Enhancing Cancer Treatment with AI: Deep Learning for Predicting Neoadjuvant Therapy Response

Abstract

Research Problem/Questions— Stomach cancer is one of the most lethal malignancies worldwide, ranking as the sixth most common cancer and the third leading cause of cancer-related deaths. Neoadjuvant therapy, which includes preoperative chemotherapy or chemoradiotherapy, is widely used to shrink tumors before surgery. However, accurately assessing a patient’s response to this treatment remains a significant challenge due to the limitations of traditional histopathological evaluation methods, which are time-consuming and subject to interobserver variability. To address this issue, artificial intelligence (AI) and deep learning models offer a more efficient, accurate, and scalable approach for evaluating treatment responses. Short Literature Review – Deep learning models have been increasingly applied in medical imaging to enhance diagnostic accuracy and decision-making. YOLOv9, a real-time object detection model, has demonstrated high precision in identifying tumor regions in histopathology slides. EfficientDet, on the other hand, enables multi-scale feature extraction and classification, allowing for a more detailed analysis of tumor characteristics and progression . Previous studies have explored AI-based image segmentation and classification in oncology, highlighting the potential of machine learning models to standardize pathology assessments and reduce diagnostic errors . Methodology—This study integrates YOLOv9 and EfficientDet to analyze histopathological tissue samples and predict the effectiveness of neoadjuvant therapy. The research methodology includes data collection from medical institutions, preprocessing and augmentation of histopathological images, transfer learning for model optimization, and performance evaluation using precision, recall, and F1-score metrics. Additionally, AI-generated pathology reports provide automated and standardized assessments for clinicians, improving decision-making in cancer treatment planning . Results and Conclusions— Preliminary findings indicate that this AI-driven approach significantly improves the accuracy of neoadjuvant therapy response prediction, surpassing conventional manual assessments. The integration of deep learning models in histopathological analysis presents a promising future for personalized medicine, early cancer detection, and optimized treatment strategies. This study highlights the potential of AI to revolutionize oncology, paving the way for faster and more reliable cancer diagnostics.

Keywords

• Artificial Intelligence
• Deep Learning
• Tumor Detection
• YOLOv9
• EfficientDet
• Neoadjuvant Therapy
• Pathology Analysis

Yapay Zeka ile Kanser Tedavisini Geliştirme: Neoadjuvan Tedavi Yanıtını Tahmin Etmek İçin Derin Öğrenme

Abstract

Mide kanseri, dünya genelinde en ölümcül hastalıklardan biri olup, kanserle ilişkili ölümler arasında üçüncü sırada yer almaktadır. Neoadjuvan tedavi, ameliyat öncesinde tümörleri küçültmek için uygulanan bir yöntemdir. Ancak, bu tedaviye verilen yanıtın doğru bir şekilde değerlendirilmesi, geleneksel histopatolojik incelemelerde zaman alıcı olması ve gözlemciler arası farklılık göstermesi nedeniyle zorluklar barındırmaktadır. Bu çalışmada, yapay zeka ve derin öğrenme tabanlı modellerin bu süreci daha hızlı, doğru ve standart hale getirme potansiyeli araştırılmıştır. Bu kapsamda, YOLOv9 ve EfficientDet modelleri kullanılarak histopatolojik görüntüler analiz edilmiştir. YOLOv9, gerçek zamanlı tümör tespiti sağlarken, EfficientDet çok ölçekli özellik çıkarımı ve sınıflandırma işlemlerinde kullanılmıştır. Çalışmada, veri toplama, görüntü işleme, transfer öğrenme ve model optimizasyonu aşamaları uygulanarak modellerin performansı F1-skora, doğruluk ve geri çağırma metriklerine göre değerlendirilmiştir. Ön bulgular, yapay zeka destekli analizlerin, neoadjuvan tedavi yanıtlarını tahmin etmede geleneksel yöntemlerden daha başarılı olduğunu göstermektedir. Yapay zeka tabanlı modellerin patolojik incelemelere entegre edilmesi, erken teşhis süreçlerini hızlandırarak kişiselleştirilmiş tedavi yöntemlerinin geliştirilmesine katkı sağlamaktadır.

Keywords

• Yapay Zeka
• Derin Öğrenme
• Tümör Tespiti
• YOLOv9
• EfficientDet
• Neoadjuvan Tedavi
• Patoloji Analizi

References

1. [1] Deng, J., Zhang, W., Xu, M., & Zhou, J. (2023). Imaging advances in efficacy assessment of gastric cancer neoadjuvant chemotherapy. Abdominal Radiology, 48(12), 3661-3676.
2. [2] Cui, Y., Zhang, J., Li, Z., Wei, K., Lei, Y., Ren, J., ... & Gao, X. (2022). A CT-based deep learning radiomics nomogram for predicting the response to neoadjuvant chemotherapy in patients with locally advanced gastric cancer: a multicenter cohort study. EClinicalMedicine, 46.
3. [3] Hörst, F., Ting, S., Liffers, S. T., Pomykala, K. L., Steiger, K., Albertsmeier, M., ... & Kleesiek, J. (2023). Histology-based prediction of therapy response to neoadjuvant chemotherapy for esophageal and esophagogastric junction adenocarcinomas using deep learning. JCO Clinical Cancer Informatics, 7, e2300038.
4. [4] Li, C., Qin, Y., Zhang, W. H., Jiang, H., Song, B., Bashir, M. R., ... & Zhong, L. (2022). Deep learning-based AI model for signet-ring cell carcinoma diagnosis and chemotherapy response prediction in gastric cancer. DOI: https://doi.org/10.1002/mp, 15437, 1535-1546.
5. [5] WongKinYiu. (n.d.). YOLOv9: Real-time object detection. Retrieved from https://github.com/WongKinYiu/yolov9.
6. [6] Wightman, R. (n.d.). EfficientDet in PyTorch. GitHub repository. Retrieved from https://github.com/rwightman/efficientdet-pytorch.
7. [7] Tan, M., Pang, R., & Le, Q. V. (2020). EfficientDet: Scalable and efficient object detection. Retrieved from https://arxiv.org/abs/1911.09070.
8. [8] TensorFlow. (n.d.). EfficientDet with TensorFlow. TensorFlow Documentation. Retrieved from https://www.tensorflow.org/lite/models/efficientdet/overview.

9. [9] Fang, M., Tian, J., & Dong, D. (2022). Non-invasively predicting response to neoadjuvant chemotherapy in gastric cancer via deep learning radiomics. EClinicalMedicine, 46.
10. [10] Ouyang, G., Chen, Z., Dou, M., Luo, X., Wen, H., Deng, X., ... & Wang, X. (2023). Predicting Rectal Cancer Response to Total Neoadjuvant Treatment Using an Artificial Intelligence Model Based on Magnetic Resonance Imaging and Clinical Data. Technology in Cancer Research & Treatment, 22, 15330338231186467.
11. [11] She, Y., He, B., Wang, F., Zhong, Y., Wang, T., Liu, Z., ... & Tian, J. (2022). Deep learning for predicting major pathological response to neoadjuvant chemoimmunotherapy in non-small cell lung cancer: A multicentre study. EBioMedicine, 86.
12. [12] Zhu, H. T., Zhang, X. Y., Shi, Y. J., Li, X. T., & Sun, Y. S. (2020). A deep learning model to predict the response to neoadjuvant chemoradiotherapy by the pretreatment apparent diffusion coefficient images of locally advanced rectal cancer. Frontiers in Oncology, 10, 574337.
13. [13] Li, Z., Zhang, D., Dai, Y., Dong, J., Wu, L., Li, Y., ... & Liu, Z. (2018). Computed tomography-based radiomics for prediction of neoadjuvant chemotherapy outcomes in locally advanced gastric cancer: a pilot study. Chinese Journal of Cancer Research, 30(4), 406.
14. [14] Janssen, B., Theijse, R., van Roessel, S., de Ruiter, R., Berkel, A., Busch, O., ... & Besselink, M. (2022). Artificial Intelligence-Based Segmentation of Residual Tumor in Histopathology of Pancreatic Cancer after Neoadjuvant Treatment. HPB, 24, S294.
15. [15] Springer. (2021). A survey of object detection models: From YOLO to EfficientDet. Retrieved from Springer database.
16. [16] TensorFlow Developers. (2020, March). AutoAugment for object detection with EfficientDet. Retrieved from https://blog.tensorflow.org/2020/03/efficientdet-object-detection-on-tpus.html.
17. [17] Towards Data Science. (2020). EfficientDet and augmentation techniques. Retrieved from https://towardsdatascience.com/efficientdet-scalable-and-efficient-object-detection-807f77a93a7.
18. [18] Zhang, Y., Chen, H., Liu, Q., & Wu, J. (2023). Best practices for training and evaluating deep learning models in medical imaging. IEEE Transactions on Medical Imaging, 42(5), 1201–1215. https://doi.org/10.1109/TMI.2023.3245678
19. [19] Litjens, G., Kooi, T., Bejnordi, B. E., Setio, A. A. A., Ciompi, F., Ghafoorian, M., ... & Sánchez, C. I. (2017). A survey on deep learning in medical image analysis. Medical Image Analysis, 42, 60–88. https://doi.org/10.1016/j.media.2017.07.005
20. [20] OpenAI. (2023). GPT-4 Technical Report. https://doi.org/10.48550/arXiv.2303.08774574337.
21. [21] Redmon, J., & Farhadi, A. (2018). YOLOv3: An incremental improvement. arXiv preprint arXiv:1804.02767. https://doi.org/10.48550/arXiv.1804.02767
22. [22] Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., & Guo, B. (2020). Deep learning-based object detection and instance segmentation in medical imaging. Computers in Biology and Medicine, 123, 103865. https://doi.org/10.1016/j.compbiomed.2020.103865
23. [23] Tan, M., Pang, R., & Le, Q. V. (2020). EfficientDet: Scalable and efficient object detection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 10781–10790). https://doi.org/10.1109/CVPR42600.2020.01080
24. [24] TensorFlow ekibi. (2024). TensorFlow (Version 2.18.0) [Computer software]. https://www.tensorflow.org
25. [25] Komura, D., & Ishikawa, S. (2018). Machine learning methods for histopathological image analysis. Frontiers in Medicine, 5, 241. https://doi.org/10.3389/fmed.2018.00241
26. [26] Steiner, D. F., MacDonald, R., Liu, Y., & Stumpe, M. C. (2020). Impact of deep learning assistance on the histopathologic review of lymph nodes for metastatic breast cancer. The American Journal of Surgical Pathology, 44(2), 220–226. https://doi.org/10.1097/PAS.0000000000001368
27. [27] Bera, K., Schalper, K. A., Rimm, D. L., Velcheti, V., & Madabhushi, A. (2019). Artificial intelligence in digital pathology — new tools for diagnosis and precision oncology. Nature Reviews Clinical Oncology, 16(11), 703–715. https://doi.org/10.1038/s41571-019-0252-y