Systems Cancer Pathology

We aim to develop a deeper understanding of heterocellular tumour ecosystems and to translate our discoveries for patient benefit.

Tumours are mixtures of different cells. The phenotypes of these cells and their interactions direct spatial arrangements, and underpin disease progression. We develop tools to precisely define cell phenotypes within intact tissues in order to model tumours as heterocellular ecosystems. Taking a ‘systems’ view of tumours as groups of diverse interacting cells will provide insights into cancer biology, and lead to new ways of diagnosing and treating the disease.

To generate detailed snapshots of human tumour ecosystems we use multidimensional molecular tissue imaging (imaging mass cytometry) to simultaneously map between forty and fifty molecules with subcellular spatial resolution. We use computational tools to extract cellular phenotypic data and to represent networks of cell-cell interactions. In a landmark study we were the first to link these multi-dimensional tumour phenotypes to somatic genomic alterations (Ali HR et al Nature Cancer 2020). Finally, we are investigating how these new insights can be translated to conventional digital pathology.

We focus on breast cancer, still a major cause of premature death. New, effective diagnostic tools and treatments are urgently needed.  

Our work encompasses three related fields:

(1)  Breast Cancer & The Tumour Microenvironment

Immunotherapies are likely to benefit some breast cancer patients and enter routine practice soon. But exactly which patients will benefit and why is largely unknown. Our work is therefore directed to understanding the evolving face of the tumour immune microenvironment during disease progression and, secondly, how it is changed by treatment (both conventional and immunotherapies). Our approach – paired sampling at different stages of the disease and longitudinal studies of neoadjuvant treatment – is designed to help understand how tumour cells modify their surroundings and which treatment induced changes of the microenvironment underlie effective therapy.

(2)  Evolutionary Dynamics

The transformation of a cell that begins the process of tumour development, is caused by genetic mutations. As tumour cells divide they acquire more mutations and compete for resources – the fittest cells survive while others die out. This process of selection occurs at the level of cell phenotype, but is encoded and propagated by the heritable genome. We also investigate the relationship between tumour genotype and phenotype to understand how mutations sculpt the cellular ecosystem and vice versa. Similarly, the specificity that defines the adaptive immune response is encoded in complementary genomes (tumour and lymphoid cells), and we aim to link these features to cellular phenotype and spatial features such as tertiary lymphoid structures. Understanding the coevolution of cancer cells and those that comprise the tumour microenvironment should enable radically new treatment strategies based on modelling population dynamics.

(3)  Digital Pathology

Tissue diagnosis of cancer is conducted by pathologists based in the clinical setting. To use our insights to help improve patient care we need methods that can be safely used in a routine histopathology laboratory. We are developing approaches to extract key features – originally learned from high resolution multidimensional data – from routine stains. These methods – encompassing wet and dry lab approaches – will enable validation of our observations in large scale studies and their prospective evaluation in clinical trials.

The Cambridge Ecosystem

We do not work in isolation but as part of the larger Breast Cancer Programme and as part of the CRUK IMAXT Grand Challenge. This brings together scientists with diverse backgrounds encompassing cell and molecular biology, genomics, chemistry, pathology, oncology, clinical epidemiology, computational biology and astronomy.