Discussion
Reduction of CNS radiation ALL patients with iCNS-R continues to be a challenge, yet remains an important goal given the late effects of CNS radiation and limited options for treatment following subsequent relapse. Excellent outcomes reported in POG 9061/9412 in patients with late CNS-R were attributed to a treatment approach reliant on the use of effective anti-leukemia agents with good CNS activity facilitated by a delay in delivery of CNS radiation that had previously limited chemotherapy delivery because of prolonged myelosuppression. Specifically, POG 9061 delayed CNS radiation for six months after relapse and POG 9412 then intensified the chemotherapy, delayed CNS radiation for 12 months after relapse and reduced the dose of CRT from 2400 to 1800 cGy. COG AALL02P2 was designed to test one year of further intensified systemic therapy, followed by CRT at a reduced dose of 1200 cGy. AALL02P2 encountered a number of challenges with lower than expected enrollment given continued success with frontline therapies and decreasing CNS-R rates, as well as a large number of enrolled patients who did not complete protocol-prescribed therapy. The study was closed before reaching planned accrual when an interim analysis crossed predefined monitoring boundaries and showed inferior EFS compared to POG 9412.
With respect to overall survival, this treatment strategy with COG AALL02P2 did prove as effective for patients with late relapse and NCI SR ALL as the predecessor POG 9412 study, however, EFS was markedly inferior for this cohort compared to the prior trial. NCI HR patients fared equally poorly on both trials, without meaningful differences in outcomes.
As in POG 9412, COG AALL02P2 continued to show a difference in outcomes between the NCI HR and SR ALL groups. The overall survival showed a trend towards better survival in the NCI SR group. Similar results are seen in this recent trial with EFS for NCI HR vs SR. Barredo et al, previously reported in POG 9412 that NCI SR ALL status was an independent favorable prognostic factor in addition to length of CR1 suggesting that the NCI/Rome criteria serve as a clinical surrogate for inherent biological differences.20 There were 11 relapses with 8 attributable to CNS relapse in POG 9412, suggesting a need for further improvement in CNS (and marrow) control. Interestingly, in AALL02P2, NCI SR ALL patients fared far worse with respect to EFS than the similar cohort in POG 9412. This may be related to several factors including differences in frontline treatments with more intensified therapies thereby rendering these patients more difficult to treat as well as a large number of patients on COG AALL02P2 who did not receive the protocol-prescribed radiation. These patients (26% of the total enrolled) that did not follow the protocol-directed therapy included 10 patients who had events prior to the timing of radiation, as well as a patients, and/or clinicians, who chose not to continue with the protocol post amendment.
Novel therapies to reduce the risk of second relapse and treatment providing safer and less toxic CNS directed therapy are needed. Consideration of other factors in addition to CNS penetration of drugs is needed; the CNS is thought to be an immunologic sanctuary not just a physical barrier.40 Subclinical seeding of the CNS from other sites, particularly the BM, is implicated in subsequent BM relapse and improved methods of eradicating disease systemically are needed. While delay of CNS irradiation has proven to be an important part of relapse therapy, this study emphasizes the necessity of ensuring adequate delivery of CRT. The results of AALL02P2 suggest that cranial radiation is a critical and necessary component of relapse therapy for iCNS-R of B-ALL in the context of chemotherapy-driven regimens. Nevertheless, emerging data using immunotherapy (chimeric antigen receptor T-cells) in the setting of extramedullary disease at diagnosis or relapse of ALL may provide an alternative strategy to decrease or eliminate CNS irradiation.41-44
While baseline bone marrow MRD data were only available from 64% (76/118) patients, MRD analysis was not predictive of either EFS or OS. Within the sensitivity of 0.01%, 19.7% of patients were MRD-positive, though it is certainly possible that rates would be higher with more sensitive MRD technologies.
The incidence of isolated CNS relapse has decreased with contemporary strategies, and several trials have shown that cranial irradiation can be omitted in newly diagnosed ALL patients.14,21,23,24 However, the individual trials conducted to date have included relatively few patients with overt CNS disease. A large study from the Ponte de Legno group aggregated data from 16,623 children (<18 years) with ALL treated by 10 cooperative groups between 1996 and 2007.24 In that study, cranial radiotherapy was associated with a reduced risk of CNS relapse only in the small subgroup of 406 patients with CNS3 status, and the overall event rate did not differ between those that did or did not receive radiotherapy. In parallel to newly diagnosed ALL, there is great interest in reducing radiotherapy dose, or eliminating it entirely, in patients with CNS-R due to the long-term adverse impact of cranial irradiation on CNS outcomes, secondary brain tumors, and quality of life.9,30,45-47 Toward this end, earlier studies from the POG and others demonstrated that it was possible to preserve outcomes in patients with late CNS-R by intensifying systemic chemotherapy while reducing radiation dose and limiting it to the cranial fossa (eliminating spinal irradiation). However, despite further intensification of systemic chemotherapy, COG AALL02P2 showed inferior EFS compared to POG 9412 when the cranial radiation dose was reduced to 1200 cGy, intended to be delivered at 12 months from onset of recurrence. Limitations in the current study included a small sample size due to increasing rarity of iCNS-R, early study closure and a substantial number of patients, 34/118 (29%), who did not receive protocol directed radiation either due to early relapse prior to 1 year or decision to not follow the treatment plan. It is also possible that cranial irradiation may need to be administered earlier after relapse, and that further intensifying systemic cytotoxic therapy is counterproductive.
The overall goal to limit cranial radiotherapy in patients with iCNS-R remains important. Future studies might also incorporate newer immunotherapies or more sensitive marrow MRD technologies and/or CNS response measures to attempt to identify which patients with iCNS-R might be cured without irradiation, and whether different subgroups might require lower or higher doses of radiotherapy. To gain adequate power to address these questions, future trials will likely require international collaboration.
Disclaimer: The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.