Volkov Kirill Andreevich (Saratov State Medical University named after V.I. Razumovsky, Saratov, Russian Federation)
Petrunkin Rodion Pavlovich (University «Reaviz», Saint Petersburg, Russian Federation)
Polidanov Maxim Andreevich (research department specialist, assistant of the Department of Biomedical Disciplines, University «Reaviz», Saint Petersburg, Russian Federation)
Dolgova Elena Mikhailovna (Candidate of Medical Sciences, Associate Professor of the Department of Public Health and Public Health Care (with courses in Law and History of Medicine), Saratov State Medical University named after V.I. Razumovsky, Saratov, Russian Federation)
Kravchenya Aliya Rimovna (Candidate of Medical Sciences, Associate Professor of the Department of Pediatric Diseases of the Faculty of Medicine, Saratov State Medical University named after V.I. Razumovsky, Saratov, Russian Federation)
Kapralov Sergey Vladimirovich (Doctor of Medicine Sciences, associate professor, Head of the Department of Faculty Surgery and Oncology, Saratov State Medical University named after V.I. Razumovsky, Saratov, Russian Federation)
| |
Objective. Consideration of predicting the occurrence of recurrence in the postoperative period in thyroid cancer using artificial intelligence. Materials and methods. During the study we analyzed the data of case histories of 106 patients who underwent surgical intervention for thyroid cancer. The average age was 43,54 years. Based on a set of examination results, we selected patients who met the following inclusion criteria: patients with thyroid cancer without confirmed metastases with disease stage from T1N0M0 to T3N0M0; absence of previous and concomitant special treatment (immunotherapy or targeted therapy); informed consent for the surgical intervention. Logistic regression, a binary classifier using a sigmoidal activation function on linear combinations of features, was used as a machine learning model. Results. The majority of patients (60,1%) underwent thyroidectomy and 39,9% underwent subtotal thyroid resection. The number of patients with occurrence of postoperative recurrence is 138. The data set was unbalanced, therefore, it was decided to take into account only the fact of presence or absence of postoperative recurrence to reduce the unbalanced data set, i.e. during training and testing the system will use the target feature divided into only two categories – «there is a recurrence» or «no recurrence». It was found that on the selected parameters (total calcium; REA; cytologic classification after TAB according to Bethesda system; parathormone after surgery) the constructed logistic regression model predicts quite well the possible occurrence of recurrences after surgical intervention for thyroid cancer. Conclusions. The obtained results show that on the basis of only 4 parameters (total calcium; REA; cytologic classification after TAB according to the Bethesda system; parathormone after surgery) it is possible to build a good enough model for predicting the occurrence of recurrences after surgical intervention for thyroid cancer on the basis of such a machine learning method as logistic regression.
Keywords:Thyroid cancer; diagnosis; recurrence prediction; machine learning; logistic regression; artificial intelligence.
|
|
| |
|
Read the full article …
|
Citation link:
Volkov K. A., Petrunkin R. P., Polidanov M. A., Dolgova E. M., Kravchenya A. R., Kapralov S. V. POSSIBILITIES OF PREDICTING THE OCCURRENCE OF RECURRENCE IN THE POSTOPERATIVE PERIOD IN THYROID CANCER WITH THE HELP OF ARTIFICIAL INTELLIGENCE // Современная наука: актуальные проблемы теории и практики. Серия: Естественные и Технические Науки. -2025. -№09. -С. 145-148 DOI 10.37882/2223-2966.2025.09.06 |
|
|
