Genomic Profiling of Cancers of Unknown Primary Site

Genomic Profiling of Cancers of Unknown Primary Site: The Next Steps To the Editor A recent article by Ross et al1 reported the results of comprehensive genomic profiling of cancers of unknown primary (CUPs) and documents frequent occurrence of clinically relevant genomic alterations. We describe herein the outcome of a case of CUP, which may serve to illustrate both the benefit of personalized therapy and potential pitfalls. A man in his 60s presented with CUP involving bones and lymph nodes. He had a history of stage II testicular seminoma, treated with orchiectomy and radiation therapy, and this disease had been in remission for 5 years. Bone pain led to the diagnosis of multiple lytic bone metastases. Excisional biopsy of a supraclavicular node demonstrated poorly differentiated adenocarcinoma. Immunohistochemistry and gene expression profiling and positron emission tomography were inconclusive for the primary site. The patient was empirically treated with carboplatin and paclitaxel with a good clinical and radiographic response. After initiation of chemotherapy, results of tumor genomic profiling became available. ErbB2 amplification and a BRCA1 mutation were found. Therefore, trastuzumab was added to the last 2 cycles of chemotherapy and continued as maintenance therapy. After completion of chemotherapy, lapatinib was prescribed in addition to trastuzumab. Two years after diagnosis of CUP, the patient remained free of disease progression or symptoms. Although this patient’s disease responded well to chemotherapy, durable remission after chemotherapy is rare in CUP. Bone metastasis is associated with poor survival in CUP.2 Therefore, it is tempting to attribute the unusually long progressionfree interval to ErbB2-directed maintenance therapy with trastuzumab and lapatinib. The tumor also harbored a mutant BRCA1 allele, which was independently detected by our institutional research study of tumor genome sequencing (UNCSeq: NCT01457196).3 The patient was referred to genetic counseling, and standard genetic testing confirmed him to be a germline carrier of a common BRCA1 founder mutation, 5385insC (formerly 5382insC).4 The wild-type allele of BRCA1 was retained in the tumor, and it is unclear whether the tumor is predisposed to respond to therapy targeting BRCA1 loss of function. Tumor profiling led to the diagnosis of his germline BRCA1 mutation, which raised serious implications for his children’s risk of cancer and their decisions regarding genetic testing. Tumor genomic profiling may yield additional treatment options with long-term clinical benefit. However, clinicians and patients should be made aware that there may be unanticipated consequences for patients and their families from detection of germline variants.