PATIENTS AND METHODS

The charts of 186 patients with GCT in first or subsequent relapse after front-line treatment with platinum-based chemotherapy who underwent HDCT with carboplatin and VP-16 as salvage therapy at Indiana University between 1986 and 1997 were reviewed. Histology included seminoma and nonseminoma of gonadal and extragonadal origin. For each patient, the following data were obtained: (1) pre-HDCT: age, histology, primary site, serum markers, type and number of cisplatin-based chemotherapy regimens, and cisplatin refractoriness; (2) HDCT: date, response, and time and pattern of progression; (3) post-HDCT: time and type of treatment; dose, schedule, and number of chemotherapy regimens; response and duration of therapy; factors prompting discontinuation; conditions at last follow-up; and date and cause of death. Relapse or progression was documented by increasing serum tumor markers or by an increasing mass on radiographic assessment. Responses were defined as follows: complete response (CR) was defined as the absence of clinically or radiographically detectable disease, including normalization of beta human chorionic gonadotropin (bHCG) and alpha-fetoprotein (AFP) levels for at least 4 weeks. Partial response (PR) with negative markers (PRm−) was defined as more than 50% reduction in the sum of the perpendicular greatest dimensions of measurable lesions, plus normalization of all prior elevated tumor markers persisting for at least 4 weeks. PR with positive markers (PRm+) was defined as more than 50% reduction in the sum of the perpendicular greatest dimensions of measurable lesions persisting for at least 4 weeks without normalization of all prior elevated tumor markers. Serologic PR was defined as greater than one-log reduction of a previously elevated bHCG level and/or greater than 50% decrease in the AFP level persisting for at least 4 weeks. Progressive disease was defined as the appearance of new lesions and/or a persistent increase in bHCG/AFP and/or a greater than 25% increase in the size of any measurable lesion. Stable disease was defined as a status not fulfilling the requirements for CR, PR, or progressive disease. Interval was defined as the time in weeks between the last HDCT and the first post-HDCT treatment. Survival was measured from the initiation of the first post-HDCT treatment. Statistical correlation between preceding variables and response to chemotherapy and survival was analyzed by Fisher’s exact test.

RESULTS

DISCUSSION

Our finding that nearly one half of the patients who experienced relapse after HDCT received further salvage attempts with single-agent or combination chemotherapy is consistent with prior observations by Pont et al.¹⁵ It also underscores the fact that physicians, patients, and families alike are often reluctant to disengage from aggressive therapy, even when cure is usually beyond reach. However, the dismal results of standard chemotherapy in these patients should stress the urgency to both physicians and patients of finding new and innovative therapeutic options. Whenever possible, these patients should be entered onto clinical trials that are testing new and promising agents rather than recycling the older agents that have produced such poor results in this patient population. Despite multiple rounds of standard and high-dose chemotherapy, these young patients frequently have relatively preserved performance status and renal and hepatic function. With a relatively similar spectrum of single-agent and combination chemotherapy, the overall response rate observed in our study (18.2%) was not very different from that reported by Pont et al¹⁵ (approximately 24%). The responses observed in our patients, however, were of short duration and were only seen with the first post-HDCT regimen, suggesting that multiple attempts to obtain an objective response may not be warranted. Although Pont et al¹⁵ observed responses with second and third post-HDCT regimens, details regarding quality and duration of response were not provided. Therefore, it is difficult to judge the antitumor and palliative value of chemotherapy in their study. Furthermore, although disease progression was the most common reason for discontinuation of treatment in both studies, chemotherapy-induced myelosuppression was a greater problem in our patients, prompting dose reduction or discontinuation of the drug in 17%. Thus the balance between the limited chances of response and the toxicity imposed on the patient remains marginal.

Of the cytotoxic drugs that were tested in our study, only VP-16 and gemcitabine showed useful antitumor activity, whereas all other combinations, including platinum-based regimens, resulted in negligible results. The three chemotherapy-induced CRs reported in this study were obtained with these two agents. Moreover, in some patients, VP-16 and gemcitabine were instrumental in providing a window of opportunity for surgical resection of residual disease, which proved to be potentially curative in three cases. Chronic oral VP-16 has been used at Indiana University as maintenance treatment for patients at high risk of relapse or progression after salvage chemotherapy¹⁶ and as a palliative regimen for patients with cisplatin-refractory disease.¹⁷ In both settings, objective responses have been observed, including in patients who experienced disease progression during treatment with intravenous cisplatin plus VP-16. Daily oral VP-16 should be considered as one of the palliative agents of choice after HDCT failure, because of its activity and ease of administration. The schedule dependency has been previously reported.^18,19

The experience with gemcitabine in relapsed GCT is limited to two recently reported phase II trials conducted at Indiana University and in Germany in heavily pretreated patients, including several who were unsuccessfully treated with HDCT.^20,21 In both studies, the drug showed definite activity and mild toxicity. Eleven of the 12 gemcitabine-treated patients reported in this study were, in fact, enrolled in the Indiana University phase II trial. The single CR and two PRs obtained in these 12 patients compare favorably with the response rate seen in a similar population with cisplatin, ifosfamide, and paclitaxel, alone or in combination.

The single response to paclitaxel that was observed in our study is inconsistent with some encouraging phase II data of paclitaxel that were previously reported in patients with refractory GCT.^22-24 Bokemeyer et al²² and Gerl et al¹⁴ reported objective responses in three of 13 patients who experienced treatment failure with HDCT. Likewise, the retrospective study of Pont et al¹⁵ reported four responses in 24 patients treated with single-agent paclitaxel, and four additional responses were seen in five patients who received combinations of paclitaxel, ifosfamide, and cisplatin, for an overall response rate of 26.2%. Furthermore, in that study, treatment with paclitaxel proved on multivariate analysis to be one of only two independent prognostic factors for survival (the other being treatment with ifosfamide). Less favorable results with paclitaxel in a small number of patients who were unsuccessfully treated with HDCT have been observed by Motzer et al²⁵ as well as in a recent phase II study of patients with relapsed GCT conducted at Indiana University.²⁶ Considering the negligible activity seen in these studies, the salvage role of paclitaxel after progression from HDCT, as a single agent or in combination, remains to be established.

It is presently unclear why a longer interval between HDCT and post-HDCT treatment was associated with a greater likelihood of response in our patients. The same correlation between duration of the interval and response was also found by Pont et al.¹⁵ Unlike late relapses of other malignancies, such as acute leukemia, Hodgkin’s Disease, and non-Hodgkin’s lymphoma, late relapses in testicular GCT are usually associated with refractoriness to further chemotherapy, rather than chemosensitivity.²⁷ It is likely that true late relapses of GCT after primary platinum-based chemotherapy are biologically different from recurrences after HDCT, thus explaining the different chemosensitivity. Because patients who experience late relapses are usually excluded from receiving HDCT at Indiana University, patient selection could represent the major reason for this observation.

It is noteworthy that there were five long-term survivors in our cohort of patients. Only one long-term survivor is described in the report of Pont et al,¹⁵ and, with that exception, cures after HDCT failure have not been previously reported in GCT. Although chemotherapy with daily oral VP-16 was essential in achieving useful PRs, the role of surgery cannot be overemphasized. Four of these five patients had detectable carcinoma in the surgical specimen, and it is predictable that all of them would have eventually relapsed without resection of their residual mass. However, one individual patient progressed both serologically and radiographically within 4 weeks of completing high-dose chemotherapy. This particular patient with choriocarcinoma had a rapidly increasing serum bHCG level and development of increased size and number of pulmonary metastases. After treatment with daily oral VP-16, his serum bHCG level returned to normal and he was left with a single pulmonary nodule. This was resected and revealed persistent choriocarcinoma. This particular patient (patient no.1 in Table 4) has been disease-free for 93 months from the start of this therapeutic approach and was almost certainly cured with this combination of daily oral VP-16 and surgery. The other patients who received daily oral VP-16 had a clinical PR with decline in formerly increasing tumor markers. However, these patients also had anatomically localized disease, and it is conceivable that they could have been cured with surgery alone rather than the combination of daily oral VP-16 followed by surgery.

Surgery should be considered for selected patients. However, unlike our prior experience with surgical salvage therapy for patients with chemotherapy-refractory GCT,¹² in which approximately 20% of patients were rendered long-term disease-free, surgery alone was overall unsuccessful in patients with post-HDCT relapse. This failure was most likely due to the fact that almost all these patients had multiple areas of metastatic disease.

We conclude that although nearly one half of the patients who experience treatment failure with HDCT for relapsed GCT go on to receive further salvage treatment with chemotherapy, objective responses are uncommon and usually of short duration. Therefore, although toxicity is moderate, careful clinical judgment should guide recommendations for treatment. Patients who experience relapse 12 weeks or more after HDCT seem to have a better chance of response. Although no single-agent or combination regimen demonstrated superior activity, a strong argument can be made against repeat cisplatin-based combinations. We did not find significant activity with paclitaxel and ifosfamide, either alone or combined with cisplatin, which is in contrast with results from other studies.^14,15,22-24 Surgical resection of residual masses after chemotherapy should be performed whenever possible. New and more active agents are clearly needed in this patient population.