---

Background: Ultrasound propagation velocity was measured experimentally in normal, fibroadenoma and ductal carcinoma breast tissues, in order to distinguish normal breast tissue from tumors.

Materials and methods: In quantitative measurements of ultrasound velocity, 403 breast tissue images were selected, comprising 130 normal breast tissue, 130 fibroadenoma, and 143 ductal carcinoma tumors. The cases were implanted in breast tissue mimicking materials and ultrasonic images (A-mode) at 35 ° C were processed and evaluated.

Results: It was observed that ultrasound propagation velocity is an important factor for distinguishing in vitro specimens of fibroadenoma and ductal carcinoma from normal tissue (P-value<0.005). Evaluation of ultrasound velocities showed that from normal breast tissue, fibroadenoma and ductal carcinoma, ultrasound velocity increases respectively. The discriminant functions of types of lesions, based on ultrasound velocity, have been formulated by discriminant analysis. The results indicate that probability of discrimination, sensitivity and specificity for tumors and normal breast tissues are 72, 60 and 100 percents at 35 ° C. With measuring ultrasound velocities, we can distinguish normal breast tissue of from ductal carcinoma and fibroadenoma masses (with the probability of 100%).

Conclusion: It is proposed that probably by measuring attenuation coefficient and ultrasound velocity on time, fibroadenoma and ductal carcinoma tumors can be differentiated well. Iran . J. Radiat. Res., 2003 1(3): 163 – 169

---

Background: We designed an algorithmic model based on the logistic regression analysis and a non-algorithmic model based on the Artificial Neural Network (ANN).

Materials and methods: The ability of these models was compared together in clinical application to differentiate malignant from benign breast tumors in a study group of 161 patients' records. Each patient’s record consisted of 6 subjective features extracted from MRI appearance. These findings were encoded as features for an ANN as well as a logistic regression model (LRM) to predict biopsy outcome. After both models had been trained perfectly on samples (n=100), the validation samples (n=61) were presented to the trained network as well as the established LRMs. Finally, the diagnostic performance of models were compared to that of the radiologist in terms of sensitiv­ity, specificity and accuracy, using receiver operating characteristic curve (ROC) analysis.

Results: The average output of the ANN yielded a perfect sensitivity (98%) and high accuracy (90%) similar to that one of an expert radiologist (96% and 92%) while specificity was smaller than that (67% verses 80%). The output of the LRM using significant features showed improvement in specificity from 60% for the LRM using all features to 93% for the reduced logistic regression model, keeping the accuracy around 90%.

Conclusion: Results show that ANN and LRM prove the relationship between extracted morphological features and biopsy results. Using statistically significant variables reduced LRM outperformed of ANN with remarkable specificity while keeping high sensitivity is achieved. Iran . J. Radiat. Res., 2004 1(4): 217-228

---

Background: Logistic discriminant method was applied to differentiate malignant from benign in a group of patients with proved breast lesions on the basis of ultrasonic parameters.

Materials and Methods: Our database include 273 patients' ultrasonographic pictures consisting of 14 quantitative variables. The measured variables were ultrasound propagation velocity, acoustic impedence and attenuation coefficient at 10 MHz in breast lesions at 20, 25, 30 and 35 º C temperature, physicsl density and age. This database was randomly divided into the estimation of 201 and validation of 72 samples. The estimation samples were used to build the logistic
discriminant model, and validation samples were used to validate the performance. Finally,
important criteria such as sensitivity, specificity, accuracy and area under the receiver operating characteristic curve (ROC) were evaluated.

Results: Our results showed that the logistic discriminant method was able to classify correctly 67 out of 72 cases presented in the validation sample. The results indicate a remarkable
diagnostic accuracy of 93%.

Conclusion: A logistic discriminantor approach is capable of predicting the probability of
malignancy of breast cancer. Features extracted from ultrasonic measurement on ultrasound imaging is used in this approach. Iran . J. Radiat. Res., 2004 2 (1): 27-34

---

 ABSTRACT

 Background: A computer aided diagnosis system was established using the wavelet transform and neural network to differentiate malignant from benign in a

  group of patients with histo-pathologically proved breast lesions based on the data derived independ­ently from time-intensity profile.

  Materials and Methods: The per­formance of the artificial neural network (ANN) was evaluated using a database with 105 patients' records each of which consisted of 8 quantitative parameters mostly derived from time-intensity profile using wavelet transform. These findings were encoded as features for a three-layered neural network to predict the outcome of biopsy. The network was trained and tested using the jack­knife method and its performance was then compared to that of the radiologists in terms of sensitiv­ity, specificity and accuracy using receiver operating characteristic curve (ROC) analysis.

  Results: The network was able to classify correctly the 84 original cases and yielded a comparable diagnostic accuracy (80%), compared to that of the radiologist (85%) by per­forming a constructive association between extracted quantitative data and correspond­ing pathological results (r=0.63, p<0.001).

Conclusion: An ANN supported by wavelet transform can be trained to differentiate malignant from benign breast tumors with a reason­able degree of accuracy.