Tissue Identification by Differential Mobility Spectrometry

Anton Kontunen

    Research output: Book/ReportDoctoral thesisCollection of Articles


    The human genome is constantly changing due to natural mutations and environmental exposure. As these changes accumulate over our lifetime, it increases the likelihood of the creation of cells that proliferate uncontrollably and ultimately invade surrounding tissue and the blood circulation or the lymphatic system. This type of malignant neoplasm, more commonly known as cancer, is a disease that either directly or indirectly affects the majority of the population as one of the leading causes of death. Cancer is a versatile disease that can affect practically any part of the body. Depending on the tissue of origin and the aggressiveness of the malignancy, the treatment options, prognosis and mortality rates can vary significantly. In general, the role of cancer as a cause of death is constantly increasing, and despite significant global financial investments and decades of research, new and better methods of treatment and diagnosis are in continuous demand.

    One particular area that requires more attention and innovation is the surgical treatment of solid cancers. The general aim of surgical treatment is to remove all malignant cells from the patient’s body – that is to say, to achieve a negative surgical margin. The resected tumour has a negative margin, when the outermost surface area has no cancerous cells. However, in a considerable number of surgeries, the removal is incomplete. The resulting residual cancer almost always triggers additional treatment steps, which often involve a reoperation. The need for a reoperation is a major detriment for the well-being of the patient, and the added healthcare costs are substantial. If the number of avoidable reoperations could be halved from their current level, the saving potential in annual global healthcare costs would already be measured in billions of dollars.

    The reason why the problem of reoperations persists despite the notable financial incentives lies in the difficulty of discriminating malignant tissue from benign, especially during a surgical procedure. The molecular contents that define the structure and function of a cell are different depending on the organ of origin, and similar differences are also present between malignant and benign cells. The biomolecules that enable the identification of the types of tissues are called
    biomarkers, and the research on this area has revealed hundreds of proteins, fatty acids and metabolic products that exhibit differences in quantities based on tissue malignancy. However, the variation of specific marker molecules is often high, and the molecular differences rarely translate into clear macroscopic differences. This means that visual assessment of the margin between benign and cancerous tissue is extremely challenging. Still, almost all surgeons rely only on visual assessment and palpation in cancer surgeries. The challenge of complete excision is further accentuated by the current resection guidelines that instruct surgeons to preserve as much non-cancerous tissue as possible. This aim and its subjective execution lead not only to high variation in positive margin rates between institutions and regions, but also to a high number of required reoperations in general. To reduce the reoperations caused by positive surgical margins, several technologies have been studied and introduced to aid in intraoperative tissue identification, but the clinical adoption has been limited due to various impeding factors involved in their use.

    In this thesis, a concept that could potentially be used in the assessment of the intraoperative surgical margin is introduced through five scientific publications that concentrate on the evolution and feasibility of the technology in tissue identification. The basis of the technology is the measurement of surgical smoke with differential mobility spectrometry (DMS). DMS is a measurement technology that provides information on the molecular content of a gaseous sample in atmospheric pressure by means of ionisation and subsequent differentiation of the ions in a high-strength asymmetric electric field. DMS is comparable to mass spectrometry (MS), and even though the analytical performance of MS is better, the reduced complexity, smaller size and lower cost of DMS make it an advantageous option. DMS has been used as a standalone measurement instrument in many types of general gas measurement applications and in some biomedical applications, such as breath analysis, but the context of use has always permitted a controlled environment and a relatively long measurement duration. Thus, the real-time application of surgical smoke measurement requires additional hardware and parameter optimisation. In addition, raw DMS measurement data do not provide directly quantifiable information on certain biomolecules, but rather a comprehensive spectrum of all contents in the sample combined. This means that the interpretation and identification of tissue type from the DMS output spectra is not trivial and involves a high number of dimensions that are most effectively analysed by means of machine learning. The interdisciplinary aspects of the system and their combined function and performance in tissue identification are the focus of this thesis.

    In the first three publications included in the thesis, the focus was on studying the overall feasibility of tissue identification and its possibilities with animal tissues and clinically relevant breast cancer samples. The results in laboratory conditions with controlled sampling were promising, and the diagnostic performance demonstrated the potential of the technology in tissue identification. In Publication IV, the system was modified to accommodate real-time measurements and to relay the classification information immediately after the measurement. The results demonstrated the feasibility of real-time tissue identification with the system, albeit in laboratory conditions and in a porcine model. In the final study, a prototype system was used intraoperatively during breast cancer surgeries. The results of this study were not comparable to the laboratory results in respect to diagnostic performance but indicated that the system can be adapted to the surgical workflow with minimal intrusiveness to provide information on the operated tissue.

    Overall, the results of this study indicate that a DMS-based tissue identification system has the potential to be used in real-time applications to identify tissue types with adequate diagnostic performance. With further development, the system presented in this thesis could fulfil the need for a surgical margin assessment device that would reduce avoidable reoperations of solid cancers and thus protect the well-being of cancer patients.
    Original languageEnglish
    Place of PublicationTampere
    ISBN (Electronic)978-952-03-2303-5
    Publication statusPublished - 2022
    Publication typeG5 Doctoral dissertation (articles)

    Publication series

    NameTampere University Dissertations - Tampereen yliopiston väitöskirjat
    ISSN (Print)2489-9860
    ISSN (Electronic)2490-0028


    Dive into the research topics of 'Tissue Identification by Differential Mobility Spectrometry'. Together they form a unique fingerprint.

    Cite this