Patients with squamous cell carcinoma (SqCa) arising in the head and neck (H/N) commonly develop solitary pulmonary metastases that mimic the clinical, radiographic, and pathologic presentation of new primary lung SqCa. Primary pulmonary and metastatic SqCas cannot be differentiated from each other histologically. However, distinguishing multiple independent primary neoplasms from a primary H/N SqCa with pulmonary metastasis has prognostic significance due to its impact on tumor stage, the most important determinant of prognosis. Since genomic instability is a common feature of cancer, we hypothesized that independently-arising neoplasms in an individual patient would exhibit measurable genomic variation, enabling discrimination of tumor lineage and relatedness. In this study, we describe a molecular approach for analysis of genetic variation among multiple tumors from a single patient that does not rely on collection of normal tissue, and which can be performed with minimal tumor samples. Genomic DNA from H/N and lung SqCas from individual patients were analyzed by microsatellite PCR to identify discordant allelic variation. This method is rapid, sensitive, does not require constitutional DNA for comparison, and can be applied to the analysis of archival tumor DNA. Our results demonstrate that microsatellite PCR can identify discordant genetic variation among multiple tumors from a single patient, facilitating the molecular discrimination of metachronous primary SqCa versus solitary pulmonary metastasis from a H/N primary SqCa.