Findings show that current quantum feature maps through the encode–measure–boost pipeline on NISQ hardware do not yet outperform a well-designed classical pipeline.
Abstract
Quantum machine learning is often proposed for richer feature representations, yet most evidence rests on idealized simulation rather than real noisy intermediate-scale quantum (NISQ) hardware. This research presents a controlled comparison of classical and quantum-enhanced diagnostic pipelines on three clinical binary classification tasks: Mammographic Mass, Anemia, and Diabetic Retinopathy. All pipelines share standardized preprocessing, principal component analysis (PCA), and a fixed extreme gradient boosting (XGBoost) classifier, so differences arise only from the feature representation. Four quantum encodings (angle, phase, basis, and the ZZ feature map) are each run on two backends: a noiseless simulator and the real IBM Heron r2 processor (156 qubits). Across nine performance metrics, compared with the classical pipeline, the performance decreases in the quantum hardware execution for all datasets, with a more significant reduction observed for the Anemia dataset. In contrast, compared with the simulated pipeline, the hardware execution shows a slight performance decrease for the Mammographic and Diabetic Retinopathy datasets. The exceptions are the angle and phase encodings for the Mammographic dataset, where the hardware result improves slightly compared with the simulator. For the Anemia dataset, the transition from simulation to real quantum hardware results in a considerable performance reduction. These findings show that current quantum feature maps through the encode–measure–boost pipeline on NISQ hardware do not yet outperform a well-designed classical pipeline.
Medical image classification is a critical component of modern healthcare; however, accurate diagnosis remains challenging due to limited annotated datasets, class imbalance, and the high dimensionality of medical imaging data. To address these challenges, a hybrid quantum-classical neural network (HQCNN) is proposed, integrating classical deep learning with variational quantum learning for medical image classification. The proposed architecture combines a five-layer convolutional neural network (CNN) for hierarchical feature extraction with a lightweight 4-qubit variational quantum circuit (VQC) incorporating quantum state encoding, superposition and entanglement mechanisms, and a quantum attention-Fourier (QAF) module. This hybrid design aims to improve nonlinear feature representation and quantum parameter efficiency while maintaining a shallow quantum circuit suitable for noisy intermediate-scale quantum (NISQ)-era constraints. Experimental evaluation on six MedMNIST benchmark datasets demonstrated competitive performance across both binary and multi-class classification tasks. HQCNN achieved 98.88% accuracy on the binary subset of PathMNIST (classes 0 vs. 1), 97.61% accuracy on the multi-class OrganAMNIST dataset, and 86.29% accuracy on BreastMNIST. Comparative experiments and statistical analyses demonstrated consistent improvements over the controlled BHQNN baseline, while component-wise ablation studies showed that the QAF module, superposition and entanglement mechanisms, and expressive parameterized rotations contributed to classification performance in a complementary and dataset-dependent manner. Moreover, HQCNN reduced the number of trainable quantum parameters by approximately 55.6% compared with the baseline hybrid quantum neural network (BHQNN). Noise-aware simulations further showed that the model retained relatively stable predictive performance under moderate depolarizing noise, supporting further evaluation under near-term quantum computing conditions. Overall, the results demonstrate that HQCNN provides a parameter-efficient hybrid quantum-classical framework for medical image classification and offers a promising foundation for further investigation of quantum-enhanced medical image analysis.
Shahjalal Khan, Jahid Karim Fahim, P. Paul et al.· International Journal of Adv...· 0 citations
The hardware available today does not support general-purpose quantum computation in medical settings, but specific subtasks such as sensing, photon-resolved detection, or image reconstruction can already be handled by quantum components embedded in classical systems. This article surveys recent developments across several branches of medical applications, with a complementary look at an analogous trend in machine learning. Nitrogen-vacancy centers in nanodiamonds enable nanoscale thermometry, high-resolution magnetocardiography, and real-time monitoring of free radicals inside living cells. In radiological imaging, photon-counting computed tomography is already entering routine clinical use, while QUBO-based image reconstruction on quantum annealers and gate-model processors points toward dose reduction in low-dose CT. Quantum-enhancedpositron emission tomography exploits polarization correlations of entangled annihilation photons to suppress background events and opens a path toward positronium-based biomarkers such as tissue hypoxia. Nuclear magnetic resonance ensemble computing is included to illustrate the scalability limits of fully quantum systems. The same hybrid pattern is then identified outside medicine, in efficient fine-tuning of large language models, where localized quantum modules reduce parameter counts without degrading task accuracy. The reviewed work differs in technical maturity, from systems already deployed clinically to proof-ofprinciple experiments such as X-ray spontaneous parametric down-conversion. Across this range, the quantum component handles one well-defined stage of an otherwise classical system.
I. Woźniak, Mateusz Jangas, Mateusz Piątek et al.· International Journal of Ele...· 0 citations
Breast cancer is a major health concern, and early detection can make a great difference in treatment and survival rates for breast cancer patients. Machine learning methods have noticeably improved prediction accuracy on high-dimensional medical datasets and have become widely used tools in medical diagnostics. The transition to a quantum computational framework has opened new directions in machine learning research. By using qubits instead of classical bits, quantum models may provide new ways to represent and process complex medical data. In medical diagnostics, this has led to a growing interest in whether quantum approaches can improve classification performance more effectively than classical methods. The purpose of this paper is to compare classical and quantum machine learning models on the task of breast cancer classification and to determine whether quantum models can achieve higher accuracy and faster prediction on a selected dataset. Alongside the main simulator-based experiment, a smaller experiment was performed on a real IBM quantum computer to demonstrate the practical execution of the same task under current hardware constraints. In the practical part of the study, the Wisconsin Diagnostic Breast Cancer (WDBC) dataset was used. This dataset consists of 569 samples with 30 numerical features extracted from digitized fine needle aspirate images and provides a reliable basis for evaluating model performance. A comparative analysis was conducted between classical and quantum machine learning models, including a support vector machine (SVM), an artificial neural network (ANN), a quantum support vector machine (QSVM), and a hybrid quantum-classical neural network (QNN). In the simulator-based experiment, both SVM and ANN achieved an accuracy of 0.956 with an F1-score of 0.965 on the test set, while QSVM reached an accuracy of 0.807 with an F1-score of 0.866 and the Hybrid QNN achieved 0.623 accuracy with an F1-score of 0.677. These quantum and hybrid models also required substantially longer training times than the classical baselines. A small hardware experiment on an IBM quantum device further illustrates both the practical feasibility and the current limitations of executing this classification task on noisy intermediate-scale quantum hardware.
Jovana Gluhovic· American Journal of Computer...· 0 citations
Extensive experiments indicate that, despite the limited number of qubits that can be simulated, QuantumBoostNet consistently outperforms state-of-the-art classical and hybrid classical-quantum models in cardiac ultrasound view identification, achieving a relative improvement over the best competitor.
Mihai Udrescu-Milosav, S. Jura, M. Udrescu et al.· 0 citations
Neuro-imaging data is necessary for timely and correct diagnosis of the disease Alzheimer’s (AD). Early clinical intervention is challenging due to the high-dimensional nature of the images, making it computationally difficult to work on representations, sparse annotated datasets and inter-subject variance. Deep convolutional VGG, ResNet, DenseNet, MobileNet and other neural networks (CNNs) have shown good performance. Diagnostic performance, but they are based on a high number of parameters and require highly trained personnel. Data can often times restrict resilience to data-skimy clinical settings. To address these challenges, the current paper presents a hybrid Quantum Machine Learning (QML) architecture for Alzheimer’s disease Detection and classification based on a combination of classical CNN-based feature extraction and Variational quantum classifier. In the suggested method, diseased, relevant, and compact features are extracted using pre-trained CNN backbones. Data from structural MRI scans is later translated into quantum states and computed. With parameterised variational quantum circuits optimised with a hybrid quantum-classical learning loop. The proposed method has been experimentally evaluated on benchmark datasets of MRI scans of Alzheimer’s patients. The hybrid CNNQML model has a mean classification accuracy of about 93% which is similar to deep CNN models, including ResNet and DenseNet (94%), and much more. Beating lightweight models like MobileNet and shallow CNNs by six percentage points. It is worth noting that the hybrid QML model performs better in conditions with limited training data. Performance degradation is observed in accuracy and AUC for all CNN variants, with up to 5% lower performance to deep CNNs. Cross-validation, confidence interval, paired significance test, and statistical validation. Effect size analysis proves the fact that the hybrid model produces statistically significant performance, as well as cases of clinically meaningful improvements over lightweight CNN baselines. Similar to the state-of-the-art deep networks with significantly reduced trainable parameters in the classification stage. These findings show that hybrid QML is a data-efficient and powerful substitute in the diagnosis of Alzheimer’s. On the whole, this is a work that sets a practical way to integrate quantum machine learning into clinical neuro-imaging pipelines and hybrid settings. QML has supplementary deep-learning enhancement to classics in next-generation Alzheimer’s Decision-support systems.
Introduction In recent years, high-dimensional clinical and genomic data have gained significant importance for prognosis and personalized medicine in breast cancer. But the use of quantum machine learning (QML) on such data is limited by the availability of few qubits, the computation time of quantum simulation, and dimensionality reduction. This work systematically compares several QML architectures for breast cancer classification in the presence of realistic and simulator constraints. Methods The experiments were performed on a dataset of METABRIC breast cancer patients (2,509 patients). After handling missing values and one-hot encoding, there were 63 features in the processed data set. The feature space was reduced by Principal Component Analysis (PCA) to 12, 4 and 2 components for the implementations of quantum computers, respectively, with 56.10 ± 0.14%, 26.78 ± 0.57% and 16.23 ± 0.33% of the variance retained. Three QML models were tested: Quantum Neural Networks (QNN), Quantum K-Nearest Neighbors (QKNN), and Quantum Support Vector Machines (QSVM), with the models being simulated. Seven classical classification models were tested: Logistic Regression, SVM with RBF kernel, K-Nearest Neighbors, Random Forest, XGBoost, LightGBM and Multilayer Perceptron, both with PCA-matched and full 63-feature representation. All primary results are reported with 5-fold cross validation. Results Among the evaluated QML architectures, QKNN using 12 principal components achieved the strongest performance, attaining an accuracy of **75.11% ± 3.76%**, an F1-score of **0.7088 ± 0.0433**, and a ROC-AUC of **0.8148 ± 0.0394**. Even though all of the QML models performed significantly poorly in comparison to classical models trained on the entire 63-feature data set, the latter models were able to achieve about **94% accuracy** with XGBoost, Random Forest, and Logistic Regression. The comparison also showed the effect of information loss due to PCA is significant in predictive performance in both classical and quantum models. Discussion The results show that for high dimensional breast cancer data, currently available simulator-based QML models can learn meaningful patterns with limited quantum resources, but are not as effective as powerful classical machine learning models when complete feature representations are available. The study does not report any sort of quantum advantage or clinical use, but rather a benchmark of current QML architectures that has been rigorously calculated and repeated, with a focus on the impact of dimensionality reduction, validation approaches, and simulator limitations, as well as outlining challenges that need to be overcome prior to practical implementation on real quantum hardware.
Saartak Allena, Smrithy G. S., Balaji Chandrasekaran· Frontiers in Artificial Inte...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
