Five abdominal computer tomography characteristics facilitate diagnosis of spontaneous perforation of pyometra in women with acute abdomen: a case control study

Background: Spontaneous perforation of pyometra is a rare clinical emergency with high mortality. As pyometra perforation is difficult to recognize pre-operatively, many cases are mis-diagnosed as gastrointestinal (GI) rupture. This study aimed to develop a strategy to improve diagnostic accuracy of spontaneous uterine perforation in female patients with acute abdomen. Methods: Twenty-two women with spontaneous rupture of the uterus (twenty cases previously reported in the literature plus two cases from the authors' hospital) were included in the study group, another 22 menopausal patients with GI perforation were selected as the control group. Radiologists carefully reviewed the CT images and reports and summarized some common signs in patients with spontaneous perforation of pyometra. GI patients' CT images were also scrutinized for these characteristic signs. The occurrence of these features was calculated in both groups. All patients underwent abdominal surgical exploration and most of patients diagnosed with spontaneous perforation of pyometra received hysterectomy and bilateral oophorectomy. Results: Five characteristics were specifically detected in patients with uterine spontaneous perforation: (1) Fluid collected in the cul-de-sac, incidence 95.5%; (2) Fluid in the uterine cavity, incidence 100%; (3) Intrauterine free air, incidence 90.9%; (4) Free air around the uterus, incidence 77.2%; (5) Defects on the uterine wall, incidence 72.7%. However, these five signs were rarely observed in control patients with GI perforation. A dignostic strategy combined clinical findings with raidiological features were developed to help identify these patients.Conclusions: The above five CT characteristics and the developed dig- nostic algorithm could be used as a tool to facilitate the preoperative diagnosis of spontaneous perforation of pyometra, although more tests are needed in large samples of cases.