Artificial intelligence tool for streamlining pathology workflow

Nearly 50,000 cases of prostate cancer are diagnosed each year in the UK. During the diagnostic process, men with suspected prostate cancer undergo a biopsy, which is analysed by pathology services. There are over 60,000 prostate biopsies performed in the UK annually, which represents a high workload for pathology teams. With increasing demand and a shortage of pathologists, tools that could help streamline this workflow would provide significant pathologist time savings and accelerate diagnoses.

To confidently diagnose prostate cancer, pathologists need to identify a number of tissue architecture and cellular cues. All biopsies are stained with Hematoxylin & Eosin (H&E), which allows the pathologist to study the size and shape (morphology) of the cells and tissue. However, in 25-50% cases, H&E staining alone does not provide sufficient evidence for a diagnosis, requiring the additional process of immunohistochemistry (IHC) to study other cellular features.

One bottleneck in the current pathology workflow is the requirement for a pathologist to review the H&E-stained biopsies to determine which require IHC. To address this need, pathologists Dr Richard Colling, Dr Lisa Browning and Professor Clare Verrill (Nuffield Department of Surgical Sciences and Oxford University Hospitals NHS Foundation Trust) teamed up with biomedical image analysts Andrea Chatrian, Professor Jens Rittscher and colleagues (Institute of Biomedical Engineering, Big Data Institute and Ludwig Oxford) to take a multidisciplinary approach.

In their paper in the journal Modern Pathology, the team used prostate biopsies annotated by pathologists at Oxford University Hospitals to train an artificial intelligence (AI) tool to detect tissue regions with ambiguous morphology and decide which cases needed IHC. The tool agreed with the pathologist’s review in 81% of cases on average. By enabling automated request of IHC based on the AI tool results, the pathologist would only need to review the case once all necessary staining had been carried out. This workflow improvement is estimated to save on average 11 minutes of pathologist time for each case, which scales up to 165 pathologist hours for 1000 prostate biopsies needing IHC.

“The NHS spends £27 million on locum and private services to make up for the shortfall in pathology service provision. By using this AI tool to triage prostate biopsies for IHC, pathologists would spend less time reviewing these cases, which would not only lead to financial savings but it would also accelerate prostate cancer diagnoses to inform patients and treating clinicians earlier.” – Professor Clare Verrill, Nuffield Department of Surgical Sciences and Oxford University Hospitals NHS Foundation Trust.

The tool will now be developed and validated further using pathology data from different locations to account for variation in IHC requests between pathologist teams and centres. This future work will continue to take advantage of the PathLAKE Centre of Excellence for digital pathology and artificial intelligence, of which Oxford is a member.

This work was supported by PathLAKE via the Industrial Strategy Challenge Fund, managed and delivered by Innovate UK on behalf of UK Research and Innovation (UKRI), the NIHR Oxford Biomedical Centre, the Engineering and Physical Sciences Research Council (EPSRC), the Medical Research Council (MRC) and the Wellcome Trust.

Cancer immunologist Lieping Chen joins Ludwig Oxford as a Visiting Professor

We are proud to announce that renowned immunologist Lieping Chen is joining the Oxford Branch of the Ludwig Institute for Cancer Research, Nuffield Department of Medicine, as a Visiting Professor of Cancer Immunotherapy. Chen, who is the United Technologies Chair in Cancer Research, a Professor of Immunobiology, Medicine and Dermatology, and co-director of the Cancer Immunology Program at Yale Cancer Center, Yale University School of Medicine, USA, is an international leader in basic T cell biology and cancer immunotherapy. His visiting professorship at Oxford opens new opportunities for collaboration and expansion in this crucial field.

Cancer immunotherapy is now poised to become a standard treatment for cancer, alongside surgery, chemotherapy, radiotherapy and targeted therapy. Chen has contributed enormously to this exciting development through his seminal work on the PD-1/PD-L1 immune suppressive pathway. The concept of turning a patient’s own immune system against a tumour was proposed decades ago, but it has only recently become a reality. The most effective immunotherapy approach to date involves the blockade of the PD-1/PD-L1 pathway.

Chen discovered PD-L1 (originally termed B7-H1) as a member of the B7 family of ligands that have immune suppressive functions. He also demonstrated that human tumours express high levels of PD-L1, and that forced expression of PD-L1 in murine tumours confers resistance to immune elimination. Further, he showed that anti-PD-L1 antibodies could block the interaction of PD-L1 with its receptor (PD-1) to undermine this immune resistance. This revealed the importance of the PD-1/PD-L1 pathway in tumour immune resistance and put this pathway on the map as a target for cancer therapy.

Chen has continued to contribute to the clinical targeting of the PD-1/PD-L1 pathway. He developed an immunohistochemistry assay for PD-L1 detection in human cancer tissues and collaboratively demonstrated that PD-L1 expression in tumours predicts a greater response to anti-PD-1/PD-L1 therapy. He also helped to initiate and organise the first-in-human clinical trial of a therapy targeting PD-1/PD-L1 pathway.

Chen has received several awards and honours in recognition of his outstanding scientific achievements, including the William B. Coley Award (2014), AAI-Steinman Award (2016), Warren Alpert Foundation Prize (2017), Giants of Cancer Care (2018) and Richard V. Smalley Award (2020). He is a member of the National Academy of Sciences USA and a Fellow of the AACR Academy, American Association for Cancer Research.

Professor David Hunter honoured by the Academy of Medical Sciences

The Academy of Medical Sciences has elected 11 University of Oxford biomedical and health scientists to its fellowship this year. All were selected for their exceptional contributions to the advancement of medical science through innovative research discoveries and translating scientific developments into benefits for patients and the wider society.

OxCODE Associate Director Professor David Hunter (Nuffield Department of Population Health) is elected as a Fellow for his leading role in HIV and later cancer research. A highly cited scientist, he has been involved in collaborative studies of nutrition and HIV pathogenesis, studied diet and cancer aetiology in large scale prospective studies, and developed a sample handling and genotyping laboratory to explore genetic associations with cancer, and gene-environment interactions.

Professor Hunter said ‘As someone born in the UK, having grown up and gone to medical school in Australia, and having spent most of my career in epidemiology in the US, I am deeply honoured to be elected to the Academy, and look forward to participating in the important work the Academy does on behalf of UK and global medical science.’

The Academy of Medical Sciences is the independent body in the UK representing the diversity of medical science. Elected Fellows are the UK’s leading medical scientists from hospitals, academia, and industry. They are recognised for their innovative research discoveries and for translating scientific developments into benefits for patients and wider society. This year, 50 Fellows were chosen from 384 candidates.

The new Fellows will be formally admitted to the Academy on 1 July 2021. For details of the other Fellows elected this year, please visit the Academy of Medical Sciences website.

New prostate cancer risk tool

Each year in the UK around 48,500 men are diagnosed with prostate cancer and 11,900 die from the disease. To improve survival, Professor Julia Hippisley-Cox (Nuffield Department of Primary Care and Health Sciences) and Professor Carol Coupland (University of Nottingham) have developed a tool to calculate personalised risk of prostate cancer using the health records of 1.45 million men in the QResearch database. The new risk prediction algorithm aims to diagnose more tumours earlier when they are easier to treat.

The tool is designed to be used for asymptomatic individuals and combines the prostate specific antigen (PSA) blood test result with factors such as age, ethnicity, body mass index, smoking status, social deprivation and family history. Compared to using the PSA test alone, the new algorithm is more accurate at predicting prostate cancer cases (68.2% compared to 43.9% using PSA-only), high-grade aggressive tumours (49.2 % versus 40.3%) and prostate cancer deaths (67% versus 31.5%).

The decision in most primary care practices to refer men who are asymptomatic is based on binary PSA thresholds, although this can lead to too many false-negative and false-positive results. Furthermore, a binary threshold does not give any indication for the patient as to their absolute risk of developing prostate cancer and/or clinically significant disease requiring immediate intervention. The results show that the risk equation provides a valid measure of absolute risk and is more efficient at identifying incident cases of prostate cancer, high-grade cancers and prostate cancer deaths than an approach based on a PSA threshold. The intended use is to provide a better evidence base for the GP and patient to improve decision-making regarding the most appropriate action, for example, reassurance, repetition of PSA test, referral for MRI, regular monitoring, referral to a urologist, or use of preventative interventions should any become available.

 – Professor Julia Hippisley-Cox (Nuffield Department of Primary Care and Health Sciences)

More research is now required to assess the best way to implement the algorithm and evaluate the health economics and the impact on prostate cancer diagnosis and subsequent survival.

Read the full publication in the British Journal of General Practice.

Read the feature in the Daily Mail

Adrian Hill elected to Royal Society

Professor Adrian Hill, Director of the Jenner Institute (Nuffield Department of Medicine), has become a Fellow of the Royal Society for his leading role in the design and development of new vaccines for globally important infectious diseases over the course of over 25 years.

One of his most important developments has been the spin out of Vaccitech, which he co-founded in 2016, to capitalise on the discovery of ChAdOx. The chimp cold virus, ChAdOx, became a weapon of choice against what the World Health Organization called “Disease X” – a hypothetical future pathogen with epidemic or pandemic potential.

ChAdOx is a viral vector which safely mimics viral infection in human cells and elicit antibody and T cell responses to pathogens and tumours. Thus far, ChAdOx has already been applied in cancers (prostate), malaria, and most recently, the AstraZeneca-Oxford vaccine for Sars-Cov-2.

Through his work, Professor Hill has demonstrated the applications of adenoviruses in immunisation regimes supporting new vaccination approaches for a variety of disease, many of which have previously not had treatment options available.

Professor Hill becomes one of 6 new Oxford researchers to join the Royal Society. Read about them here.

Understanding mutation progression to detect ovarian cancer earlier

High grade serous ovarian cancer (HGSOC) is the most common subtype of ovarian cancer, and is one of the deadliest. Over 80% of these ovarian cancers are detected at an advanced stage, such as stage III or IV, when cancers are much harder to treat. As a result, 10-year survival rates are less than 30% in the UK.

This is despite the fact that HGSOC has a latency period predicted to be between 6.5 and 40 years, whereby a precancerous lesion in the fallopian tube has developed and will go on to become a cancerous tumour. So, despite being present in the body for a long time, current methods are poor at detecting this type of ovarian cancer at an early stage once it has progressed.

This is due in part to a lack of screening techniques for ovarian cancer, such as the very successful screening programmes for other cancers like cervical or colorectal cancer, which have had a considerable impact on patient outcomes over the last decade. Ovarian cancer symptoms are also very non-specific and so make early diagnosis even more challenging, with women often presenting with bloating, abdominal pain, weight loss or weight gain. The precancerous lesions of the fallopian tubes that could develop into serous ovarian cancer are also hard to find due to their small size, and thus are hard to study. There is therefore an urgent need to find new methods of early detection.

Nina Wietek from the Ahmed Lab at the Nuffield Department of Women’s & Reproductive Health is investigating potential avenues for early detection through sequencing tumours and precancerous tissue to explore tumour initiation. To do this Nina is interrogating highly relevant samples obtained directly from patients to gain important insights into tumour development using the power of genomics. Through enhancing our understanding of these early changes, they hope to devise methods of looking for them in order to diagnose ovarian cancer at an early stage, which will have a direct impact on patient survival.

About Nina & the Ahmed lab

Nina Wietek is investigating methods of early detection and prevention of ovarian cancer at the Ahmed Lab. Publications and results from this work is expected later in 2021.

Led by Prof Ahmed Ahmed, the Ovarian Cancer Cell Laboratory in The Weatherall Institute of Molecular Medicine uses cutting-edge innovative technologies to gain deep understanding of mechanistic drivers of ovarian cancer initiation and progression. Find out more about this group here.

New study investigates how growth factors in our gut could initiate cancer

The cells that make up our tissues are strictly organised, and various differentiated cell types do different jobs in specific locations. The cell composition of tissues and the way the cells are organised is often different in pre-cancerous conditions, or even severely disrupted when they progress to tumours.

Understanding the molecular signals that cause cell differentiation and prompt the cells to find their location within the tissue, may explain the morphological changes observed in patients with pre-cancerous conditions. Ultimately, the alteration of these signals might be a driving force in tumour development and progression.

A recent paper from the Boccellato Lab at the Ludwig Institute for Cancer Research, University of Oxford, has investigated how the epithelial cell lining of our gastrointestinal tract differentiates based on different growth factors, and how this could ultimately determine how a patient progresses to precancerous conditions that could lead to stomach cancer.

Image: A picture from the published paper showing how normal gastric pits can change shape and functionality if EGF levels are altered, and eventually lead to the pre-cancerous condition Atrophic Gastritis

The team exposed healthy human gut tissue (the mucosoid cultures, patented) to a variety of growth factors, including EGF, BMP and NOGGIN. What they found is that different combinations of these factors help to determine which cells differentiate to form the gastric glands. These glands line the stomach, and contain a variety of different cells that produce digestive enzymes and gastric acid to help to digest our food, or mucus to protect the stomach lining.

For example, exposure to growth factors including EGF and BMP formed the foveolar cells that produce the mucus to line our gut, whereas inhibition of EGF induces the differentiation of cells producing gastric acid and digestive enzymes.

Patient with the pre-cancerous condition called Atrophic Gastritis have a problem with digestion due to the lack of digestive enzymes and gastric acid producing cells. In the biopsies of this pre-cancerous condition, the team have found elevated levels of EGF, which correlated with the lack of those gastric acid producing cells and with a flattened shape of the stomach tissue.

What this study has shown is that specific localisation of growth factors in the tissue microenvironment may be responsible for the differentiation process. So changing the relative quantities or localisation of these growth factors could trigger a change in the epithelium structure and cellular composition over time, potentially leading to cancer.

Building a high-resolution, dynamic map of the growth factors during cancer progression is the next step in this research. The team will also be investigating causes for these growth factor level changes. For example, long-term infection with  Helicobacter pylori bacteria is associated with increased risk of gastric cancer. Investigating how infection alters the growth factor microenvironment is essential to understand the response of the tissue and its potential aberration leading to cancer.

Dr Francesco Boccellato says:

“By better understanding the role of growth factors underlying the epithelial structures in pre-cancerous conditions, we can detect when cancers may appear and thus treat them earlier.

“This study has allowed us to draw up a new, detailed map of the signalling microenvironment in the healthy human gastric glands, which we can now draw upon in future studies as we investigate how growth factors influence cancer occurrence.”

About the Boccellato lab

The Boccellato lab is investigating oncogenic pathogens and how they contribute to cancer.  Patients infected with those pathogens have a higher chance of developing cancer, but the malignancy arises many years after the initial infection event. Cancer may develop as a result of a long battle between the pathogen that persists, hides and damages the tissue, and the host that attacks the pathogen and continuously repairs the damage caused by the infection.