Foundation

the toxicity associated with chemotherapy and radiotherapy treatments increases morbidity and mortality in cancer patients.

Objective

to design a predictive model of chemotherapy and radiotherapy toxicity in surgical cancer patients.

Methods

analytical, case-control study, in surgical oncology patients who met the inclusion criteria for the prediction of preoperative toxicity, from January to December 2022, at the María Curie Provincial Teaching Oncology Hospital in Camagüey. Using the Statistical Package for the Social Sciences, a random sample of 334 patients was selected, 197 without toxicity (control group) and 137 with toxicity (study group). Toxicity predictors were estimated using binary logistic regression. The model with the best fit was selected.

Results

the model in step three predicts an overall percentage of 83.5% with respect to the observed values. The sensitivity turned out to be 81.8; and the specificity, 84.8. The model presented good discriminative power. The variables in the equation were: arterial hypertension, left ventricular ejection fraction, and anemia. The comparison of the prediction with reality, using the Receiver Operating Characteristic curve, determined an area under the curve of 0.901.

Conclusion

a logistic regression function was obtained that allowed the estimation of the toxicity probability elective surgical cancer patients, which provided a tool for its prediction from the preoperative period.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )