Purpose: To evaluate retrospectively the results of consensus double reading of mammograms in a private practice for a period of 1.5 years (November 2001 to March 2003).
Materials and method: Two independent experts with dedicated training read all mammograms on a weekly basis. All mammograms including sonographic examinations were evaluated independently and categorized using the BI-RADS classification. The achieved consensus included a possible recommendation for recall or therapy. A total of 3936 mammograms and 1912 sonography studies were evaluated. All cases with BI-RADS 4 and 5 categories were compared with the histologic results. For a period of three months, the acceptance of double reading including a delay of the final report by one week was tested with a questionnaire and informed consent sheet.
Results: BI-RADS categories 4 and 5 were found in 57 cases, with 41 consensus results by two independent readers and 26 carcinomas verified by histology. No consensus could be reached in 16 patients, of which 10 had a final histologic result, with 5 benign lesions and 5 carcinomas of less than 1 cm in diameter. Clinical symptoms or alterations were absent in all patients. The 5 carcinomas were discovered by the double reading procedure. The result of the questionnaire (695 questionnaires) showed a refusal rate of 0.7 %, with only 5 women refusing the opportunity of double reading their mammograms.
Conclusion: Double reading of mammograms by independent experts is feasible, shows a measurable increase in quality and is accepted by almost all women.