Improvements in molecular diagnostics and image analysis for detecting and managing cervical cancer

Abstract

This thesis addresses improvements in molecular diagnostics and in image analysis, and their combination for managing cervical cancer. Screening cytology has transformed early diagnosis but is resource intensive and highly dependent on skilled cytopathologists. Current tests focus on HPV detection but lack specificity. Multiplexing HPV PCR methods with identification of tumour nucleic acid markers would increase specificity. Also, following diagnosis, staging may be supplemented by quantitative image analytics; tumour texture information can be used alongside traditional histological assessments to provide prognostic information and enable optimal management planning at the outset. I therefore explored molecular and imaging biomarkers for improved early detection and characterisation of cervical cancer. Chapter 2 focuses on selecting tumour markers and HPV markers relevant to cervical cancer and designing appropriate primers for tumour (hTERT, TERC, MYC) and HPV 16 and 18 marker detection on the lab-on-a-chip (LOC) platform. The analytical sensitivity of each of the markers was determined and found to be in the clinically relevant range. It then goes on to validate the lab-on-a-chip platform using banked tissue biopsy samples from patients with cervical cancer. Chapter 3 establishes the sensitivity and specificity of HPV 16 and 18 DNA and RNA compared to MRI for early detection of cervical cancer. Finally, it trials the LOC platform prospectively using cytology samples from patients with newly diagnosed and recurrent cervical cancer and in normal controls. Chapter 4 identifies radiomic features of cervical cancers on endovaginal MRI that differ between tumours below and above the volume threshold of eligibility for trachelectomy and determines their value in predicting lymph node metastasis and recurrence in patients in the low-volume tumour group. It shows that in patients with low-volume tumours, ADC-radiomic texture analysis is potentially useful for predicting tumour recurrence. Finally, chapter 5 indicates how molecular and imaging diagnostics may be integrated into the clinical pathway of patients with cervical cancer to achieve early diagnosis and "test of cure" at the time of treatment.