Skip to content
Blog | Jun. 05, 2019

OncoBEAM ctDNA testing (liquid biopsy) for therapy surveillance in melanoma

Research at Johns Hopkins University demonstrates potential clinical utility of targetable, real-time monitoring of response to melanoma cancer therapy

Skin cancer is the most common of all cancers (in the US affecting some 3.3 million individuals), yet “fortunately, most skin cancers are slow growing, easy to recognize, and relatively easy to treat when detected early”. (Source: American Cancer Society Fact Sheet PDF) Yet for invasive melanoma (affecting 1% of the cases or about 96,000 individuals in the US), the five-year survival rate is only 23%.

Liquid monitoring – optimizing melanoma detection using ctDNA

In a recent webinar, Dr. Evan Lipson of the Johns Hopkins University School of Medicine Sidney Kimmel Comprehensive Cancer Center presented work he titled “Liquid monitoring: optimizing melanoma detection using circulating tumor DNA (ctDNA)”. In his presentation, he laid out the obstacles for surveillance after initial metastatic melanoma treatment (typically surgical resection and follow-up adjuvant chemotherapy), the potential utility for liquid monitoring, and then showed data from what he termed targetable, real-time monitoring of disease.

The obstacles include local recurrence of metastatic disease, usually through the lymphatic system or the bloodstream, and the tools currently used for monitoring are expert radiology to examine Computerized Tomography (CT) and Magnetic Resonance Imaging (MRI) scans, as well as a protein-based marker called LDH (lactate dehydrogenase). LDH measures general tissue damage, and Dr. Lipson commented this test has poor predictive value for advanced Stage IV melanoma and limited value for monitoring disease burden in response to therapy.  Dr. Lipson highlighted that while many melanoma patients receive great benefit from immunotherapy, there is a need for clinical diagnostics to determine accurate disease clearance and residual tumor activity in order to more precisely tailor therapy.  Radiographic imaging has a limited resolution and LDH lacks the dynamic range to accurately determine micrometastatic disease. 

Dr. Lipson reviewed the molecular drivers of melanoma per work published in 2015 by The Cancer Genome Atlas network titled “Genomic Classification of Cutaneous Melanoma”, where ~50% of melanoma were driven by mutations in the BRAF gene, while another ~30% more were driven by NRAS, accounting for about 75% – 80% of melanoma cases. He called this slide “Capitalizing on hot-spot somatic gene mutations in cutaneous melanoma”, laying out the rationale for this work.

Prior work using digital PCR BEAMing technology

Citing prior work with 12 patients with metastatic melanoma undergoing immune-checkpoint blocking therapy with primary tumor analysis followed-up with every 2 to 4 weeks of regular sampling of blood plasma samples, he showed data from this Journal of Immunotherapy of Cancer article published in 2014 where individual patients’ increase of ctDNA correlate with radiographic disease progression.

Of interest was its measurement – in terms of percent of Genome Equivalents (GE) of circulating tumor DNA as a percentage of wild-type DNA and its correlation with a common standard of disease progression, the Sum of Longest Diameter (SLD, in mm) of the identified metastatic lesions. The quantitation of genomic equivalents was through use of a clinically-validated digital PCR method called BEAMing, first described in 2008 and commercialized by Sysmex Inostics.

Key goals for the current study and its methods

Dr. Lipson posed several goals in his evaluation of BEAMing technology: to estimate concordance with tissue mutation analysis, to measure the clinical limits of detection and influence of targetable tumor mutation recognition, to assess evidence of disease activity compared to radiography, and to evaluate clinical utility as a blood-based tumor marker in patients treated with systemic therapy including immune checkpoint blockers.

He then described their patient selection criteria, involvement of expert radiologists, the specific mutations in the OncoBEAM test used (for BRAF it was V600E/K, for NRAS Q61H/K/L/R), the limit of detection at 0.03% genomic equivalents, and described three experimental cohorts.

In brief the A cohort comprised of a single plasma sample of a group of 57 metastatic melanoma patients to see how many ctDNA mutations can be detected, and then calculate sensitivity and specificity of the ctDNA OncoBEAM method. The B cohort comprised of 29 patents with high risk, surgically-resected melanoma with one of the six hotspot mutations detectable with the OncoBEAM test, and serial plasma samples analysed coincident with clinical and radiographic evaluations. The last C cohort were 30 patients with unresectable or metastatic melanoma with one of the six mutations, treated with medical therapy (e.g. anti-PD-1, BRAFi etc.) and serial plasma samples analysed coincident with clinical and radiographic evaluations.

Cohort A findings: Specificity of 88.2%, Sensitivity of 95.2% at an LoD of 0.03%

Measuring concordance between tissue and ctDNA, of 57 patient samples 32 were positive in both tissue mutation and ctDNA mutation; five were positive in tissue but not detected in plasma; in the remaining 20 samples 19 were wild-type for the BRAF and NRAS genes, and 1 was wild-type in tissue but positive in plasma, and another sample insufficient tissue was available for testing but the ctDNA was positive for mutations. Overall, the sensitivity of the OncoBEAM assay was determined to be 86.8% with a specificity of 100%.

In these last two cases, interestingly they were both BRAF mutations detected in plasma-only and not in tissue; both patients received BRAFi therapy (dabrafenib and tramatenib) and both patients experienced a partial response (RECIST 1.1).

Using the aforementioned SLD measurement, they were able to determine that optimum sensitivity and specificity for tumor detection were achieved at 35.5mm SLD, and in 57 patient samples achieved a specificity of 88.2% and a sensitivity of 95.2% with a limit of detection (LoD) of 0.03%.

Cohort B findings: ctDNA may not be useful early-detection marker for high risk of local recurrence

Of the 29 high risk, resected melanoma patients with one of the six BRAF or NRAS mutations, 2 patients (7%) had locoregional recurrence via ordinary surveillance, however both patients had no measurable ctDNA.

Another 3 patients (10%) had distant recurrence, of which 2 were detected by ctDNA (liver and kidney metastasis respectively) and 1 were undetected by ctDNA (lung metastasis). Dr. Lipson suggested that ‘visceral metastases’ (in internal visceral organs) were more easily detected by ctDNA.

Cohort C findings: ctDNA detection precedes radiographic disease progression in a number of cases

Of the 30 patients with unresectable or metastatic melanoma, 17 patients (57%) experienced partial or complete response to therapy. The ctDNA measurement did not detect any evidence of disease activity after on-treatment assessment.

Of the 8 patients who developed radiographic disease progression, 4 patients had evidence of disease in ctDNA at the same time as radiographic detection of progression.

Yet in 4 patients (13%), ctDNA measurement detected disease progression by 8, 14, 25 and 38 weeks (average 21.2 weeks or over five months) before conventional radiography detected disease progression. ctDNA-based evidence of disease activity was seen in three of the four patients where expert radiographic evaluations were performed with no evidence of neoplastic disease.

Dr. Lipson said ‘ctDNA could be a helpful biomarker for radiologists, it warrants a second look at a radiographic scan’.

A case example of ctDNA predicting metastatic recurrence

He finished his presentation with a case example, a 68-year old female receiving anti-PD-1 (pembrolizumab) for metastatic melanoma, where there was detectable ctDNA at baseline of 0.95% genome equivalents MAF (minor allele fraction), but the imaging was ruled to be no evidence of disease; in retrospect, a 1.5 cm rim-enhancing centrally necrotic uterine mass is seen. After 3 months, no ctDNA was collected, but the mass grew in size, and this intra-fibroid region mass growing to 3cm ‘in an odd place for melanoma to show up” the CT images were deemed no evidence of disease.

Dr. Lipson notes the high level of experience of these radiologists, and how ‘tremendously difficult’ a region of the body it is to pick out metastasis. After 5 months, a PET-CT scan shows a 5cm hypermetabolic lesion within a fibroid uterus, coincident to a ctDNA measurement of 5.7% MAF. By biopsy, the lesion was confirmed to be metastatic melanoma.

A few conclusions

Dr. Lipson concluded that “ctDNA is a useful, non-invasive, blood-based biomarker that can detect targetable oncogenic mutations not seen in tumor analysis and can provide evidence of disease activity, predicting eventual disease progression and informing radiographic image interpretation”. In addition, he also suggested that longitudinal intrapatient blood-based detection of melanoma progression “may have implications in the setting of clinical trials where surrogate endpoints such as clinical / radiographic progression-free survival are used as markers of drug efficacy”.

You can access this published work for further details: Rowe and Lipson et al. 2018 Molecular Oncology “From validity to clinical utility: the influence of circulating tumor DNA on melanoma patient management in a real-world setting”. And for additional information about our OncoBEAM technology you can access it here, and a list of our current OncoBEAM-based ctDNA liquid biopsy tests here.


  1. Cancer Genome Atlas Network. Cell 161(7):1681-96. 2015 Genomic Classification of Cutaneous Melanoma. PubMed PMID: 26091043
  2. Lipson EJ and Diaz LA Jr. et al. J Immunother Cancer. 2(1):42. 2014 Circulating tumor DNA analysis as a real-time method for monitoring tumor burden in melanoma patients undergoing treatment with immune checkpoint blockade. PubMed PMID: 25516806
  3. Diehl F and Diaz LA Jr. et al. Nat Med. 14(9):985-90 2008 Circulating mutant DNA to assess tumor dynamics. PubMed PMID: 18670422
  4. Rowe SP and Lipson EJ et al. Mol Oncol. 12(10):1661-1672 2018 From validity to clinical utility: the influence of circulating tumor DNA on melanoma patient management in a real-world setting. PubMed PMID: 30113761