Its because, no two tumors are exactly alike, and neither are patients. Like some tumors carry specific genetic mutations that make them especially sensitive or resistant to certain drugs. For instance, in bladder cancer, rare mutations have been found in “exceptional responders” who had dramatic outcomes with standard chemo. But in other cancers like pediatric leukemia, mutations in genes like TP53 can signal poor response, while biomarkers like high tumor mutation burden (TMB) can predict better outcomes, especially when chemo is combined with immunotherapy. Secondly, in case of fast-growing or aggressive tumors (with larger size, higher grade or hormone receptor-negative), they often respond better initially. That’s because chemo targets rapidly dividing cells. But the flip side is they might also come back faster, as these tumors tend to evolve resistance. Additionally, even within the same tumor, not all cells are identical. Some differences are purely random, small fluctuations in protein levels that help a few cells survive treatment by chance. These surviving cells can eventually repopulate the tumor, making it harder to treat over time. And that’s also why personalized treatment that is based on both biology and genetics, is the future of cancer care.

I think, it may also be related to the tumor metabolism/biochemistry and the differences in people’s drug metabolism enzymes.

Also, cancers are in general grouped and considered for treatment based on localization of tumor, not by driving mutation (or at least it used to be like this, it slowly changes). There are schemes for treating breast cancer, leukemias, etc, but more attention sould be paid on thinking about driving mutations and treat cancers originating from the same type mutation regardless the localization.