The crucial goal of kidney-sparing surgical resection of a malignant tumor is complete removal of the cancerous tissue. The exact border between the cancerous and normal tissues is not always possible to identify by naked eye, therefore, a supplementary intraoperative diagnosis is needed. Unfortunately, intraoperative pathology methods used nowadays are time consuming and of inadequate quality rendering not definitive diagnosis. It has recently been shown that ATR-FTIR spectroscopy can be used for fast discrimination between cancerous and normal kidney tissues by analyzing the collected spectra of the tissue touch imprint smears. Most prominent differences are obtained in the wavenumber region from 950 cm^-1 to 1250 cm^-1, where the spectral bands due to the molecular vibrations of glycogen arise in the spectra of cancerous tissue smears. Such method of detection of cancerous tissue is limited by requirement to transfer the suspected tissue from the body to the FTIR instrument and stamp it on an ATR crystal of the spectrometer. We propose a spectroscopic tool which exploits the same principle of detection of cancerous cells as mentioned above, but does not require the tissue to be transferred from the body to the spectrometer. The portable spectrometer used in this design is equipped with fiber ATR probe and a sensitive liquid nitrogen cooled MCT detector. The design of the fiber probe allows the ATR tip to be changed easily in order to use only new sterilized tips for each measurement point of the tissue. It also enables sampling multiple areas of the suspected tissue with high lateral resolution which, in turn, increases accuracy with which the marginal regions between normal and cancerous tissues can be identified. Due to the loss of optical signal in the fiber probe the spectra have lower signal-To-noise ratio than in the case of standard ATR sampling setup. However, software for the spectral analysis used with the fiber probe design is still able to distinguish between cancerous and normal tissues with high accuracy.