tumours are the most common cancers in males between the ages of 15 and 34, with a world-wide incidence of about five per 100,000. These tumours are highly curable, with long term remission being achieved in up to 90% of patients following treatment with surgery, radiation and/or chemotherapy. The cause of disease is unknown, but there may be a genetic element, as family clusters occur. Cryptorchidism and Klinefelter's syndrome are also known predisposing factors. Testicular intraepithelial neoplasia is a precursor of invasive germ cell tumour, preceding development of invasive cancer by about five years.
Survival and prognosis are both highly dependent on TNM stage, number and extension of visceral metastases (liver, bone, pulmonary, brain), and the initial concentrations of serum tumour markers [a -fetoprotein (AFP), human chorionic gonadotrophin (hCG) and its b -subunit (hCGb ), lactate dehydrogenase (LDH), and placental alkaline phosphatase (PLAP)]. Pre-treatment serum concentrations of these markers all influence the choice of therapy.
About 95% of malignant testicular tumours originate in primordial germ cells with rare extragonadal primary sites (retroperitoneal, mediastinal, pineal) (1). Germ cell tumours are classified as seminomas (40%), non-seminomatous tumours (NSGCT, 40%), or "mixed" germ cell tumours (20%) which contain elements of both seminomatous and non-seminomatous tumours. Most NSGCT contain multiple cell types, while seminomas are most likely to produce a uniform population of cells. NSGCT may be composed of embryonal carcinoma, teratoma, choriocarcinoma and / or yolk sac carcinoma (endodermal sinus tumors) cell types.
The most useful tumour markers in germ cell tumours are AFP, hCG and hCGb , PLAP and LDH [See Table 1, pp 2790].
AFP has an upper reference limit of approximately 15 µg/L (~10 kU/L) after the first year of life, and may be elevated in benign liver and certain malignant diseases (2). Raised levels are seen most frequently in germ cell tumours and hepatocellular carcinoma, but also in gastric, colon, biliary, pancreatic and lung cancers (~20% of patients). The serum half-life of AFP is approximately 4-5 d after orchidectomy (3).
HCG is normally synthesised in syncytiotrophoblasts of the placenta, and is also produced in malignancy, in trophoblastic structures or syncytiotrophoblastic giant cells of germ cell tumours (4,5). Some non-trophoblastic tumours synthesise hCG. Very high serum hCG concentrations occur in hydatidiform moles ("molar" pregnancies) and choriocarcinomas (97%). Serum hCG levels may be increased in pancreatic adenocarcinomas and islet cell tumours, tumours of the small and large bowel, hepatoma, stomach, lung, ovarian, breast and renal cancer. For tumour marker use, measurement of both intact hCG and its free b -subunit (hCGb ) is highly desirable, as some tumours may produce only hCGb . The normal range of serum hCG for men and pre-menopausal women is up to ~5 U/L, and for post-menopausal women up to ~10 U/L. The serum half-life of intact hCG is 16-24 h.
Placental alkaline phosphatase (PLAP) is a heat-stable isoenzyme of alkaline phosphatase that is normally expressed by placental syncytiotrophoblasts. PLAP or PLAP-like activity occurs in normal tissue (e.g. testis, cervix, thymus, lung) as well as malignant tissue (e.g. germ cell, ovarian and lung tumours). Raised serum concentrations of PLAP are found in seminomas (sensitivity 51-90%) and NSGCT (sensitivity 20-36%), as well as in ovarian tumours (6). PLAP may also provide a marker for testicular intraepithelial neoplasia. In cigarette smokers, serum PLAP concentrations may be increased up to 10 times the upper reference limit (100 U/L), with considerable inter-individual variation. Consequently, serum PLAP should not be measured in smokers. The normal half-life of serum PLAP after surgical resection is between 0.6 and 2.8 days. Prolonged half-lives correlate with advanced disease. PLAP provides an excellent marker for monitoring response to chemotherapy and can provide early evidence of disease progression in non-smoking seminoma patients (7).
LDH (MW 134 kD) is an enzyme consisting of five isoforms (each with four subunits) that is expressed in cardiac and skeletal muscle as well as in other organs. The LDH-1 isoenzyme is most often found in testicular germ cell tumours (8,9). However, high levels can also occur in a number of benign conditions including skeletal muscle disease, myocardial infarction, pernicious anaemia, leukaemia, thalassaemia, and pulmonary embolism. Concentrations may also be increased artifactually (e.g. following in vitro haemolysis of blood specimens) (5). Total LDH appears to be less specific than AFP and hCG but nevertheless can be of independent prognostic value in patients with advanced germ cell tumours (sensitivity ~60% for NSGCT; ~80% for seminomas).
Relevant imaging procedures include testicular and abdominal ultrasound, chest X-ray, and (if metastasis is suspected) chest, abdominal and brain computerised tomography (CT) as well as bone scintigraphy. For final diagnosis, histological investigation of biopsy specimens are required.
There are no recommended biochemical screening tests for germ cell tumours. However, regular testicular self-examination (palpation) is recommended for all males, from puberty up to the age of about 45 years, and is particularly important for patients at higher risk of germ cell tumours (e.g. patients with cryptorchidism or Klinefelter's syndrome)
On clinical suspicion of testicular germ cell tumour (e.g. based on testicular palpation and ultrasound), recommended diagnostic procedures include chest X-ray, abdominal ultrasound, and measurement of serum AFP, hCG, PLAP and LDH concentrations. Pre-operative tumour marker determinations can help to confirm diagnosis, both of gonadal and extragonadal (e.g. mediastinal, retroperitoneal, central nervous system) germ cell tumours. [Tumour marker measurements, together with testicular ultrasound, can also help in the differential diagnosis of epididymitis in patients with painless swelling of one testis.]
In clinical Stage I disease following primary surgery, a second marker determination performed 5-6 days post-operatively allows determination of marker half-life. Stage I classification can thus be confirmed retrospectively, if marker concentrations decline according to half-life. For patients undergoing chemotherapy, marker determinations are mandatory prior to each cycle. In addition, following resection, thorough histological analysis of resected tumour specimens according to WHO criteria is obligatory (1).
Elevations of AFP (>10 kU/L) or hCG (>5 U/L) may be encountered in 80% of metastatic and 57% of Stage I NSGCT. Serum PLAP is raised in up to 80% of testicular seminomas (Stage I and metastatic) and hCG in fewer than 20% of these. Elevated serum AFP levels indicate the presence of yolk sac elements, i.e. mixed germ cell tumours (especially embryonal carcinoma and yolk sac tumours), and occur in all stages of disease [sensitivity 50-80% in metastatic disease, 70-72% in undifferentiated malignant teratoma, 60-64% in intermediate malignant teratoma, and 64% of patients with yolk sac or combination tumours]. AFP concentrations >1000 (850 kU/L) m g/L are found in 53% of undifferentiated malignant teratomas, 16% of intermediate malignant teratomas, and 26% of combination tumours (5).
Increased serum HCG concentrations occur in both seminoma and NSGCT, with a diagnostic sensitivity of 40-60% in patients with metastatic NSGCT and 15-20% in those with metastatic seminoma. hCG elevations are predominantly found in patients with tumours containing choriocarcinomatous components, syncytiotrophoblastic giant cells and (more rarely) special round cells also found in pure seminomas.
Trophoblastically differentiated teratomas usually produce hCG, while differentiated teratomas and yolk sac tumours rarely do. In choriocarcinoma, tumour mass correlates reasonably well with the degree of elevation of serum hCG values, which may be as high as several million U/L in advanced disease. A serum hCG concentration of 10 U/L corresponds to approximately 106 tumour cells (4).
Serum hCG is elevated in 55-60% of cases of undifferentiated and intermediate teratomas, 90% of concentrations ranging from 5 to 1000 U/L. The frequency, but not the extent of elevation of serum hCG, is stage-dependent in NSGCT (Stage I, ~45%; Stage II ~55%; Stage III ~85%) (6). Approximately 18% of seminomas have elevated serum hCG (10-2000 U/L) and/or hCGb concentrations. Highly elevated hCG values (>5000 U/L) are suggestive of "mixed" germ cell tumours [see above], and will influence choice of therapy.
Prior to 1997, clinical and pathological staging of germ cell tumours was dependent only on the extent of disease, according to the TNM system, requiring orchiectomy for the staging of primary tumour and radiographic assessment of chest, abdomen and pelvis to determine nodal and metastatic classification. Incorporation of pre-treatment determinations of serum AFP, hCG and LDH concentrations was recommended following a multi-center study of prognostic variables, which included data from more than 5,000 germ cell tumour patients and which was co-ordinated by the International Germ Cell Tumour Collaborative Group (10). These tumour markers are now included universally in the international germ cell tumour staging system (TNM + S0/1/2/3- category) and should be determined before and immediately after orchiectomy. If markers are raised, serial determinations should be made to allow calculation of half-life and to assess whether markers fall to within reference limits.
Staging errors may be reduced from 50% to less than 15% in Stages I and II by AFP and hCG determination (11). In addition, cerebrospinal fluid determinations of AFP and hCG may be helpful in the diagnosis and monitoring of intracranial GCT.
The most appropriate use of tumour markers will depend on disease stage, as outlined below (12).
Stage IA and IB disease. Surveillance alone is suggested [rather than retroperitoneal lymph node dissection (RPNLD)] following inguinal orchiectomy. Together with chest X-ray and clinical examination, routine measurements of tumour markers should be made, monthly during the first year post-orchiectomy, and then every second month during the second and third years. Abdominal CT scans are desirable every two to three months throughout the first three years.
If AFP or hCG remain elevated and half-lives prolonged, with no evidence of residual disease on CT, this is highly suggestive of occult metastases distant from the retroperitoneum, and systemic chemotherapy (rather than RPLND) should be considered.
Stage II disease. Following RPLND and either (a) observation followed by three or four cycles of chemotherapy at relapse, or (b) two cycles of adjuvant chemotherapy immediately following RPLND, surveillance should include tumour marker determinations with physical examination and chest X-ray every month during the first year, every second month during the second year, and every third month for the third year.
Advanced Stage II and III disease. The rate of tumour marker decline following chemotherapy predicts response to treatment. Persistently elevated marker levels or prolonged tumour marker half-lives in the first six weeks post-chemotherapy specifically indicate resistance to chemotherapy and poor prognosis. Patients with residual masses following chemotherapy may be considered for post-chemotherapy surgery, but where serum tumour markers are still elevated, salvage chemotherapy should be recommended instead, since disease is likely to be surgically unresectable.
NSGCT findings indicating poor prognosis include metastases involving liver, bone and brain, AFP >1000 kU/L and/or hCG concentration >10,000 U/L, mediastinal tumour mass >5 cm in diameter, and 20 or more pulmonary lung metastases (13). An International Germ Cell Tumour Collaborative Group has proposed use of a prognostic factor-based staging system for metastatic germ cell tumours (both seminomatous and non-seminomatous) (10). This allows classification of these tumours into three groups - good, intermediate and poor prognosis - and incorporates information from tumour marker measurements, as outlined in Table 1. The system also takes into account tumour site (testis, retroperitoneal, mediastinal) and the presence or absence of non-pulmonary visceral metastases, as described in detail elsewhere (10).
TABLE 1. Contribution of serum tumour marker measurements to the prognostic classification of metastatic germ cell tumours
|Tumour marker concentration|
|Prognostic group1||AFP (kU/L)||hCG (U/L)||LDH (multiple of RR)2|
|Good (S1)||<1000||<5000||<1.5 x (RR)|
|Intermediate (S2)||³ 1000 & £ 10,000||³ 5000 & £ 50,000||³ 1.5 x (RR) & £ 10 x (RR)|
|Poor (S3)||>10,000||>50,000||>10 x (RR)|
1S, serum tumour marker
2 LDH concentrations expressed as multiples of the upper limit of the reference range (RR)
For these three prognostic groups, progression-free and overall survival rates are respectively 89% and 92% for S1, 75% and 80% for S2, and 41% and 48% for S3 (10).
In addition, tumour markers can also provide prognostic information by contributing to the prediction of histological outcome following first-line chemotherapy. For example, the probability of necrosis in residual tumour is highest if there are no teratomatous elements in the primary tumour, if AFP and hCG levels are within their reference ranges pre-chemotherapy and only LDH is elevated, if the pre- or post-chemotherapy tumour mass is small, and if a large shrinkage of tumour mass occurs during chemotherapy (14).
Determination of the half-lives of AFP and hCG is recommended for monitoring treatment, normalisation of both markers (AFP within 5 days, hCG within 1-2 days) indicating favourable prognosis. [Half-lives may be calculated by linear regression, where y = a + bx, where x is time in days and y is the logarithm of marker concentration. If only two measurements are available these can be used to estimate marker half-life, but the linear regression approach is to be preferred. For further details of half-life calculations see reference (15).] After two cycles of chemotherapy, patients for whom half-lives are more than seven days for AFP and/or more than three days for hCG have significantly lower survival rates than those with normal tumour marker half-lives (16). Further improvement can be achieved by combining prognostic criteria of the United Kingdom Medical Research Council (MRC) with analysis of marker half-lives, resulting in three different risk groups (17). The importance of tumour marker measurements in this classification is outlined in simplified form in Table 2. Patients in the "poor risk" category may then be selected for more aggressive chemotherapy (16).
TABLE 2. Prognostic classification for metastatic germ cell tumours on the basis of UK MRC prognostic criteria and tumour marker measurements (17).
|Classification according to||Overall 10-year survival rates|
|Prognostic group||UK MRC prognostic criteria||Tumour marker half-lives|
|Good||Good risk||Normal half-lives ("good")||96%|
|Intermediate||Good risk||Prolonged half-lives ("bad")||64%|
|Poor risk||Normal half-lives ("good")|
|Poor||Poor risk||Prolonged half-lives ("bad")||28%|
Since NSGCTs may contain different cell types, it is necessary to monitor serum levels of AFP, hCG and LDH. With progression, germ cell tumours may initially produce a slow increase in marker levels. This increase may later become exponential with rapid progression. [However, false negative marker values may be present in 10-33% of NSGCT before orchiectomy.] In tumours with mixed cell types, decreases in AFP or hCG concentrations following resection reflect the degree to which the whole tumour has been resected while a decrease following chemotherapy only reflects the behaviour of the marker-positive cell type.
Following complete tumour resection, AFP levels should decrease into the reference range with a half-life of <5 days, hCG levels with a half-life of 1-2 days and PLAP with a half-life of <3 days. Twenty percent of Stage I seminomas and 30% of Stage I NSGCT will relapse following surgery alone, and close surveillance is therefore required. In follow-up, normal markers indicate absence of disease and rising levels, sustained plateau or slower clearance suggest residual active disease. Persistently raised hCG or AFP levels in non-seminomas and persistently raised hCG or PLAP levels in seminomas after orchiectomy indicate that the tumour was not restricted to the testis, following first line treatment. There is general agreement that rising concentrations of tumour markers are incompatible with tumour regression and often indicate progressive disease months before clinical evidence of recurrence (lead-time 1-6 months) (2). Increased or rising levels of either marker without radiographic or clinical findings imply active disease. These increases are regarded as sufficient cause to initiate treatment, provided known causes of false-positives and the possibility of tumour in the contralateral testis have been excluded. In follow-up of metastatic tumour, rising AFP and/or HCG levels provide the first indicator of relapse in about 50% of patients. Combined determination of AFP (cut-off 10 kU/L) and hCG (cut-off 5 U/L) yielded a diagnostic sensitivity of 86% for tumour recurrence and partial response in conjunction with 100% diagnostic specificity and positive/negative predictive values of 100% and 87%, respectively (18).
However, tumour marker levels within the reference range do not always exclude progression. Recurrent tumour may exhibit different biological behaviour, such that an initially marker positive tumour may become marker negative. In very low volume disease, or where viable metastases or mature teratomas do not produce markers, false-negative results may be obtained.
Discordant behaviour (decline in serum marker concentrations while tumour burden increases) has been reported and attributed to selective chemotherapeutic destruction of marker-producing cancer cells (5). In contrast, false positive tumour marker results may also occur transiently due to tumour lysis on initiation of chemotherapy or (for AFP) due to hepatic damage (5).
Germ cell tumours offer a unique example of a firm integration of tumour markers into diagnosis, staging, prognosis and monitoring response to therapy. Elevated serum AFP levels may provide stronger evidence than histology in distinguishing pure seminomas and tumours with trophoblastic elements, a distinction that has important therapeutic consequences. Furthermore, tumour marker measurements enable more accurate assignment of clinical stage, independent of the extent of metastases. Tumour marker measurement is mandatory in assessing the response to chemotherapy. Tumour marker kinetics during the first six weeks of chemotherapy provide important prognostic information, while increasing tumour marker levels following therapy may predict relapse months before this is clinically evident. Finally, tumour marker normalisation is a pre-requisite for successful post-chemotherapy surgery.