Specific features of designing a database for neuro-oncological 3D MRI images to be used in training artificial intelligence
The research was aimed at analyzing current approaches to the organization and design methodology of visualization database built on the basis of computer vision. Such approaches are necessary for effective development of diagnostic systems using artificial intelligence (AI). A training data set of...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | Russian |
| Published: |
Russian Academy of Sciences, Siberian Branch Publishing House
2022-12-01
|
| Series: | Сибирский научный медицинский журнал |
| Subjects: | |
| Online Access: | https://sibmed.elpub.ru/jour/article/view/934 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1839628792914509824 |
|---|---|
| author | E. V. Amelina A. Yu. Letyagin B. N. Tuchinov N. Yu. Tolstokulakov M. E. Amelin E. N. Pavlovsky V. V. Groza S. K. Golushko |
| author_facet | E. V. Amelina A. Yu. Letyagin B. N. Tuchinov N. Yu. Tolstokulakov M. E. Amelin E. N. Pavlovsky V. V. Groza S. K. Golushko |
| author_sort | E. V. Amelina |
| collection | DOAJ |
| description | The research was aimed at analyzing current approaches to the organization and design methodology of visualization database built on the basis of computer vision. Such approaches are necessary for effective development of diagnostic systems using artificial intelligence (AI). A training data set of high quality is a mandatory prerequisite for that. Material and methods. The paper presents the technology for designing an annotated database (SBT Dataset) that contains about 1000 clinical cases based on the archived data acquired by the Federal Neurosurgical Center, Novosibirsk, Russia including data on patients with astrocytoma, glioblastoma, meningioma, neurinoma, and patients with metastases of somatic tumors. Each case is represented by a preoperative MRI. The Results and discussion. The dataset was built (SBT Dataset) containing segmented 3D MRI images of 5 types of brain tumors with 991 verified observations. Each case is represented by four MRI sequences T1-WI, T1C (with Gd-contrast), T2-WI and T2-FLAIR with histological and histochemical postoperative confirmation. Tumors segmentation with verification of the tumor core elements boundaries and perifocal edema was approved by two certified experienced neuroradiologists. Conclusion. The database built during the research is comparable in its volume and quality (verification level) with the state-of-the-art databases. The methodological approaches proposed in this paper were focused on designing the high-quality medical computer vision systems. The database was used to create artificial intelligence systems with the “physician assistant” functions for preoperative MRI diagnostics in neurosurgery. |
| format | Article |
| id | doaj-art-df9d27b8e48e46fa99de96b243755d77 |
| institution | Matheson Library |
| issn | 2410-2512 2410-2520 |
| language | Russian |
| publishDate | 2022-12-01 |
| publisher | Russian Academy of Sciences, Siberian Branch Publishing House |
| record_format | Article |
| series | Сибирский научный медицинский журнал |
| spelling | doaj-art-df9d27b8e48e46fa99de96b243755d772025-07-15T11:26:37ZrusRussian Academy of Sciences, Siberian Branch Publishing HouseСибирский научный медицинский журнал2410-25122410-25202022-12-01426515910.18699/SSMJ20220606464Specific features of designing a database for neuro-oncological 3D MRI images to be used in training artificial intelligenceE. V. Amelina0A. Yu. Letyagin1B. N. Tuchinov2N. Yu. Tolstokulakov3M. E. Amelin4E. N. Pavlovsky5V. V. Groza6S. K. Golushko7Novosibirsk State UniversityNovosibirsk State University; Research Institute of Clinical and Experimental Lymphology – Branch of the Institute of Cytology and Genetics of SB RASNovosibirsk State UniversityNovosibirsk State UniversityNovosibirsk State University; Federal Neurosurgical Center of the Minzdrav of RussiaNovosibirsk State UniversityNovosibirsk State UniversityNovosibirsk State UniversityThe research was aimed at analyzing current approaches to the organization and design methodology of visualization database built on the basis of computer vision. Such approaches are necessary for effective development of diagnostic systems using artificial intelligence (AI). A training data set of high quality is a mandatory prerequisite for that. Material and methods. The paper presents the technology for designing an annotated database (SBT Dataset) that contains about 1000 clinical cases based on the archived data acquired by the Federal Neurosurgical Center, Novosibirsk, Russia including data on patients with astrocytoma, glioblastoma, meningioma, neurinoma, and patients with metastases of somatic tumors. Each case is represented by a preoperative MRI. The Results and discussion. The dataset was built (SBT Dataset) containing segmented 3D MRI images of 5 types of brain tumors with 991 verified observations. Each case is represented by four MRI sequences T1-WI, T1C (with Gd-contrast), T2-WI and T2-FLAIR with histological and histochemical postoperative confirmation. Tumors segmentation with verification of the tumor core elements boundaries and perifocal edema was approved by two certified experienced neuroradiologists. Conclusion. The database built during the research is comparable in its volume and quality (verification level) with the state-of-the-art databases. The methodological approaches proposed in this paper were focused on designing the high-quality medical computer vision systems. The database was used to create artificial intelligence systems with the “physician assistant” functions for preoperative MRI diagnostics in neurosurgery.https://sibmed.elpub.ru/jour/article/view/934mrineuro-oncologyartificial intelligencetumor segmentationclassification of brain tumors |
| spellingShingle | E. V. Amelina A. Yu. Letyagin B. N. Tuchinov N. Yu. Tolstokulakov M. E. Amelin E. N. Pavlovsky V. V. Groza S. K. Golushko Specific features of designing a database for neuro-oncological 3D MRI images to be used in training artificial intelligence Сибирский научный медицинский журнал mri neuro-oncology artificial intelligence tumor segmentation classification of brain tumors |
| title | Specific features of designing a database for neuro-oncological 3D MRI images to be used in training artificial intelligence |
| title_full | Specific features of designing a database for neuro-oncological 3D MRI images to be used in training artificial intelligence |
| title_fullStr | Specific features of designing a database for neuro-oncological 3D MRI images to be used in training artificial intelligence |
| title_full_unstemmed | Specific features of designing a database for neuro-oncological 3D MRI images to be used in training artificial intelligence |
| title_short | Specific features of designing a database for neuro-oncological 3D MRI images to be used in training artificial intelligence |
| title_sort | specific features of designing a database for neuro oncological 3d mri images to be used in training artificial intelligence |
| topic | mri neuro-oncology artificial intelligence tumor segmentation classification of brain tumors |
| url | https://sibmed.elpub.ru/jour/article/view/934 |
| work_keys_str_mv | AT evamelina specificfeaturesofdesigningadatabaseforneurooncological3dmriimagestobeusedintrainingartificialintelligence AT ayuletyagin specificfeaturesofdesigningadatabaseforneurooncological3dmriimagestobeusedintrainingartificialintelligence AT bntuchinov specificfeaturesofdesigningadatabaseforneurooncological3dmriimagestobeusedintrainingartificialintelligence AT nyutolstokulakov specificfeaturesofdesigningadatabaseforneurooncological3dmriimagestobeusedintrainingartificialintelligence AT meamelin specificfeaturesofdesigningadatabaseforneurooncological3dmriimagestobeusedintrainingartificialintelligence AT enpavlovsky specificfeaturesofdesigningadatabaseforneurooncological3dmriimagestobeusedintrainingartificialintelligence AT vvgroza specificfeaturesofdesigningadatabaseforneurooncological3dmriimagestobeusedintrainingartificialintelligence AT skgolushko specificfeaturesofdesigningadatabaseforneurooncological3dmriimagestobeusedintrainingartificialintelligence |