Liquid biopsies, using analytes such as circulating tumour DNA (ctDNA), can detect and quantify cancer in a minimally invasive way and can provide information on tumour heterogeneity. Current limitations in the liquid biopsy field are centred around the general sensitivity of the present assays and the input and logistical requirements for sensitive detection. While detection rates from liquid biopsy platforms are good when ctDNA levels and tumour burden are high, they are lacking the required sensitivity to detect cancer in the early stage and minimal residual disease setting. Here I present the INtegration of VAriant Reads (INVAR) pipeline, which can greatly enhance the sensitivity of ctDNA detection by utilising large patient specific mutation lists on sequencing data. The methodology can be applied to (custom) capture and (shallow) whole genome sequencing data and detects ctDNA down to parts per million, proving more sensitive than previously published methods. INVAR was applied to samples from 90 treatment naïve patients with early stage non-small cell lung cancer to characterise the ctDNA levels and could provide a better sensitivity than comparable cohorts. In addition, I outsourced samples from 27 patients to be analysed with the InVisionSeqTM assay, which does not require a priori tumour information but proved to be less sensitive than INVAR. I also assessed the potential for reducing the logistical burden in ctDNA analysis by sampling minimal blood volumes that do not require immediate processing. I interrogated if ctDNA can be detected from as little as a blood spot and show ctDNA detection using this approach in both, xenograft and human samples. ctDNA detection from blood spots provides a means to sample xenograft models without having to sacrifice the animal, allowing for longitudinal monitoring in this setting. It may also present an opportunity to frequently sample blood from patients and could reduce the logistical burden on sample collection and processing. In the future, after optimising the protocol, this could serve to reduce the complexity of clinical/translational studies. The removal of large DNA fragments using this protocol may also facilitate the analysis of ctDNA from archival cohorts where samples were collected under suboptimal conditions.