Computer-aided diagnosis for surgical office-based breast ultrasound

Hypothesis: The computer-aided diagnostic system is an intelligent system w ith great potential for categorizing solid breast nodules. It can be used c onveniently for surgical office-based digital ultrasonography (US) of the b reast. Design: Retrospective, nonrandomized study. Setting: University teaching hospital. Patients: We retrospectively reviewed 243 medical records of digital US ima ges of the breast of pathologically proved, benign breast tumors from 161 p atients (ie, 136 fibroadenomas and 25 fibrocystic nodules), and carcinomas from 82 patients (ie, 73 invasive duct carcinomas, 5 invasive lobular carci nomas, and 4 intraductal carcinomas). The digital US images were consecutiv ely recorded from January 1, 1997, to December 31, 1998. Intervention: The physician selected the region of interest on the digital US image. Then a learning vector quantization model with 24 autocorrelation texture features is used to classify the tumor as benign or malignant. In the experiment, 153 cases were arbitrarily selected:to be the training set of the learning vector quantization model and 90 cases were selected to eva luate the performance. One experienced radiologist who was completely blind to these cases was asked to classify these tumors in the test set. Main Outcome Measure: Contribution of breast US to diagnosis. Results: The performance comparison results illustrated the following: accu racy, 90%: sensitivity, 96.67%;specificity, 86.67%; positive predictive val ue, 78.38%; and negative predictive value, 98.11% for the computer-aided di agnostic (CAD) system and accuracy, 86.67%; sensitivity, 86.67%; specificit y, 86.67%; positive predictive value, 76.47%; and negative predictive value , 92.86% for the radiologist. Conclusion: The proposed CAD system provides an immediate second opinion. A ll accurate preoperative diagnosis can be routinely established for surgica l office-based digital US of the breast. The diagnostic rate was even bette r than the results of an experienced radiologist. The high negative predict ive rate by the CAD system can avert benign biopsies. It call be easily imp lemented on exisiting commercial diagnostic digital US machines. For most a vailable diagnostic digital US machines, all that would be required for the CAD system is only a personal computer loaded with CAD software.