Automatic deep learning method for third lumbar selection and body composition evaluation on CT scans of cancer patients
IntroductionThe importance of body composition and sarcopenia is well-recognized in cancer patient outcomes and treatment tolerance, yet routine evaluations are rare due to their time-intensive nature. While CT scans provide accurate measurements, they depend on manual processes. We developed and va...
| Main Authors: | , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2024-01-01
|
| Series: | Frontiers in Nuclear Medicine |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fnume.2023.1292676/full |
| _version_ | 1797359938543026176 |
|---|---|
| author | Lidia Delrieu Damien Blanc Damien Blanc Amine Bouhamama Fabien Reyal Fabien Reyal Frank Pilleul Victor Racine Anne Sophie Hamy Anne Sophie Hamy Hugo Crochet Timothée Marchal Pierre Etienne Heudel |
| author_facet | Lidia Delrieu Damien Blanc Damien Blanc Amine Bouhamama Fabien Reyal Fabien Reyal Frank Pilleul Victor Racine Anne Sophie Hamy Anne Sophie Hamy Hugo Crochet Timothée Marchal Pierre Etienne Heudel |
| author_sort | Lidia Delrieu |
| collection | DOAJ |
| description | IntroductionThe importance of body composition and sarcopenia is well-recognized in cancer patient outcomes and treatment tolerance, yet routine evaluations are rare due to their time-intensive nature. While CT scans provide accurate measurements, they depend on manual processes. We developed and validated a deep learning algorithm to automatically select and segment abdominal muscles [SM], visceral fat [VAT], and subcutaneous fat [SAT] on CT scans.MethodsA total of 352 CT scans were collected from two cancer centers. The detection of the third lumbar vertebrae and three different body tissues (SM, VAT, and SAT) were annotated manually. The 5-fold cross-validation method was used to develop the algorithm and validate its performance on the training cohort. Results were validated on an external independent group of CT scans.ResultsThe algorithm for automatic L3 slice selection had a mean absolute error of 4 mm for the internal validation dataset and 5.5 mm for the external validation dataset. The median DICE similarity coefficient for body composition was 0.94 for SM, 0.93 for VAT, and 0.86 for SAT in the internal validation dataset whereas it was 0.93 for SM, 0.93 for VAT, and 0.85 for SAT in the external validation dataset. There were high correlation scores with sarcopenia metrics in both internal and external validation datasets.ConclusionsOur deep learning algorithm facilitates routine research use and could be integrated into electronic patient records, enhancing care through better monitoring and the incorporation of targeted supportive measures like exercise and nutrition. |
| first_indexed | 2024-03-08T15:31:10Z |
| format | Article |
| id | doaj.art-efd2737362634356bdb49d1e08e10758 |
| institution | Directory Open Access Journal |
| issn | 2673-8880 |
| language | English |
| last_indexed | 2024-03-08T15:31:10Z |
| publishDate | 2024-01-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Nuclear Medicine |
| spelling | doaj.art-efd2737362634356bdb49d1e08e107582024-01-10T04:50:07ZengFrontiers Media S.A.Frontiers in Nuclear Medicine2673-88802024-01-01310.3389/fnume.2023.12926761292676Automatic deep learning method for third lumbar selection and body composition evaluation on CT scans of cancer patientsLidia Delrieu0Damien Blanc1Damien Blanc2Amine Bouhamama3Fabien Reyal4Fabien Reyal5Frank Pilleul6Victor Racine7Anne Sophie Hamy8Anne Sophie Hamy9Hugo Crochet10Timothée Marchal11Pierre Etienne Heudel12Residual Tumor & Response to Treatment Laboratory, RT2Lab, Translational Research Department, INSERM, U932 Immunity and Cancer, Institut Curie, Paris University, Paris, FranceQuantaCell, Pessac, FranceIMAG, Université de Montpellier, Montpellier, FranceDepartment of Radiology, Centre Léon Bérard, Lyon, FranceResidual Tumor & Response to Treatment Laboratory, RT2Lab, Translational Research Department, INSERM, U932 Immunity and Cancer, Institut Curie, Paris University, Paris, FranceDepartment of Surgical Oncology, Institut Curie, University Paris, Paris, FranceDepartment of Radiology, Centre Léon Bérard, Lyon, FranceQuantaCell, Pessac, FranceResidual Tumor & Response to Treatment Laboratory, RT2Lab, Translational Research Department, INSERM, U932 Immunity and Cancer, Institut Curie, Paris University, Paris, FranceDepartment of Medical Oncology, Institut Curie, University Paris, Paris, FranceData and Artificial Intelligence Team, Centre Léon Bérard, Lyon, FranceDepartment of Supportive Care, Institut Curie, Paris, FranceDepartment of Medical Oncology, Centre Léon Bérard, Lyon, FranceIntroductionThe importance of body composition and sarcopenia is well-recognized in cancer patient outcomes and treatment tolerance, yet routine evaluations are rare due to their time-intensive nature. While CT scans provide accurate measurements, they depend on manual processes. We developed and validated a deep learning algorithm to automatically select and segment abdominal muscles [SM], visceral fat [VAT], and subcutaneous fat [SAT] on CT scans.MethodsA total of 352 CT scans were collected from two cancer centers. The detection of the third lumbar vertebrae and three different body tissues (SM, VAT, and SAT) were annotated manually. The 5-fold cross-validation method was used to develop the algorithm and validate its performance on the training cohort. Results were validated on an external independent group of CT scans.ResultsThe algorithm for automatic L3 slice selection had a mean absolute error of 4 mm for the internal validation dataset and 5.5 mm for the external validation dataset. The median DICE similarity coefficient for body composition was 0.94 for SM, 0.93 for VAT, and 0.86 for SAT in the internal validation dataset whereas it was 0.93 for SM, 0.93 for VAT, and 0.85 for SAT in the external validation dataset. There were high correlation scores with sarcopenia metrics in both internal and external validation datasets.ConclusionsOur deep learning algorithm facilitates routine research use and could be integrated into electronic patient records, enhancing care through better monitoring and the incorporation of targeted supportive measures like exercise and nutrition.https://www.frontiersin.org/articles/10.3389/fnume.2023.1292676/fullbody compositioncomputed tomographydeep learningsarcopeniacancer |
| spellingShingle | Lidia Delrieu Damien Blanc Damien Blanc Amine Bouhamama Fabien Reyal Fabien Reyal Frank Pilleul Victor Racine Anne Sophie Hamy Anne Sophie Hamy Hugo Crochet Timothée Marchal Pierre Etienne Heudel Automatic deep learning method for third lumbar selection and body composition evaluation on CT scans of cancer patients Frontiers in Nuclear Medicine body composition computed tomography deep learning sarcopenia cancer |
| title | Automatic deep learning method for third lumbar selection and body composition evaluation on CT scans of cancer patients |
| title_full | Automatic deep learning method for third lumbar selection and body composition evaluation on CT scans of cancer patients |
| title_fullStr | Automatic deep learning method for third lumbar selection and body composition evaluation on CT scans of cancer patients |
| title_full_unstemmed | Automatic deep learning method for third lumbar selection and body composition evaluation on CT scans of cancer patients |
| title_short | Automatic deep learning method for third lumbar selection and body composition evaluation on CT scans of cancer patients |
| title_sort | automatic deep learning method for third lumbar selection and body composition evaluation on ct scans of cancer patients |
| topic | body composition computed tomography deep learning sarcopenia cancer |
| url | https://www.frontiersin.org/articles/10.3389/fnume.2023.1292676/full |
| work_keys_str_mv | AT lidiadelrieu automaticdeeplearningmethodforthirdlumbarselectionandbodycompositionevaluationonctscansofcancerpatients AT damienblanc automaticdeeplearningmethodforthirdlumbarselectionandbodycompositionevaluationonctscansofcancerpatients AT damienblanc automaticdeeplearningmethodforthirdlumbarselectionandbodycompositionevaluationonctscansofcancerpatients AT aminebouhamama automaticdeeplearningmethodforthirdlumbarselectionandbodycompositionevaluationonctscansofcancerpatients AT fabienreyal automaticdeeplearningmethodforthirdlumbarselectionandbodycompositionevaluationonctscansofcancerpatients AT fabienreyal automaticdeeplearningmethodforthirdlumbarselectionandbodycompositionevaluationonctscansofcancerpatients AT frankpilleul automaticdeeplearningmethodforthirdlumbarselectionandbodycompositionevaluationonctscansofcancerpatients AT victorracine automaticdeeplearningmethodforthirdlumbarselectionandbodycompositionevaluationonctscansofcancerpatients AT annesophiehamy automaticdeeplearningmethodforthirdlumbarselectionandbodycompositionevaluationonctscansofcancerpatients AT annesophiehamy automaticdeeplearningmethodforthirdlumbarselectionandbodycompositionevaluationonctscansofcancerpatients AT hugocrochet automaticdeeplearningmethodforthirdlumbarselectionandbodycompositionevaluationonctscansofcancerpatients AT timotheemarchal automaticdeeplearningmethodforthirdlumbarselectionandbodycompositionevaluationonctscansofcancerpatients AT pierreetienneheudel automaticdeeplearningmethodforthirdlumbarselectionandbodycompositionevaluationonctscansofcancerpatients |