Noninvasive prediction of response to cancer therapy using promoter profiling of circulating cell‐free DNA

Abstract

Dear Editor, Neoadjuvant radiotherapy and concomitant fluorouracil-based chemoradiotherapy (CRT) followed by total mesorectal excision is the conventional treatment for locally advanced rectal cancer (LARC). 1 However, there are individual differences in the sensitivity of cancer patients to cancer therapies. 2 The effectiveness prediction before treatment would assist clinical decisions, effectively avoid indiscriminately using drugs, reduce side effects, and improve the curative effect and quality of life. 3 Therefore, it is important to develop a novel noninvasive methodology to predict the effectiveness of cancer therapy before cancer treatment, which would enable timely interventions and provide a more individualized approach for better treatment outcomes. Cell-free DNA (cfDNA) is mainly derived from tumor and hematopoietic cells in cancer patients and it can reflect the characteristics of its tissue of origin. 4,5 For instance, the promoter coverage of cfDNA could be used to infer the expression status of tumor tissue. 5 As tumor expression status is closely related to patient's responses to cancer therapy, 6 we hypothesized that the promoter profiling of cfDNA could be used for pathologic complete response (pCR) prediction after neoadjuvant CRT. In this study, we first compared the local chromatin changes of cfDNA between the pCR and non-pCR groups of LARC patients. We further evaluated the potentials of promoter profiling of cfDNA for predicting the effectiveness of cancer therapy by developing classifiers for distinguishing pCR and non-pCR patients (Figure 1A). By comparing the local cfDNA signal between 10 pCR and 10 non-pCR patients, we observed a related loss of cfDNA signals in the mean coverage of transcriptional start site (TSS) in LARC patients with non-pCR (Figure 1B, P-value = 3.4 × 10 −21 , Wilcoxon rank-sum test). But the cfDNA signals around transcriptional terminal site (TTS) did not show significant difference (Figure 1C, Pvalue = 1, Wilcoxon rank-sum test). Therefore, we further compared their promoter profiling for each TSS, we iden-This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

A Deep-Learning Pipeline for TSS Coverage Imputation From Shallow Cell-Free DNA Sequencing

“…A set of deep learning-based denoising model for single cell sequencing data, including DCA ( 13 ), DeepImpute ( 15 ), and SCALE ( 14 ), inspired us for developing an autoencoder model for shallow cfDNA sequencing data. And a set of studies ( 3 , 11 , 16 , 17 ) also provide the theoretical basis for physiology and pathology status prediction using cfDNA-based nucleosome footprint.…”

“…The samples used in the study have been described in previous studies ( 16 , 17 ). The first study collected plasma cfDNA sequencing data of breast cancer patients, benign breast lesion patients, and healthy donors ( 17 ).…”

“…The first study collected plasma cfDNA sequencing data of breast cancer patients, benign breast lesion patients, and healthy donors ( 17 ). The other study collected plasma cfDNA sequencing data of rectal cancer patients ( 16 ). After discarding samples with insufficient data size and incomplete information, a total of 635 samples were used: 168 from breast cancer patients, 140 from benign breast lesion patients, 168 from rectal cancer patients, and 159 from healthy donors.…”

Noninvasive prediction of axillary lymph node status in breast cancer using promoter profiling of circulating cell-free DNA

“…The calculation of promoter profiling was similar to that used in our previous study [ 15 , 21 ]. In briefly, gene information was downloaded from RefSeq of University of California Santa Cruz [ 22 ].…”

Whole-Genome Promoter Profiling of Plasma Cell-Free DNA Exhibits Predictive Value for Preterm Birth

“…Plasma cell-free DNA (cfDNA) has been widely used in many clinical applications demonstrating its stability and applicability in clinics 6, 7 . In pregnancy, the proportion of cell-free fetal DNA (cffDNA) in maternal cfDNA is an important feature of cfDNA.…”