FDA has approved many purine analogs e.g. thioguanine, cladribine, pentostatin, mercaptopurine for various forms of leukemia and/or lymphoma, but none for non-hematological malignancies. What makes non-hematological malignancies untreatable by purine analogs?
It's not a given, that they are not useful in solid tumors (also lymphoma are accounted among the solid tumors).
There are pyrimidine-like antimetabolites, e. g. 5-FU: https://en.wikipedia.org/wiki/Fluorouracil, while some purine antimetabolites were in development, see e. g. here: https://www.ncbi.nlm.nih.gov/pubmed/4066218.
There are some approaches that try to make tumors more sensitive to purine antimetabolites, e. g. https://www.sciencedirect.com/science/article/abs/pii/S0006295219304691.
Another, older, paper that describes a purine antimetabolite in lymphoma: https://ascopubs.org/doi/abs/10.1200/JCO.19220.127.116.11
A more recent approach on the purine metabolism: https://www.frontiersin.org/articles/10.3389/fimmu.2018.01697/full