TY - JOUR
T1 - Knowledge distillation driven instance segmentation for grading prostate cancer
AU - Hassan, Taimur
AU - Shafay, Muhammad
AU - Hassan, Bilal
AU - Akram, Muhammad Usman
AU - ElBaz, Ayman
AU - Werghi, Naoufel
N1 - Funding Information:
This work is supported by research funds from the Terry Fox Foundation, United States of America Ref: I1037, Khalifa University, Ref: CIRA-2019-047, and the Advanced Technology Research Center Program (ASPIRE), United Arab Emirates, Ref: AARE20-279. Apart from this, we would also like to thank the two expert pathologists from Watim Medical & Dental College, Rawalpindi, Pakistan, for thoroughly screening the WSI biopsies under blind testing experiments.
Funding Information:
This work is supported by research funds from the Terry Fox Foundation, United States of America Ref: I1037 , Khalifa University , Ref: CIRA-2019-047 , and the Advanced Technology Research Center Program (ASPIRE), United Arab Emirates , Ref: AARE20-279 . Apart from this, we would also like to thank the two expert pathologists from Watim Medical & Dental College, Rawalpindi, Pakistan, for thoroughly screening the WSI biopsies under blind testing experiments.
Publisher Copyright:
© 2022 Elsevier Ltd
PY - 2022/11
Y1 - 2022/11
N2 - Prostate cancer (PCa) is one of the deadliest cancers in men, and identifying cancerous tissue patterns at an early stage can assist clinicians in timely treating the PCa spread. Many researchers have developed deep learning systems for mass-screening PCa. These systems, however, are commonly trained with well-annotated datasets in order to produce accurate results. Obtaining such data for training is often time and resource-demanding in clinical settings and can result in compromised screening performance. To address these limitations, we present a novel knowledge distillation-based instance segmentation scheme that allows conventional semantic segmentation models to perform instance-aware segmentation to extract stroma, benign, and the cancerous prostate tissues from the whole slide images (WSI) with incremental few-shot training. The extracted tissues are then used to compute majority and minority Gleason scores, which, afterward, are used in grading the PCa as per the clinical standards. The proposed scheme has been thoroughly tested on two datasets, containing around 10,516 and 11,000 WSI scans, respectively. Across both datasets, the proposed scheme outperforms state-of-the-art methods by 2.01% and 4.45%, respectively, in terms of the mean IoU score for identifying prostate tissues, and 10.73% and 11.42% in terms of F1 score for grading PCa according to the clinical standards. Furthermore, the applicability of the proposed scheme is tested under a blind experiment with a panel of expert pathologists, where it achieved a statistically significant Pearson correlation of 0.9192 and 0.8984 with the clinicians’ grading.
AB - Prostate cancer (PCa) is one of the deadliest cancers in men, and identifying cancerous tissue patterns at an early stage can assist clinicians in timely treating the PCa spread. Many researchers have developed deep learning systems for mass-screening PCa. These systems, however, are commonly trained with well-annotated datasets in order to produce accurate results. Obtaining such data for training is often time and resource-demanding in clinical settings and can result in compromised screening performance. To address these limitations, we present a novel knowledge distillation-based instance segmentation scheme that allows conventional semantic segmentation models to perform instance-aware segmentation to extract stroma, benign, and the cancerous prostate tissues from the whole slide images (WSI) with incremental few-shot training. The extracted tissues are then used to compute majority and minority Gleason scores, which, afterward, are used in grading the PCa as per the clinical standards. The proposed scheme has been thoroughly tested on two datasets, containing around 10,516 and 11,000 WSI scans, respectively. Across both datasets, the proposed scheme outperforms state-of-the-art methods by 2.01% and 4.45%, respectively, in terms of the mean IoU score for identifying prostate tissues, and 10.73% and 11.42% in terms of F1 score for grading PCa according to the clinical standards. Furthermore, the applicability of the proposed scheme is tested under a blind experiment with a panel of expert pathologists, where it achieved a statistically significant Pearson correlation of 0.9192 and 0.8984 with the clinicians’ grading.
KW - Deep learning
KW - Histopathology
KW - Incremental learning
KW - Instance segmentation
KW - Prostate cancer
KW - Prostate tissues
UR - http://www.scopus.com/inward/record.url?scp=85139192621&partnerID=8YFLogxK
U2 - 10.1016/j.compbiomed.2022.106124
DO - 10.1016/j.compbiomed.2022.106124
M3 - Article
AN - SCOPUS:85139192621
SN - 0010-4825
VL - 150
JO - Computers in Biology and Medicine
JF - Computers in Biology and Medicine
M1 - 106124
ER -