A deep-learning approach at automated detection of electron-dense immune deposits in medical renal biopsies
Introduction: Identification of electron-dense immune deposits in electron microscopy (EM) images is integral to the diagnosis of medical renal disease. Deep learning has the potential to augment this process, especially in areas with limited resources. Objectives: Our study explores the feasibility...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Society of Diabetic Nephropathy Prevention
2022-07-01
|
| Series: | Journal of Nephropathology |
| Subjects: | |
| Online Access: | https://nephropathol.com/PDF/jnp-11-e17123.pdf |
| _version_ | 1797828047150252032 |
|---|---|
| author | Alaa Alsadi Nasma K. Majeed Dereen M. Saeed Yash Dharmamer Manmeet B. Singh Tushar N. Patel |
| author_facet | Alaa Alsadi Nasma K. Majeed Dereen M. Saeed Yash Dharmamer Manmeet B. Singh Tushar N. Patel |
| author_sort | Alaa Alsadi |
| collection | DOAJ |
| description | Introduction: Identification of electron-dense immune deposits in electron microscopy (EM) images is integral to the diagnosis of medical renal disease. Deep learning has the potential to augment this process, especially in areas with limited resources. Objectives: Our study explores the feasibility of applying deep learning to detect electron dense immune deposits in electron microscopy images from medical renal biopsies. Patients and Methods: EM images (N=900) from native and transplant kidney biopsies were processed into 4530 tiles (512 x 512 pixels). These tiles were reviewed and classified into one of three categories: deposits absent, deposits present, and indeterminate. This classification resulted in 1255 images with consensus agreement for deposits present and deposits absent. These 1255 images were then used to train a machine learning model, using 1006 images for training, and 249 images for testing. Results: The overall accuracy on the test data was a competitive 78%, and the F1 scores for deposits absent and present was 0.76 and 0.79, respectively. Conclusion: This study demonstrated the feasibility of creating and applying a machine learning model that performs competitively in identifying electron dense deposits in EM images. |
| first_indexed | 2024-04-09T12:58:11Z |
| format | Article |
| id | doaj.art-6daac2e846fd4a2a81404d5e0822af59 |
| institution | Directory Open Access Journal |
| issn | 2251-8363 2251-8819 |
| language | English |
| last_indexed | 2024-04-09T12:58:11Z |
| publishDate | 2022-07-01 |
| publisher | Society of Diabetic Nephropathy Prevention |
| record_format | Article |
| series | Journal of Nephropathology |
| spelling | doaj.art-6daac2e846fd4a2a81404d5e0822af592023-05-13T11:03:13ZengSociety of Diabetic Nephropathy PreventionJournal of Nephropathology2251-83632251-88192022-07-01113e17123e1712310.34172/jnp.2022.17123jnp-17123A deep-learning approach at automated detection of electron-dense immune deposits in medical renal biopsiesAlaa Alsadi0Nasma K. Majeed1Dereen M. Saeed2Yash Dharmamer3Manmeet B. Singh4Tushar N. Patel5Department of Pathology, University of Wisconsin, Madison, Wisconsin, USADepartment of Pathology University of Illinois, Chicago, Illinois, USADepartment of Pathology University of Illinois, Chicago, Illinois, USADepartment of Pathology University of Illinois, Chicago, Illinois, USADepartment of Pathology University of Illinois, Chicago, Illinois, USADepartment of Pathology University of Illinois, Chicago, Illinois, USAIntroduction: Identification of electron-dense immune deposits in electron microscopy (EM) images is integral to the diagnosis of medical renal disease. Deep learning has the potential to augment this process, especially in areas with limited resources. Objectives: Our study explores the feasibility of applying deep learning to detect electron dense immune deposits in electron microscopy images from medical renal biopsies. Patients and Methods: EM images (N=900) from native and transplant kidney biopsies were processed into 4530 tiles (512 x 512 pixels). These tiles were reviewed and classified into one of three categories: deposits absent, deposits present, and indeterminate. This classification resulted in 1255 images with consensus agreement for deposits present and deposits absent. These 1255 images were then used to train a machine learning model, using 1006 images for training, and 249 images for testing. Results: The overall accuracy on the test data was a competitive 78%, and the F1 scores for deposits absent and present was 0.76 and 0.79, respectively. Conclusion: This study demonstrated the feasibility of creating and applying a machine learning model that performs competitively in identifying electron dense deposits in EM images.https://nephropathol.com/PDF/jnp-11-e17123.pdfkidneydeep learningbiopsymachine learningelectron microscopy |
| spellingShingle | Alaa Alsadi Nasma K. Majeed Dereen M. Saeed Yash Dharmamer Manmeet B. Singh Tushar N. Patel A deep-learning approach at automated detection of electron-dense immune deposits in medical renal biopsies Journal of Nephropathology kidney deep learning biopsy machine learning electron microscopy |
| title | A deep-learning approach at automated detection of electron-dense immune deposits in medical renal biopsies |
| title_full | A deep-learning approach at automated detection of electron-dense immune deposits in medical renal biopsies |
| title_fullStr | A deep-learning approach at automated detection of electron-dense immune deposits in medical renal biopsies |
| title_full_unstemmed | A deep-learning approach at automated detection of electron-dense immune deposits in medical renal biopsies |
| title_short | A deep-learning approach at automated detection of electron-dense immune deposits in medical renal biopsies |
| title_sort | deep learning approach at automated detection of electron dense immune deposits in medical renal biopsies |
| topic | kidney deep learning biopsy machine learning electron microscopy |
| url | https://nephropathol.com/PDF/jnp-11-e17123.pdf |
| work_keys_str_mv | AT alaaalsadi adeeplearningapproachatautomateddetectionofelectrondenseimmunedepositsinmedicalrenalbiopsies AT nasmakmajeed adeeplearningapproachatautomateddetectionofelectrondenseimmunedepositsinmedicalrenalbiopsies AT dereenmsaeed adeeplearningapproachatautomateddetectionofelectrondenseimmunedepositsinmedicalrenalbiopsies AT yashdharmamer adeeplearningapproachatautomateddetectionofelectrondenseimmunedepositsinmedicalrenalbiopsies AT manmeetbsingh adeeplearningapproachatautomateddetectionofelectrondenseimmunedepositsinmedicalrenalbiopsies AT tusharnpatel adeeplearningapproachatautomateddetectionofelectrondenseimmunedepositsinmedicalrenalbiopsies AT alaaalsadi deeplearningapproachatautomateddetectionofelectrondenseimmunedepositsinmedicalrenalbiopsies AT nasmakmajeed deeplearningapproachatautomateddetectionofelectrondenseimmunedepositsinmedicalrenalbiopsies AT dereenmsaeed deeplearningapproachatautomateddetectionofelectrondenseimmunedepositsinmedicalrenalbiopsies AT yashdharmamer deeplearningapproachatautomateddetectionofelectrondenseimmunedepositsinmedicalrenalbiopsies AT manmeetbsingh deeplearningapproachatautomateddetectionofelectrondenseimmunedepositsinmedicalrenalbiopsies AT tusharnpatel deeplearningapproachatautomateddetectionofelectrondenseimmunedepositsinmedicalrenalbiopsies |