Radiation Oncology Synopsis

Medulloblastoma

Background, Demographics and Epidemiology

Medulloblastomas arise in the posterior fossa, and are usually masses involving the vermis or cerebellar hemispheres. The other possible posterior fossa CNS masses include:

Recent biological studies have isolated medulloblastoma from posterior fossa PNETs by identifying isochromosome 17p.

There are 400 cases per year of medulloblastoma in the US and it is the 2nd most common pediatric CNS malignancy behind low grade gliomas. It accounts for 20 % of all pediatric CNS tumors. There is a bimodal age distribution with a median age of 7 years in children and 25 years in adults. Males are more commonly affected than females by a 2:1 ratio.

Medulloblastomas originate neuroectodermal cells that form the germinal matrix of the cerebellum or cerebellar vermis. They are embryonal tumors which are a subset of neuroepithelial neoplasms. The WHO Histopathologic Classification Scheme which also include these:

Medulloblastoma is part of a type of little round blue cell tumors. There are 8 different types of these tumors:

  1. lymphoma
  2. Ewings
  3. ALL
  4. Rhabdomyosarcoma
  5. neuroblastoma
     
  6. Neuroepithelioma
  7. Medulloblastoma
  8. Retinoblastoma

There are three histologic variants of medulloblastoma: classic, nodular/desmoplastic, large cell/anaplastic.

The desmoplastic variant is associated with LOH 9q, older age at diagnosis, and better prognosis. The large cell anaplastic variant is associated with aggresiveness.

About 5% are familial with associated genetic syndromes including Gorlin (PTCH mutation) and Turcot (APC mutation). Common genetic abberations in medulloblastoma include 17p deletion in 40% - 50%, isochrome 17q and 16q deletion.

Medulloblastomas most commonly arise in the midline cerebellar vermis with 75% of all cases originating here. The remainder arise in the cerebellar hemispheres. This derives from the origins of medulloblastoma in the neuroectodermal cells that form teh germinal matrix of the cerebellum/vermis.

Histologic grading has only recently been linked to prognosis. Extensive nodularityand desmoplastic variants have been linked to favorable prognosis. The degree of anaplasia has been linked to worse outcomes. Tumors with extraneural disease either at diagnosis or as a pattern of failure are more commonly associated with marked anaplastic histology.

Medulloblastoma is the classic tumor associated with CNS seeding. Sub-arachnoid dissemination has been reported in 20% to 35% of children at diagnosis. Leptomeningeal disease on MRI is seen in 32%: both spinal MRI and CSF cytology -- 12%, positive CSF alone 8%, MRI alone 11%.

Workup and Staging

Workup

Due to the general location of the medulloblastomas, the primary site is likely to obstruct the 4th ventricle causing increasted intracranial pressure, headache, malaise, alterned mentation due to ICP. Mass effect in the cerebellar vermis is likely to cause truncal ataxia, head bob, and diplopia due to CN 6.

The most common presenting sign is increased ICP/obstructive hydrocephalus manifest by headache and vomiting. In addition the "setting sun" sign which is a downward deviation of the gaze from increased ICP due to pressure on CN III (superior oblique) and IV abducens).

Posterior fossa mass workup includes:

Biopsy is not required by the COG ACNS0331 medulloblastoma protocol. Patients go straight to craniotomy. Bone marrow biopsy is not necessary nor is it part of the standard work up.

Frequently , due to symptomatic hydrocephalus, an VP shunt is placed. To date there has been no reported dissemination from these shunts. Other methods besides VP shunt placement include the use of steroids, acetazolamide (Diamox) to reduce elevated ICP.

In addition to the diagnostic workup, additional workup should be performed at very specific intervals during treatment.

The standard of care in medulloblastoma include post-operative staging, based on imaging of the brain to assess the completeness of resection and potential subarachnoid metastases.

On post-op Day 10-14: obtain MRI of the spine, CSF cytology. The delay to POD 10 is necessary to avoid false positives from surgical debris. The lumbar puncture is best obtained immediately after the MRI of the spine on POD 10. Subarachnoid dissemination has been reported at diagnosis in 20% - 35% of the cases.

MRI of the brain is performed pre-operatively and 24 - 48 hours post-operatively.

tMRI of the spine is performed pre-operatively or 10 - 14 days post-operatively.

Staging

Staging was developed by a radiation oncologist, Chang, in the pre-CT era. It is primarily a surgical staging system, rather than imaging based for this reason. After the introduction of modern imaging (CT, MRI) it became apparent that the imaging media were not as sensitive to brainstem invasion as was surgical identification. The two components are the T stage which is based on size and invasiveness. There are no modern data which demonstrate a role for the T stage as an independent predictor of outcome or selection of therapy. M-stage is highly significant as a prognostic factor and as a therapy selection tool, as treatment intensification strategies are based on M-stage. The M staging which is based on the degree of spread outside of the posterior fossa:

Chang Medulloblastoma Staging System
T Stage Description
T1 < 3 cm diam.
T2 > 3 cm diam.
T3a > 3 cm with extension
T3b > 3 cm with unequivocal extension into the brainstem
T4 > 3 cm with extension up beyond the Aqueduct of Sylvius and or down past the foramen magnum (ie extent beyond the posterior fossa)


Chang M Stage for Medulloblastoma
M Description
M0 No evidence of subarachnoid or hematogenous metastases
M1 Tumor cells found in CSF
M2 Intracranial tumor beyond the primary site
  • into aqueduct of Sylvius
  • in the third ventricle or foramen of Luschka or lateral ventricles
M3 Gross nodular subarachnoid spinal canal seeding
M4 Metastases outside the spinal axis, esp. bone marrow.

Risk Categories

Medulloblastoma has been divided into standard and high risk categories:

Treatment and Prognosis

The most important prognostic factor at diagnosis is the M stage. Advanced M stage (any M+) is a worse outcome and is considered a high risk disease. Other poor prognostic factors include male gender, age < 3 years old and unresectable disease or a subtotal resection. These are all considered high risk disease.

Standard Risk Disease

Standard risk disease (age > 3 years, M0 and GTR/NTR) is treated with the following management paradigm:
maximum safe surgical resection → Radiation concurrent with weekly vincristine → adjuvant chemotherapy of 8 six week cycles of CDDP/CCNU/vincristine.

Radiation therapy is given to the cranio-spinal axis to a total dose of 23.4 Gy followed by a boost to the posterior fossa of 36 Gy and then a second field size reduction to the cavity/residual disease volume or the posterior fossa to 55.8 Gy.

Recapping:

  1. Maximum safe surgical resection possible and then
  2. Radiation therapy concurrent with weekly vincristine to the cranio-spinal axis to a total dose of 23.4 Gy and then
  3. first field reduction to the posterior fossa: boost to 36 Gy and then
  4. second field reduction to the GTR/STR cavity of posterior fossa boost to 55.8 Gy and then
  5. adjuvant chemotherapy : 8 six week cycles of CDDP/CCNU/vincristine

Note, the CCG A9961 trial recently found similar outcomes when cyclophosphamide was substituted for CCNU (Packer, 2006 JCO).

There is no identified role for neoadjuvant chemotherapy in children > 3 years old. Neodadjuvant chemotherapy increases radiation toxicity, delays treatment and worsened relapse free survival based on data form the German HIT 91 trial (IJROBP, 2000).

High Risk Disease for age > 3 years old

The treatment for high risk disease (i.e. M+, unresectable or subtotal resection) for patients > 3 years is similar to standard risk, except that radiation doses are increased: CSI Dose is increased from 23.4 Gy to 36 Gy and nodular intracranial or spinal metastases are boosted to 39.6 - 50.4 Gy depending on location.

For high risk medulloblastoma patients with intracranial (M2 beyond primary site) or spinal (M3) disease the COG ACNS0332 study boosted intracranial mets to 50.4 Gy, focal spinal mets below the terminus of the cord to 50.4 Gy, nodular cord mets to 45 Gy and diffuse spinal cord mets to 39.6 Gy.

The EFS-5 for high risk medulloblastoma is 50% - 60%. For standard risk medulloblastoma 80% .

Medulloblastomas in those < 3 years old

The initial goal in treating infants < 3 years old is to delay radiation therapy as long as possible until the infant reaches at least 3 years old. Studies have shown IQ deterioration is significantly higher per year for those irradiated at very young ages than those over 7 years old. The general treatment paradigm is:

  1. Maximum safe surgical resection
  2. Chemotherapy until the child reaches 3 years old
  3. At 3 years old, consider CSI followed by more chemotherapy.
  4. If desmoplastic histology, consider omitting CSI altogether.

Delaying radiation therapy in infants < 3 years old is desireable due to its high toxicity to the developing brain. The Baby POG (Duffner 1993 NEJM) study demonstrated a means to make delay of radiation therapy reasonable, especially with desmoplastic histology until age > 3 years. The Baby POG study enrolled infants < 3 years old. Baby POG are the best data to date. (POG 8633/8634). The Baby POG Study

POG 9233/9234 failed to improve on Baby POG.

P9934 used systemic chemotherapy with second look surgery and IFRT for children ≥ 8 months and ≤ 36 months with non-metastatic medulloblastoma (M0). This trial is open, active and acruing. It is based on 2 French trials that showed that locally recurrent medulloblastoma can be salvaged with high dose chemotherapy and radiation therapy and the primary failure risk in these patients appears to be local failure.

COG 9934 Regimen is an age/risk/response adapted radiation therapy. The paradigm is: Initial surgery → induction chemotherapy x 4 months → second look surgery for identifiable/residual disease → age/risk/response based conformal radiation therapy to the posterior fossa only + primary site (NO CSI) → maintenance chemotherapy x 8 months.



COG A9934 for Medulloblastoma patients < 3 years old:

  1. Initial surgery →
  2. Induction Chemotherapy x 4 months →
  3. second look surgery for identifiable or residual disease →
    4. age/risk group adapted conformal radiation therapy to the posterior fossa + primary site (no CSI) →
    • < 24 months and Complete Response: 18 Gy to PF → boost tumor bed to 50.4 Gy or 54 Gy if PR/SD/residual disease.
    • > 24 months and CR or PR: 23.4 Gy to PF → boost tumor bed to 54 Gy.
  4. maintenance chemotherapy for 8 months.

A German baby POG style trial enrolled 43 patients < 3 years old and treated with cytoxan, vincristine, MTX, carboplatin, vp16, intrathecal MTX. PFS-5 was 58% and OS-5 was 66%. The majority were desmoplastic variant histology. The best benefit was in M0 disease.

French data (Grill, 2005, Lancet) looked at 79 patients < 5 years old and found OS-5 was best in R0M0 at 73% over 13% in those with M+ disease.

Dose

Retrospective data supports a radiation dose > 50 Gy. Local control in the posterior fossa follows a dose response above and below 50 Gy. Hughes reviewed 60 medulloblastoma cases and found that if the poterior fossa dose was < 50 Gy, local control was only 33%. If the dose was above 50 Gy, local control was 79%.

There is retrospective evidence which suggests that few failures occur in the posterior fossa outside of the tumor bed (< 5%). Fukunaga-Johnson reviewed 114 patients treated with CSI → boost to the entire posterior fossa. the solitary site of first failure within the posterior fossa but outside of the tumor bed occured in only 1/27 failures. Wolden also examined retrospective data which examined 32 patients treated with tumor bed boost only. There were 6 total failures, 5 outside the posterior fossa and one in the posterior fossa but outside of the boost volume.

COG ACNS0331 Reduced Dose Standard Risk age 3 - 7 18 Gy v. 23.4 Gy CSI. ACNS examined radiation therapy dose and extent of irradiation. This study enrolled standard risk patients (M0, GTR/NTR) and randomized to 18 Gy CSI or 23.4 Gy CSI.

Proton Beam Therapy

There is some dosimetry information that protons may imply better sparing of the cochlea and temporal lobe which is of uncertain benefit.

Altered Dose/Fractions

MSFOP 98 was a phase II trial examining hyperfractonated radiation therapy using standard risk patients treated with CSI radiationat 1 Gy BID to 36 Gy. This was followed by a tumor bed boost to 68 Gy at 1 Gy BID. OS-6 was 78% and EFS was 75% withno apparent decline in IQ compared with historical controls.

Technieques

Pediatric CSI patients are treated prone with next extended to avoid exit dose through the mouth, with shoulders inferior to allow for lateral cranial fields. Spinal fields are positioned first to allow collimator angle calculations to account for beam divergence based on spinal field lengths.

Once spinal fields are established the collimator of the lateral fields must be rotated to account for PA beam divergence from the spinal fields. The initial collimator angle is set

to match the PA spinal beam divergence. The couch must be rotated to match lateral field divergence with PA spinal field. The desired couch rotation is toward the gantry :

For spinal field borders, the superior is matched to the lateral cranial fields, the upper (cranial) spinal field's distal border and the lower spinal field's superior border must have a skin gap calculation:

d is the depth of the match point, usually the anterior cord edge. The match junctions should be feathered to insure even dose and to smooth out overlap-underlap regions. Typical feathering with a 0.5 cm gap shifts every 900 cGy. This helps reduce hot and cold spots at the field juntions. The feathering technique uses an isocenter shift on the PA spine fields with a corresponding collimator adjustment on the cranial field inferior jaw. For abutting PA spine fields, the iso center shifts and the inferior jaw moves.

The cranial isocenter is placed posterior to the eye lens to minimize divergence into the contralateral lens. Some recommend contouring the cribriform plate on imaging with a generous margin given in this area. If the entire spinal field is close but cannot be fully included, it may be possible to rotate the spinal PA field collimator to take advantage of the increased diagonal length of the collimator. If the spinal length is > 38 cm, it is likely on most accelerators that 2 fields will be required and a match/skin gap must be calculated. If two fields are required, match at L1-L2 because this region is where depth changes the most.

Dosimetry calculation points should be generated for the superior spinal PA field at the posterior edge of the mid-thoracic vertebral body, In adults this depth is generally 5 cm more or less. In the lower spinal field, the calc point is theposterior edge of mid-field vertebral body, typically 8 cm in adults.

Lateral field edges should encompass the entire vertebral body with 1 cm margin. Place the superior aspect of the field as low as possible (C4-C5) to spare the larynx and avoid divergence into the contralateral mandible as much as possible, but the shoulders need to be clear of the lateral fields. At the inferior border, a "spade" is used to insure coverage of the thecal sac and to cover the lateral sacral foramina

Follow up

"Collins" law defines the period of risk for recurrence which is defined as age at diagnosis + 9 months. I have no idea where they dug this gem up, but "Nevertheless it still moves.." G. Galeleo

Treatment Toxicity

Pediatric late effects are important considerations. The baby POG study of surgery with adjuvant chemotherapy to delay the initiation of radiation therapy is for good reason. Radiation in very young infants carries significant late effects.

Treatment toxicities of CSI and boost to the posterior fossa include growth retardation due to decreased growth hormone production, decreased IQ, ototoxicity, hypopituitarism and second malignancy. The factors that predict for a decline in IQ after CSI include:

Hoppe-Hirsh reviewed 120 medulloblastoma patients treated with CSI to 36 Gy. They found that IQ can continue to decline beyond 5 years. At 5 years, 58% had an IQ > 80, at 10 years only 1& had an IQ > 80. (1990)

The annual IQ drop depends on the age at irradiation and is 5 points/year if age < 7 years, and 1 point/year if age > 7 years. The dose to the supratentorial brain (ie temporal lobes) is the most important dose

Age at treatment with radiation therapy appears to be the most important influencing factor. Volume and dose of radiation and female gender also influence IQ scores and associated neurotoxicity.

To avoid significant injury to the cochlea, the V30 should be < 50%. The maximum dose is 35 Gy with chemotherapy.

Growth hormone deficiency is the most common hormonal deficiency and is affected by doses of approximately 10 Gy