Each image can be multiple gigabytes so it could be computationally difficult. Also those are probably all the lung cancer slides that TCGA has (edit: there seems to be 3500+ lung cancer slides at [1] so I was wrong. Maybe they're not all H&E stained)