The increasing use of automated cancer diagnosis based on histopathological images is significant because it is likely to increase the accuracy of diagnosis and decrease the workload on pathologists. This research introduces a hybrid methodology that integrates Haralick texture features with deep learning strategies to improve the automated identification of cancer in human tissue specimens. Haralick texture features, obtained from the Gray-Level Co-Occurrence Matrix (GLCM), offer essential information regarding the spatial relationships and textural characteristics present in tissue samples, which frequently signal the presence of cancerous alterations. The integration of these interpretable texture features with convolutional neural networks (CNNs) makes our approach use the strengths of both traditional texture analysis and deep learning's ability to learn complex patterns. This will process raw image data with the Haralick features leading to a powerful model that, hopefully, makes better classification along with interpretability. These features, handcrafted and capturing features like contrast, correlation, energy, and homogeneity, provide differences in the texture of the tissue that classify between normal cells and abnormal ones. Experimental results were presented in distinguishing cancerous and non-cancerous tissues with high accuracy. The diagnostic efficiency was also enhanced while at the same time providing a reliable and scalable tool that may assist pathologists during clinical decision-making, which consequently leads to efficient cancer diagnosis and patient care. © 2025 Elsevier B.V., All rights reserved.