Non-invasive Detection of Urothelial Carcinoma by Cost-Effective Low-Coverage Whole Genome Sequencing from Urine Exfoliated Cells DNA
Shuxiong Zeng,Yidie Ying,Naidong Xing,Baiyun Wang,Ziliang Qian,Zunlin Zhou,Zhensheng Zhang,Weidong Xu,Huiqing Wang,Lihe Dai,Li Gao,Tie Zhou,Jiatao Ji,Chuanliang Xu
DOI: https://doi.org/10.1158/1078-0432.ccr-20-0401
IF: 13.801
2020-01-01
Clinical Cancer Research
Abstract:1552 Background: Urothelial carcinoma (UC) is a malignancy with frequent chromosomal aberrations. The FISH assays were more sensitive as compared to cytology tests. Here we investigated cost-effective whole genome sequencing technology, which is able to detect all chromosomal aberrations for UC diagnoses. Methods: UC patients and control group are prospectively recruited in trial NCT03998371. First-morning-voided urine were freshly collected before TURBT or cystectomy. Urine Exfoliated Cells DNA was analyzed by illumina HiSeq X10, followed by genotyping by bioinformatics workflow UCAD. Results: 195 individuals were prospectively recruited. 121 UC patients and 67 non-tumor diseases were included in this study. 7 other malignancies as confirmed by pathological testing were excluded. Frequent chromosome copy number changes were found in cancer patients as compared non-tumor controls, including chromosome 3 gain, 17 gain, 7 gain and 9p loss used in FISH assays were found. In addition to that, chr9q loss, 8q gain, 5q loss, 17p loss, 11p loss, 1q gain, 8p loss, 10q loss, 6q loss, 4q loss and 11q loss were also frequent in cancer patients (AUC > 0.65). Metacentric chromosomes showed better AUC compared to acrocentric and telocentric chromosomes (P = 1.7e-03). A novel diagnosis model UCAD was built by incorporating all the chromosomal changes. The model reached performance of AUC = 0.933. At the optimal cutoff |Z| > = 3.16, the sensitivity, specificity and accuracy were 84.7%, 97.9% and 89.0% respectively. The prediction positivity was found correlated with urine microscopy visible epithelial cells (P = 0.00069), tumor invasiveness (Ta/Tis vs the other, P = 0.0048) and tumor grade (P = 0.0030), but not microscopy RBC/WBC findings, urine culture findings, smoke and drinking history. The UCAD model outperformed cytology tests by predicting all 16-cytology positive and 12 cytology negative tumors with comparable specificity. The model found 75.0% more tumors. And UCAD identified more upper urinary tract cancer (P = 0.012) and smaller tumors ( < 3cm, P = 5.9e-04). The adding of cytology to UCAD did not improve diagnosing sensitivity and specificity. UCAD reproduce the diagnoses among morning - void urine, morning, afternoon urine samples with correlation coefficient R 2 > 0.98. All the urine samples showed high concordances with matched tumor samples (R 2 > 0.85). Conclusions: UCAD could be a high specific, robust UC diagnoses method with improved sensitivity as compared to cytology tests. Clinical trial information: NCT03998371.