Effectiveness of data-augmentation on deep learning in evaluating rapid on-site cytopathology at endoscopic ultrasound-guided fine needle aspiration.
Journal
Scientific reports
ISSN: 2045-2322
Titre abrégé: Sci Rep
Pays: England
ID NLM: 101563288
Informations de publication
Date de publication:
28 Sep 2024
28 Sep 2024
Historique:
received:
02
11
2023
accepted:
05
09
2024
medline:
29
9
2024
pubmed:
29
9
2024
entrez:
28
9
2024
Statut:
epublish
Résumé
Rapid on-site cytopathology evaluation (ROSE) has been considered an effective method to increase the diagnostic ability of endoscopic ultrasound-guided fine needle aspiration (EUS-FNA); however, ROSE is unavailable in most institutes worldwide due to the shortage of cytopathologists. To overcome this situation, we created an artificial intelligence (AI)-based system (the ROSE-AI system), which was trained with the augmented data to evaluate the slide images acquired by EUS-FNA. This study aimed to clarify the effects of such data-augmentation on establishing an effective ROSE-AI system by comparing the efficacy of various data-augmentation techniques. The ROSE-AI system was trained with increased data obtained by the various data-augmentation techniques, including geometric transformation, color space transformation, and kernel filtering. By performing five-fold cross-validation, we compared the efficacy of each data-augmentation technique on the increasing diagnostic abilities of the ROSE-AI system. We collected 4059 divided EUS-FNA slide images from 36 patients with pancreatic cancer and nine patients with non-pancreatic cancer. The diagnostic ability of the ROSE-AI system without data augmentation had a sensitivity, specificity, and accuracy of 87.5%, 79.7%, and 83.7%, respectively. While, some data-augmentation techniques decreased diagnostic ability, the ROSE-AI system trained only with the augmented data using the geometric transformation technique had the highest diagnostic accuracy (88.2%). We successfully developed a prototype ROSE-AI system with high diagnostic ability. Each data-augmentation technique may have various compatibilities with AI-mediated diagnostics, and the geometric transformation was the most effective for the ROSE-AI system.
Identifiants
pubmed: 39341885
doi: 10.1038/s41598-024-72312-3
pii: 10.1038/s41598-024-72312-3
doi:
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
22441Subventions
Organisme : JSPS KAKENHI
ID : 24K18948
Organisme : JSPS KAKENHI
ID : 23K11932
Informations de copyright
© 2024. The Author(s).
Références
Yoshinaga, S. et al. Safety and efficacy of endoscopic ultrasound-guided fine needle aspiration for pancreatic masses: a prospective multicenter study. Dig. Endosc. 32, 114–126 (2020).
doi: 10.1111/den.13457
pubmed: 31166046
Dumonceau, J. M. et al. Indications, results, and clinical impact of endoscopic ultrasound (EUS)-guided sampling in gastroenterology: European Society of Gastrointestinal Endoscopy (ESGE) clinical guideline-updated January 2017. Endoscopy 49(07), 695–714 (2017).
doi: 10.1055/s-0043-109021
pubmed: 28511234
Bang, J. Y., Hawes, R. & Varadarajulu, S. A meta-analysis comparing Procore and standard fine-needle aspiration needles for endoscopic ultrasound-guided tissue acquisition. Endoscopy 48, 339–349 (2016).
pubmed: 26561917
Hawes, R. H. The evolution of endoscopic ultrasound: improved imaging, higher accuracy for fine needle aspiration and the reality of endoscopic ultrasound-guided interventions. Curr. Opin. Gastroenterol. 26, 436–444 (2010).
pubmed: 20703111
Schmidt, R. L., Walker, B. S., Howard, K., Layfield, L. J. & Adler, D. G. Rapid on-site evaluation reduces needle passes in endoscopic ultrasound-guided fine-needle aspiration for solid pancreatic lesions: a risk benefit analysis. Dig. Dis. Sci. 58, 3280–3286 (2013).
doi: 10.1007/s10620-013-2750-6
pubmed: 23824404
Matynia, A. P. et al. Impact of rapid on-site evaluation on the adequacy of endoscopic-ultrasound guided fine-needle aspiration of solid pancreatic lesions: a systematic review and meta-analysis. J. Gastroenterol. Hepatol. 29, 697–705 (2014).
doi: 10.1111/jgh.12431
pubmed: 24783248
Lewin, D. Optimal EUS-guided FNA cytology preparation when rapid on-site evaluation is not available. Gastrointest Endosc. 91, 847–848 (2020).
doi: 10.1016/j.gie.2019.12.009
pubmed: 32204817
Xu, Y. et al. Comparison of diagnostic performance between convolutional neural networks and human endoscopists for diagnosis of colorectal polyp: a systematic review and meta-analysis. PLoS ONE 16, e0246892 (2021).
doi: 10.1371/journal.pone.0246892
pubmed: 33592048
pmcid: 7886136
Lin, R. et al. Application of artificial intelligence to digital-rapid on-site cytopathology evaluation during endoscopic ultrasound-guided fine needle aspiration: a proof-of-concept study. J. Gastroenterol. Hepatol. 10, 16073 (2022).
Zhang, S. et al. A deep learning-based segmentation system for rapid onsite cytologic pathology evaluation of pancreatic masses: A retrospective, multicenter, diagnostic study. EBiomedicine 80, 104022 (2022).
doi: 10.1016/j.ebiom.2022.104022
pubmed: 35512608
pmcid: 9079232
Kebaili, A., Lapuyade-Lahorgue, J. & Ruan, S. Deep learning approaches for data augmentation in medical imaging: a review. J. Imaging 9, 9 (2023).
doi: 10.3390/jimaging9040081
Chlap, P. et al. A review of medical image data augmentation techniques for deep learning applications. J. Med. Imaging Radiat. Oncol. 65, 545–563 (2021).
doi: 10.1111/1754-9485.13261
pubmed: 34145766
Dosoviskiy, A. et al. An image is worth 16x16 words: transformers for image recognition at scale. Arxiv 2010, 11929 (2020).
Chefer Hila, Gur S, Wolf L. Transformer interpretability beyond attention visualization. 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Nashville, TN, USA. 782–91 (2021).
Soumith, C. PyTorch RandomPerspective. PyTorch https://pytorch.org/vision/main/generated/torchvision.transforms.RandomPerspective.html (2017).
Soumith, C. PyTorch RandomRotation. PyTorch https://pytorch.org/vision/main/generated/torchvision.transforms.RandomRotation.html (2017).
Soumith, C. PyTorch RandomHorizontalFlip. PyTorch https://pytorch.org/vision/main/generated/torchvision.transforms.RandomHorizontalFlip.html (2017).
Soumith, C. PyTorch RandomVerticalFlip. PyTorch https://pytorch.org/vision/main/generated/torchvision.transforms.RandomVerticalFlip.html (2017).
Soumith, C. PyTorch GaussNoise. PyTorch https://albumentations.ai/docs/api_reference/augmentations/transforms/#albumentations.augmentations.transforms.GaussNoise (2017).
Soumith, C. PyTorch RandomCrop. PyTorch https://pytorch.org/vision/main/generated/torchvision.transforms.RandomCrop.html (2017).
Soumith, C. PyTorch ColorJitter. PyTorch https://pytorch.org/vision/main/generated/torchvision.transforms.ColorJitter.html (2017).
Soumith, C. PyTorch GaussianBlur. PyTorch https://pytorch.org/vision/0.18/generated/torchvision.transforms.GaussianBlur.html (2017).
Hermsen, M. et al. Deep learning-based histopathologic assessment of kidney tissue. J. Am. Soc. Nephrol. 30, 1968–1979 (2021).
doi: 10.1681/ASN.2019020144
Sanyal, P., Mukherjee, T., Barui, S., Das, A. & Gangopadhyay, P. Artificial intelligence in cytopathology: a neural network to identify papillary carcinoma on thyroid fine-needle aspiration cytology smears. J. Pathol. Inform. 9, 43 (2018).
doi: 10.4103/jpi.jpi_43_18
pubmed: 30607310
pmcid: 6289006
de Souza, L. A. et al. Assisting Barrett’s esophagus identification using endoscopic data augmentation based on Generative Adversarial Networks. Comput. Biol. Med. 126, 104029 (2020).
doi: 10.1016/j.compbiomed.2020.104029
pubmed: 33059236
Adjei, P. E., Lonseko, Z. M., Du, W., Zhang, H. & Rao, N. Examining the effect of synthetic data augmentation in polyp detection and segmentation. Int. J. Comput. Assist. Radiol. Surg. 17, 1289–1302 (2022).
doi: 10.1007/s11548-022-02651-x
pubmed: 35678960
Nozaka, H. et al. The effect of data augmentation in deep learning approach for peripheral blood leukocyte recognition. Stud. Health Technol. Inform. 290, 273–277 (2022).
pubmed: 35673016
Monshi, M. M. A., Poon, J., Chung, V. & Monshi, F. M. CovidXrayNet: Optimizing data augmentation and CNN hyperparameters for improved COVID-19 detection from CXR. Comput. Biol. Med. 133, 104375 (2021).
doi: 10.1016/j.compbiomed.2021.104375
pubmed: 33866253
pmcid: 8048393
Hao, R., Namdar, K., Liu, L., Haider, M. A. & Khalvati, F. A Comprehensive study of data augmentation strategies for prostate cancer detection in diffusion-weighted MRI using convolutional neural networks. J. Digit Imaging 34, 862–876 (2021).
doi: 10.1007/s10278-021-00478-7
pubmed: 34254200
pmcid: 8455796
Lococo, F. et al. Implementation of artificial intelligence in personalized prognostic assessment of lung cancer: a narrative review. Cancers 16(10), 1832 (2024).
doi: 10.3390/cancers16101832
pubmed: 38791910
pmcid: 11119930
Shigeki A. Japan medical image database, https://www.radiology.jp/j-mid/ (2020).
DAMA International. Data Management Body of Knowledge (DMBOK) (2
Keisuke, H. et al. Detecting colon polyps in endoscopic images using artificial intelligence constructed with automated collection of annotated images from an endoscopy reporting system. Dig. Endosc. 34, 1021–1029 (2022).
doi: 10.1111/den.14185