RESUMO
OBJECTIVE: Innovative techniques in the field of artificial intelligence could help to resolve several methodological problems. A model taking into account all the parameters involved in a therapy can foresee the results of each type of treatment or therapeutic protocol on patients at different stages of a disease. We used a Computer Decision Support System in order to verify the reliability and efficacy of this method on chemotherapy of colorectal carcinoma. MATERIAL AND METHODS: We analyzed 8 randomized clinical trials employing 5-fluorouracil alone (5-FU) or 5-fluorouracil (5-FU) plus leucovorin (FA) in the management of advanced colorectal carcinoma. Computer Decision Support System (CDSS) was used to perform four basic tasks: data acquisition and organization; data recruitment; combination of the various principles and specific data; user-friendly display of the analysis results and responses to treatment. RESULTS: In the majority of the studies examined, the death rates were lower in patients treated with 5-FU + FA than in those on 5-FU alone, even though the difference was not statistically significant. However, there were wide fluctuations in the efficacy/tolerability ratio between the two protocols investigated, depending on the patients' clinical status. Our data showed that a strong attack using 5-FU + FA is feasible whenever the patients' clinical conditions are not particularly severe, whereas a moderate attack using 5-FU alone is recommended as the patients' clinical condition worsens. CONCLUSION: The use of CDSS in the management of colorectal carcinoma indicates which therapy is the best in terms of efficacy, overall survival and incidence of side effects.