A recent study has introduced an innovative technique for in-depth analysis of cancer cells and their surrounding tissues at the individual cell level. This advancement provides a more detailed understanding of prognoses and responses to treatment in head and neck cancers, potentially leading to improved diagnostic accuracy.
Squamous cell carcinoma in the head and neck region ranks among the ten most prevalent cancer types.
Researchers from the University of Helsinki, in collaboration with the University of Turku and the Max Planck Institute for Molecular Biomedicine in Germany, utilized machine learning to scrutinize numerous biobank patient samples at a singular cell level of detail. This groundbreaking technology merges indicators of cancer cell behavior and the tumor’s structure, as well as the surrounding healthy tissue, to create a unique ‘fingerprint’ for each patient, which can be instrumental in evaluating cancer prognosis and treatment effectiveness.
A key outcome of the research was the establishment of a novel imaging analysis approach that fuses biomarker assessments of cell behavior with structural analyses of individual cell shapes and the overall architecture of tumor tissues. This method led to the discovery of two new patient categories that had not been previously identified. One group exhibited an exceptionally favorable prognosis, while the other faced a particularly grim outlook. These differences were attributed to a specific combination of cancer cell states and the characteristics of the surrounding tissue. In the latter group, the disease’s aggressive nature was linked to communication between cancerous and neighboring healthy connective tissue, which was mediated by the epidermal growth factor (EGF).
“These findings represent a significant advancement in our understanding of cancer progression and diagnostics. For the first time, we have demonstrated that specific combinations of malignant cells with tissue cell types adjacent to what appears to be healthy tissue can notably influence cancer progression outcomes. Additionally, we pinpointed a critical signaling pathway responsible for this combined effect that could be targeted with medications, thereby potentially altering cancer development significantly,” says Research Director Sara Wickström.
“Our method also highlighted patients with particularly poor outcomes who might benefit from more aggressive treatment strategies. Conversely, we identified another group of patients with favorable prognoses for whom less aggressive interventions, such as surgery alone, might suffice. This approach could enhance the quality of life for these patients,” adds Postdoctoral Researcher Karolina Punovuori from Wickström’s team.
Diagnostic test in progress
The newly developed imaging technique lays the groundwork for precise cancer diagnoses concerning the head and neck. The research team is in the process of creating a diagnostic test aimed at improving precision in diagnosing this cancer type. They are also investigating its applicability to diagnose other cancers, including colorectal cancer. Funding from Business Finland’s Research to Business initiative supports the Multivision Diagnostics project, which is working on applications that may eventually be utilized in cancer treatment centers.
“Our research leverages cutting-edge machine learning and spatial biology analysis techniques. We examine hundreds of patient samples and millions of cells, made feasible through advanced computing and artificial intelligence. This research is a part of a transformative wave in cancer diagnostics. We believe this technology will greatly enhance cancer diagnosis and treatment strategy precision,” explains Sara Wickström.
“Currently, imaging cancer biomarkers using antibody staining is already implemented in clinical settings. Therefore, our method will not incur significant costs, requiring only our developed algorithm and a unique mix of antibodies. Given the expense of cancer treatment, this is actually quite reasonable,” she adds.
The incidence of head and neck cancers has dramatically increased over the past 30 years.
