PURPOSE: Endoluminal ultrasonography (ELUS) is accurate in the assessment of penetration through the rectal wall by carcinoma. Clinical studies were performed to determine the reliability and validity of ELUS. METHODS: The interobserver reliability among four observers with varying experience with ELUS was determined for staging the penetration of rectal cancer through the rectal wall. The ability of ELUS to change the clinical management of the referring clinician (comprehensiveness) was assessed on all referrals over a six-month period. RESULTS: The reliability of ELUS for staging rectal cancer demonstrated only fair to moderate correlation (weighted kappa range, 0.22-0.47). The accuracy of ELUS compared with surgical pathology demonstrated a learning curve proportional to the experience of the observer. In 45 percent of referrals, ELUS changed the clinical management of patients and in 76 percent of referrals the clinician's confidence in the diagnosis and management of patients was altered. ELUS was more likely to change the management of patients with pelvic pouch sepsis (70 percent) and early neoplastic lesions (57 percent) than in more advanced neoplastic lesions (40 percent), perianal Crohn's disease (40 percent), complex noninflammatory bowel disease sepsis (33 percent), and incontinence (31 percent). CONCLUSIONS: ELUS has the ability to change the clinical management of a variety of anorectal conditions. However, for neoplasia the interobserver reliability is only moderate and a learning curve exists.
PURPOSE: Endoluminal ultrasonography (ELUS) is accurate in the assessment of penetration through the rectal wall by carcinoma. Clinical studies were performed to determine the reliability and validity of ELUS. METHODS: The interobserver reliability among four observers with varying experience with ELUS was determined for staging the penetration of rectal cancer through the rectal wall. The ability of ELUS to change the clinical management of the referring clinician (comprehensiveness) was assessed on all referrals over a six-month period. RESULTS: The reliability of ELUS for staging rectal cancer demonstrated only fair to moderate correlation (weighted kappa range, 0.22-0.47). The accuracy of ELUS compared with surgical pathology demonstrated a learning curve proportional to the experience of the observer. In 45 percent of referrals, ELUS changed the clinical management of patients and in 76 percent of referrals the clinician's confidence in the diagnosis and management of patients was altered. ELUS was more likely to change the management of patients with pelvic pouch sepsis (70 percent) and early neoplastic lesions (57 percent) than in more advanced neoplastic lesions (40 percent), perianal Crohn's disease (40 percent), complex noninflammatory bowel disease sepsis (33 percent), and incontinence (31 percent). CONCLUSIONS: ELUS has the ability to change the clinical management of a variety of anorectal conditions. However, for neoplasia the interobserver reliability is only moderate and a learning curve exists.