Fourier Transform Infrared spectroscopy for the assessment of skin cancer: a potential tool in cancer diagnostics

Thesis event information

Date and time of the thesis defence

Place of the thesis defence

P117, Aapistie 5A, University of Oulu

Topic of the dissertation

Fourier Transform Infrared spectroscopy for the assessment of skin cancer: a potential tool in cancer diagnostics

Doctoral candidate

Master of Health Sciences Bijay Shakya

Faculty and unit

University of Oulu Graduate School, Faculty of Medicine, Research Unit of Health Sciences and Technology

Subject of study

Cancer diagnostics


Professor Hugh J. Byrne, Technological University Dublin


Adjunct Professor Lassi Rieppo, University of Oulu

Visit thesis event

Add event to calendar

The use of infrared spectroscopy for discriminating skin cancer cells of different stages (primary and metastatic)

Melanoma is the most researched skin cancer because of its tendency to spread quicky to nearby tissues and other vital organs, making it the most lethal form of skin cancer. Its incidence is quickly increasing in Finland, with cases being doubled in less than 10 years. Therefore, it is critical to diagnose melanoma and determine its aggressiveness (or stages) in order to improve the success rate of treatment. Importantly, this thesis demonstrated that different melanoma cells can be discriminated, and potentially their aggressiveness could be determined. In addition, this thesis proposes a protocol for standardising the measurement of skin cancer samples using infrared spectroscopy by demonstrating the mode and sample type combination that achieves the best classification result, so that the results would be reproducible and comparable across different research units in the future.
Melanoma is becoming more common globally and is responsible for the majority of skin cancer-related deaths. Primary melanomas can be treated successfully, but once it spread to distant locations (metastasize), the treatment options available typically only help to supress the cancer cells and prolong the overall survival rate. Usually, melanoma is diagnosed through histopathological inspection of biopsied tissue by a pathologist. This method is subjective and heavily reliant on the experience of the pathologist recognizing even the rarest variants and premalignant lesions. Therefore, there is a need for an additional methodology which could aid pathologist in diagnosis and characterization of skin cancers. In this thesis, machine learning was applied to infrared spectra to differentiate between primary and metastatic conditions for better characterization of the aggressiveness of the disease. The samples used in this thesis were formalin-fixed paraffin embedded to ensure that the proposed approach is compatible with the standard protocols used for biopsy samples in pathology laboratories for melanoma diagnosis.
This thesis was performed at Research Unit of Health Sciences and Technology (HST), University of Oulu in collaboration with Department of Pathology. This thesis was supported by Finnish Cultural Foundation, Tauno Tönning Foundation and Riitta and Jorma J. Takanen Foundation.
Last updated: 23.1.2024