A 67-year old male was investigated in our institution in 2012 for weight loss and abdominal pain. He was otherwise fit and well, with no significant comorbid medical history. He was not on any regular medications, had no known allergies and had no significant family history. Computed tomography (CT) revealed moderate ascites and a large splenic mass. Fine needle splenic aspirate was non-diagnostic and therefore a therapeutic/diagnostic splenectomy was performed. Metastatic melanoma was confirmed histologically, and further testing confirmed a BRAF V600R mutation via Sanger sequencing. In July 2014, he commenced dabrafenib and trametinib treatment for progressive disease but suffered unacceptable toxicity, which led to the cessation of the combined targeted therapies.
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.
Adjuvant chemotherapy for colon cancer was offered but the patient decided to continue with pembrolizumab for metastatic melanoma and declined chemotherapy. In November 2016, eleven months after curative resection of primary colorectal cancer, para-aortic nodes enlarged marginally and became intensely FDG avid on PET despite on-going pembrolizumab. Biopsy of an enlarging para-aortic node at approximately week 88 confirmed metastatic colorectal cancer. Molecular analysis of the colorectal metastasis confirmed KRAS p. G13D mutation. It is of note that the patient had no other sites of disease progression and remained in complete response from metastatic melanoma, which led to cessation of pembrolizumab treatment.
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.
ctDNA screening and monitoring
In parallel to the imaging scans, the patient was monitored for melanoma and colorectal cancer by tracking BRAF p.V600R and KRAS p.G13D mutations in ctDNA respectively. Blood samples were collected in EDTA and Streck tubes. Plasma was separated within 24 h by centrifugation at 300 g for 20 min, followed by a second centrifugation at 4700 g for 10 min, and then stored at -80 °C until extraction. Cell-free DNA (cfDNA) was extracted from 5 ml of plasma using the QIAamp Circulating Nucleic Acid Kit (Qiagen) as per the manufacturer’s instructions. Analysis of plasma ctDNA was carried out using an in-house BRAF p.V600R assays [27] and a commercial KRAS p.G13D (Bio-Rad) for droplet digital PCR (ddPCR). Protocols used for ddPCR analysis were as previously described [21, 28] and ctDNA levels were calculated based on the number of copies per millilitres of plasma (c/mL).
Plasma analysis demonstrated the presence of BRAF V600R ctDNA at baseline prior to initiating pembrolizumab, which became undetectable at subsequent follow-up (weeks 2–10.) The patient achieved sustained partial response to pembrolizumab (week 18–49) by CT and complete metabolic response by PET scan, which was supported by his corresponding ctDNA data (Fig. 1). As predicted, the patient’s blood sample at the time of colorectal cancer diagnosis (week 36) had detectable KRAS mutant ctDNA (2 c/mL). Retrospective analysis of the previous blood samples revealed detectable levels of KRAS mutant ctDNA prior to immunotherapy (3 c/mL), suggesting that colorectal cancer may have already been present at the time of stage IV melanoma diagnosis. Subsequent plasma samples (weeks 2–49) were also found to have detectable KRAS mutant ctDNA, albeit at consistently low levels that ranged from 2 to 4 c/ml, with the exception of the blood sample collected at week 64 that was negative for ctDNA. In this case, real time knowledge of detectable ctDNA following curative bowel resection, implying residual microscopic disease, and the negative BRAF mutant melanoma ctDNA may have influenced and ultimately changed the clinician and patients decision from not having adjuvant chemotherapy, to receiving it.
Increased KRAS mutant ctDNA was further observed at week 70 (6 c/mL), which provided an early indication of disease progression. Prior to the cessation of pembrolizumab corresponding to the melanoma complete response, KRAS mutant ctDNA levels was at its peak (14 c/mL). A final ctDNA assessment at week 119, revealed that BRAF V600R ctDNA continues to be undetectable which is consistent with sustained complete response of melanoma. Nevertheless, ctDNA for KRAS G13D remained high (7 c/ml) suggesting possible radiologically undetectable progression of the patient’s untreated colon cancer.
