CLASSIFICATION MODEL FOR BREAST CANCER MAMMOGRAMS

Machine learning has been the topic of interest in research related to early detection of breast cancer based on mammogram images. In this study, we compare the performance results from three (3) types of machine learning techniques: 1) Naïve Bayes (NB), 2) Neural Network (NN) and 3) Support Vector Machine (SVM) with 2000 digital mammogram images to choose the best technique that could model the relationship between the features extracted and the state of the breast (‘Normal’ or ‘Cancer’). Grey Level Co-occurrence Matrix (GLCM) which represents the two dimensions of the level variation gray in the image is used in the feature extraction process. Six (6) attributes consist of contrast, variance, standard deviation, kurtosis, mean and smoothness were computed as feature extracted and used as the inputs for the classification process. The data has been randomized and the experiment has been repeated for ten (10) times to check for the consistencies of the performance of all techniques. 70% of the data were used as the training data and another 30% used as testing data. The result after ten (10) experiments show that, Support Vector Machine (SVM) gives the most consistent results in correctly classifying the state of the breast as ‘Normal’ or ‘Cancer’, with the accuracy of 99.4%, in training and 98.76% in testing. The SVM classification model has outperformed NN and NB model in the study, and it shows that SVM is a good choice for determining the state of the breast at the early stage. ABSTRAK: Pembelajaran mesin telah menjadi topik yang diminati dalam penyelidikan yang berkaitan dengan pengesanan awal kanser payudara berdasarkan imej mamogram. Dalam kajian ini, kami membandingkan hasil prestasi dari tiga (3) jenis teknik pembelajaran mesin: 1) Naïve Bayes (NB), 2) Neural Network (NN) dan 3) Support Vector Machine (SVM) dengan 2000 imej digital mammogram hingga teknik terbaik yang dapat memodelkan hubungan antara ciri yang diekstraksi dan keadaan payudara ('Normal' atau 'Cancer') dapat diperoleh. Grey Level Co-occurrence Matrix (GLCM) yang mewakili dua dimensi variasi tahap kelabu pada gambar digunakan dalam proses pengekstrakan ciri. Enam (6) atribut terdiri dari kontras, varians, sisihan piawai, kurtosis, min dan kehalusan dihitung sebagai fitur yang diekstrak dan digunakan sebagai input untuk proses klasifikasi. Eksperimen telah diulang selama sepuluh (10) kali untuk memeriksa kesesuaian prestasi semua teknik. 70% data digunakan sebagai data latihan dan 30% lagi digunakan sebagai data ujian. Hasil setelah sepuluh (10) eksperimen menunjukkan bahawa, Support Vector Machine (SVM) memberikan hasil yang paling konsisten dalam mengklasifikasikan keadaan payudara dengan betul sebagai 'Normal' atau 'Kanser', dengan akurasi 99.4%, dalam latihan dan 98.76% dalam ujian. Model klasifikasi SVM telah mengungguli model NN dan NB dalam kajian ini, dan ia menunjukkan bahawa SVM adalah pilihan yang baik untuk menentukan keadaan payudara pada peringkat awal.

Cancer Conspicuity on Low-energy Images of Contrast-enhanced Mammography Compared With 2D Mammography

Objective Low-energy (LE) images of contrast-enhanced mammography (CEM) have been shown to be noninferior to digital mammography. However, our experience is that LE images are superior to 2D mammography. Our purpose was to compare cancer appearance on LE to 2D images. Methods In this IRB-approved retrospective study, seven breast radiologists evaluated 40 biopsy-proven cancer cases on craniocaudal (CC) and mediolateral oblique (MLO) LE images and recent 2D images for cancer visibility, confidence in margins, and conspicuity of findings using a Likert scale. Objective measurements were performed using contrast-to-noise ratio (CNR) estimated from regions of interest placed on tumor and background parenchyma. Reader agreement was evaluated using Fleiss kappa. Per-reader comparisons were performed using Wilcoxon test and overall comparisons used three-way analysis of variance. Results Low-energy images showed improved performance for visibility (CC LE 4.0 vs 2D 3.5, P < 0.001 and MLO LE 3.7 vs 2D 3.5, P = 0.01), confidence in margins (CC LE 3.2 vs 2D 2.8, P < 0.001 and MLO LE 3.1 vs 2D 2.9, P < 0.008), and conspicuity compared to tissue density compared to 2D mammography (CC LE 3.6 vs 2D 3.2, P < 0.001 and MLO LE 3.5 vs 2D 3.2, P < 0.001). The average CNR was significantly higher for LE than for digital mammography (CC 2.1 vs 3.2, P < 0.001 and MLO 2.1 vs 3.4, P < 0.001). Conclusion Our results suggest that cancers may be better visualized on the LE CEM images compared with the 2D digital mammogram.

Can artificial intelligence replace ultrasound as a complementary tool to mammogram for the diagnosis of the breast cancer?

Objectives: to study the impact of artificial intelligence (AI) on the performance of mammogram with regard to the classification of the detected breast lesions in correlation to ultrasound aided mammograms. Methods: Ethics committee approval was obtained in this prospective analysis. The study included 2000 mammograms. The mammograms were interpreted by the radiologists and breast ultrasound was performed for all cases. The Breast Imaging Reporting and Data System (BI-RADS) score was applied regarding the combined evaluation of the mammogram and the ultrasound modalities. Each breast side-was individually assessed with the aid of AI scanning in the form of targeted heat-map and then, a probability of malignancy (abnormality scoring percentage) was obtained. Operative and the histopathology data were the standard of reference. Results: Normal assigned cases (BI-RADS 1) with no lesions were excluded from the statistical evaluation. The study included 538 benign and 642 malignant breast lesions (n = 1180, 59%). BI-RADS categories for the breast lesions with regard to the combined evaluation of the digital mammogram and ultrasound were assigned BI-RADS 2 (Benign) in 385 lesions with AI median value of the abnormality scoring percentage of 10, (n = 385/1180, 32.6%), and BI-RADS 5 (malignant) in 471, that had showed median percentage AI value of 88 (n = 471/1180, 39.9%). AI abnormality scoring of 59% yielded a sensitivity of 96.8% and specificity of 90.1% in the discrimination of the breast lesions detected on the included mammograms. Conclusions: AI could be considered as an optional primary reliable complementary tool to the digital mammogram for the evaluation of the breast lesions. The color hue and the abnormality scoring percentage presented a credible method for the detection and discrimination of breast cancer of near accuracy to the breast ultrasound. So consequently, AI- mammogram combination could be used as a one setting method to discriminate between cases that require further imaging or biopsy from those that need only time interval follows up. Advances in knowledge: Recently, the indulgence of AI in the work up of breast cancer was concerned. AI noted as a screening strategy for the detection of breast cancer. In the current work, the performance of AI was studied with regard to the diagnosis not just the detection of breast cancer in the mammographic-detected breast lesions. The evaluation was concerned with AI as a possible complementary reading tool to mammogram and included the qualitative assessment of the color hue and the quantitative integration of the abnormality scoring percentage.