Generative Quantum Machine Learning Methods for Cancer Detection using Augmented Datasets

Abstract

Cancer detection using AI methods and image databases faces several challenges, including limited and imbalanced image datasets, long algorithm training times, and high-complexity feature generation. In particular, many cancer image datasets contain very few malignant examples compared to benign ones, limiting the amount of minority-class data available for training. Although classical AI and especially deep learning methods have had some success in providing cancer diagnostics, Quantum Machine Learning (QML) methods may offer advantages through high-dimensional feature representations, generative data augmentation, and the exploitation of feature correlations through quantum properties such as entanglement. In this paper, we propose using a generative QML method on the compressed breast ultrasound image dataset (BUSI) to generate synthetic images and hence augmented datasets for classifier training. In particular, we train a convolutional neural network with real, synthetic, and a combination of real and synthetic images. We test the algorithm on real images and provide comparative results for all three training cases. Our results show that the classifier trained entirely on synthetic images achieved an AUC-ROC of 0.802 (80.2%), compared with 0.760 (76.0%) for the real-data baseline.

The paper investigates whether quantum-generated synthetic images can improve breast-ultrasound cancer classification with limited, imbalanced training data. It combines a roughly 70-million-parameter variational autoencoder (VAE), which compresses images into 16-dimensional latents, with a 16-qubit IQP Quantum Circuit Born Machine (QCBM). Latents are binarized for distribution learning; sampled bitstrings are mapped back into latent space and decoded into images. QCBM training uses classical simulation and maximum mean discrepancy loss. Experiments use the 780-image BUSI dataset for malignant-versus-nonmalignant classification, with a fixed 64/16/20% training/validation/test split. The same small CNN is trained on real data, class-weighted real data, real data augmented with synthetic malignant images, or entirely synthetic class-specific images. Across ten classifier-training seeds, synthetic-only training achieves test AUC-ROC $0.802\pm0.030$, versus $0.760\pm0.008$ for real-only training, with a paired Wilcoxon $p=0.002$. AUC-PR remains essentially unchanged: $0.645\pm0.049$ versus $0.643\pm0.024$. Malignant-only augmentation yields AUC-ROC $0.752\pm0.010$ and lowers AUC-PR to $0.604\pm0.034$. The paper identifies limited synthetic-image diversity: a fixed classifier assigns synthetic malignant images nearly identical scores. Latent perturbations and replacing the QCBM with a Gaussian mixture model do not resolve this, implicating the VAE representation or decoder. A cross-reconstruction probe also detects class-specific VAE signatures, although it does not establish their use by the malignancy classifier. Evaluation uses one fixed test set of 157 real images; reported $p$-values are uncorrected for multiple comparisons. Quantum-hardware sampling and noise characterization are planned rather than reported.