Proton radiotherapy has been shown to offer a significant dosimetric advantage in cancer patients, in comparison to conventional radiotherapy, with a decrease in dose to healthy tissue and organs at risk, because the bulk of the beam energy is deposited in the Bragg peak to be located within a tumour. However, it should be kept in mind that radiotherapy of cancer is still accompanied by adverse side effects, and a better understanding and improvement of radiotherapy can extend the life expectancy of patients following the treatment of malignant tumours. In this study, the dose distributions measured with thermoluminescent detectors (TLDs) inside a tissue-equivalent adult human phantom exposed for lung and prostate cancer using the modern proton beam scanning radiotherapy technique were compared. Since the TLD detection efficiency depends on the ionization density of the radiation to be detected, and since this efficiency is detector specific, four different types of TLDs were used to compare their response in the mixed radiation fields. Additionally, the dose distributions from two different cancer treatment modalities were compared using the selected detectors. The measured dose values were benchmarked against Monte Carlo simulations and available literature data. The results indicate an increase in the lateral dose with an increase of the primary proton energy. However, the radiation quality factor of the mixed radiation increases by 20% in the vicinity to the target for the lower initial proton energy, due to the production of secondary charged particles of low-energy and short range. For the cases presented here the MTS-N TLD detector seems to be the most optimal tool for dose measurements within the target volume, while the MCP-N TLD detector, due to an interplay of its enhanced thermal neutron response and decreased detection efficiency to highly ionising radiation, is a better choice for the out-of-field measurements. The pairs of MTS-6 and MTS-7 TLDs used also in this study allowed for a direct measurement of the neutron dose equivalent. Before it can be concluded that they offer an alternative to the time-consuming nuclear track detectors, however, more research is needed to unambiguously confirm whether this observation was just accidental or whether it only applies to certain cases. Since there is no universal detector, which would allow the determination of the dosimetric quantities relevant for risk estimation, this work expands the knowledge necessary to improve the quality of dosimetry data and might help scientists and clinicians in choosing the right tools to measure radiation doses in mixed radiation fields.