Quantum Enhanced Deep Learning Bio Inspired Model for Lung Tumors Detection and Severity Analysis in Clinical CT-DICOM Images

ABSTRACT: To propose a hybrid quantum-based deep learning model to detect malignant lung nodules and accurately classify disease severity. Bio-inspired techniques are integrated to optimize the learning rate for robust generalization. A Hybrid Quantum-Enhanced Deep Neural Network (QE-DNN) combined with a Quantum Convolutional Neural Network (Q-CNN) is used to extract the multi-scale spatial patterns from high-resolution CT-DICOM images. To perform deep segmentation, Quantum Mask-RCNN is used to isolate the ROI from the images effectively. A bio-inspired Adaptive Firefly-Differential Evolution (AFDE) optimizer is employed to fine-tune the learning architecture. Quantum histogram equalization and wavelet fusion are incorporated as data pre-processing methods to retain critical edge and intrinsic features. CT-DICOM dataset is used for evaluation which consists of 25,1135 images with a resolution of 512x512 pixels. The performance is assessed in MATLAB, TensorFlow Quantum, and IBM Qiskit tools by comparing the proposed work with existing models such as SVM-WSS, GCPSO-PNN, 3D-DLCNN, and ECNNDE-BCE. The proposed QEDNN-AFDE quantum bio-inspired nodule detection strategy enhanced the generalization capability by exploring wider with proven results of 96.4% accuracy, 95.2% sensitivity, 95.8% specificity, 95.2% F1 score, 94.6% dice coefficient, 0.02 Log Loss and AUC-ROC with 0.95 TPR and 0.05 FNR. The proposed QEDNN-AFDE model strengthens the interpretability of deep learning models in medical imaging and sets a new benchmark in quantum-assisted diagnostics in precision oncology. The model shows promising performance in both classification accuracy and severity prediction and outperforms all existing models.