217 Summary and general discussion of 40 ml using the Quick-DNA Urine kit (Zymo Research, US). The urine input volume could be increased by implementing a urine concentration step into the workflow before DNA isolation (48). Reckamp and colleagues concentrated large urine volumes up to 100 ml to only 4 ml before isolation using Vivacell 100 concentrators (Sartorius, US) for EGFR mutation detection in the urine of NSCLC patients. Both sensitivity and specificity increased when using larger urine volumes (90-100 ml vs. 10-89 ml) (68). Yet, whether increasing the urine volume also enhances test accuracy remains understudied as there is a lack of large comprehensive studies that compare different input volumes of the same urine sample. Alternatively, pooling urine samples collected at different time points might be more effective in capturing the tumor signal, as shown in Chapter 7. Collecting a particular urine portion, such as the initial urine stream (also known as the first void), poses another strategy to improve test performance, as reported for highrisk human papillomavirus DNA testing (69). Designated collection devices, such as the Collipee device (Novosanis, Belgium), could ease and standardize the collection of a particular urine portion from home. For applications relying on transrenally excreted cfDNA, such as lung cancer detection in urine, the collection of the midstream urine might be preferred, as this usually contains less mucus and exfoliated cell debris. The use of an optimal cfDNA extraction kit could also maximize the urine cfDNA yield and thereby boost test performance. The origin of liquid biopsy, being a plasma or urine sample, affects both the recovery rate and minimal length of cfDNA detected. Urine samples generally tend to contain shorter fragments due to glomerular filtration and increased DNA degradation (70). Median fragment sizes are estimated at 133 base pairs in urine, as compared to ~166-169 base pairs in plasma. Tumor-derived fragments are most likely shorter with a size of 101 base pairs in urine, as compared to ~133-160 base pairs in plasma (71, 72). Interestingly, our fragment size analysis in Chapter 5 revealed even shorter tumor-derived fragments of 80 base pairs in urine samples of ovarian cancer patients. In this thesis, the column-based Quick-DNA Urine kit (Zymo Research) was used, which permits the largest urine starting volumes (40 ml) at the most affordable price. Yet, this kit seems suboptimal for the isolation of short cfDNA fragments, with the shortest extracted fragments starting at 100 nucleotides in length and also extracting long genomic fragments (73). Oreskovic et al. compared different urine DNA extraction kits and showed that the column-based Norgen Cell-Free Circulating DNA isolation kit (Norgen Biotek, Canada) exhibited the most effective extraction of short fragments of 25 base pairs. However, larger fragments of 40-150 base pairs were moderately recovered and consistent PCR inhibition was observed (70). In another comparison 8
RkJQdWJsaXNoZXIy MjY0ODMw