Computer vision to predict cell seeding coverage in re-endothelialized mouse lungs

Transplantation of donor grafts recellularized with recipient-derived or non-immunogenic universal cells is a potential means of reducing the graft rejection and post-transplant complications in lung transplantation. Achieving a fully recellularized lung, however, remains a far-off goal and has several limitations, including inadequate cell coverage of the acellular scaffold. A key parameter for evaluating recellularization efficacy is the cell seeding coverage (CSC); the ratio of seeded cell area to the total area of the lung scaffold. To calculate the CSC from a histological image, the lung scaffold and the seeded cell areas must be quantified. In this work, the ability of semantic segmentation to accurately automate the pixel-wise analysis of histological images is investigated. Specifically, the U-Net and LinkNet models are applied to re-endothelialized mouse lung images, generating pixel-wise classifications of the lung scaffold and the seeded cell areas to calculate the CSC. Model performance when trained on complete images and on image patches is compared. The patch-based U-Net model outperformed the other models, predicting CSC with a root mean square error of 2.23 ± 0.36%, in addition to classifying lung scaffold pixels and seeded cell pixels with intersection over union scores of 77.8 ± 1.4% and 69.5 ± 1.1% respectively.