Burden of extragonadal germ cell tumours in Europe and the United States

Dear Sirs,

We read with interest the recent publication by Trama et al. about the burden of testicular, paratesticular and extragonadal germ cell tumours in Europe in this journal(1). We did comparable analyses based on data from the Surveillance, Epidemiology and End Results (SEER) Program, original nine registries for the years 1973-2007(2). Many statistical figures on extragonadal germ cell tumours (EGCTs) are similar in the U.S. and Europe. For example, we also observed that mediastinal EGCTs had the worst 5-year relative survival (SEER: 58%, Europe: 53%). However, the estimated age-standardized incidence rate of EGCTs is considerably higher in the U.S. than Europe. To quantify this difference, we calculated the ratio of age-standardized incidence rates (SEER estimate, whites, 1997-2007 / European estimate 1995-2002) and corresponding 95% confidence intervals for males and females(3). Among males, the EGCT rate in the U.S. is 1.82 (95%CI: 1.58-2.10) times higher than in Europe. Among females, the EGCT rate in the U.S. is 2.46(95%CI: 1.94-3.12) times higher than in Europe. As rate estimates were quite precise in both regions, random error is an unlikely explanation for this finding.

In contrast to Trama et al., we included EGCTs of the placenta (all nondysgerminoma). 312 out of 567 EGCTs among white females (55%) during the period 1973-2007 originated from the placenta. When we exclude placental EGCTs, the age-standardized incidence of EGCT among white females forthe period 1997-2007 decreases to 1.1 per million person-years (standard error [SE]: 0.10) resulting in a US:Europe ratio of 1.59 (95%CI: 1.25-2.02). We provide some hypotheses that might explain these differences. First, EGCTs are only rarely registered and can be easily miscoded as gonadal germ cell tumours if the topography coding of tumours is not intensively monitored. In contrast to the SEER program, the European analysis was based on 64 cancer registries located throughout Europe with varying degrees of quality control and varying degrees of completeness of registration(4).Second, EGCT cancer registry reports that contain only unspecifiedhistology (e.g. carcinoma, not otherwise specified) cannotbe included in the data analysis of EGCTs. A comparison of the number of extragondal tumours with unspecifiedhistologies at typical sites of EGCTs (e.g. mediastinum, pineal gland, retroperitoneum, brain) might give clues to this hypothesis. Finally, different prevalences of the, until now, unknown risk factors of EGCTs in the U.S. and Europe may play a role.

Interestingly, in a recent unpublished pooling project of 9 population-based cancer registries in Germany of the years 1998-2008 which included 362,450,458 person-years, we estimated an age-standardized (European standard population) incidence rate for male EGCTs, female EGCTs including placental EGCTs, and female EGCTs excluding placental EGCTs of 1.79 (SE 0.1), 1.14 (SE 0.09), 0.79 (SE 0.08), respectively, which is very much in line with European-wide estimates.

Table.

Comparison of Incidence Rates of Extragonadal Germ Cell Tumours in the U.S. and Europe

	SEER 1997-2007 (Whites)	Europe 1995-2002	US:Europe Rate Ratio (95% confidence interval)
Male cases (N)	384	n.r.
Incidence rate per million/year (standard error)	3.4 (0.17)	1.87 (0.07)	1.82 (1.58-2.10)
Female Cases (N)	185	n.r.
Incidence rate per million/year (standard error)	1.7 (0.13)	0.69 (0.04)	2.46 (1.94-3.12)

Legend: n.r.: not reported; according to Trama et al., the overall number (males and females) of extragonadal germ cell tumors was 1,019; all rates are standardized to the European standard population.