Carboplatin-based regimens have demonstrated activity in unresectable low-grade glioma (LGG) in children. Despite an interesting toxicity profile, the use of these regimens has been limited by the development of carboplatin hypersensitivity reaction (HSR) in up to 30% of patients. Desensitization has been the recommended approach for HSR. However, no guidelines have existed to aid physicians when carboplatin desensitization techniques fail.
A pilot study of monotherapy with weekly vinblastine for LGG in 9 children who developed carboplatin HSR on a carboplatin and vincristine regimen was performed.
Vinblastine toxicity was moderate and readily manageable. None of the 9 patients had disease progression on therapy. Magnetic resonance imaging evaluation of tumor size from diagnosis to the end of vinblastine treatment showed 1 complete response (CR), 1 partial response (PR), 5 objective effects (OE), and 2 stable diseases (SD).
Over the last decade, evidence has been increasing that chemotherapy is a valid alternative to radiation in the management of nonresectable pediatric low-grade glioma (LGG). Although additional data are needed to assess the long-term benefit of chemotherapy, preliminary results have suggested that chemotherapy can delay or even allow avoidance of the use of radiation in a significant proportion of children.1–6 Several chemotherapy regimens have been used, but the choice of the optimal regimen is still a matter of debate. Because children treated for LGG are expected to have prolonged survival, long-term toxicity of chemotherapy should be one of the major factors influencing the choice of chemotherapeutic agents. Among the different regimens of chemotherapy used in the treatment of pediatric LGG, carboplatin alone and vincristine–carboplatin combinations have shown an attractive toxicity profile. Acute toxicity is mainly myelosuppression — neutropenia and/or thrombocytopenia — and there is no significant long-term toxicity.1 However, the limiting factor of carboplatin-containing regimens is the occurrence of carboplatin hypersensitivity reaction (HSR). When carboplatin HSR occurs, desensitization techniques have been suggested with anecdotal reports of short-term and even long-term benefit.7 However, there are no accurate predictive factors for the severity of allergic symptoms in cases of further carboplatin exposure, and most often physicians prefer to discontinue carboplatin, especially in cases of life-threatening reactions. For children who develop allergic reactions early in the treatment course or in situations where the physician considers that more chemotherapy would be beneficial for the patient, there is no established recommendation in the literature concerning potential substitution for carboplatin or chemotherapy guidelines in carboplatin intolerant patients. The current study reports a pilot experience using weekly vinblastine in LGG patients, with severe carboplatin allergy.
MATERIALS AND METHODS
Patients were identified from three institutions, the Hospital for Sick Children in Toronto, the Karolinska University Hospital in Stockholm, and the King Hussein Cancer Center in Amman.
Patients included in the study were children diagnosed with LGG who had developed an allergic reaction to carboplatin, failed carboplatin desensitization, and, subsequently, received weekly vinblastine chemotherapy. The choice of second line vinblastine chemotherapy was chosen in accord with the therapeutic policy of the hematology and oncology division of our hospital. The study was approved by the local Research Ethic Boards.
Grading of allergic reactions was defined according to the National Cancer Institute (NCI) common terminology criteria for adverse events (CTCAE version 3.0) grading system (Table 1). Management of carboplatin hypersensitivity varied between institutions and physicians. Some children were rechallenged with carboplatin using different premedication schedules (antihistamine ± steroid). Other children received a desensitization protocol following a pretreatment with antihistamine and steroids. Desensitization protocol consisted of a progressive increased rate of the carboplatin infusion within a variable period of time. Vinblastine was given as a weekly injection of 6mg/m2. The treatment was started when the absolute neutrophil count (ANC) was > 1000/mm3 and the platelet count was > 100 000/mm3. The dose of vinblastine was reduced to 4mg/m2 in cases of myelotoxicity defined as an ANC ≤ 750/mm3. The duration of vinblastine chemotherapy was at the discretion of the physician. During vinblastine administration, response was evaluated by magnetic resonance imaging (MRI), with and without gadolinium, usually every 3 months. Tumor response was recorded based on the measurement of the tumor using the 2 largest perpendicular diameters in two planes (sagittal and axial). Complete response (CR) was defined as no detectable disease. Partial response (PR) was defined as a ≥ 50% decrease of the product of the 2 largest perpendicular diameters, and objective effect (OE) was defined as a > 25% (but < 50 %) decrease in the product of the 2 largest perpendicular diameters. Progressive disease (PD) was defined as an increase of > 25% in the product of the 2 largest perpendicular diameters or the development of a new lesion. Stable disease (SD) related to all other situations.
Table 1. Allergic Grading Toxicity according to the NCI Common Terminology Criteria for Adverse Events (CTCAE), Version 3.0
Transient skin rash
Face & neck swelling
Dyspnea, mild bronchospasm (cough without wheezing)
Hypotension tachycardia, cardiac arrest, loss of conscious
Nine patients diagnosed with LGG treated with second line vinblastine following carboplatin allergy were identified in the 3 institutions. There were 5 girls and 4 boys. Median age at diagnosis was 2.1 years (0.6–11.4 yrs). Seven tumors were located in the optic pathway, one in the spinal cord, and one patient had a diffuse leptomeningeal low-grade oligoastrocytoma. Six patients had surgery at the time of presentation (four biopsies, two partial resections). All patients were initially treated with the vincristine–carboplatin combination. Carboplatin was given in a monthly schedule (550 mg/m2 intravenously [i.v.] over 1 hour) in 4 patients8 and in a weekly schedule (175 mg/m2 i.v. over 1 hour) in 5 patients.1 The median time to first allergic reaction was 4.2 months (range, 1.8–9.1 mos) for a median number of 8 injections of carboplatin (range, 4–23 injections). The cumulative dose of carboplatin at the time of first allergic reaction was 3330 mg/m2 (range, 1225–4900 mg/m). The initial allergic reaction was a Grade 1 HSR in six patients, a Grade 3 in two patients and a Grade 4 in one patient. Eight of 9 patients were rechallenged with further carboplatin injections. Three of these patients received additional carboplatin with premedication only. One of these three patients tolerated 3 more cycles, and the two others experienced acute systemic reaction that required discontinuation of the infusion at the end of the first subsequent cycle.
Six patients received further carboplatin administration using various premedication schedules in addition to a carboplatin desensitization protocol. The desensitization enabled the administration of only 1 additional carboplatin infusion in 3 patients, 3 additional infusions in 1 patient, and 17 infusions in 1 patient who was treated with the weekly carboplatin schedule. In the sixth patient, the first attempt at desensitizing was unsuccessful, and carboplatin administration had to be interrupted during the infusion. With the use of desensitization techniques and premedication, the total cumulative dose of carboplatin ranged from 1550 to 5040 mg/m2 (median, 3550 mg/m2) with a median gain of 550 mg/m2 between the first reaction and carboplatin discontinuation (Table 2).
Table 2. Patient characteristics
Age at diagnosis (yrs)
Cumulative CB dose at allergic reaction (mg/m2)
No. of CB injections/No. expected at time of 1st CB HSR
All the carboplatin-sensitive patients were treated with weekly vinblastine (6mg/m2). Six patients completed treatment after a median duration of 7.4 months (range, 1.3–9.9 mos) for a median cumulative dose of 165 mg/m2 of vinblastine (range, 36–264 mg/m2). Three patients are currently still on treatment, at 3 and 6 months after initiation of vinblastine.
The treatment was well tolerated in all patients. The main toxicity was myelosuppression with six patients experiencing neutropenia (two Grade 1 to 2 and four Grade 3 to 4 ANC toxicity). No patient required blood or platelet transfusions. The vinblastine dose was reduced to 4 mg/m2 in two patients because of hematologic toxicity. Two patients experienced uncomplicated fever and neutropenia episode. No bacterial documentation of these episodes was found. One patient had Grade 2 alopecia; four patients Grades 1 and 4 had no evidence of hair loss. Two patients, who had initial alopecia, had significant hair growth during the course of vinblastine. No allergic reaction was noticed. No nausea and vomiting were recorded. No preventive antiemetic treatment was necessary before the administration of vinblastine.
Our group evaluated tumor response before initiation of vinblastine, recognizing that as vinblastine therapy was always initiated immediately after discontinuation of previous regimen, the evaluation of the tumor response was difficult. Before initiation of vinblastine, one patient had shown PR to carboplatin–vincristine, two had OE, and six had SD. Subsequent serial evaluation of tumor size from diagnosis to the end of the vinblastine administration showed one CR, one PR, five OEs, and two SDs (Table 3). The tumor response of one of three patients still on treatment is illustrated in Figure 1.
Table 3. Evaluation of Tumor Response
Total no. of VBL injections
Cumulative dose of VBL (mg/m2)
Tumor response to VCR/CB from diagnosis
Tumor response on VBL from VCR/CB evaluation
Overall tumor response from diagnosis
Follow-up (mos) from diagnosis
VBL: vinblastine; VCR; vincristine; CB: carboplatin; OPG: optic pathway glioma; OE: objective effect; SD:; OPHG: optic pathway and hypothalamic glioma; PR: partial response; SCT: spinal cord tumor; CR: complete response; Σevaluation of vinblastine treatment at 6 mos (Patients 5, 6) and 3 mos (Patient 9).
Patients have been followed with a median followup of 22 months (range, 6.0–45.9 mos) from diagnosis and 10 months (range, 3.0–32.4 mos) from initiation of vinblastine. One patient who achieved a CR on vinblastine experienced a local recurrence 9 months after the end of the vinblastine treatment (27 mos after diagnosis) and is currently on a third line of chemotherapy.
Carboplatin allergic hypersensitivity is well described both in adults and in children.9–11 In the pediatric population, different schedules of carboplatin are associated with different incidences of allergic reactions. The highest incidence is currently reported among the group of children with LGG treated with either carboplatin alone or with the vincristine–carboplatin combination. The reported incidence of this complication has increased since the first report from Packer et al. and ranges 7–32%. It is most likely that the initial report of 7% in children with LGG treated with vincristine and carboplatin has underestimated the real incidence or only reported severe allergic events.1 As this complication becomes better recognized and reported, incidence in the more recent reports reaches more consistently 30%.4, 12–13
Carboplatin HSR appears to be associated with the cumulative number of injections rather than the cumulative dose of the drug. Schiavetti et al. has reported a cumulative risk of 2% at 6 cycles rising to 47% for greater than 12 cycles of carboplatin.14 There is also some suggestion that the weekly schedule and a lower dose per infusion could be associated with a higher incidence of carboplatin HSR compared with monthly infusion at higher doses.15 Nevertheless, there is no reliable predictive factor to identify the patients at risk of HSR. Furthermore, the severity of the HSR is unpredictable, ranging from mild skin rash to life-threatening anaphylactic reaction.16–17 There is also no clear clinical or biologic factor that can predict an individual reaction to further carboplatin exposure.18–19 To overcome this problem, various protocols of desensitization have been proposed.7, 12, 20–22 Although some success has been reported with this strategy, reexposure to carboplatin is not without risk. Most reports mention the need to perform desensitization techniques in the intensive care unit (ICU) setting. Moreover, the gain in terms of additional courses of carboplatin allowed by this desensitization is generally disappointing, leading at the best to a delay of carboplatin discontinuation.7, 12 In our experience, the median additional gain of carboplatin obtained with this technique was 550 mg/m2. There is no large pediatric series to confirm the effectiveness of this approach; however, a recent report from the HIT-SIOP group suggests that the majority of physicians confronted with carboplatin HSR ultimately discontinue carboplatin administration.13 In our experience, only six out of the eight rechallenged patients were actually able to receive further carboplatin infusions (1–3 infusions). Only 1 patient treated with the weekly carboplatin schedule was able to receive 17 additional infusions, but eventually this patient developed a reaction of increasing severity in addition to a cushingoid appearance because of the frequent need for steroid premedication. None of the patients completed their planned vincristine–carboplatin treatment.
Despite the high incidence of carboplatin HSR among the LGG pediatric population, there is no guideline for the management of these patients when further administration of carboplatin is considered unsafe. A physician may consider that, at the time of the carboplatin HSR, the patient has received enough chemotherapy and decide to closely observe this patient without further chemotherapy. However, most of these chemotherapy protocols are 12 to 16 months in duration, and the risk of early chemotherapy discontinuation is unknown, particularly when children develop carboplatin HSR during an early course of their treatment. In our experience, six patients developed their first carboplatin HSR less than half way through the course of their treatment. For the three patients with a late onset of carboplatin HSR, the decision to continue chemotherapy was made because of evidence of aggressive tumor behavior (two patients) or late/slow response to chemotherapy (one patient). Another alternative would be to consider the occurrence of HSR as a chemotherapy failure and proceed to radiation. However, the majority of physicians will consider radiation as salvage treatment in the eventuality of tumor progression rather than as an alternative to carboplatin in case of HSR. The substitution of another agent for carboplatin is an option. Gnekow et al. have recently reported in the HIT-LGG study an incidence of 32% allergic reaction.13 Among their 27 reacting patients, only 1 patient received further carboplatin at reduced dose. In 21 patients, cisplatin or cyclophosphamide was substituted for carboplatin. Three other patients who were near completion of their planned chemotherapy did not receive further chemotherapy. Gnekow et al. do not provide a separate analysis of the outcome of this group of allergic patients.
The natural history of nonresectable LGG is often unpredictable and the optimal management of these tumors is a source of passionate debate. There is growing evidence that there is no straightforward treatment for these tumors, and their management should be considered in the context of a chronic condition with emphasis on long-term toxicity. Therefore, the choice of chemotherapy agents needs to take into account the long-term perspective ensuring a minimum of late side effect. Carboplatin-containing regimens fulfill these criteria with excellent activity and little evidence of long-term toxicity. However, many other regimens used in children with LGG contain agents associated with significant long-term toxicity. Regimens including alkylating agents such as CCNU, procarbazine, or temozolomide are theoretically associated with increased risk of mutagenesis.23 Similarly, the prolonged use of DNA topoisomerase II inhibitors, such as etoposide, is associated with risk of secondary leukemia.24 Despite an interesting 70% of objective response rate, the cisplatin–etoposide combination used in LGG is associated with a high incidence of ototoxicity, which can be devastating in the context of children with optochiasmatic glioma and severe visual impairment.6 In the current study, our group took the option of a substitution using a weekly schedule of vinblastine.
Vinblastine has been widely used in oncology. In children, it is used for the treatment of several malignant conditions including Hodgkin disease, anaplastic large cell lymphoma, and germ cell tumors.25–27 It is also commonly used in the management of chronic conditions such as Langerhans cell histiocytosis, aggressive desmoid tumors, and fibromatosis.28, 29 The toxicity profile of vinblastine is excellent with no evidence of long-term toxicity, even after prolonged exposure. Central nervous system (CNS) penetration of vinblastine is poor, and this drug has only been used in combination with other agents for the treatment of intracranial germ cell tumors. However, vinblastine has shown activity against human glioma xenografts.30 In a pilot study, Bouffet et al. reported promising activity with vinblastine in children with recurrent and refractory LGG.31 In the Bouffet study, all children had previously received chemotherapy and 6 out of 11 children treated with weekly vinblastine showed a response. Toxicity was moderate to absent.
The mechanism of action of vinblastine in LGG is not known. Preclinical data have demonstrated an antiangiogenic activity of low-dose vinblastine in models in vitro.32, 33 These observations suggest potential antiangiogenic effect of vinblastine on LGG, as the poor CNS penetration of this agent and the low mitotic activity of pediatric LGG do not support a direct cytotoxic effect.
In the current study, six patients experienced transient and moderate hematologic toxicity. No other significant toxicity was noted. With a median followup of more than 2 years, no patient has shown any evidence of long-term toxicity.
In the context of our pilot study, the assessment and interpretation of tumor response is difficult, as all patients were already treated with chemotherapy at the time of initiation of vinblastine. Because of previous treatment with vincristine and carboplatin, tumor shrinkage may be interpreted as a continuous benefit of the initial chemotherapy. However, it is noteworthy that none of the nine patients showed progression. Of 9 patients, 7 had > 25% tumor shrinkage, and 2 had SD during the treatment with vinblastine. One patient experienced a recurrence of a spinal cord astrocytoma 9 months after completion of vinblastine.
The current study suggests that vinblastine is a valuable option in children with LGG, treated with carboplatin containing regimen, who develop HSR to carboplatin and fail desensitization techniques. The weekly administration of single agent vinblastine is well tolerated with moderate hematologic toxicity and no evidence of long-term toxicity. Further studies should be conducted to confirm the benefit of vinblastine in children with LGG.