I understand.. everyone has their subjective idea of “worth while” 
I am going to continue to be very particular about definitions here. When you say “reliable”, to most people that means: “Will I be able to get a temerpature reading from the chip reliably, or will I have problems reading the chip?” .. so the answer is yes, you will be able to reliably read the chip with a capable reader like the this;
If you mean “Will I be able to get temperature data from the chip that will always be within N degrees of my body’s core temperature?” then the answer to this is - maybe. It will depend on a lot of factors.
For example, if you implanted it into your rectal tissue, it didn’t migrate away during healing, and you are willing to sit on a reader capable of reading FDX-B chips and decipher the “app data” to get a temperature reading, then absolutely yes Anywhere else, even the arm pit, will depend a lot on exact placement and how you are attempting to get a read each time. If you place it perfectly up in your arm pit area (maybe consult a doctor), keep your arm down for at least 1 minute, and take the temperature reading without opening your arm, or taking it very quickly as soon as you open your arm.. then yes it’s possible to get consistent readings based on a repeatable procedure that controls variables as much as possible.
Now that your placement and procedure can produce consistent readings, the documentation for the Destron Fearing microchip thermal sensor states the chip itself has an error margin of +/- 0.2°C. Considering the human body’s “normal” range is 36.1°C to 37.2°C, an error margin of only 0.2°C is not that much.
With all these factors at play (perfect placement, a good consistent read procedure, etc.) it should be possible to create calibration such that you could apply a correction to the chip based on a linear or % based scale, depending on initial test data.. but you would have to perform this long term testing and come up with your calibration method since each placement is going to be unique.