Clinical and technical evidence
A literature search was carried out for this briefing in accordance with the interim process and methods statement. This briefing includes the most relevant or best available published evidence relating to the clinical effectiveness of the technology. Further information about how the evidence for this briefing was selected is available on request by contacting email@example.com.
Five studies are summarised in this briefing. There are several other studies that have not been summarised in this briefing. We have focused on published comparative and validation studies with the best methodological quality and largest study populations.
Three retrospective analyses of 269 people (Perrot et al. 2018), 133 people (Martinez-Lopez et al. 2014) and 125 people (Takamatsu et al. 2017) with multiple myeloma are included. One prospective cohort study of 110 children (Faham et al. 2012) with newly diagnosed ALL in the US is summarised. An analytical evaluation of 66 people (Ching et al. 2020) with ALL, CLL and multiple myeloma is also included.
The clinical evidence and its strengths and limitations is summarised in the overall assessment of the evidence.
The evidence base for the clonoSEQ assay consists of a few large studies with many smaller studies. Several analytical studies have been done to determine the specificity and sensitivity of the technology when used as a diagnostic tool. The primary outcomes in the studies mainly assess survival and time to progression based on MRD status. Many comparative studies have been done between NGS and the 2 current methods for MRD assessment, multiparametric flow cytometry (MFC) and allele-specific oligonucleotide-polymerase chain reaction (ASO‑PCR). None of the studies are in a UK or NHS context, but some have been done in Europe and the US. Most of the studies are retrospective analyses or cohort studies, no randomised controlled trials were found.
Studies on sequencing assays have been done across all 3 of the patient groups specified by the company: multiple myeloma, ALL and CLL. In addition, both adults and children are included in the evidence base. Overall, the evidence suggests that NGS is a useful method for MRD assessment, with higher levels of sensitivity than methods it has been compared with. The technology has been shown to be accurate when more than 1 in 1,000,000 cells are malignant. The evidence base consists of independent studies and studies funded by the company.
Retrospective analysis of bone marrow samples from 269 people with multiple myeloma to detect MRD using NGS and MFC. The study was done in France, Belgium and Switzerland.
MRD assessment was completed in 224 people at the start of maintenance therapy and 183 people after completion. Before maintenance therapy, MRD status was found to predict progression-free survival (PFS; hazard ratio 0.22, p<0.001) and overall survival (OS; hazard ratio 0.24, p=0.001). After maintenance therapy MRD was also predictive of PFS (p<0.001) and OS (p=0.008). People who tested negative for MRD had a higher probability of prolonged PFS than people with detectable residual disease, regardless of treatment group or risk profile at diagnosis. People with the deepest level of MRD‑negativity (<10-6) had longer PFS than people with higher levels of disease burden (both before and after maintenance; p<0.001). People who still tested negative for MRD after maintenance had better PFS and OS than people who had MRD before and after maintenance or tested positive for MRD after maintenance. There were 429 samples (75%) that were tested with clonoSEQ that had the same result as testing with MFC. There were 143 results that were different between the 2 methods. Of these, 133 samples tested positive for MRD with clonoSEQ and negative with MFC, while 10 samples tested negative for MRD with clonoSEQ and positive with MFC.
A significant limitation of this study was that NGS was only compared with MFC by assessing whether results were the same using both tests. No analysis was done to follow the OS or PFS of people who had results that differed between clonoSEQ and MFC. This would have shown whether NGS or MFC MRD assessment is a more reliable indicator of patient outcomes. The follow-up period used was median 50 months after initial maintenance therapy, which is relatively long. The study also used a large sample size, with 562 samples from 269 people.
People who were found to not have MRD with clonoSEQ had a significantly longer time to tumour progression (TTP) than those who had MRD (median 80 months compared with 31 months, p<0.0001) and better OS (median not reached compared with 81 months, p=0.02). When classifying different levels of MRD in people the TTP medians were:
MRD above 10-3, 27 months
MRD 10-3 to 10-5, 48 months
MRD below 10-5, 80 months.
Similarly, MRD below 10-5 was associated with significantly longer OS compared with MRD above 10-3 (median not reached compared with 55 months, p=0.002).
MRD information from MFC and ASO‑PCR analysis was available in 99 and 41 people, respectively. Results were concordant in 83% of samples between NGS and MFC, and in 85% of samples between NGS and ASO‑PCR. To assess the clinical significance of difference between methods, a comparison was made for the 99 people who had MRD results using both sequencing and MFC. There were 82 people who had the same results (60 double‑positives and 22 double‑negatives), whereas 12 tested negative for MRD with MFC and positive with clonoSEQ. For the comparison with clonoSEQ negative cases (median TTP not reached), MFC negative or clonoSEQ positive cases had a TTP of median 50 months (p=0.05). Of the 5 remaining people who tested positive for MRD with MFC and negative with clonoSEQ, only 1 person had disease progression during the study follow-up period.
A significant limitation of this study was that time to progression was only compared between NGS and MFC, despite 41 people having both NGS and ASO‑PCR results available. Also, the length of the follow-up period was not stated. A strength of this study was the focus on TTP as an outcome, which is an important consideration for people with multiple myeloma. Classifying results by amount of MRD is another significant strength, because this provides further evidence of the high sensitivity of NGS when comparing outcomes between the 3 groups.
Technical performance of the assay was tested using diagnostic samples from 12 of the 110 people with ALL. Serial dilutions of leukaemic cells ranging from less than 1 in 1,000,000 cells to less than 1 in 1,000 cells were prepared and analysed. The assay was highly quantitative for frequencies above 10-5, random error increased at clonotype frequencies below 10-5. The assay showed high r2 values (median 0.9991, range 0.977 to 0.996) between each of the expected and measured clonotype frequencies. The cell dilution experiments showed that the assay was quantitative with sensitivity levels at or below 1 malignant cell in 1,000,000.
MRD was assessed in follow-up samples during therapy from 106 children. Concordance was tested between MRD results obtained by the assay and flow cytometry, which was used in 105 of the 106 children. The 2 methods gave concordant positive or negative MRD results in 95 out of 105 samples (90%). In 10 samples (10%), MRD was positive with clonoSEQ but undetectable by flow cytometry. MRD levels ranged between 0.00004% and 0.011% by sequencing in 9 of the 10 samples, 7 of these were also positive with ASO‑PCR. Results between the assay and ASO-PCR were concordant in 102 of 106 follow-up samples (96%). In 3 samples, MRD was positive with clonoSEQ but undetectable by ASO‑PCR. The remaining sample was detected as 0.002% by ASO‑PCR but undetectable by sequencing.
A strength of this study is that it tested samples from children rather than adults, across the evidence base there are no comparative studies that use a large population of children. The comparison between the assay and 2 different methods was done without knowledge of the results beforehand, this aspect of blinding reduces the chance of bias. The dilution study used to test the sensitivity of the assay is also useful in this study population.
There were 68 autograft and 25 bone marrow samples assessed with both MRD assessment methods. There were 35 samples that tested positive for MRD with NGS and tested negative for MRD with ASO‑PCR. This was consistent with the higher sensitivity of the NGS method (10-6 compared with 10-4 to 10-5). A high correlation was seen between NGS and ASO‑PCR results at MRD levels of 10-5 or higher (r=0.618, p=0.005).
The researchers investigated whether people whose autografts tested positive for MRD with NGS and negative with PCR have different prognosis from those whose autografts tested negative with NGS. PFS was compared in 7 NGS‑MRD negative autograft cases (group 1) with 11 NGS‑MRD positive and ASO‑PCR MRD negative cases (group 2). People in group 1 showed significantly better PFS than people in group 2 (p=0.018). There was no difference in OS between the 2 groups, which was 100% for both. PFS was also compared in the 7 NGS‑MRD positive and ASO‑PCR MRD negative cases (group 2) with 12 NGS‑MRD negative bone marrow cases (group 3). People in group 3 showed significantly better PFS than people in group 4 (p=0.001), but there was no difference in OS (both 100%).
The major limitations of this study are the relatively small sample size and retrospective nature of the analysis, rather than prospective. Also, some of the authors are employees of the company. A strength of the study is the comparison of MRD assessment between NGS and ASO‑PCR in 2 different types of tissue, autograft and bone marrow. MRD status is most often used as a predictor of OS or PFS, the 2 primary outcomes that were investigated in this comparison.
The analysis aimed to determine the sensitivity and specificity of clonoSEQ by assessing the limit of detection (LoD), limit of quantitation (LoQ) and the limit of blank (LoB). LoD was defined as the malignant cell count at which the assay would detect MRD in 95% of samples. LoQ was defined as the lowest clonoSEQ sample MRD frequency that could be determined within 70% relative total error. LoB was the probability that a non-malignant clone would not be excluded, which could lead to false detection or an inflated estimate of MRD. Results were based on combined data from ALL, CLL and multiple myeloma samples at 2 DNA input levels (500 nanograms and 20 micrograms).
The LoD of the clonoSEQ assay was estimated to be 1.903 malignant cells at an input level of 20 micrograms of DNA and the LoQ was 2.39 malignant cells. Follow-up studies confirmed the LoD and LoQ across total DNA inputs ranging from 200 nanograms to 40 micrograms. The LoB of the assay was 0 at both 500 nanograms and 20 micrograms of healthy donor DNA. This confirmed that less than 5% of MRD measurements in healthy samples produce non‑zero values. Comparison between the clonoSEQ assay and MFC measurements showed similar quantitative accuracy across the tested range, particularly at MRD frequencies above 10-4 (r2=0.98). Relative bias between disease burden increased at lower cell inputs, a test range that spans clonoSEQ's LoQ (2.39 cells).
A significant strength of this study is that sensitivity and specificity of the clonoSEQ assay was assessed in all 3 disease groups: ALL, CLL and multiple myeloma. A comparative analysis with MFC showed that clonoSEQ identified cells at lower input levels. The authors stated that a subgroup of people with B‑cell precursor ALL need other methods of MRD monitoring. A limitation of the study is that it was wholly funded by the company.
The clonoSEQ Watch Registry: a prospective, multi-centre, observational study of adults with a diagnosis of ALL, multiple myeloma, CLL or non-Hodgkin lymphoma. ClinicalTrials.gov identifier: NCT04545333. Status: ongoing, recruiting. Indications: multiple myeloma, ALL, CLL, non‑Hodgkin lymphoma. Device: clonoSEQ. Estimated completion date: April 2024. Location: the US.
DNA sequencing-based monitoring of MRD to predict clinical relapse in aggressive B-cell non-Hodgkin lymphomas. ClinicalTrials.gov identifier: NCT02633111. Status: active, not recruiting. Indication: B-cell non‑Hodgkin lymphoma. Device: clonoSEQ. Estimated completion date: October 2023. Location: the US.
Study to assess for MRD in multiple myeloma to determine if MRD-negativity allows people to stop post-transplant maintenance therapy after having at least 1 year of maintenance therapy. ClinicalTrials.gov identifier: NCT04108624. Status: active, recruiting. Indication: multiple myeloma. Device: clonoSEQ. Estimated completion date: December 2024. Location: the US.