Quantum Dilated Convolutional Neural Networks for Thyroid Cancer Recurrence Prediction

Description

QD-CNN is a deep learning framework designed for accurate thyroid cancer recurrence prediction using patient-level clinical and pathological data. The model employs dilated convolutional neural networks to capture long-range dependencies and multiscale feature patterns without increasing computational complexity.

To enhance learning efficiency and predictive accuracy, QD-CNN integrates quantum-inspired feature representation and optimization mechanisms, enabling effective exploration of complex feature spaces and improved generalization across diverse patient cohorts. This combination allows the model to identify subtle recurrence-related patterns that may be missed by conventional CNN approaches.

QD-CNN provides a robust AI-driven decision support system for recurrence risk assessment and clinical follow-up planning in thyroid cancer management.

Input

Age,Gender,History of Radiotherapy,Adenopathy,Pathology,Focality,Risk category,T,N ,M,Overall Stage

Output

• Predicted Thyroid Cancer Recurrence Status

·No Recurrence (0)

·Recurrence (1)

Refrences

1.Chen, Y., & Fang, W. (2021). Quantum Dilated Convolutional Neural Networks for Image Recognition. IEEE Access. This work introduces quantum dilated CNNs, extending dilated convolution to hybrid quantum-classical architectures, which is foundational for your QD-CNN concept.

2.Kil, J., Kim, K. G., Kim, Y. J., et al. (2020). Deep Learning in Thyroid Ultrasonography to Predict Tumor Recurrence in Thyroid Cancers. Journal of the Korean Society of Radiology. This study demonstrates the application of convolutional neural networks for predicting thyroid cancer recurrence, supporting the recurrence-prediction context of your work

3.Ahmad, M. A.-S., & Haddad, J. (2024). An Explainable AI Model for Predicting the Recurrence of Differentiated Thyroid Cancer. arXiv. This paper uses AI/deep learning for recurrence prediction and emphasizes interpretability—useful for motivating advanced models like your QD-CNN.