Advice

Clinical and technical evidence

Clinical and technical evidence

A literature search was carried out for this briefing in accordance with the interim process and methods statement for medtech innovation briefings. 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 mibs@nice.org.uk.

Published evidence

There are 6 studies summarised in this briefing, including a total of 5,532 people.

The evidence includes 3 secondary analyses, of which 2 were derived from randomised controlled trials; 1 analytical validation; 1 prospective case series; and 1 prospective single arm phase 2 trial. The clinical evidence and its strengths and limitations is summarised in the overall assessment of the evidence.

There are 3 published secondary analyses that are not presented below (van Beers et al. [2017], Chng et al. [2016]; Kuiper et al. [2012]). These studies showed that SKY92 (formerly known as EMC92) performed well compared with other gene expression classifiers and showed improved performance when combined with other gene expression profile signatures. Kuiper et al. (2012) also outlines the development of the SKY92 (EMC92) gene expression signature.

Overall assessment of the evidence

The evidence for MMprofiler is of moderate quality. Most studies evaluated the clinical or analytical validity of the technology, with 3 studies comparing the technology with other prognostic markers. Studies showed consistent findings supporting the performance of MMprofiler (known as EMC92 or SKY92 in clinical research) in isolation or when combined with other markers. The predominance of secondary analyses meant the methodological information included in these papers were limited in detail compared with primary studies. Two studies examined differential treatment effects based on people's risk classification while 1 study explored the impact of SKY92 risk scores on treatment decisions. One study examined response rates to an intensive treatment strategy in people with ultrahigh-risk multiple myeloma determined by genetic screening, including SKY92. These studies show that SKY92 risk classification can potentially affect treatment decisions and outcomes. Further evidence is needed evaluating the use of MMprofiler in treatment decisions and its effect on clinical outcomes compared with standard care markers. Further prospective comparative research is also needed on risk-stratified treatment for multiple myeloma in clinical practice.

Biran et al. (2021)

Study size, design and location

Prospective case series of 147 people with multiple myeloma enrolled from 5 centres in the US. This paper reports initial findings from the ongoing PROMMIS trial.

Intervention and comparator

MMProfiler SKY92 compared with routine clinical practice.

Key outcomes

This study describes the clinical relevance of MMprofiler in helping healthcare professionals with treatment decisions. MMprofiler classified 43 out of 147 (29%) people as having SKY92 high-risk. The risk distribution by R‑ISS (revised International Stating System [ISS]) was 33% (44/133) stage 1, 58% (77/133) stage 2, and 9% (12/133) stage 3. Before knowing SKY92 risk classification, healthcare professionals classified 73 (50%) people as having high-risk multiple myeloma, 46 of whom were classified by MMprofiler as having SKY92 standard-risk. Review of SKY92 risk scores in these cases resulted in healthcare professionals reclassifying 30 people as having standard-risk multiple myeloma. For the 74 people initially regarded by healthcare professionals as having standard-risk multiple myeloma, MMprofiler identified 16 as having SKY92 high-risk. After reviewing the SKY92 risk scores, healthcare professionals agreed with reclassification in all cases. For 131 (89%) people, the final risk classification assigned by healthcare professionals matched the SKY92 result. Changes in risk classification affected proposed treatment plans, especially options after autologous stem cell transplant. Treatment plans were de-escalated in the 30 people reclassified as having standard-risk multiple myeloma and escalated in 15 of the 16 people reclassified as having high-risk multiple myeloma. Healthcare professionals also found SKY92 scores helpful in confirming risk classifications and treatment plans in concordant cases. MMprofiler affected treatment decisions in 37% (54/147) of cases, which was above the predefined threshold of clinical relevance of 15% (p<0.001). SKY92 results also reportedly increased healthcare professionals' confidence in their proposed treatment plan.

Strengths and limitations

This study is a prospective multicentre study comparing MMprofiler with routine clinical practice used by 30 healthcare professionals (haemato-oncologists). Routine clinical practice was not defined and therefore the risk classification methods used may have varied across healthcare professionals and centres. Healthcare professionals were blinded to MMprofiler results for their initial risk assessment and treatment planning. Authors suggested there may have been some recruitment bias because the number of people classed as having high-risk multiple myeloma in the cohort (29%) was higher than reported in the literature (15% to 25%). It is possible that healthcare professionals selected people with perceived higher risk multiple myeloma to be included in the study. The study initially recruited 250 people but 103 were excluded because of screen failure including smouldering (not active) multiple myeloma or low-quality bone marrow sample. The PROMMIS study was sponsored by SkylineDx (the company behind MMprofiler) and several members of the research team disclosed employment with or financial interest in the company.

Kaiser et al. (2021) [abstract]

Study size, design and location

Prospective single arm phase 2 trial of 107 people with ultrahigh-risk, newly diagnosed multiple myeloma or plasma cell leukaemia recruited from 39 hospitals in the UK. This abstract reports initial findings from the UK optimum/MUKnine trial which aimed to determine if a novel treatment combination was sufficiently active to take forward to a phase 3 trial.

Intervention

People with ultrahigh-risk, newly diagnosed multiple myeloma (detected by central trial genetic or SKY92) or plasma cell leukaemia had up to 6 cycles of daratumumab, cyclophosphamide, bortezomib, lenalidomide, dexamethasone (Dara-CVRd) induction, augmented high-dose melphalan, and autologous stem cell transplantation augmented with bortezomib, followed by Dara-CVRd consolidation for 18 cycles and daratumumab with lenalidomide (Dara-R) maintenance.

Key outcomes

Median follow up was 22.2 months (95% confidence interval 20.6 to 23.9). Two people died during induction because of infection. Responses in the intention-to-treat population at the end of induction were 94% overall response rate with 22% complete response, 58% very good partial response, 15% partial response, 1% progressive disease, and 5% timepoint not reached. Reponses at day 100 after autologous stem cell transplant were 83% overall response rate with 47% complete response, 32% very good partial response, 5% partial response, 7% progressive disease, and 10% timepoint not reached. Minimal residual disease status after induction was 41% minimal residual disease negative, 40% minimal residual disease positive, and 19% not evaluable. Minimal residual disease status at day 100 after autologous stem cell transplant was 64% minimal residual disease negative, 14% minimal residual disease positive, and 22% not evaluable. Authors concluded that response rates were high, with toxicity comparable to other induction regimens.

Strengths and limitations

This study seems to be the first to use genetic screening including SKY92 to prospectively identify people with ultrahigh-risk, newly diagnosed multiple myeloma to be offered an intensive treatment schedule. Ultrahigh-risk, newly diagnosed multiple myeloma is rare, occurring in about 20% of people with multiple myeloma. Recruitment of people from this patient group was achieved through multicentre involvement across 39 hospitals in the UK. The study reports high response rates in treating this difficult-to-treat population with the intensive treatment strategy. This provides some early support for the idea of risk-stratified treatment in multiple myeloma and highlights the need for further comparative research in this area. This study was reported in abstract and was therefore limited in detail.

van Beers et al. (2021)

Intervention and comparator

SKY92 gene assay; no comparator.

Key outcomes

SKY92 was found to be an appropriately sensitive test, producing robust results with varied levels of RNA input in line with the recommended minimum tumour content for the assay. The SKY92 result was not affected by the presence of interfering substances with the test demonstrating specificity in detecting high-risk multiple myeloma. The test also showed good repeatability and intermediate precision when analysed for the effect of differing reagents, microarrays, instruments, and operators. None of the precision tests done exceeded the maximum allowed standard deviation of 0.45, with class switching because of imprecision below 10%. The SKY92 array showed high reproducibility in clinical samples across 3 independent labs. Based on these findings, the SKY92 assay was said to meet or exceed the requirement for a prognostic test used in routine clinical practice.

Strengths and limitations

This study provides analytical validation of the procedural parameters to run the MMprofiler SKY92 assay in clinical settings. Analysis was guided by relevant Clinical and Laboratory Standards Institute standards and pre-defined statistical acceptance criteria. Several members of the research team disclosed employment with or financial interest in SkylineDx.

Shah et al. (2020)

Intervention and comparator

SKY92 gene signature compared with chromosomal aberrations assessed using quantitative reverse transcriptase polymerase chain reaction (qRT‑PCR) and multiplex ligation-dependent probe amplification.

Key outcomes

SKY92 identified 81 of 329 (24.6%) people as having SKY92 high-risk. People classed with SKY92 high-risk had shorter progression-free survival (median 16.0 months compared with 33.8 months; p<0.001) and overall survival (median 36.7 months compared with not reached; p<0.001). SKY92 and chromosomal high-risk markers were combined to produce 4 risk groups: SKY92 and double-hit chromosomal high-risk markers (9.7%), SKY92 or double-hit chromosomal high-risk markers (23.4%), single chromosomal high-risk marker (24%), and no risk marker (42.9%). Progression-free survival and overall survival rates showed significant improvement across higher to lower risk groups. Differential treatment effects were found for the different risk groups. Lenalidomide single-agent maintenance extended progression-free survival in people with the single chromosomal high-risk marker or no risk marker when compared with observation. This benefit was not seen in people with SKY92 or double-hit chromosomal high-risk markers.

Strengths and limitations

This study provides clinical validation of the MMprofiler SKY92 gene signature and shows its independent prognostic prediction when compared with chromosomal high-risk markers. It shows the utility of combining MMprofiler with chromosomal markers to detect high-risk multiple myeloma. Some evidence was provided to suggest benefit in using prognostic markers in treatment decision making.

Kuiper et al. (2020)

Study size, design and location

Secondary analysis of 180 people with previously untreated symptomatic multiple myeloma enrolled in a randomised controlled trial in the Netherlands, Denmark, Sweden and Norway. The larger trial (n=636) compared treatment with melphalanprednisone-thalidomide followed by thalidomide maintenance (MPT‑T) to melphalan-prednisone-lenalidomide followed by lenalidomide maintenance (MPR‑R).

Intervention and comparator

SKY92 gene classifier compared with SKY‑RISS (SKY92 combined with revised ISS [R‑ISS]).

Key outcomes

Combining SKY92 with R‑ISS (n=168) resulted in 3 risk groups: low-risk (SKY‑RISS 1; 15%), intermediate-risk (SKY‑RISS 2; 74%), and high-risk (SKY‑RISS 3; 11%). The 3‑year progression-free survival rates for these groups were 54%, 27% and 7%, respectively (p<0.001). The respective 3‑year overall survival rates were 88%, 66% and 26% (p<0.001). SKY‑RISS was independent of other prognostic markers for progression-free survival and overall survival. A differential treatment effect was seen in the SKY‑RISS 3 group (n=18), with MPR‑R treatment resulting in longer overall survival compared with MPT‑T (57% compared with 0%; median overall survival 55 months compared with 14 months; p=0.007). No difference in overall survival was found between treatment arms for SKY‑RISS 1 (n=124) or SKY‑RISS 2 (n=26). These treatment effects were also seen in the R‑ISS 3 and SKY92 high-risk groups but were less pronounced. Combining SKY92 with R‑ISS resulted in more accurate risk classification compared with either individual prognostic marker.

Strengths and limitations

This study validates the use of the MMprofiler SKY92 gene classifier to predict progression-free survival and overall survival in multiple myeloma. The findings suggest a benefit of using SKY-RISS as a prognostic test to assist in treatment decisions for people with high-risk multiple myeloma. The study was a secondary analysis with a small number of people classified as having high-risk multiple myeloma in the treatment arms.

Kuiper et al. (2015)

Intervention and comparators

EMC92 compared with 7 gene expression classifiers (UAMS17, UAMS70, UAMS80, IFM15, MRCIX6, HM19, and GP150) and standard prognostic markers (ISS, fluorescence in-situ hybridisation [FISH]).

Key outcomes

Gene expression classifiers appeared to show better risk separation than ISS and FISH. The percentage of high-risk classifications varied between gene expression classifiers, with EMC92 identifying more people with high-risk multiple myeloma (18%) than the other classifiers (8% to 12%). The EMC92‑ISS compound risk marker had the best median rank score when ranked on performance against all single and combined risk markers. EMC92 was the best single marker (ranked 7th out the 32 markers), while ISS was ranked 23rd. EMC92‑ISS classifies people into 4 risk groups: low risk (38%; median overall survival not reached after 96 months), intermediate to low risk (24%; median overall survival 61 months), intermediate to high risk (22%; median overall survival 47 months), and high risk (17%; median overall survival 24 months). EMC92‑ISS showed utility in identifying both high- and low-risk multiple myeloma. This was described as an advantage over FISH markers which identified only high risk.

Strengths and limitations

This study suggests MMprofiler combined with ISS is a strong prognostic marker for multiple myeloma. Analysis included multiple independent data sets resulting in the validation of findings in a large sample of people with multiple myeloma. However, the secondary analysis of these data sets limits details of the methodology and quality of the primary studies.

Sustainability

MMprofiler is a single-use prognostic test. The company claims the technology will have lower environmental affect compared with current UK practice. There is no published evidence to support these claims.

Recent and ongoing studies