structure is likely to be positioned. Furthermore, obtaining these datasets is resource intensive so training such models using limited data may be required. With a small dataset size, differing patient anatomy is likely not well represented causing epistemic uncertainty which should also be estimated so it can be determined for which cases the model is effective or not. METHODS We use a 3D probabilistic U-Net to train a model from which several segmentations can be sampled to estimate the range of uncertainty seen between multiple observers. To ensure that regions where observers disagree most are emphasised in model training, we expand the Generalised Evidence Lower Bound (ELBO) with a Constrained Optimisation (GECO) loss function with an additional contour loss term to give attention to this region. Ensemble and Monte-Carlo dropout (MCDO) uncertainty quantification methods are used during inference to estimate model confidence on an unseen case. We apply our methodology to two radiotherapy clinical trial datasets, a gastric cancer trial (TOPGEAR, TROG 08.08) and a post-prostatectomy prostate cancer trial (RAVES, TROG 08.03). Each dataset contains only 10 cases each for model development to segment the clinical target volume (CTV) which was defined by multiple observers on each case. An additional 50 cases are available as a hold-out dataset for each trial which had only one observer define the CTV structure on each case. Up to 50 samples were generated using the probabilistic model for each case in the hold-out dataset. To assess performance, each manually defined structure was matched to the closest matching sampled segmentation based on commonly used metrics. RESULTS The TOPGEAR CTV model achieved a Dice Similarity Coefficient (DSC) and Surface DSC (sDSC) of 0.7 and 0.43 respectively with the RAVES model achieving 0.75 and 0.71 respectively. Segmentation quality across cases in the hold-out datasets was variable however both the ensemble and MCDO uncertainty estimation approaches were able to accurately estimate model confidence with a p-value < 0.001 for both TOPGEAR and RAVES when comparing the DSC using the Pearson correlation coefficient. CONCLUSIONS We demonstrated that training auto-segmentation models which can estimate aleatoric and epistemic uncertainty using limited datasets is possible. Having the model estimate prediction confidence is important to understand for which unseen cases a model is likely to be useful.
Collapse
Affiliation(s)
- Phillip Chlap
- University of New South Wales, South Western Sydney Clinical School, Sydney, Australia; Ingham Institute for Applied Medical Research, Sydney, Australia; Liverpool and Macarthur Cancer Therapy Centres, Department of Radiation Oncology, Sydney, Australia.
| | - Hang Min
- University of New South Wales, South Western Sydney Clinical School, Sydney, Australia; Ingham Institute for Applied Medical Research, Sydney, Australia; CSIRO Australian e-Health Research Centre, Herston, Australia
| | - Jason Dowling
- University of New South Wales, South Western Sydney Clinical School, Sydney, Australia; Ingham Institute for Applied Medical Research, Sydney, Australia; CSIRO Australian e-Health Research Centre, Herston, Australia
| | - Matthew Field
- University of New South Wales, South Western Sydney Clinical School, Sydney, Australia; Ingham Institute for Applied Medical Research, Sydney, Australia; Liverpool and Macarthur Cancer Therapy Centres, Department of Radiation Oncology, Sydney, Australia
| | - Kirrily Cloak
- University of New South Wales, South Western Sydney Clinical School, Sydney, Australia; Ingham Institute for Applied Medical Research, Sydney, Australia; Liverpool and Macarthur Cancer Therapy Centres, Department of Radiation Oncology, Sydney, Australia
| | - Trevor Leong
- Peter MacCallum Cancer Centre, Melbourne, Australia; The Sir Peter MacCallum Department of Oncology, University of Melbourne, Melbourne, Australia
| | - Mark Lee
- Ingham Institute for Applied Medical Research, Sydney, Australia
| | - Julie Chu
- Peter MacCallum Cancer Centre, Melbourne, Australia
| | - Jennifer Tan
- Peter MacCallum Cancer Centre, Melbourne, Australia
| | - Phillip Tran
- Peter MacCallum Cancer Centre, Melbourne, Australia
| | - Tomas Kron
- Peter MacCallum Cancer Centre, Melbourne, Australia
| | - Mark Sidhom
- University of New South Wales, South Western Sydney Clinical School, Sydney, Australia; Liverpool and Macarthur Cancer Therapy Centres, Department of Radiation Oncology, Sydney, Australia
| | | | - Sarah Keats
- Liverpool and Macarthur Cancer Therapy Centres, Department of Radiation Oncology, Sydney, Australia
| | - Andrew Kneebone
- University of Sydney, Institute of Medical Physics, Sydney, Australia; Northern Sydney Cancer Centre, Sydney, Australia
| | - Annette Haworth
- University of Sydney, Institute of Medical Physics, Sydney, Australia
| | - Martin A Ebert
- School of Physics, Mathematics, and Computing, The University of Western Australia, Crawley, Australia; Department of Radiation Oncology, Sir Charles Gardiner Hospital, Nedlands, Australia; School of Medicine and Population Health, University of Wisconsin, Madison, WI, USA
| | - Shalini K Vinod
- University of New South Wales, South Western Sydney Clinical School, Sydney, Australia; Ingham Institute for Applied Medical Research, Sydney, Australia; Liverpool and Macarthur Cancer Therapy Centres, Department of Radiation Oncology, Sydney, Australia
| | - Lois Holloway
- University of New South Wales, South Western Sydney Clinical School, Sydney, Australia; Ingham Institute for Applied Medical Research, Sydney, Australia; Liverpool and Macarthur Cancer Therapy Centres, Department of Radiation Oncology, Sydney, Australia; University of Sydney, Institute of Medical Physics, Sydney, Australia
| |
Collapse
|
4
|
Wahid KA, Kaffey ZY, Farris DP, Humbert-Vidan L, Moreno AC, Rasmussen M, Ren J, Naser MA, Netherton TJ, Korreman S, Balakrishnan G, Fuller CD, Fuentes D, Dohopolski MJ. Artificial Intelligence Uncertainty Quantification in Radiotherapy Applications - A Scoping Review. MEDRXIV : THE PREPRINT SERVER FOR HEALTH SCIENCES 2024:2024.05.13.24307226. [PMID: 38798581 PMCID: PMC11118597 DOI: 10.1101/2024.05.13.24307226] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Grants] [Track Full Text] [Subscribe] [Scholar Register] [Indexed: 05/29/2024]
Abstract
Background/purpose The use of artificial intelligence (AI) in radiotherapy (RT) is expanding rapidly. However, there exists a notable lack of clinician trust in AI models, underscoring the need for effective uncertainty quantification (UQ) methods. The purpose of this study was to scope existing literature related to UQ in RT, identify areas of improvement, and determine future directions. Methods We followed the PRISMA-ScR scoping review reporting guidelines. We utilized the population (human cancer patients), concept (utilization of AI UQ), context (radiotherapy applications) framework to structure our search and screening process. We conducted a systematic search spanning seven databases, supplemented by manual curation, up to January 2024. Our search yielded a total of 8980 articles for initial review. Manuscript screening and data extraction was performed in Covidence. Data extraction categories included general study characteristics, RT characteristics, AI characteristics, and UQ characteristics. Results We identified 56 articles published from 2015-2024. 10 domains of RT applications were represented; most studies evaluated auto-contouring (50%), followed by image-synthesis (13%), and multiple applications simultaneously (11%). 12 disease sites were represented, with head and neck cancer being the most common disease site independent of application space (32%). Imaging data was used in 91% of studies, while only 13% incorporated RT dose information. Most studies focused on failure detection as the main application of UQ (60%), with Monte Carlo dropout being the most commonly implemented UQ method (32%) followed by ensembling (16%). 55% of studies did not share code or datasets. Conclusion Our review revealed a lack of diversity in UQ for RT applications beyond auto-contouring. Moreover, there was a clear need to study additional UQ methods, such as conformal prediction. Our results may incentivize the development of guidelines for reporting and implementation of UQ in RT.
Collapse
Affiliation(s)
- Kareem A. Wahid
- Department of Imaging Physics, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
| | - Zaphanlene Y. Kaffey
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
| | - David P. Farris
- Research Medical Library, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
| | - Laia Humbert-Vidan
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
| | - Amy C. Moreno
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
| | | | - Jintao Ren
- Department of Oncology, Aarhus University Hospital, Denmark
| | - Mohamed A. Naser
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
| | - Tucker J. Netherton
- Department of Radiation Physics, University of Texas MD Anderson Cancer Center, Houston, TX, USA
| | - Stine Korreman
- Department of Oncology, Aarhus University Hospital, Denmark
| | | | - Clifton D. Fuller
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
| | - David Fuentes
- Department of Imaging Physics, The University of Texas MD Anderson Cancer Center, Houston, Texas, USA
| | - Michael J. Dohopolski
- Department of Radiation Oncology, The University of Texas Southwestern Medical Center, Dallas, Texas, USA
| |
Collapse
|
5
|
Wahid KA, Sahlsten J, Jaskari J, Dohopolski MJ, Kaski K, He R, Glerean E, Kann BH, Mäkitie A, Fuller CD, Naser MA, Fuentes D. Harnessing uncertainty in radiotherapy auto-segmentation quality assurance. Phys Imaging Radiat Oncol 2024; 29:100526. [PMID: 38179210 PMCID: PMC10765294 DOI: 10.1016/j.phro.2023.100526] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 11/05/2023] [Accepted: 12/13/2023] [Indexed: 01/06/2024] Open
Affiliation(s)
- Kareem A. Wahid
- Department of Imaging Physics, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
| | - Jaakko Sahlsten
- Department of Computer Science, Aalto University School of Science, Espoo, Finland
| | - Joel Jaskari
- Department of Computer Science, Aalto University School of Science, Espoo, Finland
| | - Michael J. Dohopolski
- Department of Radiation Oncology, The University of Texas Southwestern Medical Center, Dallas, TX, USA
| | - Kimmo Kaski
- Department of Computer Science, Aalto University School of Science, Espoo, Finland
| | - Renjie He
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
| | - Enrico Glerean
- Department of Neuroscience and Biomedical Engineering, Aalto University School of Science, Espoo, Finland
| | - Benjamin H. Kann
- Artificial Intelligence in Medicine Program, Brigham and Women’s Hospital, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA
| | - Antti Mäkitie
- Department of Otorhinolaryngology, Head and Neck Surgery, University of Helsinki and Helsinki University Hospital, Research Program in Systems Oncology, University of Helsinki, Helsinki, Finland
| | - Clifton D. Fuller
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
| | - Mohamed A. Naser
- Department of Radiation Oncology, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
| | - David Fuentes
- Department of Imaging Physics, The University of Texas MD Anderson Cancer Center, Houston, TX, USA
| |
Collapse
|
6
|
Lorenzen EL, Celik B, Sarup N, Dysager L, Christiansen RL, Bertelsen AS, Bernchou U, Agergaard SN, Konrad ML, Brink C, Mahmood F, Schytte T, Nyborg CJ. An open-source nnU-net algorithm for automatic segmentation of MRI scans in the male pelvis for adaptive radiotherapy. Front Oncol 2023; 13:1285725. [PMID: 38023233 PMCID: PMC10654998 DOI: 10.3389/fonc.2023.1285725] [Citation(s) in RCA: 0] [Impact Index Per Article: 0] [Reference Citation Analysis] [Abstract] [Key Words] [Track Full Text] [Figures] [Journal Information] [Subscribe] [Scholar Register] [Received: 08/30/2023] [Accepted: 10/17/2023] [Indexed: 12/01/2023] Open
Abstract
Background Adaptive MRI-guided radiotherapy (MRIgRT) requires accurate and efficient segmentation of organs and targets on MRI scans. Manual segmentation is time-consuming and variable, while deformable image registration (DIR)-based contour propagation may not account for large anatomical changes. Therefore, we developed and evaluated an automatic segmentation method using the nnU-net framework. Methods The network was trained on 38 patients (76 scans) with localized prostate cancer and tested on 30 patients (60 scans) with localized prostate, metastatic prostate, or bladder cancer treated at a 1.5 T MRI-linac at our institution. The performance of the network was compared with the current clinical workflow based on DIR. The segmentation accuracy was evaluated using the Dice similarity coefficient (DSC), mean surface distance (MSD), and Hausdorff distance (HD) metrics. Results The trained network successfully segmented all 600 structures in the test set. High similarity was obtained for most structures, with 90% of the contours having a DSC above 0.9 and 86% having an MSD below 1 mm. The largest discrepancies were found in the sigmoid and colon structures. Stratified analysis on cancer type showed that the best performance was seen in the same type of patients that the model was trained on (localized prostate). Especially in patients with bladder cancer, the performance was lower for the bladder and the surrounding organs. A complete automatic delineation workflow took approximately 1 minute. Compared with contour transfer based on the clinically used DIR algorithm, the nnU-net performed statistically better across all organs, with the most significant gain in using the nnU-net seen for organs subject to more considerable volumetric changes due to variation in the filling of the rectum, bladder, bowel, and sigmoid. Conclusion We successfully trained and tested a network for automatically segmenting organs and targets for MRIgRT in the male pelvis region. Good test results were seen for the trained nnU-net, with test results outperforming the current clinical practice using DIR-based contour propagation at the 1.5 T MRI-linac. The trained network is sufficiently fast and accurate for clinical use in an online setting for MRIgRT. The model is provided as open-source.
Collapse
Affiliation(s)
- Ebbe Laugaard Lorenzen
- Laboratory of Radiation Physics, Department of Oncology, Odense University Hospital, Odense, Denmark
- Department of Clinical Research, University of Southern Denmark, Odense, Denmark
| | - Bahar Celik
- Laboratory of Radiation Physics, Department of Oncology, Odense University Hospital, Odense, Denmark
| | - Nis Sarup
- Laboratory of Radiation Physics, Department of Oncology, Odense University Hospital, Odense, Denmark
| | - Lars Dysager
- Department of Oncology, Odense University Hospital, Odense, Denmark
| | | | | | - Uffe Bernchou
- Laboratory of Radiation Physics, Department of Oncology, Odense University Hospital, Odense, Denmark
- Department of Clinical Research, University of Southern Denmark, Odense, Denmark
| | - Søren Nielsen Agergaard
- Laboratory of Radiation Physics, Department of Oncology, Odense University Hospital, Odense, Denmark
| | - Maximilian Lukas Konrad
- Laboratory of Radiation Physics, Department of Oncology, Odense University Hospital, Odense, Denmark
| | - Carsten Brink
- Laboratory of Radiation Physics, Department of Oncology, Odense University Hospital, Odense, Denmark
- Department of Clinical Research, University of Southern Denmark, Odense, Denmark
| | - Faisal Mahmood
- Laboratory of Radiation Physics, Department of Oncology, Odense University Hospital, Odense, Denmark
- Department of Clinical Research, University of Southern Denmark, Odense, Denmark
| | - Tine Schytte
- Department of Clinical Research, University of Southern Denmark, Odense, Denmark
- Department of Oncology, Odense University Hospital, Odense, Denmark
| | | |
Collapse
|