Effectiveness of transfer learning for enhancing tumor classification with a convolutional neural network on frozen sections

• Published in: Scientific reports Vol. 10; no. 1; p. 21899
• Main Authors: Kim, Young-Gon, Kim, Sungchul, Cho, Cristina Eunbee, Song, In Hye, Lee, Hee Jin, Ahn, Soomin, Park, So Yeon, Gong, Gyungyub, Kim, Namkug
• Format: Journal Article
• Language: English
• Published: London Nature Publishing Group UK 14.12.2020; Nature Publishing Group; Nature Portfolio
• Subjects: 631/67/1347; 639/166/985; Adult; Aged; Aged, 80 and over; Biopsy; Classification; Datasets; Deep learning; Female; Frozen Sections; Humanities and Social Sciences; Humans; Image Interpretation, Computer-Assisted; Lymph nodes; Lymphatic Metastasis; Male; Metastases; Middle Aged; multidisciplinary; Neoplasms - classification; Neoplasms - pathology; Neural networks; Neural Networks, Computer; Retrospective Studies; Science; Science (multidisciplinary); Sentinel Lymph Node Biopsy; Transfer learning
• ISSN: 2045-2322; 2045-2322
• DOI: 10.1038/s41598-020-78129-0

Fast and accurate confirmation of metastasis on the frozen tissue section of intraoperative sentinel lymph node biopsy is an essential tool for critical surgical decisions. However, accurate diagnosis by pathologists is difficult within the time limitations. Training a robust and accurate deep learning model is also difficult owing to the limited number of frozen datasets with high quality labels. To overcome these issues, we validated the effectiveness of transfer learning from CAMELYON16 to improve performance of the convolutional neural network (CNN)-based classification model on our frozen dataset (N = 297) from Asan Medical Center (AMC). Among the 297 whole slide images (WSIs), 157 and 40 WSIs were used to train deep learning models with different dataset ratios at 2, 4, 8, 20, 40, and 100%. The remaining, i.e., 100 WSIs, were used to validate model performance in terms of patch- and slide-level classification. An additional 228 WSIs from Seoul National University Bundang Hospital (SNUBH) were used as an external validation. Three initial weights, i.e., scratch-based (random initialization), ImageNet-based, and CAMELYON16-based models were used to validate their effectiveness in external validation. In the patch-level classification results on the AMC dataset, CAMELYON16-based models trained with a small dataset (up to 40%, i.e., 62 WSIs) showed a significantly higher area under the curve (AUC) of 0.929 than those of the scratch- and ImageNet-based models at 0.897 and 0.919, respectively, while CAMELYON16-based and ImageNet-based models trained with 100% of the training dataset showed comparable AUCs at 0.944 and 0.943, respectively. For the external validation, CAMELYON16-based models showed higher AUCs than those of the scratch- and ImageNet-based models. Model performance for slide feasibility of the transfer learning to enhance model performance was validated in the case of frozen section datasets with limited numbers.