Skip to main content
Emiliano Spezi  FInstP PhD FIPEM CPhys

Professor Emiliano Spezi



Professor of Healthcare Engineering
Director of Research

School of Engineering

+44 29208 76521
Queen's Buildings, Room S/2.05, 5 The Parade, Newport Road, Cardiff, CF24 3AA
Available for postgraduate supervision


A second landmark paper from the Image Biomarker Standardisation Initiative (IBSI) on standardizing convolutional filters for reproducible radiomics is now published in the Radiological Society of North America (RSNA) Radiolgy journal  

This is another milestone in the standardisation, reproducibility and usability of quantitative imaging biomarkers and I am particularly proud of the contribution of my SPAARC team ( at Cardiff University School of Engineering to this particular work.

Also note the accompanying Radiology editorial from Merel Huisman & Tugba Akinci D’Antonoli explaining why this work matters and Cardiff University newsletter describing this work to a lay audience.

Interdisciplinary Precision Oncology Cardiff Hub (IPOCH)

We run an EPSRC-DTP Interdisciplinary Doctoral Training Hub in Precision Oncology at Cardiff University. The projects in biomedical imaging, pathology and genomics are delivered by our fantastic students across the Schools of Engineering, Computer Science and Medicine.

Find out more on the IPOCH research website.


Founding Member, Medical Engineering Research Group

Team Leader, Life Imaging and Data Analytics Research Team

Chair, School of Engineering Research Ethics Committe

Associate Editor, European Journal of Medical Physics

Review Editor, Frontiers in Medicine | Nuclear Medicine

Past chair, Task Group No. 363 - Guidelines for harmonizing the validation of tumor PET auto-segmentation algorithms

Professor Emiliano Spezi is the Director of Research at Cardiff University School of Engineering, Chair of the School Research Ethics Committee and Leader of the Life Imaging and Data Analytics team. He is a state registered Clinical Scientist with 15 years working experience in Research and Development in the National Health Service, where he now holds honorary position with Velindre University NHS Trust and Velindre Cancer Centre.  His current research interests centre on three main areas: (1) Quantitative Imaging Biomarkers and Radiomics, (2) Image Guidance for Precision Medicine, (3) Modelling in Radiation Oncology. He receives grants from National and International funding bodies and from the Industry.

Examples of application of the research generated in Professor Spezi’s laboratory include: (1) use of advanced image segmentation methods in the PEARL radiotherapy trial as featured in BBC News, (2) innovative use of AI with Intel Corporation to improve accuracy and efficiency of radiotherapy treatments (MedWales Lifestories Magazine, pp. 26), (3) development of federated learning methods to aid cancer research.

























Book sections

  • Spezi, E. and Bray, M. 2024. Foreword. In: Proceedings of the Cardiff University Engineering Research Conference 2023. Cardiff: Cardiff University Press, (10.18573/conf1.a)
  • Sykes, J., Alaei, P. and Spezi, E. 2017. Imaging dose in radiation therapy. In: Mijnheer, B. ed. Clinical 3D Dosimetry in Modern Radiation Therapy. CRC Press, pp. 561-588.





Life Imaging and Data Analytics

Life Imaging and Data Analytics is a team with multi-disciplinary skills established at the Cardiff University School of Engineering. LIDA is led by Dr Emiliano Spezi (Professor of Healthcare Engineering) and consists of post-doctoral research associates and PhD students. The research interest of the group spans from advanced medical image processing and radiomics to advanced computer modelling in radiation oncology. In the field of medical imaging the team has developed ATLAAS, an award-winning machine learning based tool which can be used to select the optimal Positron Emission Tomography automated segmentation method for radiotherapy treatment planning.

Furthermore, we have established a programme of image analysis techniques, including segmentation, texture, shape and intensity analysis and wavelet analysis, combining this with clinical and genomic data to produce diagnostic, prognostic and predictive models. We have a very intensive programme of development of radiomics algorithms, which are advanced imaging techniques that allow non-invasive, high-throughput, three-dimensional extraction of large numbers of descriptive features from any volume of interest. LIDA is a founding member of the Image Biomarker Standardisation Initiative (IBSI) to develop standardised radiomics algorithms and reporting guidelines that can make radiomics analyses reproducible and comparable. We developed SPAARC radiomics, a tool for multimodal quantitative image analysis incorporating 164 features all compliant and validated in accordance with the IBSI recommendations. Features include morphology, intensity-based statistics, intensity and intensity volume histograms and grey level matrixes. This is a selected list of publications featuring SPAARC radiomics:

  • Palumbo, al. 2021. Prediction of early distant recurrence in upfront resectable pancreatic adenocarcinoma: A multidisciplinary, machine learning-based approach. Cancers13(19), article number: 4938. (10.3390/cancers13194938)
  • Mori, al. 2020. Training and validation of a robust PET radiomic-based index to predict distant-relapse-free-survival after radio-chemotherapy for locally advanced pancreatic cancer. Radiotherapy and Oncology153, pp. 258-264. (10.1016/j.radonc.2020.07.003)
  • Zwanenburg, al. 2020. The Image Biomarker Standardization Initiative: standardized quantitative radiomics for high throughput image-based phenotyping. Radiology295(2), pp. 328-338. (10.1148/radiol.2020191145)
  • Piazzese, al. 2019. Discovery of stable and prognostic CT-based radiomic features independent of contrast administration and dimensionality in oesophageal cancer. PLoS ONE14(11), article number: e0225550. (10.1371/journal.pone.0225550)
  • Whybra, al. 2019. Assessing radiomic feature robustness to interpolation in 18F-FDG PET imaging. Scientific Reports9(1), article number: 9649. (10.1038/s41598-019-46030-0)

Watch a video on radiomics research: radiomics reseach

(credit: MAASTRO Clinic).

In addition, we are building an IT infrastructure and standardised algorithms that can be included in any machine learning training and validation process. The idea is that any developed and validated prognostic/predictive model can be included in libraries as part of a Decision Support System that can be used in real time by clinicians in the clinic. LIDA and Velindre University NHS Trust were the recipients of the NHS Innovation Award 2018 (Welsh Government, Efficiency Through Technology programme). As part of the “AI Solutions for Personalised Radiotherapy” (ASPIRE) project LIDA, Velindre and Intel Corporation are working on a project aimed at training and validating AI software for the automated delineation of tumour volumes on anatomical and functional imaging modalities. The training is designed to be performed on a large retrospective dataset of labelled clinical scans and will be validated on a prospective dataset acquired throughout the duration of the project. In addition to reducing dramatically the workload in radiotherapy planning, we expect AI auto-segmented volumes to be of equal quality and more consistent than those outlined manually, which our research group have investigated extensively. The integration of AI in the clinical radiotherapy workflow will prepare the ground for future developments related to high throughput medical image analysis (radiomics), development of a fully automated radiotherapy workflow (from AI-based automated segmentation to AI-based automated planning), and development of decision support system for clinicians to use in clinical practice. LIDA and Velindre are also partners on a project (FAST-RTP2) aimed at including AI in the process of automating the preparation of external radiotherapy plans.

Machine learning applications for personalized medicine are highly dependent on access to sufficient data. Large datasets from a broad range of different populations representing the variation in the entire cancer patient population need to be acquired and used to learn prediction models. In LIDA we use a distributed learning approach which was designed to address ethical and legal boundaries and to limit the impact of data privacy collaboration between research institutes. Cardiff University and The Christie NHS Foundation Trust are the only two UK centers participating to the European Computer Assisted Theragnostics project (EuroCAT) project and to the Community in Oncology for Rapid Learning (CORAL). CORAL includes almost 30 cancer centres worldwide (Netherlands, USA, UK, India, China, South-Africa, Australia, Italy, Germany, Belgium, Canada, Denmark). CORAL is based on a distributed approach in which data do not cross the firewall, where data is made semantically interoperable locally, and where centres allow applications to enter their firewall and use their data to answer a particular research questions, without any patient identifiable information being exposed/shared. Applications are focused on machine learning and modelling to predict for instance overall survival in a particular tumour site.

Watch a video on distributed learning in radiation oncology and on personal health trains:

distributed learning in radiation oncology

personal health train

(credit: MAASTRO Clinic).

Members of the LIDA team

Dr Philip Whybra (Research associate)

Miss Iona Foster (PhD candidate)

Mr Emad Alsyed (PhD candidate)

Miss Elisabetta Cagni (PhD candidate)

Mr Kerim Duman (PhD candidate)

Associated members of the LIDA team

Prof John Staffurth (Professor Cinical Oncology, Cardiff University School of Medicine and Velindre Cancer Centre)

Dr Kieran Foley (Consultant Radiologist, Velindre Cancer Centre)


ASPIRE: AI Solutions for Personalised Radiotherapy (Efficiency Through Technology programme) Co-PIWelsh Government and Intel Corp198,0302018-2020
DOTATER+: Advanced Personalised 3D Dosimetry for a clinical trial in peptide radionuclide therapyPICancer Research Wales80,3412015-2018
ARENA: Extension of RTTQA outlining activity into the educational arena Co-PI       Velindre NHS Trust340,2002018-2021
PEARL: PET-based Adaptive Radiotherapy Clinical Trial Co-ICancer Research Wales720,0002017-2021
FAST-RTP2: Implementing automated techniques in radiotherapy treatment planningCo-IVelindre NHS Trust76,4302018-2021
STORM_GLIO: Developing Radiomics as an Imaging Biomarker in High Grade GliomaCo-PIVelindre NHS Trust37,8352018-2021
TEXRAD: Establishing image derived prognostic and predictive biomarkers of radiotherapy treatments and assessing treatment response using texture analysis
Co-PI   Velindre NHS Trust - Moondance Foundation81,5402017-2020

Past grants and contracts

Title: Informatics Platform for Advanced Cancer Imaging Research
Value: £18,638
Role: Principal Investigator
Period: 2017
Funding body: Data Innovation and Research Institute

Title: Raydose GUI development
Value: £26,646
Role: Principal Investigator
Period: 2015-2016
Funding body: EURAMET - Velindre NHS Trust

Title: Advanced FDG PET-CT target volume delineation in Intensity Modulated Radiotherapy planning for Head and Neck cancers
Value: £45,000
Role: Principal Investigator
Period: 2014-2015
Funding body: Cancer Research Wales

Title: 3-D Printed sources for High-Resolution Molecular Imaging
Value: £10,000
Role: Principal Investigator
Period: 2014-2015
Funding body: Velindre NHS Trust

Title: RAYDOSEPLAN a multimodality platform for treatment planning research in molecular radiotherapy
Value: €242,000
Role: Principal Investigator
Period: 2012-2015
Funding body: European Association of National Metrology Institutes (EURAMET)

Title: Adaptive Image-Guided Radiotherapy Strategies for Bladder and Cervical Cancer to Enable Dose Escalation and Reduce Late Toxicity
Value: £60,000
Role: Co-Investigator
Period: 2014-2015
Funding body: Cancer Research Wales

Title: XVI5IEC: Evaluation of patient dose to skin and eye lens for default CBCT settings
Value: £8,000
Role: Principal Investigator
Period: 2013
Funding body: Elekta Ltd Crawley UK

Title: XVIoptimal5: Evaluation of patient dose reduction and image quality for new Cone Beam CT settings
Value: £8,100
Role: Principal Investigator
Period: 2013
Funding body: Elekta Ltd Crawley UK

Title: MOZART: Parameters affecting tumour control and toxicity in oesophageal cancer: a multi-dimensional outcome analysis
Value: £67,000
Role: Principal Investigator
Period: 2012-2015
Funding body: Cancer Research Wales

Title: POSITIVE: Optimisation of positron emission tomography based target volume delineation in Head and Neck radiotherapy
Value: £81,000
Role: Principal Investigator
Period: 2011-2014
Funding body: Cancer Research Wales

Title: RAYDOSE: Assessment of patient dose using novel radioisotopes in Molecular Targeted Radiotherapy
Value: £170,000
Role: Principal Investigator
Period: 2010-2012
Funding body: Wales Office of Research and Development for Health and Social Care (WORD)

Title: A Comparison of Convolution/Superposition and Monte Carlo methods for conformal radiotherapy
Value: £80,000
Role: Co-Investigator
Period: 2009-2014
Funding body: Cancer Research Wales

Title: XVIctdi: Cone Beam CT dosimetry using an optimised phantom
Value: £20,000
Role: Principal Investigator
Period: 2009-2014
Funding body: Elekta Ltd Crawley UK

Title: Extending the RTGrid portal to the wider user community
Value: £52,000
Role: Co-Investigator
Period: 2009
Funding body: JISC ENGAGE e-Infrastructure programme

Title: MRI in radiotherapy planning
Value: £10,000
Role: Principal Investigator
Period: 2009
Funding body: Velindre NHS Trust

Title: The application of GafChromic film in routine and non routine quality control methods in radiotherapy physics
Value: £4,000
Role: Principal Investigator
Period: 2007
Funding body: Velindre NHS Trust


In addition to providing undergraduate project, dissertation and essay supervision, I am Module Organiser for the following modules at the School of Engineering:

  • EN4505: Medical Image Processing (MEng)
  • EN4506: Clinical Engineering 2 (MEng)

Between 2016 and 2021 I was also module organiser for the following module at the School of Physics and Astronomy:

  • PX3247: Radiation for Medical Therapy (BSc)



2003: PhD (Medical Physics), University of Wales College of Medicine, Cardiff, UK
1998: MSc (Clinical Scientist), University of Bologna, Bologna, Italy
1996: Laurea (Physics), University of Bologna, Bologna, Italy

Honours and awards

Winner of the European Radiology ESGAR Silver Award 2018 (awarded in 2019)

  • Foley et al Development and validation of a prognostic model incorporating texture analysis derived from standardised segmentation of PET in patients with oesophageal cancer. Eur Radiol. 2018 Jan;28(1):428-436

Winner of the ESTRO-Varian Research Award 2019

  • Deist et al Distributed learning on 20 000+ lung cancer patients, Radiother Oncol (2019) Vol 133 Supp. 1, S287-8

Winner of the journal’s Best Paper Award 2018: European Journal of Nuclear Medicine and Molecular Imaging Physics

  • Sjögreen Gleisner et al Variations in the practice of molecular radiotherapy and implementation of dosimetry: results from a European survey, EJNMMI Physics (2017) 4:28

Winner of the Burgen Scholarship Award 2016: Academia Europaea

  • Computational Models in Funtional Imaging and Radiation Therapy

Winner of the Best Physics Poster Award 2016: ESTRO

  • Berthon et al Towards standardisation of PET-autosegmentaion with the ATLAAS machine learning algorithm, Radiother. Oncol. (2016) 119 (Supp 1): S452

Winner of the Manufactures' Award for Innovation 2015: IPEM

  • ATLAAS: an automatic decision tree-based learning algorithm for advanced image segmentation in positron emission tomography, Phys. Med. Biol. (2016) Jul 7;61(13):4855-69

Winner of IPEM/AAPM Travel award: IPEM/AAPM

  • A Monte Carlo investigation of the accuracy of intensity modulated radiotherapy, Med. Phys. (2004)

Professional memberships

Fellow Member of the Institute of Physics and Engineering in Medicine (IPEM)

Chartered Member of the Instutte of Physics (IoP)

State Registered Clinical Scientist: Health and Care Professions Council (HCPC)

Academic positions

External examiner (research degrees): University of The Free State (South Africa), University of Cape Town (South Africa), Swansea University, Swansea (United Kingdom), Universitat Politècnica de Catalunya, Barcelona (Spain), National University of Ireland, Galway (Ireland), University of London, London (United Kingdom), Niels Bohr Institute, Copenhagen (Denmark).

Committees and reviewing

Current duties

Chair of the Research Ethics Committe of Cardiff University School of Engineering

Associate Editor of the European Journal of Medical Physics (Physica Medica), Elsevier

Member of the Welsh Government - Welsh Scientific Advisory Committee

Past duties

Vice-Chair of the American Association of Physicists in Medicine (AAPM) Task Group 211 Classification, Advantages and Limitations of the Auto-Segmentation Approaches for PET

Chair of theNational Cancer Research Institute (NCRI) Radiotherapy Trials QA Database solutions and IT subgroup

Member of the NCRI Clinical and Translational Radiotherapy Research Working Group (CTRad) Workstream 4: New Technology, Physics and Quality Assurance


Areas of interest

I am available to supervise Post Graduate Research students in the following areas:


PubMed search of publications by Professor Spezi (link).


Opportunities for Post Graduate Research

If you are interested in enrolling on Postgraduate Research at the School og Engineering, contact the PGR Enquiries Team to know more about all current opportunities.


Current PhD projects







Radiomics enhanced deep learning-based classifier to improve survival in glioblastoma multiforme




Full time

Main supervisor

Microstructural imaging of the tumour microenvironment: towards virtual biopsy of prostate cancer.  

MITRA Solanki



Full time

Main supervisor

Integration of Engineered and Deep Learning Radiomics Imaging Features to Characterise Tumour heterogeneity in Non-Small Cell Lung Cancer

LI Mengcheng



Full time

Main supervisor

Artificial Intelligence with Human In The Loop for Automated Medical Image Contouring. 




Full time


Non-invasive characterisation of brain cancer tissue microstructure from MRI using Deep Learning 




Full time


Artificial Intelligence assisted grading of prostate cancer progression in patient biopsies with novel tissue labelling biomarkers. 




Full time


Current supervision

Abdulkerim Duman

Abdulkerim Duman

Graduate Demonstrator

Mengcheng Li

Mengcheng Li

Graduate Demonstrator

Solanki Mitra

Solanki Mitra

Research student

Adam Threlfall

Adam Threlfall

Research Student

Michail Papachristos

Michail Papachristos

Research student

Faye Warren

Faye Warren

Research student

Past projects

Implementing Automated Techniques in Radiotherapy Treatment Planning, PhD, Iona Foster

PET Image Texture Analysis and Radiotherapy, PhD, Emad Alsyed

Automated Image-Guided Radiotherapy Planning, PhD, Elisabetta Cagni

Application of Texture Analysis to Identify Prognostic Biomarkers for the Optimisation of Radiotherapy Treatments, PhD, Philip Whybra

Advanced Personalised 3D Dosimetry for a Clinical Trial in Peptide Radionuclide Therapy, PhD, Salvatore Berenato

Advanced Automated Segmentation of PET in Radiotherapy, PhD, Craig Parkinson

Parameters Affecting Tumour Control and Toxicity in Oesophageal Cancer: a Multi-dimensional Outcome Analysis, PhD, Rhys Carrington

Outlining Variantion in Upper Gastrintestinal Cancer Radiotherapy Clinical Trials, MD, Sarah Swynne

Development of Techniques for Verification of Advanced Radiotherapy by Portal Dosimetry, PhD, Yasmin Radzi

Radiotherapy dose calculation in oesophageal cancer: comparison of analytical and Monte Carlo methods, PhD, Dewi Johns

Optimisation of Positron Emission Tomography Based Target Volume Delineation in Head and Neck Radiotherapy, PhD, Beatrice Berthon