Breast cancer detection depends on accurate imaging techniques that capture the full complexity of the breast tissue. Angular breast images, particularly the mediolateral oblique and mediolateral projections, play a vital role in traditional mammography, offering comprehensive views that include difficult-to-visualize areas. However, infrared imaging, which measures surface temperature distribution, provides a complementary approach by highlighting thermal anomalies that may indicate underlying malignancies. While frontal infrared images are commonly used, they often fail to capture the entire thermal profile of the breast, particularly in areas obscured from the frontal view. To address this limitation, angular infrared images can be integrated to provide a more complete analysis, although this approach introduces challenges in image processing and interpretation. In this paper, we present a model to address the problem of breast cancer detection specifically on angular images, consisting of a 3D convolutional neural network (CNN) that integrates lateral and oblique views into a single volumetric representation, allowing for comprehensive analysis of the breast. The results of our ablation experiment demonstrates an architecture that effectively classifies breast cancer, achieving 93.82% accuracy, 93.82% sensitivity, and an area under the receiver operating characteristic curve of 97.22%, which highlights the model’s efficacy in distinguishing between healthy and cancerous cases. The use of a 3D CNN for this problem is well-suited given the volumetric nature of the data, making it a promising approach for further exploration in medical applications.