Prior to the development of effective cisplatin-based chemotherapy, no effective therapy was available for nonseminomatous mediastinal germ cell tumors. Local treatment modalities were ineffective because of the high percentage of patients with metastatic disease and the relative radioresistance of these tumors.

The use of intensive cisplatin-based chemotherapy developed for the treatment of advanced nonseminomatous testicular neoplasms has improved the previously dismal outlook in patients with mediastinal nonseminomatous germ cell tumors. Although overall cure rates remain lower than those achieved in the treatment of testicular cancer, a compilation of results from reported series containing 10 or more patients and using optimal chemotherapy indicates a 41% long-term survival rate (

Table 96-2). The large bulk of most mediastinal germ cell tumors at the time of diagnosis contributes to these relatively poor results. Comparable long-term survival rates of 40 to 50% have been reported when testicular germ cell tumors with far advanced, bulky metastases are treated with similar cisplatin-based regimens. However, recent reviews of survival and prognostic factors have suggested that inherent biologic differences between mediastinal and testicular germ cell tumors also play a role in determining the relatively low cure rate.

At present, treatment for all histologic subtypes of mediastinal nonseminomatous germ cell tumor should follow the same guidelines. Early reports of unfavorable outcome for patients with pure endodermal sinus tumors have not been confirmed when modern cisplatin-based chemotherapy is used. Pure mediastinal chemocarcinoma probably has a worse prognosis; however, there is currently no rationale for treating these patients differently from other patients with nonseminomatous mediastinal germ cell tumors.

The treatment for mediastinal nonseminomatous germ cell tumors should follow guidelines for poor-prognosis testicular cancer. Initial treatment with four courses of bleomycin, etoposide, and cisplatin is considered standard therapy. As in the treatment of testicular cancer, administration of chemotherapy at full doses and on schedule is important in obtaining optimal results. Following the completion of therapy, patients should be restaged with repeat serum tumor markers and CT scans of the chest and abdomen.

Subsequent management is determined by the response to initial chemotherapy. Patients with normal CT scans and tumor marker levels should receive no further therapy. Approximately 20% of these patients subsequently relapse, with almost all relapses occurring during the first 2 years after completion of therapy. Standard follow-up of these patients includes monthly physical examination, chest radiography, and serum tumor marker determinations during the first year, and similar evaluations every 2 months during the second year following therapy.

Patients with persistent elevations of either HCG or α-fetoprotein following completion of initial therapy have residual active carcinoma and a very poor prognosis. Salvage regimens, such as VIP (etoposide or vinblastine, ifosfamide, and cisplatin) have been effective second-line treatments in 20 to 30% of patients with refractory testicular cancer, but results with this approach in mediastinal germ cell tumors have been very poor. In a group of 79 patients with refractory mediastinal nonseminomatous germ cell tumor reviewed retrospectively, the salvage rate was only 12%. Outcome was not substantially different in patients receiving high-dose second-line regimens versus standard-dose salvage regimens.

Surgical intervention is often necessary in patients who have residual mediastinal abnormalities and normal serum tumor markers following initial combination chemotherapy. In this setting, consideration for surgical resection should not be limited to patients with normal serum levels of HCG and α-fetoprotein. In a recent series, 10 of 18 patients with persistent marker abnormalities (usually α-fetoprotein < 100 ng/mL) had only necrotic tumor or teratoma at resection. Overall, approximately 75% of patients have either nonviable tumor or benign teratoma at the time of resection. Patients with a large component of teratocarcinoma in the original biopsy are more likely to have residual benign teratoma. Unresected benign teratoma can cause further problems, either by slow local growth or by subsequent malignant degeneration; resection is, therefore, therapeutic. Because resection of necrotic tumor or fibrosis is not therapeutic, the proper timing of surgery following chemotherapy has been debated. In some patients, residual necrotic tumor continues to decrease in size for several months after completion of therapy. Follow-up with serial scans in such patients sometimes results in the avoidance of a major operative procedure. It is reasonable to delay resection for 2 to 3 months in patients with partial radiographic response, as long as serial tumor shrinkage is observed on follow-up radiographs. Tumors that fail to decrease in size should be resected.

Patients with no viable tumor found at the time of surgical resection have the same low risk for subsequent relapse as do patients achieving complete remission with chemotherapy alone. These patients should be monitored without further therapy. Patients with resected teratoma have an intermediate risk for subsequent relapse, while the risk of relapse is high (> 60%) if residual carcinoma is resected. According to guidelines established for patients with testicular carcinoma, patients with residual carcinoma resected should receive two additional courses of cisplatin-based chemotherapy.

Although the prognosis for patients with nonseminomatous mediastinal germ cell tumors has been markedly improved by the development of cisplatin-containing chemotherapy regimens, approximately half of these patients still die of uncontrolled tumor. Patients who do not achieve complete remission with primary therapy have extremely poor prognoses and are candidates for investigational approaches. Early use of high-dose therapy is being investigated. Paclitaxel and gemcitabine are active agents in patients whose cancer is refractory to cisplatin and etoposide, and may improve the efficacy of salvage regimens. Further improvements in therapy will probably parallel the development of increasingly effective treatment for patients with poor-prognosis testicular germ cell neoplasms.