THE ROLE OF CT IN PREDICTING THE SURGICAL FEASIBILITY OF EXENTERATIONIN RECURRENT CARCINOMA OF THE CERVIX

In women with recurrent cervical cancer, there is a 52% 5-year surviva l rate when exenteration of the centrally placed tumor is performed. I f not exenterated, the 2-year survival rate is less than 2%. The purpo se of this study is to ascertain whether CT can predict inoperability reliably leading to a reduction in operations which do not have a cura tive outcome. The records of patients with recurrence of cervical canc er who underwent laparotomies for exenteration over the last 10 years under the care of one surgeon were reviewed. The CT scans were retrosp ectively assessed by two radiologists without knowledge of the subsequ ent outcome of the laparotomy. CT scans on 31 patients were reviewed, 21 of whom underwent a radical surgical procedure. Using CT criteria, 9 cases were felt to be operable with a curative intent and the remain ing 22 cases were deemed to be inoperable. The sensitivity of CT predi ction of inoperability is 93% (95%Cl:66-100%) and the specificity is 4 7% (95%Cl:23-72%). In three cases ascites was the only abnormal findin g other than the central pelvic mass and in all of these cases a radic al procedure with clear resection margins was possible. When the group with ascites alone was not considered to have peritoneal disease unde rlying the ascites, the sensitivity of CT prediction of inoperability is 93% (95%Cl:66-100%) and the specificity of 65% (95%Cl:38-86%). If, in addition, lymphadenopathy is not taken as definite evidence of inop erability, the specificity rises to 82% (95%Cl:57-96%). A high quality CT scan is highly specific for predicting inoperability based on exte nsion of the tumor to the pelvic side walls, encasement of adjacent ve ssels or ureteric dilatation and so should be a major tool in assessin g women for radical surgical treatment of recurrent cervical cancer. I f ascites is the only abnormal finding other than the central pelvic m ass then exploration should be undertaken.