Dados do Trabalho

Título

USE OF ARTIFICIAL INTELLIGENCE FOR IDENTIFICATION OF TOXICITY PATTERNS IN PATIENTS WITH RENAL CANCER TREATED WITH TYROSINE QUISANE INHIBITORS

Introdução

Sunitinib, the first tyrosine kinase inhibitor (TKI) aproved for the systemic treatment of patients with metastatic renal cell carcinoma (mRCC), showed a longer progression-free survival and higher responses rates when compared with interferon alfa in a phase III trial. A high proportion of patients treated with sunitinib have a significant clinical and laboratory toxicities, which impact the quality of life of patients and limit option treatment. The use of new computational technologies, such as Artificial Intelligence (AI), allows the identification of patients at higher risk of developing toxicity, which may allow the application of individual preventive measures to reduce these toxicities.

Objetivo

Identification of toxicity patterns in patients with mRCC when treated with TKI

Método

Retrospective study of patients mRCC undergoing treatment with TKI in the period 2008-2020. For data analysis, the R program was used to generate statistical data. The computational models chosen for pattern recognition are those established in the machine learning area: Support Vector Machines, Logistic Regression, Linear Discriminant Analysis, Decision Trees and Bayesian Networks.

Resultado

37 patients diagnosed with RCC were included; 24 males; 13 female; Median age of 57 years for both sex; The median progression-free survival was 19 months (CI95 12.8 to 31.2 months) and the median for overall survival was aproximately 65 months. Anemia grade II - 7(18,9%) grade III- 6(16,2%); Neutropenia grade III - 6(16,2%); grade IV - 6(16,2%); Lymphopenia grade II-9(24,3%) grade III - 1(2,7%); grade IV - 1(2,7%); Nephrotoxicity with grade I- 14 (37,8%); grade II-18 (48,6%); grade III - 5(13,5%); Hepatotoxicity grade I - 17 (47,2%); grade II - 2 (5,6%).

Conclusão

The mean age and gender prevalence of patients were in agreement with the literature. All patients had some degree of toxicity during treatment with sunitinib, with haematological and nephrotoxicity being the most prevalent. The correlation between the occurrence of hematological toxicities, especially neutropenia and lymphopenia, showed a high correlation. Neutropenia seems to be the toxicity with the highest correlation with the others, which may be an important factor in the evaluation of AI models.

Palavras-chave

Clear cells renal cell carcinoma; Sunitinib; Toxicity;