This paper provides a practical evaluation of few-shot segmentation methods, most of which were originally designed for natural image segmentation. Our evaluation focuses on histopathology images, which differ significantly from natural images in their characteristics. We assess the effectiveness of these methods in segmenting small cell boundaries, addressing the unique challenges of medical imaging data, such as irregular object shapes. While these methods achieve competitive performance on natural images, we demonstrate that they face significant limitations when applied to histopathology, particularly in accurately distinguishing small, densely packed cells. Our findings highlight the need for domain-specific adaptations to improve the reliability and precision of segmentation methods for histopathology images.