Circulating tumour DNA (ctDNA) as a biomarker in metachronous melanoma and colorectal cancer- a case report

In recent years, tumour-derived cell free DNA (ctDNA) has emerged as a promising biomarker of disease status for metastatic cancer [1‐3]. Plasma ctDNA are short nucleic acid fragments (~ 166 bp) thought to be released in the systemic circulation as a result of tumour cell apoptosis and/or necrosis [4, 5]. Previous studies have shown that ctDNA carries genetic information from the entire tumour genome and can therefore provide insights into clonal heterogeneity and evolution of all solid cancers present at any one time [6, 7]. As analysis of ctDNA can be tailored for different cancers by targeting specific mutations, it provides detailed information via a minimally invasive ‘liquid biopsy’, eliminating the morbidity associated with serial sampling of tumours for monitoring patients with any advanced solid cancers.

Various studies in breast, lung and colorectal cancers have demonstrated the potential clinical application of ctDNA analysis at each stage of cancer management: early diagnosis [5, 8], molecular profiling [6, 9‐11], prognostication [5, 12, 13], detection of residual disease [14, 15], monitoring response and clonal evolution [16‐20]. In melanoma, several studies have also shown the efficacy of utilising ctDNA for monitoring patients with BRAF mutant tumours, particularly in the context of treatment response and identification of mechanisms of resistance to BRAF inhibitors [7, 21‐26]. These studies provide credence to the utility of ctDNA for patient monitoring only in the context of singular cancer. To date, ctDNA remains unutilised in clinical management of patients with multiple tumour types and/or those metachronous cancers where new primary tumours arise that are unrelated to the original malignancy. In this case study, we demonstrated the efficiency of ctDNA to delineate the different status of both melanoma and colorectal cancers in a single patient.

At progression the patient was subsequently treated with four doses of ipilimumab (3 mg/kg three weekly) but was found to have disease progression on the first response assessment CT scan. Confirmed progression in lung metastases and the intra-abdominal nodal disease led to commencement of anti PD-1 therapy (pembrolizumab 2 mg/kg three weekly) in March 2015 (week 2, Fig. 1). He completed 28 cycles (week 94) of pembrolizumab and achieved a complete metabolic response on PET at six months in all the previously identified metastatic sites. He tolerated treatment well with vitiligo as the sole side effect.

However, PET at 32 weeks identified a new FDG avid lesion within the sigmoid colon. This was investigated with colonoscopy and tissue biopsy confirmed a low grade sigmoid adenocarcinoma. He proceeded to a subtotal colectomy, ilio-sigmoid anastomosis and lymph node dissection in January 2016 (week 46). Histopathology confirmed a stage III (T4N1M0 AJCC 7th edition) low grade sigmoid adenocarcinoma with 3/33 lymph nodes involved. The tumour had no mismatch repair deficiency. Molecular analysis using next generation sequencing via the Illumina Trusight tumour panel showed the primary tumour to be KRAS p. G13D mutant, NRAS and BRAF wild type. Post-operative CEA measurements were negative.

The recurrence was unresectable and the patient was offered palliative FOLFOX chemotherapy with bevacizumab (B) but chose to undergo observation with three monthly clinical and radiological reviews. His imaging demonstrated RECIST (Response Evaluation Criteria in Solid Tumours) stable disease for 18 months and he then progressed with new liver and lung lesions. He has recently commenced B-FOLFOX chemotherapy with response assessment pending.

This case study highlights the evolving role of ctDNA in detecting metachronous cancers. Development of new primary tumours that are unrelated to the original malignancy have become a significant adverse effect in patients with metastatic cancers who achieved sustained or durable disease control in response to targeted and/or systemic therapy. Thus, oncologists should be vigilant to the possibility that patients are at continued risk for new and separate malignancies.

Previous studies have demonstrated the clinical utility of ctDNA as a biomarker of disease status in patients with metastatic cancers, particularly in the context of singular tumour types. However, to date there has been no report of the clinical utility of ctDNA to delineate status of different tumour types within patients with multiple cancers. In this case study, we demonstrated the efficiency of ctDNA, by targeting tumour-specific mutations to specifically inform treatment response and tumour status in a patient with both melanoma and colorectal cancer. Given the increased risk of cancer patients to develop other malignant tumours, this study supports the potential clinical use of ctDNA for profiling of other emerging lesions and identification of their origin. Plasma ctDNA may be useful for accurate stratification of treatment response in patients with two or more different tumour types, providing better perspective of disease status for more informed treatment options. In this setting, pan-cancer ctDNA testing can aid on the early detection of metachronous cancers.

The low melanoma derived ctDNA at baseline may be the result of partial disease control by the previous ipilimumab therapy, although not evident by the CT scan performed. The immediate drop and undetectable levels of BRAF mutant ctDNA during pembrolizumab treatment indicated the response of the patient’s melanoma tumour to this treatment. On the other hand, colon cancer derived ctDNA (KRAS) was also detectable at baseline, and given that it is at similar concentrations as BRAF, the patient’s melanoma and colorectal tumour burden may be relatively similar. Studies have shown that ctDNA is readily detectable in early stages colon cancer patients [29, 30] which may explain the detectability of colon cancer ctDNA in this patient. Nevertheless, we like to note that we observed fluctuations of the level of KRAS mutant ctDNA at the time of pembrolizumab treatment. Interestingly, the KEYNOTE-164 clinical trial has demonstrated durable anti-tumour activity of pembrolizumab in colorectal cancer patients, particularly those with high microsatellite instability (MSI) [31]. We hypothesise that pembrolizumab may have exerted some level of control on the colorectal tumour. However, further investigation is needed, particularly identifying the MSI status of the patient, to determine if he may have benefited from pembrolizumab treatment. The variability on detection of ctDNA across multiple cancers and tumour locations, also remains a topic of investigation in the field of liquid biopsy research.

In conclusion, ongoing close surveillance of melanoma patients who achieved complete response to BRAF inhibition and/or immune-checkpoint inhibitors is paramount to monitor potential recurrent disease. Emergent of new malignant lesions in this population should be regarded as a metastasis only after detailed evaluation, including a biopsy where feasible; otherwise there is a possibility of missing a secondary malignancy. Plasma ctDNA may aid in clarifying disease status of patients with metachronous cancer.