Germ cell tumours comprise a heterogeneous group of neoplasms, which can be found at different, although restricted anatomical locations. In the testis three groups of germ cell-derived tumours are distinguished:

• I- teratomas and yolk sac tumours of infants;
• II- seminomas and nonseminomas of adolescents and adults;
• III- spermatocytic seminomas of the elderly.

These groups are defined by epidemiological characteristics, histological composition and chromosomal constitution (Table 1). Designation of tumours to these groups is clinically relevant because they require different strategies for treatment

That the different types of germ cell tumours of the testis are derived from cells belonging to the germ cell lineage, is established, although the actual non-malignant counterparts are still a matter of debate. It is likely that the teratomas and yolk sac tumours of infants originate from an embryonic germ cell, while this is a spermatogonial/spermatocyte-type of cell for spermatocytic seminomas. In contrast, it is established that the precursor of seminomas and nonseminomas is carcinoma in situ (CIS), also referred to as intratubular germ cell neoplasia undifferentiated (IGCNU) or testicular intratubular neoplasia (TIN). CIS is composed of tumour cells located on the basal membrane at the inner side of the seminiferous tubules, under the tight junction, where normally the spermatogonia reside. It has been suggested that the normal counterpart of CIS, i.e., a primordial germ cell/gonocyte is present within the gonad around the 7th to 10th week gestational age. This is supported by the epidemiological finding that the incidence of seminomas and nonseminomas show a lower incidence in cohorts of men born during the period of the second word war in Denmark, Norway and Sweden. An alternative model has been proposed, in which the cell of origin is a pachytene spermatocyte.

During the first few years of life, the only types of germ cell tumours diagnosed in the testis are teratomas and yolk sac tumours. They are evidently unrelated to puberty. In contrast, the seminomas, nonseminomas, and spermatocytic seminomas are clinically manifest during or after puberty, therefore likely related to sexual maturation. Spermatocytic seminomas are predominantly found in patients of 50 years and older. While most patients with a seminoma present in their 4th decade of life, this is in the 3rd decade for patients with a nonseminoma. An increasing incidence (in between 6-11/100.000) has been reported both for seminomas and nonseminomas during the last decades in white populations throughout the world, with an annual increase of 3-6%. Although in general rare, accounting for 1-2% of all malignancies in males, seminomas and nonseminomas are the most common cancer in young Caucasian males. In some European countries, i.e., Denmark and Switzerland, the life time risk for seminoma or nonseminoma is up to 1%. However, the increase seems to stabilise to date. In contrast to whites, blacks have a significantly lower, not increasing, incidence for seminomas/nonseminomas, although histology and age-distribution are the same. No significantly increasing incidence has been reported for teratomas and yolk sac tumours of infants and spermatocytic seminomas.
The incidence of CIS, the precursor of both seminomas and nonseminomas, in the general population is similar to the life time risk to develop a seminoma/nonseminoma. This indicates that CIS will always progress to invasiveness. About 5% of patients with a unilateral seminoma or nonseminoma have contralateral CIS.

CIS cells show similarities to embryonic germ cells, like their positivity for alkaline phosphatase, the stem cell factor receptor (c-KIT), and their glycogen content. These cells are frequently found in the adjacent parenchyma of an invasive seminoma and nonseminoma, of which a representative example is given in Figure 1. Histologically and immunohistochemically, seminoma cells mimick CIS. Lymphocytic infiltrations in the supportive stroma are a consistent feature of these tumors (Figure 2). So far, no differences have been found between CIS and seminoma cell, except the invasive growth of the latter. In contrast to the homogeneity of CIS and seminomas, nonseminomas can be composed of different elements, including embryonal carcinoma (the undifferentiated, stem cell, component), teratoma (the somatically differentiated component), yolk sac tumour and choriocarcinoma (the components of extra-embryonal differentiation) (see Figure 3-6). These different histological elements can be identified using immunohistochemistry for different markers, like CD30 for embryonal carcinoma, alpha fetoprotein (AFP) for yolk sac tumour, and human chorionic gonadotropin (hCG) for choriocarcinoma (see illustrations). Most nonseminomas are mixtures of these different elements. About 50% of germ cell tumors of adolescents and adults are pure seminomas, and 40% pure or mixed nonseminomas. Tumours containing both a seminoma and a nonseminoma component are classified as combined tumours according to the British Classification system, and as nonseminomas according to the World Health Organisation (WHO) classification. These tumours present at an age in between that of pure seminoma and nonseminoma.
The spermatocytic seminomas are histologically uniform and composed of three cell types, small, intermediate and large cells, that are evenly distributed (Figure 7). The immunohistochemical markers for CIS/seminoma are overall negative in spermatocytic seminomas. So far, no specific markers have been reported for spermatocytic seminomas, Histologically and immunohistochemically, the teratoma and yolk sac tumour components found in the infantile testis are indistinguishable from those elements found in nonseminomas of the adult testis. However, they differ in chromosomal constitution (see Table 1 and below), and the first lack CIS in the adjacent parenchyma.

In spite of the fact that it is generally accepted that CIS is the precursor for seminoma and nonseminoma, the relationship(s) between these histological elements is still a matter of debate. It has been shown, especially using cell lines derived from nonseminomas, that embryonal carcinoma is the undifferentiated stem cell of all differentiated nonseminomatous components. So far, no cell lines for seminoma or CIS are available. Nonseminomas mimick embryonal development to a certain level. However, it is unproven so far whether seminomas may also progress into nonseminoma, although various observations, both biological and clinical, may support this model. It has been suggested that CIS present in the adjacent parenchyma of an invasive seminoma or nonseminoma is only one step behind in the progression of the cancer, which is supported by molecular findings.

The teratomas of infants, and the spermatocytic seminomas are generally benign. Therefore, orchidectomy alone is mostly curative. However, spermatocytic seminomas may progress to sarcoma, a highly malignant tumour. When the yolk sac tumour component of infants is metastatic, it can be cured in the majority of patients using chemotherapy. Seminomas are highly sensitive to irradiation, while nonseminomas are overall highly sensitive to cisplatin-based chemotherapy, with cure rates of up to 90%. Criteria have been developed to distinguish nonseminoma patients with a good, intermediate and poor response (Table 2). Although these parameters are not informative on an individual basis, they separate the three groups as a whole. Seminoma patients always fall in the good and intermediate prognostic group.
Stage I disease might be treated by orchidectomy followed by a wait and see strategy. Alternatively, retroperitoneal lymph node dissection (nerve sparing) and/or irradiation (in case of pure seminoma) can be performed. Moreover, a single dose cisplatin-based chemotherapy is tested in an experimental set up. These issues are of interest, because the risk of occult metastases in clinically stage I nonseminomas is about 30%. Established factors predicting metastastic disease are lymphovascular space invasion and percentage of embryonal carcinoma. For nonseminomas there is no consensus on the best method to define the risk of occult metastases and on how the information can be used for the clinical management of patients. In clinically stage I seminoma patients occult metastases are predicted by vascular invasion and tumor size. More recently, the mean nuclear volume has been reported to be an informative parameter. Obviously these parameters could serve to define a group of patients that could benefit from surveillance. Patients with refractory disease might benefit from high-dose chemotherapy.
Because CIS is formed during intra-uterine growth, and the treatable cancer in most cases becomes clinically manifest after puberty, methods for early diagnosis and treatment might prevent progression of CIS to an invasive seminoma or nonseminoma, thereby preventing possible progression to refractory disease. A number of putative parameters have been reported, although none of them have been tested in a clinical setting thus far. Moreover, it has been shown that CIS can be effectively eradicated using local irradiation, with limited side effects.
The presence of CIS in the contralateral testis in 5% of patients with a seminoma or nonseminomas has led to the routine of a contralateral biopsy in some countries. However, in most countries the clinicians prefer a closely wait and see strategy. Patients with cryptorchidism, atrophic testis, or prior infertility have a higher risk of CIS in the contralateral testis. The exact numbers are unknown, but it is estimated that high-risk patients comprise 40-50% of the population with CIS. Altogether about 50-60% of patients with a unilateral testis tumor will have no other risk factors for CIS.

The isochromosome 12p can be identified on interphase nuclei by fluorescent in situ hybridization, using simultaneously a probe specific for the centromeric region and the short am of chromosome 12. The use of the centromeric probe only was not found to be informative.

Familial predisposition About 2% of the patients with a seminoma or nonseminoma have an affected family member, indicating a genetic component in the development of this cancer. So far, two genome-wide linkage analyses have been performed. The first showed linkage to regions of chromosome 1, 4, 5, 14 and 18, while the second found no evidence for linkage to chromosome 1, and a weaker indication for involvement of region 2 of chromosome 4 (4cen-q13). For the other region on chromosome 4 (p14-p13), and for chromosome 5, similar results were obtained. Both studies indicated linkage to chromosome 18. The latter study found linkage to the short arm of chromosome 2, and the telomeric region of 3q. A telomeric region of the long arm of chromosome 12 showed linkage when the results of both studies were combined, while no linkage was found in the separate studies. An other finding of interest is the fact that bilateral occurrence of the tumor is more frequent in familial than in sporadic cases (15 versus 5%). Indeed, most recently linkage to Xq27 has been found for cryptorchidism and bilateral germ cell tumors, although the gene involved is still unknown.

In spite of several suggestions about a possible role of a number of genes in the development of teratomas and yolk sac tumours of the infantile testis, the seminomas and nonseminomas and the spermatocytic seminomas, actual prove for their involvement is missing so far.