Liu, Wing Kin; (2025) Radiological tracking of lung cancer metastases from relapse to death using longitudinal imaging and ctDNA. Doctoral thesis (Ph.D), UCL (University College London).

Text Liu_10212787_Thesis.pdf Access restricted to UCL open access staff until 1 September 2026. Download (17MB)

Abstract

Metastatic non-small cell lung cancer (NSCLC) shows significant heterogeneity in clinical progression and treatment response. The prognostic impact of the distribution and growth dynamics of metastases from relapse to death remains poorly understood. Characterising these patterns on a lesion level may reveal the biological mechanisms underlying metastatic growth. Longitudinal 3D volumetric data was compiled for 704 metastases, alongside multi-region sequencing from primary and metastatic tumours from 151 patients with early-stage NSCLC who relapsed after surgical resection in the TRACERx (TRAcking Cancer Evolution through therapy (Rx) and PEACE (Posthumous Evaluation of Advanced Cancer Environment) studies. I integrated radiological data with tumour genomics to uncover genomic and transcriptomic features driving metastatic dissemination and treatment resistance. Additionally, I explored the clinical utility of circulating tumour DNA (ctDNA) as a marker for tracking changes in tumour volume during the disease course. I observed that the anatomical nature of metastatic disease progressed from relapse to death in distinct temporal patterns. Extrathoracic metastases, including bone, liver, and adrenal glands emerged later in the disease course and were associated with poor prognosis. Despite significant heterogeneity in lesion growth rates within patients, rapid metastatic growth at first relapse was associated with poor prognosis. Transcriptomic and genomic features of the primary tumour, including proliferation, immune suppression, and chromosomal instability (CIN) were associated with rapid metastatic growth. A mixed response to treatment was associated with worse prognosis and extrathoracic metastases demonstrated lower response rates than intrathoracic metastases, potentially due to higher CIN in the metastasis. ctDNA tracked with the total metastatic burden for most patients, with liver metastases showing the highest ctDNA shedding per unit of tumour volume, suggesting organ-specific differences in ctDNA release. By integrating longitudinal imaging with genomic, transcriptomic, and ctDNA analyses I provide real-world insights into the molecular features that underpin phenotypic lesion-specific heterogeneity in metastatic NSCLC.

Download activity - last month

Export as XMLCSVJSON

Download activity - last 12 months

Export as XMLCSVJSON

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item