Photoacoustic Imaging of Head and Neck Tumours

Overview

Cancer is the commonest cause of death in the UK, and a national and international healthcare priority. Survival in the UK is relatively poor vs. European comparators1, meaning early tumour detection and accurate clinical assessment is particularly important to improve outcomes. Treatment fundamentally depends on tumour staging, both of the local cancer and of draining lymph nodes (LN), as well as distant spread of disease i.e. TNM stage (tumour (T), node (N) and metastases (M). However, current non-invasive pre-operative imaging technologies of ultrasound (US), computed tomography (CT) and magnetic resonance imaging (MRI) are limited in sensitivity and specificity for nodal assessment, either missing disease or subjecting patients to unnecessary additional invasive biopsies or surgery. A simple, rapid, non-invasive tool to assess primary tumours and LN involvement would be of great clinical value. One candidate technology is photoacoustic tomography (PAT), a relatively novel modality that combines exquisite spatial resolution with the ability to image multiple biological tissues, including blood, water and lipid. To date, PAT has been most successful in imaging the vasculature, which is of particular interest for oncological imaging because one of the key hallmarks of cancer is the development of new, abnormal blood vessels (neoangiogenesis). The high sensitivity for superficial imaging with PAT means that head and neck tumours and neck LN are readily amenable for assessment. In this cohort of patients, those with oral cavity tumours, in particular tumours arising from the lining of the tongue, would be readily accessible for direct scanning.

Cancer accounts for almost 30% of all deaths in England and Wales. Head and neck cancer is the 8th most common cancer in the UK, and over the last decade the incidence and mortality rates have increased by 24% and 14% respectively. Squamous cell carcinoma (SCC) or a variant is the commonest histologic type in these cancers. The prognosis and treatment fundamentally depend on the TNM stage. For example, the 1- and 3-year survival rates for patients with early stage oral cavity cancer (Stage 1 and 2) are around 90% and 65%; however, for late stage disease (Stage 3 and 4), this drops to around 80% and 45% respectively. Patients with advanced disease often require more aggressive treatment such as pre-operative (neo-adjuvant) chemotherapy, radiotherapy or both; and usually need more extensive surgery / radiotherapy when definitive treatment is instituted. The eighth edition of the TNM classification of malignant tumours, which provides internationally agreed-upon standards to stage cancer, now includes the depth of invasion (DOI) of oral cancer for accurate T staging, with 5mm being the cut off between Stages 1 and 2, and 10mm the cut off for Stages 2 and 3 disease. Not only does the precise extent of local growth determine whether or not a limited local resection can be achieved rather than a more radical excision, but the DOI dimension, which is thought to reflect the proximity to underlying lymph and vascular structures, is a good predictor of LN involvement. As described, the survival rate in these patients is related to the stage of the disease therefore imaging is a crucial determinant of prognosis. There is a paucity of data on the reliability of radiologic and clinical staging of DOI as well as the subsequent concordance of these measurements against the pathologic i.e. post-surgical, DOI. The imaging modality used to assess the T stage/DOI of oral cavity cancer varies between US, CT and MRI depending on the preference of the centre. Small studies evaluating how accurate each of these imaging modalities are at determining DOI have been performed with promising results, however they are limited. Limitations in radiologic and clinical staging of the neck LN are also well known and elective neck dissection (END), which involves resecting the entire regional LN group, is still considered the most accurate diagnostic procedure for N staging. The reported mean incidence of subclinical occult nodal metastasis detected at END in patients with early stage tongue cancer is around 25% - implying these were missed by conventional imaging. However, this is an invasive surgical technique with significant associated morbidity e.g. neuropathic pain and reduced shoulder movement, and there is uncertainty as to whether it provides any meaningful clinical benefit over observation in patients with early stage disease. An alternative to END in radiologic/clinical N0 SCC of the oral cavity is sentinel lymph node biopsy (SNLB), which involves identifying and excising the initial LN(s) to which a primary tumour drains, thereby attempting to limit the surgical dissection. Even so, this still requires a surgical procedure with the associated risks, costs and patient inconvenience, is not widely-available and still misses \>15% of diseased nodes. A rapid and easy-to-use, non-invasive and well-tolerated test that could improve local and regional cancer staging, as well as help to standardize this assessment across centres would therefore be of great clinical value. Photoacoustic tomography (PAT) is a relatively novel technology that may be able to help address these needs. PAT relies upon the absorption of laser-generated light of specific wavelengths (often in the infra-red spectrum) by intrinsic components of the imaged tissue. Such absorption results in the emission of ultrasound waves, which can be reconstructed into images in a similar manner to conventional ultrasound scanners. By imaging at multiple wavelengths, the tissue distribution of water, lipid and haemoglobin (in red blood cells and therefore blood vessels) can be mapped with extremely high resolution (\~100 microns), raising the possibility that PAT can depict the small volume tumour that existing techniques cannot.