There are a number of these broad spectrum, interesting, often low toxicity candidates, drugs and nutraceuticals, that may not have much specific CRC data or experience, much less individually specific prognostic data.
What we really need is a blood test that personally shows us some of the effect of these off label drugs have on a single person's cancer in 1-7 days, either a big reduction in cancer stem cell count or a big debris field. I've done this to an extent with CEA, CA199 etc, but it is slow and painstaking over weeks and months. Also I'm probably better at conditioning the patient, following inflammation levels, keeping "noises" low, and things steady for multiple measurements (e.g. 3-4 CEA data points in a flat or a monotone declining line). In reality, I usually worked with binary blood tests (a before day blood test, and an after day in 10-30 days) with multiple panels in a constant improvement mode, with some classical 3 - 12 flat data points periods.
Let's say Signatura is capable of picking up the cancer cells that are sensitive to a DIY off label trial drug. We need a before and after treatment reading. The bone I have is that Signatura is not too sensitive with a 7 mL sample to give a high enough resolution answer and probably needs to process 50 - 250 mL to be useful. Even better, lab services with personalized volume say, 50- 100 mL before, and 200 - 400 mL after. If Signatura only shows 1-2 cells (or zero) to start with, there's no room for meaningful reduction measurement. Even a large blood sample (blood donor ~ 470 mL, a pint=473 mL, divided into two proximate sample days) would not be a problem, once we had solved the cancer/chemo anemia problem, Hgb 10 --> 14, a la the Liverwurst.