diarrhea, electrolyte abnormalities, elevations of hepatic transaminases and 
serum creatinine. The non-hematologic drug limiting toxicity was mucositis. 
Although this was primarily a Phase I study, antitumor responses were noted. 
There was one complete remission lasting 373 days and 16 partial remissions. 
In this study we propose to estimate the efficacy and further characterize the 
toxicity of carboplatin/ etoposide and autologous marrow support in pediatric 
patients with re;a[sed/refractory neuroblastoma who lack marrow involvement. 
In addition, we plan to use gene marking of an aliquot of the harvested 
marrow to study the mechanism of relapse and of autologous reconstitution. 
2.3 Mechanism of Relapse 
Although ABMT may offer advantages over conventional chemotherapy, the 
major cause of treatment failure remains relapse or disease progression. When 
ABMT is undertaken for solid tumors, relapse is generally at the site of 
original disease, implying that supralethal chemo-radiotherapy has not 
eradicated the malignancy. The mechanism of relapse is less clear following 
ABMT for malignancies such as neuroblastoma which may involve the BM. 
As the incidence of relapse in these diseases is often higher after autologous 
BMT than it is after allogeneic marrow transplantation, two explanations have 
been offered. The first is that the alloreactive T lymphocytes present in 
allogeneic marrow have a major role in recognition and elimination of residual 
host malignant cells. Since these cells are absent from autologous marrow, 
relapse is more likely after autologous BMT. This explanation suggests that 
relapse occurs for the same reason as relapse after ABMT for marrow sparing 
solid tumors - persistence of residual disease in the host . An alternative 
explanation is that even though cryopreserved autologous marrow is harvested 
in remission, it nonetheless contains residual malignant cells. In this case, 
relapse is due to residual disease not in the host, but in the rescuing marrow . 
It is likely that both factors contribute to relapse, although the relative 
contribution of each in any given disease is unknown. 
The mechanism of relapse following ABMT for malignancy derived from or 
involving the marrow is an important issue to resolve. The fear that stored 
marrow eontains residual malignant cells has led to intensive investigation of 
the value of marrow purging prior to storage and subsequent reinfusion. 
Purging may be undertaken with monoclonal antibodies, with cytotoxic agents 
active in vitro, or by physicochemical means. In animal and pre-clinical human 
studies all of these methods reduce contamination with malignant cells if these 
are deliberately added to the marrow, but none have consistently been shown 
to reduce the risk of relapse in naturally occurring disease. Marrow for 
Recombinant DNA Research, Volume 14 
[711] 
