Conventional object detection models are typically trained to detect objects in images devoid of sequential correlation. However, in the context of MRI imaging, where data is inherently sequential, valuable information lies within the adjacent slices for cancer detection. While 3D Neural Networks present a logical avenue to exploit this sequential nature, their high parameter counts, and complex training dynamics pose significant challenges. Additionally, mutual attention mechanisms, although effective, often require a large volume of training data, which may not always be readily available in medical settings. In response to these challenges, our proposed method efficiently harnesses sequential characteristics of MRI data by integrating information from preceding, current, and succeeding slices. Notably, this integration is strategically incorporated into specific segments of the network architecture to optimize performance. We provide empiric results and give practical advice for using our method with existing architectures. By seamlessly integrating this approach into the YOLOv5 framework, we achieved a notable 2.8% enhancement in the mean Average Precision (mAP).