Automated detection of microsatellite status on unstained samples in early colon cancer

Artificial intelligence (AI)-integrated infrared imaging is able to identify microsatellite status with high sensitivity in unstained samples of early colon cancer, a German study showed.

Label-free Quantum Cascade Laser (QCL)-based infrared imaging combined with deep learning provides spatially and molecularly resolved alterations of the genome and proteome in unstained cancer tissue sections. This technique was shown to be able to distinguish between microsatellite instability-high (MSI-H) and microsatellite-stable (MSS) status of sporadic colorectal cancer (CRC) [1]. To verify the method, tissue samples from the prospective, multicentre AIO CPP registry study were analysed.

In detail, images of tissue sections taken in 20 min with QCL infrared microscopes were classified by convolutional neural networks (CNN). An in-house developed segmenting CNN (U-Net) localised tumour regions and a second CNN (VGG-Net) subsequently classified the microsatellite status. Endpoints were area under curve of receiver operating characteristic (AUROC) and area under precision recall curve (AUPRC). Dr Frederik Großerüschkamp (Ruhr-Universität Bochum, Germany) presented the results [2].

The multicentre clinical cohort included 491 patients, of which 100 tumour-free and 391 with tumour. Baseline characteristics of age, sex, stage, location, including BRAF mutation status were equally distributed among test cohorts. The U-Net was verified on 294 patients serving as training dataset, 100 as test dataset, and 97 as validation dataset. An AUROC of 0.99 was achieved for the validation dataset. Tumours are thereby precisely spatially resolved in the sections. The microsatellite status classification of the identified tumour regions was verified on 391 patients: 245 served as training dataset, 73 as test dataset, and 73 as validation dataset. In the current study, an AUROC of 0.83 and an AUPRC of 0.64 were achieved.

Based on these results, Dr Großerüschkamp concluded that “MSI-H was identified with high sensitivity but low specificity and demands therefore longer training phases and larger sample numbers for training. Both are currently under work.”

1. Kallenbach-Thieltges A, et al. Sci Rep. 2020;10:10161.
2. Großerüschkamp F, et al. Automated detection of microsatellite status in early colon cancer (CC) using artificial intelligence (AI) integrated infrared (IR) imaging on unstained samples from the AIO ColoPredictPlus 2.0 (CPP) registry study. Abstract 385O, ESMO Congress 2021, 16–21 September.

 

Copyright ©2021 Medicom Medical Publishers



Posted on 19 November 202117 May 2024