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. 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
Five studies are summarised in this briefing, including 1,138 tumour samples analysed from 1,460 patients.
The evidence consists of 2 studies comparing the MammaTyper test to IHC as a part of an RCT and 2 further studies confirming the reliability of the test with different PCR machines, machine operators and pathology laboratories. One further large UK RCT examined using different breast cancer tests and compared their clinical and cost effectiveness.
Table 1 summarises the clinical evidence as well as its strengths and limitations.
Overall assessment of the evidence
The evidence shows that the MammaTyper test is a reliable way to subtype breast cancer compared with conventional IHC testing and can provide clinical information to guide subsequent chemotherapy. It also shows there is little intra- and inter-site variability in the test results, which can happen with IHC testing.
The size of the studies are appropriate to show the validity of the assay but are from non-NHS sources. The most relevant study for the NHS shows that different breast cancer tests relying on testing for the upregulation of different genes give different results in defining breast cancer subtypes. Further clinical trials on clinical outcomes for adjuvant chemotherapy treatments chosen on the basis of MammaTyper testing would be useful.
Table 1 Clinical evidence
|
Study size, design and location |
769 archived breast cancer tumour samples from a cohort of 1,010 women with breast cancer in a RCT in Finland (the FinHer trial). |
|
Intervention and comparator(s) |
RCT comparing docetaxel plus FEC and vinorelbine plus FEC subtyping of the breast cancer was done using MammaTyper (RT-qPCR) and IHC. |
|
Key outcomes |
Both IHC and MammaTyper assessed ER and PGR accurately and gave a prognosis for DDFS (HR=0.42 [95 % CI 0.25 to 0.71]) p=0.001; and HR=0.56 (0.37 to 0.8), p=0.005, respectively]. Women with luminal B-like subtype identified using MammaTyper had a better DDFS and OS when treated with docetaxel-FEC (DDFS, HR 0.52 (95 % CI 0.29 to 0.94) p=0.031), compared with those identified using IHC. |
|
Strengths and limitations |
Breast cancer subtypes defined using RT-qPCR and IHC show good concordance, but cancer MKI67 mRNA content correlated better with DDFS than Ki-67 expression. This shows MammaTyper could be used to more reliably identify patients likely to benefit from specific adjuvant therapy. |
|
Study size, design and location |
101 core needle biopsies from a 105-patient RCT on primary invasive breast cancer, Germany. Company sponsored. |
|
Intervention and comparator(s) |
Ki-67 assessed by manual scoring of slides stained by vIHC compared with automatic scoring using digital image analysis (qIHC) or MKI67 gene expression with RT-qPCR (MammaTyper). |
|
Key outcomes |
Overall agreement for measuring ER and PGR between the 3 methods was high, with qIHC and MammaTyper showing slightly higher agreement than manual IHC and MammaTyper (all >90%). A moderate correlation was seen for qIHC and RT-qPCR for Ki-67/MKI67 (Spearman's r=0.50, p=0.0001). There is no agreed cut off for measuring Ki-67 using IHC, so the authors set sensitivity=100%. Specificity for the prediction of pathological complete response was higher for MammaTyper (measuring mRNA) compared with IHC (measuring protein) (68.9% compared with 22.2%). MammaTyper showed patients whose disease achieved pathological complete response had much greater variability in proliferation levels compared with qIHC or vIHC. The authors conclude that digital image analysis can be used for assessing ER, PR and Ki-67. |
|
Strengths and limitations |
This study shows that, while results from IHC and MammaTyper show good concordance for ER and PR, correlation for Ki-67 is only moderate and MammaTyper is more specific in measuring Ki-67 compared with IHC. This is a relatively small study but on the back of a carefully defined group of patients in a RCT for neoadjuvant therapy for breast cancer. |
|
Study size, design and location |
Multicentre prospective study to test inter- and intra-site reproducibility of the MammaTyper test, 10 international pathology labs in Switzerland, Germany, China, France, Italy, US, Finland and Canada. |
|
Intervention and comparator(s) |
Testing samples from 1 tumour sample multiple times across 10 sites over several days and testing samples from 16 different tumour samples across 10 sites and several days using MammaTyper on both centrally and locally extracted RNA from FFPE breast cancer specimens. |
|
Key outcomes |
ICC=0.980 to 0.998 (excellent agreement) for quantitative measurements. Positive/negative single-marker results were reproducible and subtype agreement was excellent (kappa=0.90 to 1.00). |
|
Strengths and limitations |
This shows the MammaTyper test is precise and reproducible for measuring breast cancer biomarkers and identifying cancer subtypes in pathology laboratory sites and across different sites. |
|
Study size, design and location |
Test accuracy study across 3 sites, different device operators, different PCR devices, standard RNA samples and 16 subtyped breast cancer tumour samples. |
|
Intervention and comparator(s) |
MammaTyper and RNA standards, no comparators. |
|
Key outcomes |
The individual PCR machines showed good (>90%) between-site concordance for all 4 markers: ERBB2, ESR1, PGR, MK167 and low inter-site variability (SD=<0.7 Cqs). The intra-site agreement was high (>97%) for individual markers and was stable up to a 64-fold sample dilution. The assay was reliable using different RNA isolation methods and with samples containing up to 80% of non-tumour tissue/in situ carcinoma. |
|
Strengths and limitations |
The mRNA expression rates of the ERBB2 ESR1, PGR and MKI67 genes measured using MammaTyper allows cancer subtyping of the luminal A-like or B/HER2 positive or negative/triple-negative subgroups. This study shows MammaTyper has robust results across different PCR devices, different sites and operators and the initial RNXtract RNA kit used on the tissue before MammaTyper, gives reliable results. |
|
Study size, design and location |
313 women with early breast cancer, RCT, UK. NIHR-funded. |
|
Intervention and comparator(s) |
Standard (chemotherapy and endocrine therapy) or test-directed (chemotherapy if Oncotype DX recurrence score >25) treatment. Risk stratification was also determined with Prosigna (PAM50), MammaPrint/BluePrint, MammaTyper, NexCourse Breast (IHC4-AQUA), and conventional IHC4 (IHC4). |
|
Key outcomes |
Data were available for 302 patients and results from all tests were available for 236 patients. The 5 tests producing risk estimates (not including MammaTyper) showed modest agreement when defining high compared with low/intermediate risk: 119 (39.4%) tumours were clearly either high or low/intermediate risk and 183 (60.6%) had different risk estimates depending on the test used. Discordant subtyping between tests was seen in 123 (40.7%) tumours. MammaTyper showed moderate agreement with BluePrint, kappa=0.39 (0.29 to 0.50) and Prosigna, kappa=0.44 (0.34 to 0.54). The authors conclude that evidence provided by different tests suggests that they present broadly equivalent risk information for women with ER-positive breast cancers and individual patients may benefit from some of the specific information provided by different tests. |
|
Strengths and limitations |
This was a large RCT centrally funded in an NHS population of women with breast cancer and so represents high-quality, relevant evidence. The tests compared all rely on different combinations of genes and use different technologies (IHC, PCR, array-based technologies) and had been shown in computer-based comparisons to be discordant. This study tested direct in vivo comparisons and showed the disagreement between the tests is genuine. |
|
Abbreviations: CI, confidence interval; Cqs, quantification cycles; DDFS, distance disease-free survival; ER, oestrogen receptor; FEC, fluorouracil [5FU] epirubicin and cyclophosphamide; FFPE, formalin-fixed, paraffin-embedded; HR, hazard ratio; ICC, inter-class correlation coefficient; IHC, immunohistochemistry; mRNA, messenger ribonucleic acid; NIHR, National Institute for Health Research; OS, overall survival; PCR, polymerase chain reaction; PR, progesterone receptor; qIHC, IHC scoring using digital image analysis; RCT, randomised controlled trial; RNA, ribonucleic acid; RT-qPCR, quantitative real-time polymerase chain reaction; SD, standard deviation; vIHC, visual IHC scoring. |
|