Abstract 2284: Patient-specific circulating tumor DNA monitoring using esophageal squamous cell cancer gene panel and digital PCR

• Published in: Cancer research (Chicago, Ill.) Ill.), 2019-07, Vol.79 (13_Supplement), p.2284-2284
• Main Authors: Iwaya, Takeshi, Endo, Fumitaka, Sasaki, Yasushi, Sasaki, Noriyuki, Yaegashi, Mizunori, Akiyama, Yuji, Sasaki, Akira, Masuda, Mari, Yamada, Tesshi, Tokino, Takashi, Nishizuka, Satoshi S.
• Format: Article
• Language: English
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.AM2019-2284

Background: Esophageal squamous cell carcinoma (ESCC) is belligerent cancer with poor prognosis. At present, no specific ESCC biomarkers have been established for therapeutic monitoring. This study aimed to monitor ESCC therapeutic response using circulating tumor DNA (ctDNA) using digital PCR (dPCR) with individual ESCC tumor-specific mutations as identified from the originally-designed ESCC-specific next generation sequencing (NGS) panel. Methods: Thirty-five ESCC patients were enrolled in our study (stage I/II/III/IV: 8/3/19/5, respectively). First-line therapies included surgery, chemotherapy with cisplatin/5-FU, and docetaxel/cisplatin/5-FU for 8, 3, and 24 patients, respectively. Mutation screening of primary ESCCs was performed with amplicon sequencing using ESCC panel, targeting 31 genes. The level of ctDNAs was evaluated by dPCR using 58 originally-designed sets by Hypercool Primer & ProbeTM based on mutations of the individual patient. A total of 458 blood samples were examined for ctDNA with a median follow-up time of 392 days. NGS screening of TP53 gene in plasma DNA was also performed. Results: Among mutations whose variant allele frequency (VAF) was higher than 5%, the ESCC panel identified 5.1 mutations per sample. The most frequent mutated gene was TP53 (33 of 36, 92%). Among the 58 mutations identified from primary tumors analyzed by dPCR in pretreatment blood, 44 (76%) were detectable as ctDNA. The VAFs of plasma DNA in advanced stages were significantly higher than those of early stages (p < 0.0001). The rate of positive ctDNA was 12.5% (1/8) in stage I, whereas that of stage II or higher was 92.3% (24/26). In stage II or higher, the average tumor volume was 65,572 mm3 (range 3,816-521,818 mm3). There was a low correlation between tumor volume and ctDNA-VAF (r=0.51, 95%CI: 0.13-0.76). Sensitivity/specificity and positive/negative predictive value of ctDNA in plasma were 73.5%/93.5% and 96.2%/70.0%, respectively. These data suggest that ctDNA has higher credibility than conventional serum tumor makers, such as CEA, SCC, and CYFRA. In chemotherapy patients, those with a detectable level of ctDNA (n = 12) showed poorer survival than those with an undetectable level of ctDNA (n = 13) (hazard ratio: 0.12, 95% CI, 0.03-0.49, p = 0.001, log-rank). The VAFs of primary tumor mutations by both NGS and dPCR were strongly correlated (n = 59, r = 0.97), whereas the correlation between NGS and dPCR for VAFs of ctDNA for TP53 mutations in pretreatment