Cell Maps Representation For Lung Adenocarcinoma Growth Patterns Classification In Whole Slide Images (2311.15847v2)
Abstract: Lung adenocarcinoma is a morphologically heterogeneous disease, characterized by five primary histologic growth patterns. The quantity of these patterns can be related to tumor behavior and has a significant impact on patient prognosis. In this work, we propose a novel machine learning pipeline capable of classifying tissue tiles into one of the five patterns or as non-tumor, with an Area Under the Receiver Operating Characteristic Curve (AUCROC) score of 0.97. Our model's strength lies in its comprehensive consideration of cellular spatial patterns, where it first generates cell maps from Hematoxylin and Eosin (H&E) whole slide images (WSIs), which are then fed into a convolutional neural network classification model. Exploiting these cell maps provides the model with robust generalizability to new data, achieving approximately 30% higher accuracy on unseen test-sets compared to current state of the art approaches. The insights derived from our model can be used to predict prognosis, enhancing patient outcomes.
- World Health Organization, “Cancer,” 2022, Accessed: November 8, 2023.
- Andrew G. Nicholson et al., “The 2021 WHO classification of lung tumors: Impact of advances since 2015,” vol. 17, no. 3, pp. 362–387, 2022.
- Andre L. Moreira et al., “A grading system for invasive pulmonary adenocarcinoma: A proposal from the international association for the study of lung cancer pathology committee,” vol. 15, no. 10, pp. 1599–1610, 2020.
- Kris Lami et al., “Overcoming the interobserver variability in lung adenocarcinoma subtyping: A clustering approach to establish a ground truth for downstream applications,” vol. 147, no. 8, pp. 885–895, 2022.
- “Ai slipping on tiles: Data leakage in digital pathology,” Pattern Recognition. ICPR International Workshops and Challenges, p. 167–182, 2021.
- “Pathologist-level classification of histologic patterns on resected lung adenocarcinoma slides with deep neural networks,” vol. 9, no. 1, pp. 3358, 2019.
- “Growth pattern fingerprinting for automatic analysis of lung adenocarcinoma overall survival,” 2023.
- “Tailoring pretext tasks to improve self-supervised learning in histopathologic subtype classification of lung adenocarcinomas,” vol. 166, pp. 107484, 2023.
- Apaar Sadhwani et al., “Comparative analysis of machine learning approaches to classify tumor mutation burden in lung adenocarcinoma using histopathology images,” vol. 11, no. 1, pp. 16605, 2021.
- Gabriele Campanella et al., “H&e-based computational biomarker enables universal egfr screening for lung adenocarcinoma,” 2022.
- Simon Graham et al., “Hover-net: Simultaneous segmentation and classification of nuclei in multi-tissue histology images,” vol. 58, pp. 101563, 2019.
- “PanNuke: An open pan-cancer histology dataset for nuclei instance segmentation and classification,” in Digital Pathology, 2019, Lecture Notes in Computer Science, pp. 11–19.
- John N. Weinstein et al., “The cancer genome atlas pan-cancer analysis project,” vol. 45, no. 10, pp. 1113–1120, 2013.
- “CoAtNet: Marrying convolution and attention for all data sizes,” in Advances in Neural Information Processing Systems, 2021, vol. 34, pp. 3965–3977.
- Weiming Hu et al., “GasHisSDB: A new gastric histopathology image dataset for computer aided diagnosis of gastric cancer,” vol. 142, pp. 105207, 2022.
- Arkadiusz Gertych et al., “Convolutional neural networks can accurately distinguish four histologic growth patterns of lung adenocarcinoma in digital slides,” Scientific Reports, vol. 9, no. 1, 2019.