TY - JOUR
T1 - Deep learning with limited data
T2 - Organ segmentation performance by U-net
AU - Bardis, Michelle
AU - Houshyar, Roozbeh
AU - Chantaduly, Chanon
AU - Ushinsky, Alexander
AU - Glavis-Bloom, Justin
AU - Shaver, Madeleine
AU - Chow, Daniel
AU - Uchio, Edward
AU - Chang, Peter
N1 - Funding Information:
Funding: This research is supported by a Radiological Society of North America Medical Student Research Grant (RMS1902) and additionally by an Alpha Omega Alpha Carolyn L. Kuckein Student Research Fellowship.
Publisher Copyright:
© 2020 by the authors. Licensee MDPI, Basel, Switzerland.
PY - 2020/8
Y1 - 2020/8
N2 - (1) Background: The effectiveness of deep learning artificial intelligence depends on data availability, often requiring large volumes of data to effectively train an algorithm. However, few studies have explored the minimum number of images needed for optimal algorithmic performance. (2) Methods: This institutional review board (IRB)-approved retrospective review included patients who received prostate magnetic resonance imaging (MRI) between September 2014 and August 2018 and a magnetic resonance imaging (MRI) fusion transrectal biopsy. T2-weighted images were manually segmented by a board-certified abdominal radiologist. Segmented images were trained on a deep learning network with the following case numbers: 8, 16, 24, 32, 40, 80, 120, 160, 200, 240, 280, and 320. (3) Results: Our deep learning network’s performance was assessed with a Dice score, which measures overlap between the radiologist’s segmentations and deep learning-generated segmentations and ranges from 0 (no overlap) to 1 (perfect overlap). Our algorithm’s Dice score started at 0.424 with 8 cases and improved to 0.858 with 160 cases. After 160 cases, the Dice increased to 0.867 with 320 cases. (4) Conclusions: Our deep learning network for prostate segmentation produced the highest overall Dice score with 320 training cases. Performance improved notably from training sizes of 8 to 120, then plateaued with minimal improvement at training case size above 160. Other studies utilizing comparable network architectures may have similar plateaus, suggesting suitable results may be obtainable with small datasets.
AB - (1) Background: The effectiveness of deep learning artificial intelligence depends on data availability, often requiring large volumes of data to effectively train an algorithm. However, few studies have explored the minimum number of images needed for optimal algorithmic performance. (2) Methods: This institutional review board (IRB)-approved retrospective review included patients who received prostate magnetic resonance imaging (MRI) between September 2014 and August 2018 and a magnetic resonance imaging (MRI) fusion transrectal biopsy. T2-weighted images were manually segmented by a board-certified abdominal radiologist. Segmented images were trained on a deep learning network with the following case numbers: 8, 16, 24, 32, 40, 80, 120, 160, 200, 240, 280, and 320. (3) Results: Our deep learning network’s performance was assessed with a Dice score, which measures overlap between the radiologist’s segmentations and deep learning-generated segmentations and ranges from 0 (no overlap) to 1 (perfect overlap). Our algorithm’s Dice score started at 0.424 with 8 cases and improved to 0.858 with 160 cases. After 160 cases, the Dice increased to 0.867 with 320 cases. (4) Conclusions: Our deep learning network for prostate segmentation produced the highest overall Dice score with 320 training cases. Performance improved notably from training sizes of 8 to 120, then plateaued with minimal improvement at training case size above 160. Other studies utilizing comparable network architectures may have similar plateaus, suggesting suitable results may be obtainable with small datasets.
KW - Artificial intelligence
KW - Convolutional neural network
KW - Deep learning
KW - Segmentation
KW - Training size
KW - U-Net
UR - http://www.scopus.com/inward/record.url?scp=85090624224&partnerID=8YFLogxK
U2 - 10.3390/electronics9081199
DO - 10.3390/electronics9081199
M3 - Article
AN - SCOPUS:85090624224
SN - 2079-9292
VL - 9
SP - 1
EP - 12
JO - Electronics (Switzerland)
JF - Electronics (Switzerland)
IS - 8
M1 - 1199
ER -