The answer is clearly NO. We do not understand much about the mode of action of anti-cancer drugs, Only in a few rare cases, such as cisplatin these things have been studied in detail and were largely understood. Here you can try to predict anticancer activity on related Pt(II) complexes. However, at the same time there are tousands of Pt(II) complexes with completely different structures, which are either more toxic than cisplatin or not toxic at all.
Medicine is still largely trial and error in this field since apoptosis is very complicated.
For other, less complicated things such as channel blocking the modelling methods have reached a very high quality and pharma companies rely on these calculations when starting searches for new drugs.
For literature I can refer you to my book as a good starting point:
Trying to quantitatively predict a drug's activity is not realistic, especially when it comes down to unpredictable diseases such as cancer. The activity of a certain drug can only be apprehended after trials in test objects.