A promising deep learning-assistive algorithm for histopathological screening of colorectal cancer.

Colorectal cancer is one of the most common cancers worldwide, accounting for an annual estimated 1.8 million incident cases. With the increasing number of colonoscopies being performed, colorectal biopsies make up a large proportion of any histopathology laboratory workload. We trained and validated a unique artificial intelligence (AI) deep learning model as an assistive tool to screen for colonic malignancies in colorectal specimens, in order to improve cancer detection and classification; enabling busy pathologists to focus on higher order decision-making tasks. The study cohort consists of Whole Slide Images (WSI) obtained from 294 colorectal specimens. Qritive's unique composite algorithm comprises both a deep learning model based on a Faster Region Based Convolutional Neural Network (Faster-RCNN) architecture for instance segmentation with a ResNet-101 feature extraction backbone that provides glandular segmentation, and a classical machine learning classifier. The initial training used pathologists' annotations on a cohort of 66,191 image tiles extracted from 39 WSIs. A subsequent application of a classical machine learning-based slide classifier sorted the WSIs into 'low risk' (benign, inflammation) and 'high risk' (dysplasia, malignancy) categories. We further trained the composite AI-model's performance on a larger cohort of 105 resections WSIs and then validated our findings on a cohort of 150 biopsies WSIs against the classifications of two independently blinded pathologists. We evaluated the area under the receiver-operator characteristic curve (AUC) and other performance metrics. The AI model achieved an AUC of 0.917 in the validation cohort, with excellent sensitivity (97.4%) in detection of high risk features of dysplasia and malignancy. We demonstrate an unique composite AI-model incorporating both a glandular segmentation deep learning model and a classical machine learning classifier, with excellent sensitivity in picking up high risk colorectal features. As such, AI plays a role as a potential screening tool in assisting busy pathologists by outlining the dysplastic and malignant glands.

Analysis of deferrals on static telepathology consultation service.

The objective of this study is to analyze the deferrals in static telepathology consultation service. A store and forward approach is used to transmit cases from two remotely located rural centers to Tata Memorial Hospital. A total of 346 tele-surgical pathology cases were accessioned for second opinion and were reported from January 2002 to August 2005. The glass slides and paraffin blocks were reviewed at a later date and the telepathology diagnosis was compared with the final diagnosis rendered on light microscopy. Of all 251 teleconsults referred from one of the referring centers, a telepathology diagnosis was rendered in 205 cases and 46 cases were deferred. The reasons for deferral were as follows: the requirement for ancillary studies (40 cases), clinical details (5 cases) and poor quality sections and images (1 case). In all these deferred cases, a probable diagnosis was rendered by the telepathologist and was compared with the final diagnosis after paraffin block evaluation. In 47% of the cases, the "probable" diagnosis on telepathology matched the final diagnosis.

An analysis of 46 static telecytology cases over a period of two years.

We analysed 46 telecytology cases sent from two rural hospitals about 500 km from a tertiary cancer centre. The cases were submitted for second opinion over a period of two years and evaluated using a static store and forward telecytology approach. A total of 715 digital images were studied (average 15 per case). Forty-one of the 46 cases (89%) were reported within 3 days and 54% of cases were reported within one working day. The aspiration smears and images were found to be of diagnosable quality in 89 and 93% of the cases, respectively. The diagnostic concordance was assessed by comparing the telecytology diagnosis, glass slide diagnosis and final histopathology diagnosis (when available). A clinically useful diagnosis was rendered in 91% cases with 74% complete concordance. Five out of 46 cases (11%) were deferred for glass slide review. Store and forward telecytology using the Internet is a rapid and effective method of providing expert diagnosis in cytology.