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
Three studies are summarised in this briefing including a total of 1,025 people with stage 1, 2A or 2B melanoma. Two of the studies (Ewen et al. 2021 and Labus et al. 2020) appear to sample from the same centres. It is unclear if they both included the same sample of people with stage 2 melanomas.
All studies were retrospective validation studies, with 2 reported in abstracts and posters only. The clinical evidence and its strengths and limitations are summarised in the overall assessment of the evidence.
There are additional publications on AMBLor or the AMBRA1 and loricrin biomarkers that are not summarised in this briefing. Andrew et al. (2021) explored the use of artificial intelligence to quantify the expression of AMBRA1 and loricrin. This study was excluded because this algorithm is not included with the AMBLor kit described in this briefing. Cosgarea et al. (2021) used the biomarkers from the AMBLor kit to evaluate the impact of melanoma paracrine transforming growth factor beta2 signalling on the loss of AMBRA1. This study was excluded because it did not evaluate the AMBLor biomarkers but focused on the biology underlying epidermal loss.
Overall assessment of the evidence
The evidence on AMBLor consists of retrospective validation studies that were mostly reported in abstracts or posters. The company provided additional information on the design of these studies, which was not included in the abstracts. Based on this added information, the studies appear to be well-designed retrospective studies using prospective sampling from existing diagnostic specimens. Pathologists were blinded to patient outcomes when scoring slides with AMBLor, and data of clinical outcomes was only provided from sites once scoring was completed.
The evidence suggests that AMBLor risk classification of stage 1 or 2 melanomas may be associated with disease progression up to 12 years' follow up. AMBLor is reported to have good sensitivity and negative predictive value. The studies support the use of AMBLor as a prognostic marker for identifying low risk of progression in stage 1, 2A or 2B melanoma. But it is not suitable for identifying high-risk melanoma because of its high rate of false positives. This is in line with AMBLor's intended use as a rule-out test.
More evidence is needed on the use and value of AMBLor in stratifying low-risk stage 1, 2A or 2B melanomas. This would ideally include prospective studies evaluating the effect of AMBLor on clinical decision making and outcomes for low-risk stage 1 to 2B melanomas. Evidence is also needed comparing AMBLor with standard care to determine the patient and system benefits of its addition to the care pathway. This should include outcomes related to resource use, including number of follow-up appointments and sentinel lymph node biopsies.
Ewen et al. (2021a)
Study size, design and location
People were recruited from centres in the US (n=241) and Australia (n=170). Clinical follow-up data ranged from 60 to 287 months. Cohorts were powered for metastasis rates of 10% for stage 1 and up to 20% for stage 2 melanoma.
Findings are reported in both abstract and poster (Ewen et al. 2021b) and both are summarised in the key outcomes.
Key outcomes
Retained expression of AMBRA1 and loricrin (n=70) was associated with 97% disease-free survival compared with 87% for people with loss of these protein markers (n=341; hazard ratio [HR] 0.20; 95% confidence interval [CI] 0.09 to 0.42; p=0.01). Within 5 years, 46 people had disease recurrence. Assay sensitivity was 96% with a negative predictive value of 97%. There were only 2 false-negative results. AMBLor returned 297 false positives, with low positive predictive value (11%) and specificity (19%). The authors concluded that AMBLor is a marker to identify low-risk subsets of stage 1 or 2 melanomas, but it is not suitable for identifying high-risk melanomas.
Strengths and limitations
This study was reported in abstract and poster only. The poster included additional findings not reported in the abstract. The retrospective study design meant that AMBLor was not used in treatment decisions and its effect on clinical outcomes was not measured. The 5‑year period for measuring disease recurrence and assay effectiveness aligns with the standard follow-up period in the NHS. The study included both stage 1 and 2 non-ulcerated melanomas. All stages were combined for analysis, with no results provided on stage 1 or stage 2 separately. The study reported a high rate of false positives. The company commented that the standard treatment for early-stage melanoma is a wide layer excision, which is curative in most cases. The company claims that this explains the high rate of false positives.
Labus et al. (2020)
Study size, design and location
This study is also reported in another abstract (Ellis et al. 2019a) and poster (Ellis et al. 2019b). All findings are summarised in the key outcomes.
Key outcomes
AMBLor high- and low-risk classifications were associated with differences in disease-free survival in people with non-ulcerated stage 2A or 2B melanomas. High-risk melanoma (n=138) was associated with 56% disease-free survival at 12 years compared with 89% for low-risk tumours (n=21; HR 4.83; 95% CI 2.29 to 10.14; p=0.015). Multivariate analysis of a subgroup of 80 people with non-ulcerated stage 2B melanoma also showed lower disease-free survival (68%) at 12 years in people with high-risk tumours (n=70) compared with low-risk (83%, n=10). Assay sensitivity ranged from 95% to 96%, with negative predictive values ranging 90% to 91% and positive predictive values ranging 34% to 37%.
Further subgroup analysis of people with sentinel lymph node negative stage 2A or 2B melanomas found 36% disease-free survival at 10 years in people with high-risk melanoma (n=37) compared with 100% in low-risk (n=5). Assay sensitivity for this group was 100%, with a negative predictive value of 100% and a positive predictive value of 46%. Authors concluded that AMBLor can be used as a stratifying biomarker for adjuvant immunotherapy in people with non-ulcerated stage 2 sentinel lymph node negative melanomas.
Strengths and limitations
This study was reported in abstracts and a poster, with details of study design and findings missing from some of the publications. It was a multicentre study using cohorts across different countries. No demographic information was reported. AMBLor was assessed on previously collected diagnostic samples and was not used to guide treatment decisions.
Ellis et al. (2019c)
Study size, design and location
Peritumoral AMBRA1 expression was evaluated in an initial discovery cohort (n=76). Then, AMBRA1 and loricrin expression was correlated with clinical outcomes in 2 validation and qualification cohorts (n=379).
Key outcomes
Initial analysis showed that people with peritumoral AMBRA1 expression loss (n=54) had 82% disease-free survival at 7 years compared with 100% in people with retained expression (n=22; p=0.081). Sensitivity for identifying people at risk of disease progression was 100%, but specificity was 33%.
Analysis of the combined expression of AMBRA1 and loricrin found that people identified with low risk (n=239) had 98% disease-free survival compared with 85% for high-risk melanomas (n=140; p<0.001). This combined analysis had a sensitivity of 83%, negative predictive value of 98%, specificity of 66%, and positive predictive value of 14%. Subgroup analysis of stage 1B melanomas found AMBRA1 and loricrin expression was a stronger predictor of disease-free survival (HR 4.04; 95% CI 1.69 to 9.66; p=0.002) than Breslow depth (HR 2.97; 95% CI 0.93 to 9.56; p=0.07). Authors concluded that AMBRA1 and loricrin expression is a prognostic marker that can stratify risk better than American Joint Committee on Cancer (AJCC) staging alone.
Strengths and limitations
This was an earlier validation study of the immunohistochemical analysis of AMBRA1 and loricrin. Study design followed the Cancer Research UK prognostic biomarker roadmap, with methods and statistical analysis clearly reported. The cohorts were statistically powered to detect a HR of more than 4.0 at p=0.05. The study evaluated the use of antibodies to detect AMBRA1 and loricrin biomarkers. The company advises that these are used in the current AMBLor kit, but there have been changes to the earlier scoring methods used in the current technology and its instructions for use.
Sustainability
The company claims the technology could reduce the number of follow-up appointments, resulting in less travel and emissions. There is no published evidence to support these claims.
Recent and ongoing studies
The company reported that there is an ongoing clinical validation study on AMBLor in about 450 retrospective biopsies in Northern Ireland and Spain. The company is also planning a prospective observational study of 500 people with stage 1 or 2 melanoma. This is planned to start in October 2022 with a 5‑year follow up.