Positron emission tomography (PET) with long axial field-of-view (LAFOV) scanners generates unprecedentedly large datasets, posing major challenges for conventional reconstruction algorithms. Recent advances in deep learning have improved PET image quality, yet scaling to high-dimensional data remains computationally demanding. Quantum computing (QC) offers new opportunities to address such challenges. This work presents a pilot study on image denoising using a hybrid quantum-classical autoencoder. The model employs a classical encoder, a quantum bottleneck implemented via a parameterized quantum circuit, and a classical decoder. Using simulated experiments on the medical modified national institute of standards and technology (MedMNIST) dataset with 28 × 28 grayscale images and on PET scan images with 64 × 64, results demonstrated that the quantum bottleneck can effectively recover structural details from noisy inputs. While scaling to large-volume PET data will require advances in quantum hardware, circuit design, and error mitigation, these results highlight the potential of hybrid quantum–classical approaches to complement deep learning in PET reconstruction.
Positron emission tomography (PET) with long axial field-of-view (LAFOV) scanners generates unprecedentedly large datasets, posing major challenges for conventional reconstruction algorithms. Recent advances in deep learning have improved PET image quality, yet scaling to high-dimensional data remains computationally demanding. Quantum computing (QC) offers new opportunities to address such challenges. This work presents a pilot study on image denoising using a hybrid quantum-classical autoencoder. The model employs a classical encoder, a quantum bottleneck implemented via a parameterized quantum circuit, and a classical decoder. Using simulated experiments on the medical modified national institute of standards and technology (MedMNIST) dataset with 28 × 28 grayscale images and on PET scan images with 64 × 64, results demonstrated that the quantum bottleneck can effectively recover structural details from noisy inputs. While scaling to large-volume PET data will require advances in quantum hardware, circuit design, and error mitigation, these results highlight the potential of hybrid quantum–classical approaches to complement deep learning in PET reconstruction.
Deep learning has transformed radiology, but the computational cost of training and deploying modern models is outpacing what classical hardware can support. Quantum computing has been proposed as one response to this bottleneck: an n-qubit system can represent 2n states simultaneously, although practical speedups require efficient algorithms, data encoding, and measurement strategies rather than superposition alone. This review introduces radiologists to five core quantum concepts, qubits, superposition, entanglement, quantum gates, and quantum circuits, and to the superconducting, trapped-ion, and photonic hardware platforms that define the current noisy intermediate-scale quantum (NISQ) era. We then summarize the quantum machine learning architectures most often applied to medical imaging (quantum support vector machines, quantum neural networks, variational quantum circuits, and quantum convolutional neural networks) and quantum image representations such as the flexible representation of quantum images (FRQI) and the novel enhanced quantum representation (NEQR). Reported applications include brain tumor classification and segmentation on MRI, pneumonia detection on chest radiographs, skin and breast lesion classification, CT and PET reconstruction, and ultrasound denoising. Nearly all published systems are hybrid quantum-classical pipelines evaluated on small, downsampled benchmark datasets or simulators, and clinical deployment is not yet feasible. We conclude with current limitations and translational priorities for radiology.
Multi-modal biomedical imaging has become central to modern diagnostics, offering complementary anatomical, functional, and molecular information through modalities such as MRI, CT, PET, ultrasound, and optical imaging. Yet integrating these heterogeneous datasets remains computationally demanding due to differences in spatial resolution, noise profiles, acquisition dynamics, and high-dimensional feature distributions. This paper investigates AI–quantum hybrid diagnostic systems as an emerging paradigm for multimodal image fusion, reconstruction, and disease classification. By combining deep learning architectures with quantum-enhanced algorithms—including variational quantum circuits, quantum feature encoders, and quantum kernel methods—the hybrid framework aims to accelerate image processing, improve cross-modal consistency, and enhance diagnostic precision. The analysis highlights advancements in quantum-accelerated denoising, multi-modal registration, probabilistic inference, and high-dimensional pattern recognition. It also evaluates workflow integration challenges, such as NISQ-era noise, hardware scalability, and clinical interpretability. Overall, AI–quantum hybrid systems represent a promising frontier in medical imaging, offering potential improvements in speed, sensitivity, and personalized diagnostic accuracy.
Building a 3D scene from a limited number of images remains a challenging problem in computer vision, particularly for applications such as augmented reality, robotics, medical imaging, and autonomous navigation. Existing deep learning models like LRM and vFusion3D achieve high-quality reconstructions but require significant computational resources and long inference times, making them difficult to deploy on edge devices. This research introduces QuantumVision, a hybrid quantum-classical framework that integrates quantum image-processing techniques into a traditional 3D reconstruction pipeline. The framework consists of three modular components: quantum-assisted preprocessing, amplitude-based feature encoding, and a variational quantum optimization stage. These modules are designed to integrate with classical workflows while leveraging the potential advantages of Noisy Intermediate-Scale Quantum (NISQ) devices. Experimental simulations using Qiskit demonstrate improved feature extraction, edge detection, and reconstruction fidelity compared to classical baselines. The proposed approach highlights the potential of hybrid quantum computing to enhance computational efficiency and accuracy in partial-view 3D reconstruction tasks while remaining compatible with current quantum hardware limitations.
Abstract Background Cancer is a leading cause of death worldwide. While routine diagnosis of cancer is performed mainly with biopsy sampling, it is suboptimal to accurately characterize tumor heterogeneity. Positron emission tomography (PET)-driven radiomic research has demonstrated promising results when predicting clinical endpoints. This study aimed to investigate the added value of quantum machine learning both in simulator and in real quantum computers utilizing error mitigation techniques to predict clinical endpoints in various PET cancer patients. Methods Previously published PET radiomics datasets including 11C-MET PET glioma, 68GA-PSMA-11 PET prostate and lung 18F-FDG PET with 3-year survival, low-vs-high Gleason risk and 2-year survival as clinical endpoints respectively were utilized in this study. Redundancy reduction with 0.7, 0.8, and 0.9 Spearman rank thresholds (SRT), followed by selecting 8 and 16 features from all cohorts, was performed, resulting in 18 dataset variants. Quantum advantage was estimated by Geometric Difference (GD Q ) score in each dataset variant. Five classic machine learning (CML) and their quantum versions (QML) were trained and tested in simulator environments across the dataset variants. Quantum circuit optimization and error mitigation were performed, followed by training and testing selected QML methods on the 21-qubit IonQ Aria quantum computer. Predictive performances were estimated by test balanced accuracy (BACC) values. Results On average, QML outperformed CML in simulator environments with 16-features (BACC 70% and 69%, respectively), while with 8-features, CML outperformed QML with + 1%. The highest average QML advantage was + 4%. The GD Q scores were ≤ 1.0 in all the 8-feature cases, while they were > 1.0 when QML outperformed CML in 9 out of 11 cases. The test BACC of selected QML methods and datasets in the IonQ device without error mitigation (EM) were 69.94% BACC, while EM increased test BACC to 75.66% (76.77% in noiseless simulators). Conclusions We demonstrated that with error mitigation, quantum advantage can be achieved in real existing quantum computers when predicting clinical endpoints in clinically relevant PET cancer cohorts. Quantum advantage can already be achieved in simulator environments in these cohorts when relying on QML.
Quantum computing is being investigated as a computational approach for medical image analysis, but its potential role in cardiovascular imaging remains unclear. This review examines gate-based, quantum annealing, and hybrid quantum-classical techniques for cardiovascular image enhancement, image reconstruction, image segmentation, image feature extraction, and image-based disease classification, and it treats quantum-inspired classical algorithms as a separate category. The literature was identified through a structured search of four bibliographic databases and one preprint server, using stated eligibility criteria and study-level data extraction. Applications are considered in echocardiography, computed tomography, cardiac magnetic resonance imaging, nuclear imaging, and coronary angiography, and direct cardiovascular evidence is distinguished from methods developed for other medical imaging tasks. One adjacent cardiac study based on chest radiography is reported separately because it falls outside these five modalities. For each study, the imaging data, qubit count, encoding strategy, execution on a simulator or physical hardware, classical baseline, and validation design are tabulated. The review also defines the forms of quantum advantage that a study could claim, quantifies the qubit and state-preparation costs of common image encodings for realistic cardiac image sizes, and compares quantum methods with classical machine learning and deep learning in terms of computational requirements, training, interpretability, and reported performance. The existing evidence is exploratory and relies mostly on small datasets, compressed image representations, and quantum simulators. Some results are encouraging, but no consistent advantage over well-tuned classical approaches has been established. The key obstacles are image encoding, hardware noise, circuit depth, trainability, end-to-end computational cost, and the lack of external validation. Hybrid designs offer a practical setting for testing selected quantum components within current imaging pipelines. Future progress requires reproducible cardiovascular benchmarks, evaluation on physical quantum devices, and staged clinical validation. Quantum computing is therefore an emerging field of study whose role in cardiovascular imaging has to be defined through rigorous technical and clinical comparison.