AI in Radiobiology & Radiopharmaceuticals: April 2026 Update

Research Date: 2026-04-04
Category: AI-Radiobiology-Radiopharmaceutical
Focus: AI-driven theranostics dosimetry, precision radiotherapy frameworks, and radiopharmaceutical discovery advances

1. AI-Enhanced Theranostics Dosimetry: Comprehensive 2025 Review

A systematic review in Nuclear Medicine and Molecular Imaging (August 2025) examined deep learning applications in theranostic radiopharmaceutical dosimetry across three critical domains: image quality enhancement, dose estimation, and organ segmentation [1].

Deep Learning Architectures

  • U-Net-based models: Primary architecture for organ segmentation, achieving Dice similarity coefficients >0.90 in benchmark challenges [1]
  • Generative Adversarial Networks (GANs): Used for PET image synthesis and quality enhancement; Jyoti et al. achieved PSNR 32.83 and SSIM 77.48 for synthetic brain PET representing Alzheimer’s disease stages [1]
  • Hybrid transformer networks: Emerging for multi-task dosimetry workflows combining segmentation and dose prediction [1]

PET Image Synthesis Innovation

  • Wang et al. demonstrated 3D U-Net synthesis of synaptic density (¹¹C-UCB-J) and amyloid deposition (¹¹C-PiB) PET from widely available ¹⁸F-FDG scans [1]
  • Mean region-of-interest biases within ±2% across Alzheimer’s disease and cognitively normal groups [1]
  • Applications: overcoming short-lived radionuclide imaging limitations, reducing radiation exposure, enabling delayed-time-point dosimetry without additional scans [1]

Dosimetry Software Integrating AI

  • QDOSE: Supports AI-based semi- and fully-automated organ segmentation, single time-point dosimetry, one-click hybrid dosimetry [1]
  • MIM Software: Voxel-level dosimetry with AI-enhanced segmentation capabilities [1]
  • VoxelDose, BigDose, RMDP: Additional voxel-based dosimetry packages incorporating ML components [1]

Critical Challenges Identified

  • Accurate dose estimation from theranostic pairs (diagnostic/therapeutic imaging correlation) [1]
  • Lack of standardized imaging datasets for DL training [1]
  • Radionuclide decay chain modeling complexity for multi-emitter isotopes [1]
  • Need for optimization and standardization of AI models for clinical reliability [1]

2. Precision Radiotherapy Implementation Framework: Semantic AI Analysis

A PMC-published study (2025) applied AI-driven semantic and temporal analysis to 3,343 unique articles (1964–2025) from Scopus, PubMed, and Web of Science, mapping radiotherapy-radiobiology-oncology evolution [2].

Historical Phase Analysis

  • Phase 1 (1964–1990): Limited production, foundational radiobiology research [2]
  • Phase 2 (1991–2010): Moderate growth, clinical radiotherapy standardization [2]
  • Phase 3 (2011–2024): Exponential expansion; publication peaks in 2020 and 2023 [2]

LDA Topic Modeling: Two Principal Axes

  • Clinical-Anatomical Axis: Cancer sites, treatment modalities, prognosis modeling [2]
  • Mechanistic-Molecular Axis: DNA repair pathways, radiosensitivity mechanisms, biomarker discovery [2]

Five Operational Classes Defined (2014–2025 Case Synthesis)

  1. DNA repair and molecular response pathways [2]
  2. Precision oncology and genomic modeling [2]
  3. Individual radiosensitivity prediction [2]
  4. Mechanisms of radioresistance [2]
  5. Advanced technologies: FLASH radiotherapy, hadron therapy, voxel-level analytics [2]

Voxel-Based Dose Painting Applications

  • Li et al. (2025) demonstrated hypoxia-guided dose painting feasibility in lung cancer [2]
  • Multimodal imaging identifies resistant niches (hypoxic, highly metabolically active subvolumes) [2]
  • Selective dose escalation to resistant regions while sparing normal tissues [2]
  • Enables spatial biology-driven treatment personalization [2]

3. Graph Neural Networks & Transformers in Radiopharmaceutical Discovery

A ScienceDirect review (September 2025) systematically investigated AI technical principles for radiopharmaceutical discovery and molecular imaging [3].

Target Identification

  • Graph Neural Networks (GNNs): Model molecular structures as graphs; capture binding site topology for target-ligand interaction prediction [3]
  • Applications: PSMA, somatostatin receptor, and novel target identification for radiopharmaceutical development [3]

Ligand Design Optimization

  • Generative Adversarial Networks: Generate novel ligand candidates with optimized pharmacokinetic properties [3]
  • Transformer models: Sequence-based molecular generation with attention mechanisms for binding affinity optimization [3]

Pharmacokinetic Optimization

  • AI models predict renal clearance, hepatic metabolism, and tumor-to-background ratios [3]
  • Integration with Monte Carlo simulations for absorbed dose estimation [3]

Image Reconstruction & Enhancement

  • Deep learning models for PET/SPECT image quality improvement [3]
  • Noise reduction and super-resolution modeling for low-dose imaging protocols [3]

4. Theranostics Digital Twin Framework

A Nature npj Drug Discovery article (October 2025) described theranostic digital twin (TDT) frameworks for optimizing radiopharmaceutical and immunotherapeutic interventions [4].

Framework Components

  • Patient-specific anatomical and physiological modeling [4]
  • Real-time treatment response simulation [4]
  • Drug dosing and scheduling optimization based on tumor growth dynamics and immune interactions [4]

Integration with Cancer Immunomodulation

  • AI simulations support precision cancer immunomodulation therapy optimization [4]
  • TDT enables radiopharmaceutical-immunotherapy combination strategy design [4]

5. AI for Simplified Dosimetry: Patient-Friendly Approaches

An arXiv preprint accepted in PET Clinics 2025 addressed AI solutions for key limitations in current radiopharmaceutical therapy dosimetry [5].

Key Advances

  • AI-based simplified dosimetry reducing computational burden [5]
  • Patient-friendly protocols minimizing imaging time points [5]
  • Deep learning models for single-time-point dose estimation [5]

6. Geant4 Simulations + AI Integration

A November 2025 review in Discovery highlighted emerging nuclear medicine trends integrating Geant4 Monte Carlo simulations with AI [6].

Combined Approach Benefits

  • Geant4 provides physics-based radiation transport simulation [6]
  • AI accelerates Monte Carlo computation for clinical feasibility [6]
  • Hybrid frameworks enable patient-specific absorbed dose calculations [6]

7. FDA-Approved AI Systems in Oncologic Imaging

The Theranostics review (March 2024) cataloged 520+ FDA-approved AI/ML algorithms, predominantly for radiological and oncological applications [7].

Current Status

  • None dedicated solely to theranostics applications [7]
  • Breast cancer AI systems dominate FDA approvals (BU-CAD, Koios DS, MammoScreen, Lunit INSIGHT MMG, Saige-Q, Transpara, ProFound AI) [7]

Opportunity Gap

  • Significant opportunity for theranostics-specific AI tool development and regulatory approval [7]

Key Research Gaps & Future Directions

  1. Dataset standardization: Lack of large, annotated theranostic imaging datasets for DL training [1,3]
  2. Radionuclide decay modeling: Multi-emitter isotopes (e.g., ¹⁷⁷Lu decay chain) require specialized AI architectures [1]
  3. Clinical validation: Multi-center prospective trials needed for AI dosimetry tools [1,2]
  4. Regulatory pathway: FDA approval framework for theranostics-specific AI needs development [7]
  5. Real-time integration: AI-assisted intraoperative radiopharmaceutical guidance unexplored [3]
  6. Generative drug discovery: AI-designed radiopharmaceutical candidates not yet clinically validated [3,7]

References

[1] Springer Nature. “The Role of Artificial Intelligence in Advancing Theranostics Dosimetry for Cancer Therapy: a Review.” Nuclear Medicine and Molecular Imaging. August 23, 2025. DOI: 10.1007/s13139-025-00939-9. https://link.springer.com/article/10.1007/s13139-025-00939-9

[2] PMC. “From Semantic Modeling to Precision Radiotherapy: An AI Framework Linking Radiobiology, Oncology, and Public Health Integration.” Biomedicines. 2025. https://pmc.ncbi.nlm.nih.gov/articles/PMC12730318/

[3] ScienceDirect. “Artificial intelligence for radiopharmaceutical and molecular imaging.” September 27, 2025. https://www.sciencedirect.com/science/article/pii/S2211383525006471

[4] Nature npj Drug Discovery. “Integrating artificial intelligence into small molecule development for precision cancer immunomodulation therapy.” October 1, 2025. DOI: 10.1038/s44386-025-00029-y. https://www.nature.com/articles/s44386-025-00029-y

[5] arXiv. “Accepted in PET Clinics 2025: AI for simplified dosimetry toward patient-friendly radiopharmaceutical therapy.” https://www.arxiv.org/pdf/2510.12714

[6] R Discovery. “Emerging trends in nuclear medicine: breakthrough in radiopharmaceuticals, Geant4 simulations, and AI integration.” November 12, 2025. https://discovery.researcher.life/article/emerging-trends-in-nuclear-medicine-breakthrough-in-radiopharmaceuticals-geant4-simulations-and-ai-integration/00f39f5547033f6aac4143c6ce2c58f7

[7] Theranostics. “Theranostics and artificial intelligence: new frontiers in personalized medicine.” Volume 14, Issue 6, pages 2367-2378. March 25, 2024. DOI: 10.7150/thno.94788. https://www.thno.org/v14p2367.htm