Challenge for Diagnostic Assessment of Deep Learning Algorithm for Metastases Classification in Sentinel Lymph Nodes on Frozen Tissue Section Digital Slides in Women with Breast Cancer

Abstract

Assessing the status of metastasis in sentinel lymph nodes (SLNs) by pathologists is an essential task for the accurate staging of breast cancer. However, histopathological evaluation of SLNs by a pathologist is not easy and is a tedious and time-consuming task. The purpose of this study is to review a challenge competition (HeLP 2018) to develop automated solutions for the classification of metastases in hematoxylin and eosin-stained frozen tissue sections of SLNs in breast cancer patients. Materials and Methods A total of 297 digital slides were obtained from frozen SLN sections, which include postneoadjuvant cases (n=144, 48.5%) in Asan Medical Center, South Korea. The slides were divided into training, development, and validation sets. All of the imaging datasets have been manually segmented by expert pathologists. A total of 10 participants were allowed to use the Kakao challenge platform for 6 weeks with two P40 GPUs. The algorithms were assessed in terms of the area under receiver operating characteristic curve (AUC). Results The top three teams showed 0.986, 0.985, and 0.945 AUCs for the development set and 0.805, 0.776, and 0.765 AUCs for the validation set. Micrometastatic tumors, neoadjuvant systemic therapy, invasive lobular carcinoma, and histologic grade 3 were associated with lower diagnostic accuracy. Conclusion In a challenge competition, accurate deep learning algorithms have been developed, which can be helpful in making frozen diagnosis of intraoperative SLN biopsy. Whether this approach has clinical utility will require evaluation in a clinical setting.

WITHDRAWN: Few-shot Learning on the Diagnosis of Lymphatic Metastasis of Lung Carcinoma

“…The clinical-grade pathology using was trained by weakly supervised deep learning with over thousands of slides dataset including prostate core biopsy, breast cancer metastatic lymph node, and skin lesion separately; and the area under the curve (AUC) can be 0.965 on the test of breast axillary lymph node metastasis [8]. The accuracy of the weakly supervised deep learning may be affected by the labels [9] and the size of the dataset which could be improved by upgrading the algorithm iteration [11][12][13]. Atsushi Teramoto improves 4% AUC by using a two-step supervised strategy for the classification of lung cytological images with 60 cases dataset [14].…”

Artificial intelligence and frozen section histopathology: A systematic review

“…In a challenge using the same FFS lymph node biopsy specimens, teams prediction times and accuracies varied significantly despite using the same data. The strongest performing team (AUC 0.805) took 10.8 min per slide to process while the next highest (AUC 0.776) took only 0.6 min 20 …”

“…The strongest performing team (AUC 0.805) took 10.8 min per slide to process while the next highest (AUC 0.776) took only 0.6 min. 20 Models have also been shown to have strong performance in differentiating types of cancer from one another, even when trained on FFPE images and tested on FFS images. 21 This can be aided by adding rules which predispose a model to be more conservative, calling uncertain tissue malignant in an attempt to reduce false negatives.…”

“…33 While some of the models presented here were likely to be too slow to be clinically useful, others were quite rapid, at less than a minute per slide. 20 The goal of models here is clinical utility; rendering useful results in a reasonable time. The time and manpower required to simply create the data on which these models would predict is likely a substantial current barrier to their implementation in practice.…”

Deep learning-based intraoperative differentiation of primary CNS lymphoma and glioma: a discovery, multicenter validation, and proof-of concept study

“…Histological artefacts present in frozen sections can hinder rapid diagnostic assessments during surgery 14 , but a deep-learning algorithm may improve the quality of whole-slide images (WSIs) from H&E-stained frozen sections, leading to more accurate tumor classi cation by pathologists 15 . Recent studies have also shown the ability of deep learning models to diagnose thyroid nodules 16 and determine the metastatic status of sentinel lymph nodes in breast cancer 17 from the conventional intraoperative frozen sections, highlighting the potential of frozen samples in developing deep learning models. Thus, we hypothesize a deep learning approach can facilitate the intraoperative diagnosis of brain tumors.…”