Logistic models for prediction of enteric morbidity in the treatment of ovarian and cervical cancers

Roger A. Potish, Leo B. Twiggs, Leon L. Adcock, Konald A. Prem

Research output: Contribution to journalArticle

34 Scopus citations

Abstract

To identify patients at high risk for the development of enteric injury, an analysis was made of 212 women who received extensive pelvic and abdominal radiotherapy at the University of Minnesota from 1970 through 1981. One hundred one patients with cervical carcinomas received 8,000 to 8,500 rads to point A, 6,000 rads to point B, and 4,500 to 5,075 rads to the periaortic lymph nodes. One hundred eleven women with ovarian cancers received 2,000 rads to the entire abdomen, followed by an additional 2,975 to 3,000 rads to the pelvis. The overall complication rate was 6.6% ( 14 212). The only patients who sustained chronic radiation morbidity had thin physiques. The maximum likelihood multiple logistic model was utilized to predict the probability of enteric injury as a function of thin physique, previous operations, and hypertension for individual patients. The usefulness of predictive models is discussed, and possible reasons for the susceptibility of thin women to ionizing radiation are explored.

Original languageEnglish (US)
Pages (from-to)65-72
Number of pages8
JournalAmerican journal of obstetrics and gynecology
Volume147
Issue number1
DOIs
StatePublished - Sep 1 1983

ASJC Scopus subject areas

  • Obstetrics and Gynecology

Fingerprint Dive into the research topics of 'Logistic models for prediction of enteric morbidity in the treatment of ovarian and cervical cancers'. Together they form a unique fingerprint.

  • Cite this