Purpose: In cases of multiple non-small cell lung cancer, clinicians must decide whether patients have independent tumors or metastases and tailor treatment accordingly. Decisions are currently made using the Martini and Melamed criteria, which are mostly based on tumor location and histologic type. New genomic tools could improve the ability to assess tumor clonality.
Experimental design: We obtained fresh-frozen tumors specimens from patients who underwent surgery on at least two occasions for presumptively independent NSCLC. We did array comparative genomic hybridization (aCGH), mutational profiling of select genes, and detailed clinicopathologic review.
Results: We analyzed a total of 42 tumors from 20 patients (6 patients with synchronous tumors, 14 patients with metachronous tumors, 24 potential tumor pair comparisons); 22 tumor pairs were evaluable by aCGH. Surprisingly, classification based on genomic profiling contradicted the clinicopathologic diagnosis in four (18%) of the comparisons, identifying independent primaries in one case diagnosed as metastasis and metastases in three cases diagnosed as independent primaries. Matching somatic point mutations were observed in these latter three cases. Another four tumor pairings were assigned an "equivocal" result based on aCGH; however, matching somatic point mutations were also found in these tumor pairs. None of the tumor pairs deemed independent primaries by aCGH harbored matching mutations.
Conclusion: Genomic analysis can help distinguish clonal tumors from independent primaries. The development of rapid, inexpensive, and reliable molecular tools may allow for refinement of clinicopathologic criteria currently used in this setting.