Multimodal spectral histopathology for quantitative diagnosis of residual tumour during basal cell carcinoma surgery

Boitor, Radu (2019) Multimodal spectral histopathology for quantitative diagnosis of residual tumour during basal cell carcinoma surgery. PhD thesis, University of Nottingham.

[img] PDF (Thesis - as examined) - Repository staff only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader
Download (171MB)


Basal cell carcinoma (BCC) is the most common type of cancer in the world. The preferred method of treating BCC is surgical excision followed by histopathological investigation of the removed tissue. Mohs micrographic surgery is a highly specialised surgical procedure which involves the removal of sequential layers of skin until a tumour is completely excised. Each removed skin layer is analysed through frozen section histopathology to assess whether residual BCC is present on the resection margin and if so, a subsequent tissue layer is removed. Mohs surgery is only used to treat BCCs in high risk locations (such as the head and neck area) or recurrences. Mohs surgery is not available universally, as it requires highly specialised personnel, is labour intensive and expensive. New imaging modalities are therefore being developed to investigate whether they can accompany or replace frozen section histopathology in assessing the surgical margin of removed skin specimens within the Mohs clinical workflow.

This thesis describes the development and optimisation of a prototype instrument that utilises a multimodal spectral histopathology (MSH) approach to investigate the surgical margins of excised skin tissue samples. Multimodal spectral histopathology combines autofluorescence confocal microscopy and Raman spectroscopy to objectively determine whether the margins of an investigated sample are BCC-positive.

The prototype instrument was developed to measure tissue samples automatically (without requiring user input), and to display the measurement results as easy-to-interpret tissue maps. The maps do not require specialist interpretation, meaning that MSH can be used by non- spectroscopy users, resulting in improved objectivity of diagnosis. The MSH measurement and data analysis software, which enables the automated operation of the instrument, was optimised by measuring 97 frozen skin samples obtained after Mohs surgery of 70 patients. The MSH software was shown to be transferable between similar instruments, demonstrating the potential to produce such devices at scale.

The performance of the prototype instrument for frozen skin samples was tested by measuring an independent set of 10 samples from 9 patients. The prototype was shown to correctly diagnose all samples based on their corresponding adjacent histopathology sections.

The prototype instrument was then moved into the Mohs clinic where it was used to measure fresh tissue samples intra-operatively. In order to adapt the prototype instrument to the clinical setting, 50 fresh tissue samples from 18 patients were investigated. The MSH software and tissue handling procedure were adapted to account for the abundant presence of blood and surgical positioning ink. The measurement time was also shortened to 30 minutes per Mohs layer, to increase compatibility with the Mohs workflow. The performance of the prototype instrument for fresh samples was assessed by measuring an independent set of 30 samples from 12 patients, after the software changes were finalised. Fresh sample diagnoses were shown to be in accordance with Mohs frozen sections for 96.4% of the test samples. These results were enabled by the use of a multinomial logistic regression classification model capable of distinguishing between BCC and healthy spectra with a 94.3% sensitivity and a 95.3% specificity.

The prototype described in this thesis is, therefore, the first fully-automated instrument based on Raman spectroscopy for intra-operative microscopic imaging of surgical margins during cancer surgery, suitable to be used by a non-specialist user in a clinical environment.

Item Type: Thesis (University of Nottingham only) (PhD)
Supervisors: Notingher, Ioan
Williams, Hywel
Keywords: basal cell carcinoma, Mohs surgery, multimodal spectral histopathology, spectroscopy
Subjects: Q Science > QC Physics > QC350 Optics. Light, including spectroscopy
R Medicine > RB Pathology
R Medicine > RD Surgery
Faculties/Schools: UK Campuses > Faculty of Science > School of Physics and Astronomy
Item ID: 56911
Depositing User: Boitor, Radu
Date Deposited: 11 Oct 2023 12:00
Last Modified: 11 Oct 2023 12:00

Actions (Archive Staff Only)

Edit View Edit View