Developing microstructural and metabolic magnetic resonance imaging to address the diagnostic and prognostic unmet needs in breast cancer

2015 saw over 55,000 new cases of breast cancer in the UK. However, not all of these breast cancers are the same. Breast cancers that look identical on standard imaging tests like x-ray mammography and ultrasound incorporate both those that are slow-growing and those that are destined to spread rapidly and claim lives. The most life-threatening tumours need aggressive treatment early but the question is how to discriminate these from those that are slower growing? Today, we do this by biopsying the tumour and examining its grade, protein expression, and if hormone receptors are present. However, a biopsy does not represent the whole tumour and current predictive methods lack precision.

Hyperpolarised C13 imaging is a very novel technique that combines magnetic resonance imaging (HPMRI) with metabolic information (i.e. how the cancer cells utilise energy, for example glucose). Specifically, HPMRI allows us to measure how cells convert pyruvate to lactate, which reflects metabolic activity. Cancers have more metabolic activity than normal cells and some cancers are more metabolically active than others. For example, we know from animal studies that the most aggressive tumours convert pyruvate faster and in greater volumes than less aggressive tumours. HPMRI may therefore identify the most aggressive cancers but the test is so novel that this has not been researched to any great degree in man. Furthermore, the equipment needed for HPMRI is extremely expensive and restricted to just a few centres worldwide.

There is also another problem that needs addressing, namely that almost 60% of women that have biopsy to see if there is a tumour actually have a negative result, so we really need a better way to tell if something was tumour or not. We have developed novel MRI techniques (VERDICT, T2-mapping and Fat-fraction MRI) that we are using to find tumours and avoid unnecessary biopsies in patients being investigated for prostate cancer. We think these techniques could be applied to women being investigated for breast cancer and in a similar fashion could help avoid many unnecessary breast biopsies.

Our research aims to develop these new imaging techniques in women with breast cancer. We will do so in a robust, safe, validated manner by following the "Imaging Biomarker Roadmap for Cancer Studies" (Nature Reviews Clinical Oncology 2017). We will do experiments to technically validate these biomarkers, followed by biological and clinical validation, to plan for clinical implementation. We will determine if these biomarkers are reproducible and what they tell us about a breast cancer. Ultimately, if they prove sufficiently promising, we ultimately aim to add their information to existing prognostic models to detect aggressive breast cancer. A prognostic model is a collection of information known about a patient that predicts what will happen to them in the future, i.e. whether their tumour will spread.

We at UCL are especially well-placed to do this work. The Medical Research Council and Cancer Research UK and other funders have already invested >£30 million to support our imaging research. For example, an MRC Clinical Research Infrastructure award funded installation and associated equipment necessary for hyperpolarised metabolic imaging and digital histopathology. We have an assembled a world-class multidisciplinary research team that is working currently in prostate cancer. A successful CARP award will allow us to turn our attention to breast cancer by funding research time for an existing NHS consultant radiologist (Dr. Abeyakoon) to join our team. She has extensive prior research experience in breast cancer imaging (having gained a PhD; unusual in radiology).

The results of our research will be disseminated to the scientific community via conference proceedings and peer reviewed papers, the public via public patient engagement forums at UCL/H, and funders via grant applications to support the next phase of research

The CARP awardee will develop novel MRI microstructural (VERDICT/T2-mapping/fat-fraction MRI) and metabolic (DNP 13C-Pyruvate Hyperpolarised MRI) methods in breast cancer, by:

(a) Optimising novel microstructural MRI: Digital histopathology slides will be segmented (using software algorithms available at UCL) to define proportions of cellular/luminal/glandular/fat within normal/cancer regions. This will inform MRI sequence development of a 10-minute breast VERDICT sequence (as previously described (Invest Radiol. 2015 Apr;50(4):218-27)). T2-mapping and fat-fraction imaging will be performed in patients to define best image resolution achievable.

(b) Optimising HP-MRI: Together with collaborators in Keil (Prof. Jan-Bernd Hoevener) UCL is exploring a balanced steady state free precession (bSSFP) pulse sequence to provide full 3D coverage (Mol Imaging Biol 2018 Dec;20(6):902:918). Patients with breast cancer will be scanned and iterations made to the bSSFP sequence.

(c) Biological/technical validation of microstructural MRI: 47 patients with breast cancer scheduled for wide local excision (WLE) will be scanned. The WLE specimen will be scanned ex-vivo to facilitate in-vivo image-histopathlogy matching as described (Radiology 2009 May 251(2):369-379). Histopathology segmented components, histological subtype and molecular staging will be correlated to related MRI metrics. 20 patients will be invited for repeat imaging with two consecutive scans in the same sitting and then return for one further scan at 7-10 days.

(d) Technical validation of HP-MRI in phantoms: data generated as part of a linked PhD project will be shared with the awardee.

(e) Generation of clinical value preliminary data: 6 patients scheduled for mastectomy, 3 with triple negative (likely non-curative) and 3 with luminal A (likely treatable) status will be scanned (with all techniques). Preliminary results used to power/design/apply for subsequent funding for a clinical utility study.

Patients with breast cancer will be the greatest beneficiaries of the research proposed. They will benefit in the following ways: improved diagnosis by avoiding unnecessary biopsies/anxiety and potential reduction in harmful diagnostic radiation exposure; and better prediction of prognosis, potentially enabling strategies for active surveillance over the current treatment for all management approach.

The NHS and beyond will benefit by efficiencies both in time and expenditure conferred by improved diagnostics (through cost reduction through avoidance of unnecessary secondary tests and associated clinic visits).UK HEIs will benefit as a result of the opportunities for discovery that are generated from having validated novel MRI methods that have the potential be shared between institutions (through the CRUK National Cancer Imaging Translational Accelerator) and applied to other areas of research.

Industry will benefit by the opportunity that will result from creating a novel MRI technique that can be incorporate onto specific MRI scanner platforms leading to potential marketing opportunities.The above are not guesses but reasonable predictions based on what UCL has achieved so far. UCL is fortunate in having the most comprehensive portfolio of prostate cancer diagnostic studies globally, nearly all of which involve MRI. Many hundreds of patients are currently benefiting from this augmented diagnostic pathway. Millions of men are now to benefit through the subsequent revision of the 2019 Prostate Cancer NICE guidelines on diagnosis and management (https://www.nice.org.uk/guidance/ng131) which now specifically advocate use of multi-parametric (mp) MRI prior to biopsy in men suspected with prostate cancer. The revision is specifically attributed within the guidelines to work led by UCL (Ahmed et al. Lancet 2017; 389:815-822 and Kasivisvanathan V et al. N Engl J Med 2018; 378(19):1767-1777) and economic benefits to the NHS have also been recognised (Faria R et al Eur Urol 2018;73(1):23-30). There is now significant interest in further research on prostate mp-MRI, namely development of AI for automated/assisted diagnosis, thus driving forward further UK HEI led research efforts/funding and spawning a multitude of industrial opportunities for new start-up SME's and established MR vendors.

It is by this established and demonstrated pathway to impact that the novel MRI breast research proposed within the CARP application, if successful, will achieve maximum benefit.Contrary to the case highlighted at UCL, most novel imaging biomarkers of cancer fail transition into clinical practice, as data are over-interpreted, studies underpowered, the long times to study initiation, problems in acquiring consistent data and inability of studies to recruit. UCL and other HEI's (KCL, Imperial, Oxford, Cambridge, Manchester and ICR) in the UK have specifically recognised these deficiencies in the translational pipeline for cancer imaging biomarkers. They have set-up a CRUK funded national infrastructure (CRUK C42780/A27066: UCL PI Punwani £10m, 2019-2024) to provide for gaps and facilitate the systematic translation of new (predominantly MRI) techniques from concept to the clinic. This National Cancer Imaging Translational Accelerator (NCITA) has adopted this CARP application for study delivery.

Linked to this, perhaps the immediate beneficiaries from will be those others actively working in the field of imaging biomarker validation, both at UCL, the other HEI's linked to NCITA and more globally through new linkages facilitated by the CRUK International Early Cancer Detection (ICED) Alliance (PI: Emberton, Co-I: Punwani: £3.5m 2019-2023 and £30m reserved for ICED alliance projects) which links 3 UK HEI's (UCL, Cambridge and Manchester) with 3 US institutions (Stanford, Oregan Health and Science University and Massachusetts Institute of Technology). Successful developed novel imaging techniques will rapidly disseminate throughout these.