Neuromotor, musculoskeletal, metabolic, and oncological pathologies

In this field we began looking at classic pathologies such as osteoarthritis joint replacement, osteoporosis bone fractures, and back pain. We then expanded into diabetes, looking at neuropathies, also as possible means of early type 2 diabetes diagnosis (Prof Simon Heller); into geriatrics and ENT in relation to the risk of falling (Dr. Jaydip Ray); into the functional assessment of patients with neurodegenerative diseases, such as Amyotrophic Lateral Sclerosis (Prof Chris McDermott); or the functional evaluation of the pelvic floor muscle after parturition, as a predictor of future prolapse and incontinence (Prof Dilly Anumba).

Predictive medicine works very well with children, because the personalisation helps to compensate for differences from adult physiology and pathology: for example, we are working on the detection of abuse bone fractures, and on the risk of bone fracture in obese children. We are now doing more work in oncology, first in a contiguous problem (bone fracture in metastatic patients), and more recently with a general oncological problem (stratification of therapy) in collaboration with the CHIC international consortium.

Here are specific areas of clinical impact, defined through some concrete examples of active research:

Risk of bone fracture in osteoporotic women

In the VPHOP (Osteoporotic Virtual Physiological Human) project we developed a complete technology to predict the risk of low-energy impact femoral and vertebral fracture in osteopenic or osteoporotic patients.

In the MultiSIM project we completed a retrospective accuracy study on 200 patients, which confirmed the validity of this approach.

We can immediately provide an online service for bone strength estimation from CT data, to be included in any clinical trial (osteoporosis interventions, bone metastases, other bone metabolic diseases, even paediatric).

Using the technologies developed in the NMS Physiome project we can provide an online service for patient-specific neuromusculoskeletal modelling based on MRI and gait analysis data, to be used in every clinical study where a detailed quantification of the neuromuscular condition is required (dystrophies, neurodegenerative diseases, cachexia, sarcopenia, balance pathologies, biomechanical determinants of osteo-degenerative conditions, etc.).

Technical PI: Prof Marco Viceconti, Mechanical Engineering; Clinical PI: Prof Richard Eastell, Prof Eugene McCloskey, Dr Jennifer Walsh, Human Metabolism.

Stratification of back-pain treatments

The technology developed in the MySpine project, tested for usability on nearly 200 patients, assists the spinal surgeon in the treatment decision (conservative, discectomy, vertebral fusion) by generating patient-specific models from MRI and CT data.

The service could be provided freely as a non-clinical research tool to explore its efficacy in a large multicentre trial.

Technical PI: Prof Damien Lacroix; Clinical PI: Prof Peter Varga, National Centre for Spine Disorder, Budapest.

Aetiology of Juvenile Idiopathic Arthritis

In the MD-Paedigree project we are testing the hypothesis that the severity and the location of Juvenile Idiopathic Arthritis (JIA) is correlated with patient-specific biomechanical determinants. If this hypothesis is confirmed, we will provide a stratification tool for JIA. If it is rejected, we will still be able to provide, for other pathologies, a research pipeline to explore the role of patient-specific biomechanical determinants in the aetiopathogenesis, or in the progression of other paediatric musculoskeletal conditions.

Technical PI: Prof Marco Viceconti, Mechanical Engineering; Clinical PI: Prof Alberto Martini, Giannina Gaslini Children’s Hospital, Genova.

Prediction of type 2 diabetes onset

In the MissionT2D project we are developing an early diagnosis tool for type 2 diabetes that detects early signs of neuropathy. The technology is currently under development.

Technical PI: Dr Claudia Mazzà, Mechanical Engineering; Clinical PI: Prof Simon Heller, Human Metabolism.

Genotype/phenotype stratification of Parkinson’s patients

Parkinson’s disease is perfect example of the complex interaction between genetic and disease phenotype traits; while some mutations are associated with a very high probability of developing Parkinson’s, many patients do not exhibit any mutation, but show the same clinical symptoms.

The combination of clinical genomics, neuromotor biomechanics, and clinical neurology can open whole new avenues in the early diagnosis and the therapy of this complex condition.

Technical PI: Dr Claudia Mazzà; clinical PI: Dr Alisdair McNeil, Neuroscience; Prof Oliver Bandmann, Neuroscience.

Prediction of the risk of vertebral facture in patients with metastases

A number of tumours induce bone metastases; some like breast and lung cancer frequently induce metastases in the spine. These lesions are usually painful, but sometime they are so extensive that the vertebral body collapses, with significant pain, need for immobilisation, and a potential risk of neurological lesions and potentially paraplegia. The clinicians need to make difficult treatment decisions between radiotherapy, which reduces the pain, but not the risk of bone fracture, and vertebroplasty, an interventional procedure, that also reduces the risk of fracture.

We are developing a CT-based patient-specific modelling workflow that will provide accurate prediction of the risk of bone fracture as a function of the localisation and extension of the metastatic lesion, providing an effective decision-support system for treatment stratification.

Technical PI: Dr Enrico Dall’Ara, Human Metabolism; Clinical PI: Prof Rob Coleman, Oncology.

Stratification by therapy for solid tumours patients

Some solid tumours today can be treated with multiple therapies, as a replacement for or alongside surgery – radiotherapy, chemotherapy, hormonal therapies, immune therapies, etc. In many cases, different patients with the same type of cancer respond differently to the same therapy, because of the specific genome of the cancer, or because of disease specific phenotype traits, such as cellularity, vascularisation, localisation, etc.

In the CHIC project we are collaborating on the development of an in silico simulator that uses patient-specific models to predict the response each patient would have for each of the therapeutic options available. We are currently targeting lung cancer, glioblastoma, and Wilms’ tumour.

Technical PI: Dr Dawn Walker, Computer Science; Clinical PI: Prof Norbert Graf, Prof Rainer Bohle, Saarland University, Saarbrücken; Prof Stefaan van Gool, UZ leuven.