Evidence Based Questions

Question 1

In MB, how are NTR, STR, and “Bx only” defined?

Answer 1

NTR: <1.5 cm2 residual on postop MRI

STR: ≥1.5 cm2 residual on postop MRI

Bx only: No attempt at definitive resection

Question 2

In MB, is there a difference b/t NTR vs. GTR in terms of EFS? How about STR and GTR?

Answer 2

Retrospective studies suggest that pts who obtain an NTR and GTR have similar outcomes. (Gajjar A et al., Ped Neurosurg 1996) However, 5-yr EFS is worse in STR pts (54%) compared to GTR/NTR pts (78%). (Zeltzer PM et al., JCO 1999)

Question 3

What chemo agent improved DFS and OS according to MB studies in the 1990s?

Answer 3

Cisplatin. Prior to the introduction of cisplatin, several studies (SIOP I and CCG 942) failed to show improved OS with the addition of adj chemo.

Question 4

What 2 studies demonstrated the need for chemo with reduced-dose CSI (23.4 Gy) for standard-risk MB?

Answer 4

There has been no RCT comparing reduced-dose CSI +/– cisplatin-based chemo. The need for cisplatin-based chemo is inferred from the following 2 studies:

1. POG 8631/CCG 923: randomized standard-risk MB to 36 Gy vs. 23.4 Gy CSI alone (no chemo). There was a trend toward ↓ EFS and OS in the 23.4-Gy arm. (Thomas PR et al., JCO 2000)
2. CCG 9892 (phase II): standard-risk MB treated with 23.4 CSI with concurrent weekly vincristine → 55.8-Gy boost to PF → adj cisplatin/CCNU/vincristine. 5-yr PFS was 79%, which was similar to historical controls treated with 36 Gy CSI and similar chemo. (Packer R et al., JCO 1999) Basis for POG A9961 reduced dose CSI.

Question 5

Can RT be delayed for MB pts <3 yo by using chemo alone? What studies support this?

Answer 5

Yes. Given the toxicity of RT in pts <3 yo, it is reasonable to delay RT until 3 yo, especially with desmoplastic histology.

Baby POG (Duffner PK et al., Neurooncol 1999, NEJM 1993): <3 yo, 206 pts, high-/low-risk MB + other PNET, chemo alone (Cytoxan + Vincristine (VCR) × 2→ cisplatin + etoposide) × 2 yrs if <2 yo, × 1 yr if 2–3 yo. 5-yr OS was 40%, and PFS was 32%.

German BTSG data (Rutkowski S et al., NEJM 2005): <3 yo, 43 pts, high-/low-risk MB, chemo (Cytoxan, vincristine, Mtx, carboplatin, VP-16, intrathecal Mtx). 5-yr PFS was 58%, and OS was 66%. The majority of pts had a desmoplastic variant histology. The benefit was best in M0 pts (5-yr PFS of 68% and OS of 77%)

SFOP data (Grill J et al., Lancet Oncol 2005): <5 yo, 79 pts. 5-yr OS was best in R0M0 (73%) vs. 13% with M+.

Question 6

What was the Tx regimen on COG A9934 for MB pts <3 yo?

Answer 6

Initial Sg → induction chemo × 4 mos with Cytoxan, vincristine, cisplatin, etoposide → 2nd Sg for identifiable or residual Dz → age/risk group/response-adapted conformal RT to PF + primary site (no CSI) → maintenance chemo × 8 mos. Enrolled children were older than 8 mos but younger than 3 yrs, all M0 MB.

Age/risk/response-adapted RT:

(Ashlet DM et al., JCO 2012)

If <24 mos and CR: 18 Gy to PF → tumor bed boost to 50.4 Gy, or 54 Gy if PR/Stable disease/+ residual

If >24 mos and CR or PR: 23.4 Gy to PF → tumor bed boost to 54 Gy

Question 7

What evidence supports the use of >50 Gy total doses in MB?

Answer 7

Retrospective data suggest that LC in the PF varies with dose above and below 50 Gy. In 60 MB cases, if the PF dose was >50 Gy, the LC was 79%. However, if the PF dose was <50 Gy, the LC was 33%. (Hughes EN et al., Cancer 1988)

Question 8

In MB pts, does the entire PF need to be boosted to >50 Gy?

Answer 8

Retrospective evidence suggests that few failures occur in the PF outside the tumor bed (<5%).

Fukunaga-Johnson et al. reviewed 114 pts treated with CSI → boost to the entire PF. The solitary site of the 1st failure within the PF but outside the tumor bed occurred in 1 of 27 failures. (IJROBP 1998)

Wolden et al. reviewed 32 pts treated with tumor bed boost only. There were 6 total failures: 5 outside the PF and 1 within the PF but outside the boost volume. (JCO 2003)

Merchant et al. conducted a prospective phase II trial of 23.4 Gy CSI + PF boost to 36 Gy and primary site to 55.8 Gy with dose-intensive chemo. 5-yr EFS was 83%, and PF failure was 5%. Reduced doses to temporal lobes, cochlea, hypothalamus. (IJROBP 2008)

Question 9

What are the RT technique questions being addressed in COG ACNS0331?

Answer 9

In ACNS0331, standard-risk pts 3–7 yo are randomized to CSI to 18 Gy vs. 23.4 Gy. For the 18 Gy arm, all pts got a PF boost to 23.4 Gy. All standard-risk pts 3–7 yo underwent a 2nd randomization: CD to 54 Gy to whole PF vs. tumor bed only. Standard-risk pts 8–22 yo: 23.4 Gy CSI → randomization to CD to 54 Gy to PF vs. tumor bed only.

Question 10

What was the rationale for 18-Gy CSI in ACNS0331?

Answer 10

CSI doses in excess of 20 Gy still pose a significant risk for cognitive and growth outcomes, particularly in young children. Pilot study in 10 children with PNET of the PF showed comparable outcomes to higher doses. (Goldwein J et al., IJROBP 1996)

Question 11

What do the preliminary results of ACNS0331 show?

Answer 11

For pts with standard-risk MB, boosting the tumor bed alone is sufficient, but decreasing the CSI dose to 18 Gy is associated with higher risk of recurrence and is not recommended. (Michalski JM et al., IJROBP 2016)

Question 12

What question does ACNS0334 attempt to address?

Answer 12

Phase III trial in children <3 yrs with high-risk MB or PNET. Trial addresses the addition of high-dose Mtx to the 4-drug induction chemo regimen of VCR, etoposide, Cytoxan, cisplatin → 2nd Sg, consolidation, and peripheral blood stem cell rescue. RT is at the discretion of the institution.

Question 13

What study is examining molecular risk-adapted Tx?

Answer 13

SJMB12 is a St. Jude trial examining risk-adapted escalation and de-escalation of radiotherapy and chemo based on molecular subtype (WNT, SHH, and Non-WNT/SHH [Group 3/4]), cytogenetics, histology, and extent of resection.

Question 14

Is there a role for pre-RT chemo in MB pts >3 yo?

Answer 14

No. In MB pts >3 yo, intensive chemo prior to RT (vs. RT then chemo) is associated with ↑ RT toxicity, RT Tx delays, and worsened RFS. (German HIT 91: Kortmann RD et al., IJROBP 2000)

Question 15

What benefit does proton therapy have in the Tx of MB?

Answer 15

Retrospective data suggest that proton plans have ↓ dose to the cochlea/temporal lobe compared to IMRT (0.1%–2% vs. 20%–30%), and virtually no exit dose to the abdomen, chest, heart, and pelvis. Recent study suggests less morbidity, including GI and heme toxicity (although this is in adults). (Brown AP et al., IJROBP 2013)

Question 16

Is there a role for hyperfractionated RT to reduce cognitive sequelae of MB Tx?

Answer 16

MSFOP 98, a phase II trial, evaluated hyperfractionated RT in MB and showed promising results. 48 standard-risk pts were treated with CSI 1 Gy bid to 36 Gy → tumor bed boost 1 Gy bid to 68 Gy. 6-yr OS was 78%, and EFS was 75%. Decline in IQ appeared less pronounced than in historical controls. (Carrie C et al., JCO 2009)

Question 17

What did early studies attempting to optimize MB treatment show regarding the use of CHT?

Answer 17

CCG 94230 and SIOP 131 were early studies evaluating the addition of post-RT CHT to CSI (at 36 Gy) in an unselected pt population. Both ultimately showed that CHT did not confer a survival benefit among all pts, though, on subset analysis, did show a benefit to those with T3-T4 disease and M1-3 disease. Thus, several subsequent studies were performed without the incorporation of CHT as described in the following.

Question 18

Due to these results showing that CHT did not improve outcomes, can RT alone be modified to offer optimal EFS and OS while minimizing toxicity?

Answer 18

Acknowledging the neurocognitive toxicity of 36 Gy to the craniospinal axis, French investigators attempted to reduce RT volume by delivering RT to the infratentorium only; however, results were terrible, with <20% 6-yr EFS and 64% failure in the supratentorium.32 With this study clearly demonstrating that RT should be delivered to both supratentorial and infratentorial regions, the POG/CCG collaborative group modified the dose rather than the volume with their RCT (CCG 923/POG 8631) randomizing pts to a CSI dose of either 23.4 Gy or 36 Gy. The trial closed prematurely due to initial early relapses on the low-dose arm, though on longer FU 5-yr PFS was no different.33 A companion JCO publication by Mulhern et al. reported on neuropsychologic testing in long-term survivors (>6 years), finding significantly less neuropsychologic toxicity in those treated to 23.4 Gy rather than 36 Gy with the difference most pronounced in those <9 years of age.

Question 19

What trials ultimately led to the reincorporation of CHT in the management of MB?

Answer 19

Several trials continued to evaluate the role of CHT. In a prospective multi-institutional study published in 1994, Packer et al. evaluated 63 pts (both with average-risk and high-risk disease) with a treatment regimen incorporating weekly vincristine during RT followed by eight 6-week cycles of cisplatin/CCNU/vincristine, with authors concluding that “chemotherapy has a definite role in the management of children with medulloblastoma.”35 Later, the PNET 3 PRT randomized pts s/p GTR/NTR to RT alone versus CHT followed by RT.36 Initial results showed improved 5-yr EFS of 60% versus 74% (p = .037) with no difference in OS, though a later update showed a reduction in health status (including hearing, speech, vision, ambulation, dexterity, emotion, cognition) in those who received CHT.

Question 20

What studies led to the use of reduced-dose CSI in average-risk disease?

Answer 20

The same Packer study referenced earlier treated standard-risk pts to 23.4 Gy with the use of concurrent weekly vincristine (in the same study, they treated high-risk pts to 36 Gy with weekly vincristine) and found favorable outcomes.35 This study was then expanded into CCG 9892, a phase II trial limited to standard-risk pts again delivering 23.4 Gy to the CSI followed by a 55.8 Gy PF boost with concurrent weekly vincristine, followed by subsequent CHT with CCNU/vincristine/cisplatin. This resulted in a 3-yr PFS of 88% and 3-yr OS of 85%, rates authors argued were comparable to previous studies using higher dose CSI. Authors concluded that reduced-dose CSI with concurrent and adjuvant CHT is a feasible approach for M0 disease.

Question 21

Can cyclophosphamide replace CCNU in the adjuvant CHT portion of treatment?

Answer 21

COG A9961 was a large PRT randomizing average-risk MB pts (all of whom were post-op and received 23.4 Gy CSI) to two different adjuvant CHT regimens, one with cyclophosphamide and another with CCNU. The rationale was that data supporting the use of CCNU in pediatric tumors was scant whereas xenograft and early clinical data for the use of cyclophosphamide was more promising.38 Ultimately, there was no significant difference between the two regimens, with 5-yr OS about 85% in both arms. Authors concluded that though neither CHT regimen was superior, the favorable outcomes seen with both regimens offer additional support to the use of reduced-dose CSI.

Question 22

When treating average-risk MB, does hyperfractionation of CSI affect outcomes or reduce toxicity?

Answer 22

MSFOP 98 was a phase I/II average risk trial using hyperfractionated therapy, 36 Gy CSI with boost to 68 Gy to the tumor bed at 1 Gy/fx BID.39 It showed excellent long-term EFS in the absence of CHT and full-scale intelligence quotient (IQ) drop was less pronounced compared to other standard RT reports. This led to a large European PRT (HIT-SIOP PNET-4), which enrolled average-risk MB pts who were randomized to standard fractionation (23.4 Gy CSI with PF boost to 54 Gy at 1.8 Gy/fx) versus hyperfractionation (36 Gy CSI with PF boost to 60 Gy and 68 Gy tumor bed at 1 Gy/fx BID with 8-hr interfx interval).40 Results published in the JCO in 2012 showed equivalent outcomes for EFS and OS and no difference in ototoxicity; IQ measurements were not reported in their final publication. Based on these results, hyperfractionation is typically not employed in average risk MB.

Question 23

What is the optimal dose to deliver to the PF?

Answer 23

This has never been prospectively studied, but a 1988 Harvard RR showed better LC if the PF dose was >50 Gy (LC 79% vs. 33% if less than 50 Gy; p < .02).

Question 24

Does the boost volume need to include the entire PF or is tumor bed + margin sufficient?

Answer 24

Two RRs showed that the PF failure rate was ≤5% with the use of IF-directed boost. These reports in part provided the rationale for ACNS 0331.

Question 25

Can pts with average-risk MB who are most vulnerable to the neurocognitive effects of CSI receive a lower dose? Can any average-risk pt receive IF boost rather than whole PF boost?

Answer 25

ACNS0331 was developed to answer both questions. Multiple studies have shown that CSI doses of >20 Gy can damage neurocognitive and growth outcomes, prompting investigators to determine whether lower doses could confer favorable outcomes with less toxicity.44,45 In 1989, Goldwein et al. reported on their prospective cohort study of 10 MB pts delivering 18 Gy CSI with a PF boost of 50.4 to 55.8 Gy with weekly vincristine and subsequent CDDP/VCR/CCNU CHT showing favorable cure rates with this approach.46 Additionally, regarding IF versus PF boost, Wolden’s and Merchant’s studies (see the previous discussion) finding low PF failure rates with the use of tumor bed boost alone suggested that IF-only boost may be appropriate. These results help set the stage for ACNS 0331.

Question 26

What data initially supported the use of CHT in high-risk disease?

Answer 26

CCG 942 (discussed earlier in the average-risk section) and SIOP I were both PRTs evaluating postoperative pts who received CSI and were then randomized to CHT or no CHT.30,31 For both studies, there was no difference in outcome between the two groups, but when limited to those with more advanced disease (T3-T4, M+ or STR), an improvement in EFS was observed.

Question 27

Can outcomes be improved by intensifying the CHT regimen with additional agents?

Answer 27

CCG 921 was performed in pts with a variety of high-risk pediatric brain tumors to see if “8 in 1” CHT (8 types of CHT in 1 day: cisplatin, procarbazine, CCNU, vincristine, cyclophosphamide, methylprednisolone, hydroxyurea, cytarabine) was better than a combination of vincristine/CCNU/prednisone (VCP). 421 children were enrolled, of which 203 had MB. Subset analysis of this group showed better outcomes with VCP than 8 in 1 CHT (5-yr PFS 63% vs. 45%, p = .006).

Question 28

What about altering the sequence of CHT (e.g., delivering CHT immediately post-op followed by RT)?

Answer 28

Four PRTs have evaluated this question: SIOP II, SIOP III, POG 9031, and HIT 91 from Germany. All of them except for SIOP III showed no benefit to immediate post-op CHT; both POG 9031 and SIOP II showed 5-yr EFS to be about 60% to 70% and 5-yr OS about 75% in both groups with no significant difference between them and HIT 91 actually showed an improvement in 3-yr EFS with immediate RT (78% vs. 65%, p = .03).5,23,47 SIOP III is the only exception, showing improved 3- and 5-yr EFS with up-front CHT.4 Therefore, with three of these four studies showing no benefit to up-front CHT, standard of care is to perform maximal safe resection followed by RT (with concurrent vincristine) followed by adjuvant CHT.

Question 29

Is there any detriment to interrupting RT?

Answer 29

Interestingly, SIOP III also showed better 3-yr OS when RT was delivered within 50 days, suggesting that avoiding RT interruption can lead to better outcomes.4 These results confirm earlier findings from the University of Florida that prompt completion of RT confers better outcomes (at a cut point of 45 days in that 1998 study).

Question 30

Does the use of carboplatin as a radiosensitizer during CSI lead to better outcomes?

Answer 30

This was addressed in COG 99701, a phase I/II trial that evaluated the role of adding radiosensitizing carboplatin to vincristine during CSI.49 Note that the adjuvant therapy offered changed slightly once the recommended dose of carboplatin was determined, but among all pts on this study, 5-yr OS was approximately 75%. Authors concluded that CRT with vincristine + carboplatin followed by 6 months of maintenance CHT produced outcomes at least as good as (if not better than) other prior trials using higher dose CSI or higher dose alkylator-based therapy. This carboplatin-containing regimen is being tested in an ongoing phase III PRT, ACNS 0332. This trial is evaluating the intensification of systemic therapy in high-risk MB (M+, STR and/or diffuse anaplasia) and is utilizing two randomizations: (a) concurrent carboplatin during RT and (b) isotretinoin during and after maintenance therapy.

Question 31

Is there a role for re-irradiation?

Answer 31

Recurrent MB is rarely cured and has a dismal prognosis with 2-year OS historically <25%. However, a number of salvage treatments have been considered including surgical resection, brachytherapy, radiosurgery, high-dose CHT with autologous stem cell transplant and re-irradiation. Re-irradiation may be a reasonable option to consider in both standard-risk and high-risk pts.

Question 32

What is the recommended treatment for infants (<3 years)?

Answer 32

The mainstay of treatment for MB involves maximal safe resection followed by CSI + boost. However, CSI can lead to significant neurocognitive toxicity, which is not only dose dependent but also dependent on the age of the pt (younger is worse).50,51 Therefore, CHT has been evaluated as a stopgap to delay RT. Baby POG#1 showed this was possible with a 5-yr OS of 40%. An unintended consequence of this study was complete parent refusal of radiation, which showed that in a select group of MB infants, there maynot be a need for RT at all. As described earlier, the CCG group tried eight CHT drugs (“8-in-1”) with worse outcomes but confirmed the approach of CHT before RT was feasible. Follow-up trials including the Head Start I and II for infant MB used intensive CHT and used radiation only in a salvage setting.52This approach eliminated craniospinal XRT in 52% of pts and may preserve quality of life and intellectual function. However, the intense CHT is not without cost and 4 of 21 infants died of treatment-related death. The findings of Baby POG#1 have been further bolstered by Rutkowski et al. using surgery and subsequent CHT, with 5-yr OS for those s/p GTR of 93%, 56% if STR, and 38% for those with macroscopic metastases.

Question 33

What is the appropriate management for medulloblastoma patients less than 3 years old?

Answer 33

Maximal safe resection → chemo until patient reaches age 3 (at which point can consider CSI → more chemo). In a Children’s Cancer Group study, patients were treated with one of two four-drug combination chemotherapy regimens, and RT was withheld unless there was evidence of tumor progression. At 5 years, 10 of 38 children who had M0 disease were alive and event-free without having received RT.

Geyer, JR, Sposto, R, Jennings, M, et al. Multiagent chemotherapy and deferred radiotherapy in infants with malignant brain tumors: a report from the Children’s Cancer Group. J Clin Oncol. 2005;23:7621-7631.

Question 34

What did the CCG 921 high-risk medulloblastoma trial investigate?

Answer 34

In CCG 921, high-risk patients (classified as age 1.5–21, or M1-4, or T3-4, or residual >1.5 cm2) were randomized to CSI36 Gy/PF 54 Gy/spinal mets 50.4–54 Gy (age <3 received CSI 23.4 Gy/PF 45 Gy) + concurrent vincristine → Vincristine, lomustine, and prednisone (VCP). x8 versus “8 in 1” (vincristine, prednisone, lomustine, hydroxyurea, procarbazine, cisplatin, cyclophosphamide, and cytarabine) chemotherapy x2 → RT → “8 in 1” chemo x8. Better 5-year PFS with VCParm (65% vs. 45%, p = .006).

Zeltzer, PM, Boyett, JM, Finlay, JL, et al. Metastasis stage, adjuvant treatment, and residual tumor are prognostic factors for medulloblastoma in children: conclusions from the Children’s Cancer Group 921 randomized phase III study. J Clin Oncol. 1999;17:832-845.