| Title |
Mortality rate estimation models of patients with prostate cancer diagnosis / |
| Authors |
Ruzgas, Tomas ; Kraujalis, Vytautas ; Milonas, Daimantas |
| DOI |
10.15388/DAMSS.12.2021 |
| ISBN |
9786090706732 |
| eISBN |
9786090706749 |
| Full Text |
|
| Is Part of |
DAMSS 2021: 12th conference on data analysis methods for software systems, Druskininkai, Lithuania, December 2–4, 2021.. Vilnius : Vilnius University press, 2021. p. 62-63.. ISBN 9786090706732. eISBN 9786090706749 |
| Abstract [eng] |
Prostate cancer is one of the most frequent type of male cancer all around the world, including Lithuania. Prostate cancer diagnosis takes up to 25% of all Lithuanian men cancer diagnosis in the country. Every year, ~3000 new prostate cancer cases are diagnosed and ~500 Lithuanian men die from this illness. It is necessary for a medical professional to be able to distinguish a fatal and non-fatal cancer in time, statistical methods and models could be implemented to help with this case and in our work we will try to implement those methods and models using data from “Kauno Klinikos” clinic (Kaunas, Lithuania). During the research we used well known Kaplan-Meier survival curves as well as compared 2 best known hazard estimation models: Cox and Fine-Gray models. In this dataset, there were 56 deaths reported from cancer specific causes and 294 from other causes, the median age of a patient was 64 years (n - 2410). During the analysis of Kaplan-Meier survival curves, we found the worst survival prognosis associated with patients, who’s lymph nodes were damaged by cancer or patients with 5 metastatic lymph nodes. It was also discovered that patients with 1 or 2 metastatic lymph nodes are much more likely to experience death from one of the causes – cancer specific cause or other causes than men with 4 metastatic lymph nodes. Significant hazard ratio was also found between men, who’s cancer is developing in the prostate area and men, who’s cancer has already spread outside the prostate cancer, with the later one having the worse survival prognosis. Patients with cancer damaged lymph nodes have a higher mortality rate than men with untreated lymph nodes only from cancer specific causes while the hazard ratio linked with man’s age was found significant only in deaths from other causes. Fine and Gray hazard estimation model distinguished less significant risk factors and usually the hazard ratios were reported smaller than the ones in the Cox model. Training and testing datasets were used to test the performance of both models and Cox model was found to be optimal on both datasets in response to the ROC curves analysis. |
| Published |
Vilnius : Vilnius University press, 2021 |
| Type |
Conference paper |
| Language |
English |
| Publication date |
2021 |
| CC license |
|
