Now I asume you have the plastic tube (with termocouple inside for reference - not to be used in the production version) and the termocouple or thermistor (or some integrated sensor) firmly pressed to the outer of the tube (some general sticky tape, isn't it? :-)
The sensing element is now heated by the tube and cooled by the outer air. The tube is heated by the liquid flowing inside and cooled by the outer air.
You need to get rid of cooling. Insulate very well (some sleeve - foam? This might simultaneously press the sensor against the tube wall. In final, it needsto be openable because of the disposable tube).
With perfect thermal insulation, you will have the same temperature on the sensor as inside - given enough liquid (to heat up/cool the thermal inertial mass of the tube wall and the sensor) and time (given the limited thermal conductivity of all the plastics). Both can be reduced by reducing the tube wall thickness. You need to decide, if the temporal response is good enough.

Jan Waclawek