Leukemia comprises a diverse group of malignancies which is accompanied with genetic disorderliness in hematopoietic cells. We evaluated effective risk factors in recovery process of under treatment patients suffering...Leukemia comprises a diverse group of malignancies which is accompanied with genetic disorderliness in hematopoietic cells. We evaluated effective risk factors in recovery process of under treatment patients suffering from acute myeloblastic leukemia （AML）. This study conducted a cross-sectional descriptive-analytical study on a population of 76 samples obtained non-randomly from patients in Taleghani Hospital （Tehran, Iran）. 30.3% patients resulted in death. According to logistic regression results, sexes [OR = 6.40, 95% CI = （0.27, 3.45）], ALT [OR = 1.03, 95% CI = （0.01, 0.05）] and HCT [OR = 0.55, 95% CI = （-1.12, -0.06）] were recognized as significant in prognoses. We predicted the probability of death with an error of 20.03% based on a prognoses system using support vector machine （SVM） classifier. Using this theory, we experienced an error of 20.03%. 46.6% patients with a positive and 20.8% patients without positive drug history resulted in death, which shows a significant correlation between patients＇ drug history and their death.展开更多
文摘Leukemia comprises a diverse group of malignancies which is accompanied with genetic disorderliness in hematopoietic cells. We evaluated effective risk factors in recovery process of under treatment patients suffering from acute myeloblastic leukemia （AML）. This study conducted a cross-sectional descriptive-analytical study on a population of 76 samples obtained non-randomly from patients in Taleghani Hospital （Tehran, Iran）. 30.3% patients resulted in death. According to logistic regression results, sexes [OR = 6.40, 95% CI = （0.27, 3.45）], ALT [OR = 1.03, 95% CI = （0.01, 0.05）] and HCT [OR = 0.55, 95% CI = （-1.12, -0.06）] were recognized as significant in prognoses. We predicted the probability of death with an error of 20.03% based on a prognoses system using support vector machine （SVM） classifier. Using this theory, we experienced an error of 20.03%. 46.6% patients with a positive and 20.8% patients without positive drug history resulted in death, which shows a significant correlation between patients＇ drug history and their death.
