Dramatic improvements in treatment for germ cell tumors occurred in the late 1970s with the introduction of cisplatin-based combination chemotherapy. For the first time, cure was achievable for the majority of patients treated for metastatic disease. 1 As a medical oncology trainee in the mid-1980s, I was taught the prevailing dogma that relapses virtually never occurred more than 2 years after completion of this chemotherapy. However, as early as 1983, reports began to emerge of “late” relapses after cisplatin-based combination chemotherapy. 2 By the mid-1990s, investigators at Indiana University reported 81 cases of late relapse, including 34 patients experiencing relapse 2 to 5 years after treatment and 47 patients experiencing relapse more than 5 years after treatment. 3 The authors concluded that late relapse was associated with relative resistance to chemotherapy, and surgery was recommended as the preferred mode of therapy. There have been multiple subsequent publications reporting experience with late relapse, mostly single-institution series from large referral centers. Interpretation of these reports is difficult because of variability of the cases considered with respect to prior therapy, histologic group, stage at initial diagnosis, primary site, and inclusion of cases with prior relapse. A recent comprehensive review of published literature concluded that the risk of late relapse was 3.2% in nonseminoma and 1.2% in seminoma. 4 The accepted definition of late relapse was relapse beyond 2 years after completing therapy or after achieving complete response, in the absence of a new primary tumor. The most common site of relapse was the retroperitoneum in both seminoma and nonseminoma. In this issue ofJournalofClinicalOncology, Sharp et al 5 report the single-institution experience of late relapse at Memorial SloanKettering Cancer Center (New York, NY), analyzing prognostic factors for cancer-specific survival (CSS). As with many prior series, they have included chemotherapy-naive patients experiencing relapse after orchiectomy followed by surveillance, retroperitoneal lymph node dissection, or radiation therapy. They report 75 cases of late relapse, comprising 70 patients with nonseminoma and five patients with pure seminoma. The time to relapse was 2 to 5 years in 28 patients (38%), 5 to 10 years in 19 patients (26%), and more than 10 years in 27 patients (36%). Eighteen patients (24%) had not received any chemotherapy at the time of relapse. The 5-year CSS was 93% in chemotherapy-naive patients but only 49% in those who had received prior chemotherapy. Complete resection of disease at relapse was associated with improved CSS. In multivariate analyses, considering only those who had received prior chemotherapy, the presence of symptoms and multifocal disease at relapse were each associated with an inferior CSS. The authors state “complete surgical resection is central to the management of late relapse” but qualify this elsewhere with the statement “we believe that the definition of late relapse should be limited to patients