Breast cancer is a type of cancer with the highest incidence worldwide in 2021, with early diagnosis and rapid treatment intervention being the reasons for the decreasing mortality rate associated with the disease. The major challenge faced by clinicians and pathologists is the lack of accuracy in the histopathological analysis of biopsy or resection samples, leading to classification misdiagnosis and compromising the prognosis of patients. Spectral histopathology has provided great advances regarding cancer diagnosis, especially through the use of FTIR spectroscopy, proving to be a valuable complement to histopathological analyses. In this study unstained formalin-fixed paraffin embedded breast cancer tissue samples, collected from patients undergoing surgery and mounted on glass slides, were probed through FTIR and Raman microspectrocopy. Two classification models were constructed using the AdaBoost algorithm, both achieving >90% accuracy and successfully discriminating invasive breast carcinoma from surrounding normal tissue. Chemical maps from the interfaces of invasive breast carcinoma-surrounding normal tissue were also generated. This study showed the potential of spectral histopathology, in particular FTIR, for daily use in pathology laboratories, introducing few disruptions to the routine workflow while increasing the sensitivity, specificity and accuracy of the diagnoses.