Deep learning-based grading of ductal carcinoma in situ in breast histopathology images

Abstract

Ductal carcinoma in situ (DCIS) is a non-invasive breast cancer that can progress into invasive ductal carcinoma (IDC). Studies suggest DCIS is often overtreated since a considerable part of DCIS lesions may never progress into IDC. Lower grade lesions have a lower progression speed and risk, possibly allowing treatment de-escalation. However, studies show significant inter-observer variation in DCIS grading. Automated image analysis may provide an objective solution to address high subjectivity of DCIS grading by pathologists. In this study, we developed and evaluated a deep learning-based DCIS grading system. The system was developed using the consensus DCIS grade of three expert observers on a dataset of 1186 DCIS lesions from 59 patients. The inter-observer agreement, measured by quadratic weighted Cohen’s kappa, was used to evaluate the system and compare its performance to that of expert observers. We present an analysis of the lesion-level and patient-level inter-observer agreement on an independent test set of 1001 lesions from 50 patients. The deep learning system (dl) achieved on average slightly higher inter-observer agreement to the three observers (o1, o2 and o3) (κo1,dl = 0.81, κo2,dl = 0.53 and κo3,dl = 0.40) than the observers amongst each other (κo1,o2 = 0.58, κo1,o3 = 0.50 and κo2,o3 = 0.42) at the lesion-level. At the patient-level, the deep learning system achieved similar agreement to the observers (κo1,dl = 0.77, κo2,dl = 0.75 and κo3,dl = 0.70) as the observers amongst each other (κo1,o2 = 0.77, κo1,o3 = 0.75 and κo2,o3 = 0.72). The deep learning system better reflected the grading spectrum of DCIS than two of the observers. In conclusion, we developed a deep learning-based DCIS grading system that achieved a performance similar to expert observers. To the best of our knowledge, this is the first automated system for the grading of DCIS that could assist pathologists by providing robust and reproducible second opinions on DCIS grade.

Automated Diagnosis of Cervical Intraepithelial Neoplasia in Histology Images via Deep Learning

“…Consequently, the mean accuracies of the CNN models for three-class classification were increased than those for four-class classification. In other similar studies, the imbalanced datasets were used without a balancing strategy [ 33 , 34 ]. In the light of previous other studies and our experimental results, we concluded that the imbalance was not a critical issue in our study.…”

A deep learning model for breast ductal carcinoma in situ classification in whole slide images

“…Deep learning has been widely applied in computational histopathology, with applications such as cancer classification in WSIs, cell detection and segmentation, and the stratification of patient outcomes Yu et al (2016); Hou et al (2016); Madabhushi and Lee (2016); Litjens et al (2016); Kraus et al (2016); Korbar et al (2017); Luo et al (2017); Coudray et al (2018); Wei et al (2019); Gertych et al (2019); Bejnordi et al (2017); Saltz et al (2018); Campanella et al (2019); Iizuka et al (2020). For breast histopathology in particular, deep learning has been applied for classification of cancer in WSIs Bayramoglu et al (2016); Sharma and Mehra (2020); Hameed et al (2020); Mi et al (2021); Sohail et al (2021); Wetstein et al (2021). As for histopathological grading of differentiation level for DCIS, a deep learning-based DCIS grading system that achieved a performance similar to expert observer pathologists was reported Wetstein et al (2021).…”

“…For breast histopathology in particular, deep learning has been applied for classification of cancer in WSIs Bayramoglu et al (2016); Sharma and Mehra (2020); Hameed et al (2020); Mi et al (2021); Sohail et al (2021); Wetstein et al (2021). As for histopathological grading of differentiation level for DCIS, a deep learning-based DCIS grading system that achieved a performance similar to expert observer pathologists was reported Wetstein et al (2021). IDC classification has also been recently investigated Kanavati and Tsuneki (2021a).…”

Breast Invasive Ductal Carcinoma Classification on Whole Slide Images with Weakly-Supervised and Transfer Learning

“…The applications from cancer cells classification and segmentation and patient outcome predictions for a variety of organs and diseases [5][6][7][8][9][10][11][12][13][14][15][16][17][18]. Machine learning has been previously applied to various applications of breast histopathology classification [19][20][21][22][23][24].…”