Mathematical processing of primary data on the survival of patients with locally advanced oral squamous cell carcinoma

The purpose of the primary data analysis is to identify trends and general principles, study the relationships between phenomena, and describe the data structure. Materials and methods. We conducted a prospective study based on the FAHI NR Scientific Research Institute of Clinical Oncology Nizhny Novgorod Regional Clinical Oncology Dispensary (formerly, until 2024, FBHI NR NRCOD). The study included clinical observations of primary patients (230/100%) with stage III–IV locally advanced squamous cell carcinoma of the oral mucosa and recurrent oral mucosa cancer treated between 2016 and 2021. Results. The lack of normality in the distribution is associated with the consecutive enrolment of patients in the prospective study based on the diagnosis of primary locally advanced stage III–IV oral squamous cell carcinoma or the registration of recurrences after anticancer treatment of these patients. The reason for the lack of normal distribution of this parameter in the groups of primary patients is the presence of outliers, which are primarily of interest for clinical research–a long survival period after the anticancer treatment initiation. To compare independent samples for one of the variables, which is continuous (life expectancy), the dependent variable, the estimated feature, which is a grouping categorical variable, nonparametric analysis methods were used depending on the amount of data being evaluated. The data obtained on the highest frequency of fatal outcomes in primary patients undergoing radiation and drug treatment without surgery (89/77%) confirms the need for surgical treatment of primary patients with stage III–IV locally advanced oral squamous cell carcinoma as part of combination therapy. Discussion. The structure of the prospective study model under investigation is determined by primary data on life expectancy from the clinical observations of patients who were consecutively included in the study due to different timing of disease detection. Under this condition, the structure of the model under investigation is not specified when developing the dissertation plan. The values of the recorded life expectancy after anticancer treatment vary greatly from the median as a measure of life expectancy of all patients included in the study (n=230). Such a high variation in the extreme values of the studied characteristic of life expectancy, from a minimum of 6 months to a maximum of 79 months, indicates that it is impossible to assert the identification of a pattern in the conclusions. Life expectancy parameters were chosen as the main evaluation criterion in this study. First, factors that had a known significant impact on life expectancy were evaluated. Based on an analysis of studies on the organization of research planning, the authors obtained different results depending on the starting timepoint for calculating life expectancy—from the date of pathological confirmation of cancer diagnosis to the date of randomization. Since the study is prospective with consecutive inclusion of patients, and no randomization was performed, the start of anticancer treatment with registration was chosen as the baseline date. Conclusions. This study is dedicated to the treatment of patients with stage III-IV squamous cell carcinoma of the oral mucosa, as well as patients who have undergone unsuccessful previous treatment and experienced recurrence of the disease. Surgery is a lifesaving measure for patients with recurrent disease, since non-surgical treatment methods have already been exhausted or are palliative in nature, especially in cases of tumor resistance to radiation therapy and chemotherapy. The goal of the math is to justify different approaches to the sequence of anticancer treatment components with statistically significant results among subgroups of primary patients (193/83.9%) and patients with relapses (37/16.1%). The correct sequence of statistical analysis of primary data using mathematically sound calculation methods is the basis for obtaining reliable results, on the basis of which conclusions are made in clinical research. An in-depth analysis of the dependence of life expectancy indicators on various factors, with the identification of dependencies, can lead to the improvement of the most variable surgical treatment in terms of a combined and comprehensive approach. Due to the statistically significant absence of normality in the distribution of the “life expectancy” characteristic in the groups, the conditions for the applicability of the parametric Student’s t-test cannot be met.
Keywords: squamous cell carcinoma, oral mucosa, combination therapy, GCP, Shapiro–Wilk test, Kruskal–Wallis test, life expectancy, normal distribution
Conflict of interest. The authors have no conflicts of interest.
Funding. Absent. Informed consent. Patients signed informed consent forms to participate in the study and get their data published.