Measuring circulating tumor (ct)DNA in the blood, a so-called liquid biopsy, may be used to detect minimal residual disease (MRD) after definitive treatment, as well as to predict disease relapse throughout the course of treatment. Other emerging applications include assessment of response, detection of treatment resistance, early detection of cancer, and tumor classification, said Ash A. Alizadeh, MD, PhD, Assistant Professor of Medicine, Divisions of Oncology and Hematology, Stanford Cancer Institute, at the 2018 Gastrointestinal Cancers Symposium.
Blood samples can be studied through cell-free DNA molecules for amplifications and deletions, translocations, point mutations, and chromosomal abnormalities.
Through whole-exome sequencing, whole-genome sequencing, single-mutation next-generation sequencing, and multimutation next-generation sequencing, limits of ctDNA detection have been pushed lower “to be able to address the difficult disease questions in patients with lower and lower disease burden,” Dr Alizadeh said.
However, whole-exome or whole-genome sequencing is not sensitive enough to detect ctDNA in most patients with cancer. Deep sequencing (CAPP-Seq) can overcome the challenge of limited input molecules in detecting ctDNA from the blood by using a panel of mutations known to occur in a particular cancer.
Dr Alizadeh and colleagues first targeted a panel of 300 genes recurrently mutated in non–small-cell lung cancer (NSCLC) and small-cell lung cancer.
In an initial cohort of 13 patients with stage II to IV NSCLC, “We could see high sensitivity even in a small number of patients with early-stage lung cancer at very low cell-free DNA concentrations [1.9 to 226 pg/mL],” Dr Alizadeh said. The detection threshold of ctDNA was about 0.02%.
CAPP-Seq cannot accurately sequence very small quantities of DNA and, therefore, requires amplification, which carries the potential to introduce errors in the sequence. By developing a technique called integrated digital error suppression, which tags circulating double-stranded DNA molecules with bar codes that mark each original molecule, Dr Alizadeh and colleagues were able to decrease the background error rate.
Detection of Minimal Residual Disease
There are no reliable assays to assess MRD for most solid tumors, unlike with hematologic malignancies. Detection of ctDNA MRD, which has recently been demonstrated in breast and colon cancer using personalized assays, could allow for personalization of adjuvant therapy administration and assessment of response, Dr Alizadeh said.
A prospective study of patients with NSCLC conducted at Stanford University showed that MRD detection after treatment was associated with disease recurrence, whereas the disease was unlikely to recur in patients without detectable disease. In this same study, tracking multiple mutations improved the likelihood of detecting MRD.
In the United States, the majority of patients with stage II colorectal cancer (CRC) do not receive adjuvant chemotherapy, but patients with stage III disease do. The Stanford team used a surveillance and monitoring panel of 197 genes to investigate whether ctDNA could predict outcomes after tumor resection in a group of 145 patients with stage II or stage III NSCLC.
They found that time to disease recurrence, relapse-free survival, and overall survival were all highly predicted by a single assay performed 10 days after resection. The assay’s sensitivity was 57.1%, the specificity was 100%, the positive predictive value was 100%, and the negative predictive value was 93.2%. The MRD prognostic value held up for stage II and stage III tumors.
“On multivariate analysis, preoperative ctDNA levels were the dominant source of the variation associated with risk of recurrence,” Dr Alizadeh said.
Evidence suggests, he said, that patients with MRD can potentially be selected for adjuvant therapy, which may possibly be avoided in patients who are MRD-negative. These possibilities need to be assessed in prospective clinical trials, Dr Alizadeh noted.
Early ctDNA dynamics have been found to be strongly predictive of outcomes for curative chemotherapy of diffuse large B-cell lymphoma (DLBCL). In one study of ctDNA with CAPP-Seq in patients with DLBCL, an early 2-fold log drop in ctDNA levels predicted an increased likelihood of radiographic complete remission, event-free survival, and overall survival with 1 cycle of chemotherapy.
CTCs Accurately Identify Early-Stage Colorectal Cancer
Results of a new analysis presented at the meeting showed that another type of liquid biopsy used to detect circulating tumor cells (CTCs) in the blood was able to identify all stages of CRC with 88% accuracy.
This finding comes from a study of 620 individuals who presented for routine colonoscopy or had a confirmed diagnosis of CRC, and had the CellMax CTC blood test (or liquid biopsy) using the CellMax biomimetic platform. The test requires 2 mL of blood, said lead investigator Wen-Sy Tsai, MD, PhD, Assistant Professor, Department of Colon and Rectal Surgery, Chang Gung Memorial Hospital, Taipei, Taiwan.
CRC is often diagnosed at a late stage, he said, but if diagnosed at an early, localized stage, the 5-year survival rate is 91%.
Of the 620 study participants, 111 had adenomas or polyps, 327 had stage I to stage IV CRC, and 182 were healthy controls. The CellMax CTC or liquid biopsy results were compared with colonoscopy findings to determine the predictive accuracy of the CellMax CTC assay.
The specificity of the assay was 97.3% in patients with precancerous lesions and in those with CRC. The false-positive rate was 3.3% in the control group, and the false-negative rate was 15.8% among patients with precancerous lesions or with CRC.
The sensitivity ranged from 77% for detection of CTCs in precancerous lesions to 86.9% for patients with CRC. For precancerous lesions, the sensitivity of CellMax was superior to other screening methods, except colonoscopy, with a sensitivity of 76% to 94%.
Approximately 33% of Americans have never been screened for CRC. Some 87% of unscreened individuals prefer blood tests to stool-based screening. “Our results may point to a solution for people who are reluctant to get an initial screening colonoscopy or are not compliant in returning stool-based test kits,” Dr Tsai asserted.