Pediatric Head and Neck Malignancies

Abstract

• Head and neck malignancies account for approximately 5% of all pediatric tumors.

• Although relatively rare, pediatric head and neck malignancies are challenging from a treatment-planning standpoint given proximity to critical structures and developing normal tissue.

• This chapter provides practical contouring guidelines for common pediatric head and neck tumors including malignant tumors such as rhabdomyosarcoma, Ewing sarcoma, salivary gland tumors, and the benign, but locally aggressive, nasopharyngeal angiofibroma.

• Rhabdomyosarcoma and Ewing sarcoma are covered more broadly in Chaps. 8 and 9.

15.1 Background

• Head and neck malignancies account for approximately 5% of all pediatric tumors.

• Although relatively rare, pediatric head and neck malignancies are challenging from a treatment-planning standpoint given proximity to critical structures and developing normal tissue.

• This chapter provides practical contouring guidelines for common pediatric head and neck tumors including malignant tumors such as rhabdomyosarcoma, Ewing sarcoma, salivary gland tumors, and the benign, but locally aggressive, nasopharyngeal angiofibroma.

• Rhabdomyosarcoma and Ewing sarcoma are covered more broadly in Chaps. 8 and 9.

15.2 Rhabdomyosarcoma

15.2.1 Background

• Rhabdomyosarcoma (RMS) accounts for 350 cases of childhood cancer in the USA each year, comprising the most common soft tissue sarcoma in this patient population [1].

• RMS of the head and neck accounts for approximately 25% of all RMS cases, with parameningeal sites being more commonly involved [2].

• RMS of the orbit accounts for approximately 10% of all RMS cases [2].

15.2.2 Histology

• Favorable histologies include embryonal RMS and botryoid or spindle cell variants of embryonal RMS, while alveolar and undifferentiated RMS are unfavorable histologies [2, 3].

• Embryonal histology is the most common histology found in the head and neck region [2, 3].

15.2.3 Staging

• RMS is subclassified into prognostic categories for low-, intermediate-, and high-risk patients based on the probability of treatment failure. Prognostic categories are determined by a combination of histology, clinical group, and stage ([4]; Table 15.1).

Table 15.1 Rhabdomyosarcoma risk stratification

15.2.4 Anatomy and Patterns of Spread

• RMS of the head and neck include parameningeal, non-parameningeal, and orbit.

• Parameningeal sites include the infratemporal fossa, middle ear, mastoid, nasal cavity, nasopharynx, paranasal sinuses, parapharyngeal space, and pterygopalatine fossa.

• The incidence of lymph node metastasis is negligible for orbital primaries [5].

• Well-lateralized primary sites of other regions of the head and neck may have involvement of the ipsilateral jugular cervical, preauricular, occipital, or supraclavicular lymph nodes, while central primary sites may have bilateral involvement.

• The incidence of lymph node metastasis has been reported between 7 and 20% for non-orbital RMS head and neck primary sites [5,17,3,6,8,7].

• The majority of patients with parameningeal RMS primary sites present with either cranial base bony erosion or cranial nerve palsy due to direct local extension, and approximately 25–50% will have intracranial extension [8].

15.2.5 Imaging for Radiotherapy Volume Delineation

• At initial presentation, the patient should undergo imaging with CT with IV contrast and preferably an MR of the head and neck region. PET/CT is recommended but not required.

• In addition, parameningeal sites require an MR brain to assess for intracranial extension and base of skull erosion [8].

15.2.6 Assessment of Lymph Node Status

• FDG PET is recommended for the assessment of lymph node involvement as it has been demonstrated to have sensitivity of 94% and specificity of 100% for detecting nodal disease, in addition to improving the detection of bone, bone marrow, and soft tissue metastases [9].

• FDG PET-detected abnormalities should be confirmed by an additional imaging modality (MR or CT) or through pathological confirmation.

• Patients with head and neck primary sites do not require prophylactic cervical neck lymph node dissection.

• Sampling through an open biopsy is recommended, and in some situations through a needle aspiration, for all clinically or radiographically enlarged lymph nodes.

15.2.7 Treatment Strategy

• Induction chemotherapy followed by concurrent chemoradiotherapy is the current standard of care for any patient with unresectable disease (group III), positive margins, or lymph node involvement (group II) and all patients with alveolar histology based on the improvement in progression-free survival that has been demonstrated with the use of radiotherapy in these groups [10, 11].

• Outcomes for orbital RMS are excellent with the use of definitive radiotherapy and chemotherapy; thus, biopsy alone generally followed by chemotherapy and radiation is the standard of care for orbital stage I, group III disease.

• For RMS non-parameningeal primaries, wide excision is appropriate when feasible, but the possibility of achieving wide margins is generally restricted to those patients with relatively superficial lesions.

• Craniofacial resection for anterior RMS skull-base tumors of the nasal areas, paranasal sinuses, temporal fossa, and other such sites should be reserved to those surgical teams expert in its performance. These regions will generally be unresectable, and radiotherapy will likely be the primary form of local control [12, 13].

15.2.8 Indications for Radiotherapy

• Radiotherapy timing has varied among cooperative group protocols. No benefit has been demonstrated with the early initiation of radiotherapy and beginning treatment at approximately weeks 9–13 after the initiation of chemotherapy remains the accepted standard for most patients.

• Early radiotherapy for cranial nerve involvement, base of skull involvement, or intracranial extension is no longer recommended, and emergent radiotherapy is reserved for neurologic symptoms such as vision loss or symptomatic cord compression occurring secondary to tumor compression [14].

• Administration of radiosensitizing agents dactinomycin and doxorubicin is generally avoided during radiotherapy.

• Radiotherapy dose varies with surgical group status (Table 15.2).

Table 15.2 Radiotherapy dose according to histology, clinical group, and site

15.2.9 Simulation

• The patient should be simulated in the supine position with hyperextension of the neck and the use of a long thermoplastic mask extending to the shoulder region for immobilization.

• An anesthesiologist may be necessary for sedation of young children.

• A 3D simulation should be performed. The use of IV contrast and ≤5-mm-thick slices is recommended.

• Use of a bite block may be considered for sparing dose to the uninvolved mucosal surfaces of the oral cavity.

• For orbital primaries, the patient may be asked to keep a steady focused gaze on an external object during treatment in the direction that maximizes treatment to the tumor and reduces dose to the cornea, lens, and/or lacrimal gland.

15.2.10 Target Volume Delineation

• All radiotherapy volumes will be at least initially targeting the extent of disease prior to surgical resection and chemotherapy; thus, fusion of CT simulation imaging with imaging studies at presentation including a CT with IV contrast and MR T1 post-contrast-/T2-weighted imaging is imperative for the delineation of the initial gross tumor volume (GTV). Fusion with PET, if available, may also be helpful. Fusion may be limited by differences in neck extension.

• If the tumor has responded to chemotherapy and the normal tissues have returned to their normal positions, the GTV excludes the volume that extends into the normal tissue or cavity. The GTV must include all infiltrative disease detected at initial presentation.

• The clinical tumor volume (CTV) is defined as GTV+ a 1 cm margin in current Children’s Oncology Group (COG) protocols. The CTV does not extend outside patient and is anatomically constrained by structures that serve as a barrier to spread (i.e., uninvolved bone), and thus, smaller margins may be used in head and neck primary sites.

• In situations where a pushing rather than infiltrative margin by the tumor is present, reduced radiotherapy volumes can be planned after 36Gy and can be considered for patients whose total dose will be 50.4 Gy. Boost volumes may be planned to encompass any visible or palpable residual tumor at the time of radiotherapy planning as assessed by CT, MR, PET, or physical exam. All sites of infiltrative disease detected at initial presentation (i.e., bone involvement) should be included in the boost volume. In the case of lymph node involvement, the involved lymph node region should be targeted in the CTV, and current COG protocols recommended inclusion of the entire cervical chain.

• In general, elective nodal irradiation is not recommended.

• For surgically resected group II disease, the GTV should include the preoperative tumor volume or involved lymph node confined to the operative bed, excluding normal tissues that may have shifted after surgery. The CTV should include GTV + 1 cm anatomically confined margin and the entire operative bed or involved lymph node region. There are no dose reductions for group II disease.

• Examples of target volume delineation for a stage I, group III orbital RMS case (Fig. 15.1); stage III, group III nasopharynx RMS case with lymph node involvement (Fig. 15.2); and stage III, group III maxillary sinus RMS with no lymph node involvement (Fig. 15.3) are shown.

Fig. 15.1

Axial CT simulation planning images (column a) and corresponding pre-chemotherapy MR T1 and T2 imaging (columns b and c, respectively) for a patient with a right-sided orbital embryonal rhabdomyosarcoma stage I, group III, with a good partial response to VAC alternating with VA chemotherapy. Local control with radiotherapy took place at week 12. Pre-chemotherapy MR T1 pre-contrast imaging was chosen in this particular case given better contrast between tumor and orbital fat and, along with MR T2 imaging, was fused with CT planning images. Note that planning accuracy can be limited by the quality of fusion between components. The initial tumor had intraconal and extraconal components and was situated between the superior rectus muscle, optic nerve, and superior oblique. There was involvement of the nasolacrimal duct. There was no intracranial involvement or extension medially into the ethmoid sinus. The extent of tumor on pre-chemotherapy T1/T2 imaging is delineated to form GTV1 (light blue) and CTV1 (dark blue), which does not include the entire orbit but only tumor volume with a margin. The CTV1 included areas at risk of microscopic spread and does not extend outside of the bony orbit when there is no bone erosion present and was contoured on each slice separately taking into consideration anatomic constraints. Tumor present at the time of simulation was delineated to form a GTV2 (light red). A CTV2 (not shown) was delineated and, in this case, represented the extent of pre-chemotherapy disease and was equal to GTV1. Please note that the CTV to PTV margins depend on patient- and institution-dependent factors, and therefore, recommendations for uniform expansions cannot be made. In the case of treatment with proton therapy, no PTV margin is applied. Please note that these are representative slices and not all slices have been included. The patient was asked to keep her eyes slightly open to look to the patient’s right-hand side in order to shield the lens and cornea more adequately. In orbital cases, every effort should be made to shield the lens, cornea, and lacrimal gland appropriately

Fig. 15.2

Axial CT simulation planning images (columns a and b) and corresponding pre-chemotherapy MR T1 post-contrast imaging (column c) for a patient with embryonal rhabdomyosarcoma of the nasopharynx stage III, group III with involvement of retropharyngeal and bilateral level II lymph nodes with a near-complete response to VAC alternating with VI chemotherapy. Local control with radiotherapy took place at week 12. MR T1 post-contrast was fused with CT planning images. MR T2 imaging can also be helpful for tumor delineation but is not included here. Note that planning accuracy can be limited by the quality of fusion between components. The initial tumor involved the bilateral nasopharynx and parapharyngeal space with involvement of both right medial and lateral pterygoid plates with extension anteriorly to the nasal cavity and inferiorly to the posterior oral cavity and oropharynx. CT imaging bone windows showed involvement of the hard palate. PET showed avidity of a right-sided retropharyngeal lymph node. The extent of tumor on pre-chemotherapy T1/T2 imaging is delineated to form GTV1 (light blue) and CTV1 (dark blue). The GTV1 excludes normal tissue that returns to normal locations following chemotherapy. Small additional bilateral retropharyngeal lymph nodes and cervical lymph nodes felt to be suspicious for involvement were included in the treatment volume; thus, the entire bilateral retropharyngeal and level II lymph nodes were included in CTV1. The extent of post-chemotherapy residual disease, confined to the right nasopharynx, was delineated to form GTV2 (red) and CTV2 (dark red). In this case, CTV1 was treated to 41.4Gy (RBE) and CTV2 was boosted to a total of 50.4Gy (RBE). Please note that the CTV to PTV margins depend on patient- and institution-dependent factors, and therefore recommendations for uniform expansions cannot be made. In the case of treatment with proton therapy, no PTV margin is applied. Please note that these are representative slices and not all slices have been included. The patient was treated with a bite block to decrease dose to uninvolved mucosal surfaces of the oral cavity

Fig. 15.3

Axial CT simulation planning images (left column) and corresponding initial diagnostic T1 post-contrast fat-suppressed MRI sequence MRI (middle column) and post-induction chemotherapy post-contrast MRI at the time of treatment planning (right column) for a stage III, group III, alveolar RMS of the left maxillary sinus. Initial diagnostic MRI revealed a heterogeneously enhancing expansile mass centered in the left maxillary sinus invading into the left nasal cavity, left pterygopalatine fossa, left orbit through the orbital floor and medial orbital wall, left ethmoid sinus and left frontal sinus with a focal area of dehiscence through the left ethmoid air cells, and cribriform plate suggestive of intracranial extension. There was no evidence of regional lymphadenopathy by MRI or CT with contrast. Biopsy revealed alveolar rhabdomyosarcoma. The patient was treated with VAC/IE chemotherapy. After 9 weeks of chemotherapy, MRI was repeated (right) and revealed significant interval decrease in size of the soft tissue lining the expanded maxillary sinus with resolution of the previous mass effect on the orbit and no further evidence of intraorbital or intracranial extension. The pre-chemotherapy GTV (light blue) was defined on the initial diagnostic MRI (middle). The CTV (dark blue) is defined by a 5–10 mm expansion from the GTV and was anatomically confined by anatomic borders or potential spread and cropped off of normal critical structures that have shifted in place due to decrease mass effect of the tumor after chemotherapy (e.g., the eye). An additional PTV margin of 3 mm was used (not shown) in this case with daily AP and lateral kV on-board imaging for IGRT. This initial tumor volume was treated to 36 Gy in 1.8 Gy per fraction. The post-induction chemotherapy tumor volume (red) was defined on the T1 post-contrast MRI sequence obtained after 9 weeks of chemotherapy (right), and similar CTV margins were used to create the post-induction chemotherapy CTV (dark red). Because the tumor was locally invasive, despite a good response to chemotherapy, the post-induction chemotherapy CTV is only slightly smaller than the initial CTV and is particularly reduced in size where the initial tumor was pushing rather than invasive, such as into the orbit, and all sites of initial gross tumor volume remained covered. Where the post-induction chemotherapy CTV is not shown, it is the same as the preinduction chemotherapy CTV. This volume received an additional 14.4 Gy in 1.8 Gy per fraction to bring the cumulative total dose to 50.4 Gy

15.2.11 Treatment Planning

• For orbital primaries, radiotherapy techniques include the use of 3D conformal radiotherapy, IMRT/VMAT, or proton beam therapy [15]. A wedge pair may provide adequate conformality with well-lateralized tumors.

• For other head and neck primary sites, radiotherapy techniques also include IMRT/VMAT or protons [16].

• There is no universally applicable planning target volume (PTV) margin, since this will depend on institution practice and the use of image-guided radiotherapy (IGRT).

• When using IMRT or VMAT, it is recommended that the PTV be cropped 3–4 mm away from the skin when uninvolved to avoid excess dose to the skin surface. If target volumes do include the overlying skin, bolus may be required.

• In the case of protons, PTV varies with each individual field and coverage and will require additional adjustment to the lateral margins, smearing of compensator, range of beam (depth of penetration), and modulation (number of required Bragg peaks). Adjustments to any of these will be based on the setup error determined for the particular body site as the individual proton institution.

• In the case of protons, a single beam that stops within a critical organ should not be used.

• Organs within the irradiated volume and organs at risk should be contoured.

• A DVH should be prepared to determine target coverage and to evaluate dose to normal tissue.

• Relevant organs at risk include the spinal cord, optic chiasm, optic nerves and lens/cornea, lacrimal gland, cochlea, brainstem, hypothalamus and pituitary gland, temporal lobes, oral cavity (when uninvolved), parotid and submandibular glands, temporomandibular joint, esophagus, and thyroid gland. Normal tissue constraints for head and neck OARs can be found by references such as QUANTEC [17]. See Table 15.3 for a representative set of dose constraint guidelines included in national cooperative group protocols.

Table 15.3 Pediatric head and neck organs at risk dose recommendations

15.2.12 Outcomes

• Long-term failure-free survival for low- and intermediate-risk RMS using multimodality therapy is 85% and 65%, respectively [2].

• Orbital primaries have 98% and 90% rates of local control and overall survival, respectively [13].

• Non-parameningeal head and neck primaries have 90% and 80% rates of local control and overall survival, respectively [2].

• Parameningeal head and neck primaries generally fare the worst with 85% and 75% rates of local control and overall survival, respectively [12].

15.3 Ewing Sarcoma

15.3.1 Background

• Ewing sarcoma (EWS) accounts for 200 cases of childhood cancer in the USA each year, comprising the second most common pediatric bone tumor following osteosarcoma [1].

• EWS can present with osseous or extraosseous involvement.

• Primary EWS of the head and neck region is extremely rare.

• Skull-based EWS comprises 2% of all osseous primaries. Extraosseous EWS of the head and neck accounts for 18% of all extraosseous primaries [18].

15.3.2 Histology and Cellular Classification

• The Ewing family of tumors include Ewing sarcoma and primary neuroectodermal tumors (PNET). Both EWS and PNET are small round blue cell tumors and are treated in the same manner.

• Presence of the translocation involving the EWSR1 gene on chromosome 22 band q12 and any number of partner chromosomes is the hallmark feature in the diagnosis of EWS [19].

• 95% of EWS stain positive for CD99 [20].

15.3.3 Prognosis

• There is no official staging system for EWS, but prognostic factors include size, age, gender, elevated LDH levels [21], presence of metastases [22], site [23], and poor response to chemotherapy [24].

15.3.4 Anatomy and Patterns of Spread

• Common skeletal EWS primary sites of the head and neck include the calvarium, maxilla, and mandible.

• Extraskeletal EWS primary sites of the head and neck may include the paranasal sinuses, oral cavity, and soft tissue of the neck.

• Patients with extraskeletal primary EWS are more likely to have lymph node involvement as compared to patients with skeletal primary EWS.

• The incidence of regional lymph node involvement has been estimated at 12% versus 3% for patient with extraskeletal and skeletal EWS primary sites, respectively [25].

15.3.5 Imaging for Radiotherapy Volume Definition

• At initial presentation, the patient should undergo imaging with a CT with IV contrast, MR of the head and neck region, and bone scan. In addition, PET/CT may be a valuable staging study in Ewing sarcoma for detecting lymph node involvement and metastatic disease and can modify treatment strategy and planning.

• EWS enhances on MR T1 imaging.

• At the completion of induction chemotherapy, the patient should be restaged in a similar manner as was performed at initial presentation. These may be important studies for guiding delineation of boost volumes, if utilized.

15.3.6 Assessment of Lymph Node Status

• Patients with clinically or radiographically suspicious-appearing lymph nodes should preferably undergo open biopsy or, in certain situations, needle biopsy or fine needle aspiration. Elective nodal sampling is generally not recommended.

15.3.7 Indications for Radiotherapy

• Although surgery is the preferred local therapy in patients with Ewing sarcoma, approximately 35% of patients receive radiotherapy as definitive or adjuvant treatment. Local control should aim to eradicate local tumor while best preserving function.

• Local control takes place at week 13, after recovery from the sixth cycle of induction chemotherapy. Timing is the same for definitive or preoperative radiotherapy.

• Radiotherapy is delivered concurrently with consolidation chemotherapy. Adriamycin is held during radiotherapy.

• Definitive radiotherapy may be required for skull-based or facial bone primaries where functional impairment by surgery would be high, but consultation with a surgical oncologist or neurosurgeon experienced in tumors of these sites should be made.

• Preoperative radiotherapy may be given if it is felt to help achieve negative margins with surgical resection. Surgery is generally pursued within 2 weeks of completion of radiotherapy.

• If adjuvant radiotherapy is given, as is the case of microscopic close or residual margins, gross residual disease, or intraoperative tumor spillage, it is scheduled as soon as recovery from surgery permits. It can also be considered in cases of poor tumor response [26].

15.3.8 Simulation

• Please refer to Sect. 2.9 for a discussion of simulation.

15.3.9 Target Volume Delineation

• All radiotherapy volumes will initially target the pre-chemotherapy extent of bone and/or soft tissue disease; thus, fusion of CT simulation imaging with imaging studies at presentation including CT with IV contrast (± bone window) and MR T1 post-contrast-/T2-weighted imaging is imperative. Fusion with PET, if available, can also be helpful. Fusion may be limited by differences in neck extension.

• If the tumor has responded to chemotherapy and the normal tissues have returned to their normal positions, the GTV1 excludes the soft tissue volume that extends into the cavity.

• CTV = GTV + 1 cm but does not extend outside patient and is modified to account for specific anatomic barriers to tumor spread.

• In the case of clinically or pathologically involved lymph nodes, the entire cervical chain should be included in the CTV1. In general, elective nodal irradiation is not recommended.

• In the case of bone involvement, radiotherapy targeting the tumor plus a margin is recommended (i.e., radiotherapy to the entire bone is not necessary).

• GTV2, when implemented, includes delineation of the pre-chemotherapy extent of bone and post-chemotherapy extent of soft tissue disease.

• In the case of an EWS extraosseous head and neck primary with a complete response to chemotherapy or following gross total resection, there will only be a single GTV.

• See Table 15.4 for radiotherapy dosing and target volumes.

• An example of EWS of the nasal cavity (Figure 15.4) is shown.

Table 15.4 Ewing sarcoma radiation dose guidelines

Fig. 15.4

Axial CT simulation planning images (column a) and corresponding pre-chemotherapy MR T1 post-contrast and T2 imaging (columns b and c, respectively) for a patient with an extraosseous Ewing sarcoma of the nasal cavity status post chemotherapy and gross total resection. MR T1 post-contrast and T2 imaging was fused with CT planning images. Note that planning accuracy can be limited by the quality of fusion between components. The initial tumor involved the nasal cavity and bilateral ethmoid air cells, bilateral sphenoid sinuses, bilateral orbit, and left maxillary sinus. There was also intracranial extension with involvement of the dura. The extent of soft tissue tumor on pretreatment T1/T2 imaging is delineated to form GTV (light blue) and CTV1 (dark blue), excluding tissue that returned to normal after treatment. Given gross total resection, there was only a single GTV treated to a total dose of 50.4Gy (RBE). Please note that the CTV to PTV margins depend on patient- and institution-dependent factors, and therefore recommendations for uniform expansions cannot be made. In the case of treatment with proton therapy, no PTV margin is applied. Please note that these are representative slices and not all slices have been included

15.3.10 Treatment Planning

• Please refer to Sect. 2.11 for a discussion of treatment planning.

• Relevant OARs include the spinal cord, optic chiasm, optic nerves, lens/cornea, and cochlea. Normal tissue constraints for head and neck OARs can be found by references such as QUANTEC [17]. See Table 15.3 for a representative set of dose constraint guidelines included in national cooperative group protocols.

15.3.11 Outcomes

• Long-term event-free survival and overall survival for nonmetastatic EWS are 75% and 90%, respectively, and are similar for outcomes for EWS of the head and neck region [27].

15.4 Salivary Gland Tumors

15.4.1 Background

• Malignant salivary gland tumors are rare in the pediatric population, comprising approximately 10% of all pediatric head and neck neoplasms [28, 29].

• The majority of salivary gland tumors are located in the parotid gland with approximately 15% occurring in the submandibular gland or minor salivary glands.

15.4.2 Histology

• The most common histology for malignant salivary gland tumors in children is mucoepidermoid carcinoma followed by acinic cell carcinoma and adenoid cystic carcinoma [29].

• Pleomorphic adenoma is a common benign salivary gland tumor.

15.4.3 Anatomy and Patterns of Spread

• Salivary gland tumors carry a high risk for metastasis to the cervical lymph node chain. Lymph node involvement at diagnosis is present in approximately 25–40% of patients and varies with location and histology; the highest rates are typically seen for mucoepidermoid carcinoma and parotid gland primary site [29, 30].

• Salivary gland tumors also commonly demonstrate perineural involvement and may spread along cranial nerves to the base of the skull.

15.4.4 Imaging for Radiotherapy Volume Definition

• CT with contrast is essential for anatomical imaging of the primary tumor.

• PET/CT and MRI provide additional value if local regional and systemic staging and evaluation of perineural tumor spread.

15.4.5 Treatment Strategy

• The primary treatment for salivary gland tumors is surgical resection and selective neck dissection under the care of a head and neck surgeon.

• Adjuvant radiation therapy is indicated for high-risk features such as positive surgical margins, perineural invasion, lymphovascular space invasion, advanced tumor stage, high-grade histology, and the presence of lymph node metastases as adjuvant radiation has been demonstrated to significantly improve local control for patients with these features [29, 30].

• Elective nodal irradiation, as a component of adjuvant therapy even for those with pathologically negative lymph nodes, has been demonstrated to reduce the risk of nodal relapse for patients with advanced T-stage disease and high-grade histologies such as mucoepidermoid carcinoma, squamous cell carcinoma, adenocarcinoma, and undifferentiated carcinoma [31].

15.4.6 Simulation

• Please refer to Sect. 2.8 for a discussion of simulation.

15.4.7 Target Volume Delineation

• Fusion of the preoperative diagnostic CT, PET, and/or MRI to the CT simulation for identification of the preoperative tumor volume and any clinically involved lymph nodes prior to surgery is recommended.

• The GTV should include the gross tumor volume as defined on preoperative imaging that is confined to the operative bed and excludes normal tissues that may have shifted during surgery.

• The CTV should include the GTV with a 1 cm anatomically confined margin limited by anatomical barriers of spread and not extending outside of the patient but including the entire operative bed, cervical lymph node regions pathologically involved, and sites of involved cranial nerves.

• Doses used typically follow that used in the adult population with 60 Gy in 1.8–2 Gy per day prescribed to the high-risk CTV as outlined above, while higher doses of 64.8–66 Gy are reserved for positive margins or nodal regions at sites of pathologic extranodal extension.

• A lower-dose volume prescribed to 50–56 Gy in 1.6–2 Gy per fraction should include regions of elective nodal irradiation, surgical scar, other areas adjacent to the primary tumor considered to be at risk for spread, and named cranial nerves at risk for perineural tumor spread.

• An example of adenoid cystic carcinoma of the parotid gland (Figure 15.5) is shown.

• Elective nodal irradiation (ENI):

  • For parotid primary tumors, cervical lymph node levels II–III are at greatest risk for subclinical involvement and should be targeted when elective nodal irradiation is used. Level IV is also often included in ENI for adult patients [32], though risk of isolated nodal relapse in level IV is low [31], and so the risk of the added morbidity of extending the volume in young children should be weighed in consideration of the patients’ unique disease characteristics. Level Ib, IV, and V and the retrostyloid space should be considered for inclusion in ENI target volumes in patients who are lymph node positive.

  • For submandibular or sublingual primary tumors, lymph node regions at risk are levels IB–III (± IV per above) with inclusion of the contralateral level I if the ipsilateral level I is involved and the retrostyloid space in lymph node positive patients.

  • Contouring guidelines for CTV definition of LN volumes can be found in the following references [33, 34].

• Perineural Coverage

  • Elective perineural coverage is recommended when there is gross involvement of a named cranial nerve and may also be of benefit in patients’ microscopic perineural tumor spread and/or for tumors with high risk of perineural spread such as adenoid cystic carcinoma.

  • Contouring guidelines for CTV definition of nerves at risk for perineural tumor spread can be found in the following references [34, 35].

  • For tumors of the parotid gland, perineural tumor coverage should include the facial nerve to the stylomastoid foramen or through the petrous portion of the temporal bone if known to be involved. There is also risk for spread along the auriculotemporal nerve (a branch of V₃).

  • For tumors of the submandibular gland, perineural tumor spread may occur along the lingual nerve (branch of V3), the chorda tympani of the facial nerve (which eventually joins V3), or the hypoglossal nerve.

Fig. 15.5

Axial CT simulation with contrast (left) and treatment planning MRI T1 with contrast (middle) and T2-weighted (right) sequences for a teenager undergoing adjuvant radiation therapy for a resected adenoid cystic carcinoma of the parotid gland. At the time of surgery, there was evidence of close margins at less than 1 mm and no lymph node involvement. No high-quality preoperative imaging was available for treatment planning purposes, so postoperative imaging was utilized. The GTV (light blue) is defined as the residual soft tissue and the parotid bed. The CTV (dark blue) is defined by a 1 cm anatomically confined margin. In this case, there was no elective nodal irradiation, but because adenoid cystic histology has a propensity for perineural spread, cranial nerves at risk for perineural spread were included in the initial target volume. The facial nerve (fuschia) is contoured from the petrous portion of the temporal bone through the stylomastoid foramen and into the parotid bed. The mandibular nerve is contoured from the foramen ovale, coursing between the lateral and medial pterygoid muscles, behind the condyle of the mandible and into the parotid bed. The greater auricular nerve (green), which arises from the C2/C3 cervical roots to course around the sternocleidomastoid muscle and arises along the surface of the SCM beneath the platysma to innervate the parotid gland and the overlying skin, is also illustrated within the parotid bed. In this case, the cranial nerves and parotid bed CTV (dark blue) were treated to a dose of 54Gy (RBE) in 1.8Gy per fraction, and then the parotid bed CTV was treated to an additional 10.8Gy (RBE) to bring the cumulative dose to 64.8Gy (RBE), all using 3D conformal passively scattered proton therapy

15.4.8 Treatment Planning

• Conformal radiotherapy techniques, including 3D conformal photon therapy, IMRT/VMAT, or proton therapy, may be used.

• For parotid bed treatment without elective nodal irradiation or perineural tumor coverage, a 3D conformal wedge pair technique may provide adequate conformality, though the use of more advanced techniques such as IMRT, VMAT, or protons is recommended in most clinical scenarios.

• Please refer to Sect. 2.11 for a detailed discussion of treatment planning.

15.4.9 Outcomes

• The prognosis for children with salivary gland tumors treated with surgical resection and adjuvant therapy as indicated is good, with reported overall survival rates of 95% at 10 years and greater than 80% at 20 years [29, 36].

15.5 Juvenile Nasopharyngeal Angiofibroma

15.5.1 Background

• Juvenile nasopharyngeal angiofibroma (JNA) is a rare, benign, highly vascularized tumor, which accounts for less than 0.5% of head and neck tumors [37].

• JNA almost exclusively presents in adolescent males and includes symptoms such as nasal obstruction, recurrent epistaxis, and a mass in the roof of the nasopharynx.

• JNA can be locally aggressive and associated with significant morbidity and occasional mortality due to intracranial extension and hemorrhage [38].

• Extranasopharyngeal angiofibroma is extremely rare, occurring more often in older females, and tends to be less vascular and aggressive than JNA [39].

15.5.2 Histology

• JNA is a histologically benign vascular malformation, usually encapsulated and composed of vascular tissue and fibrous stroma.

• JNA is comprised of vessels that are thin walled, lack elastic fibers, and have absent or incomplete smooth muscle, which are prone to hemorrhage.

15.5.3 Staging

• Staging takes into account extension outside of the nasopharynx and involvement of paranasal cavities and the base of skull structures (Table 15.5; [38]).

Table 15.5 Current staging system for juvenile nasopharyngeal angiofibroma (JNA)

15.5.4 Anatomy and Patterns of Spread

• JNA originates in close proximity to the posterior attachment of the middle turbinate near the superior border of the sphenopalatine foramen [40].

• Larger tumors extend beyond the nasopharynx and nasal cavities to intracranial structures by posterior and lateral routes [41].

• Intracranial extension occurs in approximately 20% of cases [41].

• The most common sites of intracranial invasion include the pituitary, anterior, and middle cranial fossa [41].

• Four routes by which intracranial extension occurs have been described and include [42, 43]:

  • Extension from the infratemporal fossa through the floor of the middle cranial fossa

  • Extension from the pterygomaxillary fissure and infratemporal fossa into the superior and inferior orbital fissures, where proptosis and optic nerve atrophy may occur

  • Direct erosion of the sphenoid sinus into the region of the sella turcica and cavernous sinus

  • Rarely, extension from the horizontal lamina of the ethmoids and cribriform plate into the anterior cranial fossa

• The base of the skull is the most frequent site of recurrence [38].

15.5.5 Imaging for Radiotherapy Volume Delineation

• At initial presentation, and following a through history and physical exam focusing on symptoms and cranial nerve deficits, the patient should undergo imaging with CT with IV contrast to identify the presence and extent of bone destruction.

• An MRI with and without contrast with T1 and T2 sequences should be performed to identify the tumor relationship to adjacent soft tissue structures, including the orbits, optic nerves, pterygomaxillary space, buccal and masticator space, carotid arteries, and anterior cranial fossa.

• A biopsy should be avoided unless there is evidence to question a diagnosis of JNA.

15.5.6 Treatment Strategy

• Surgery paired with preoperative embolization is the preferred management of JNA [44].

• There are a variety of surgical approaches available depending on the size and the extent of tumor involvement. An endoscopic approach is preferred for uncomplicated, noninvasive tumors [44, 45].

• Involvement of the base of the skull or intracranial structures may make gross total resection challenging and cause potential morbidity and mortality [41].

• Radiotherapy is a viable treatment strategy for patients who undergo subtotal resection or in the recurrent setting [41, 46, 47].

• Nontraditional treatment approaches include chemotherapy and hormonal manipulation and may be less efficacious than primary surgery or radiotherapy [48, 49].

15.5.7 Indications for Radiotherapy

• Radiotherapy as the primary treatment modality for advanced JNA has resulted in high rates of local control [41, 46, 50].

• Definitive radiotherapy doses have ranged from 30 to 46 Gy in recent series [46, 50].

15.5.8 Simulation

• Please refer to Sect. 2.9 for a discussion of simulation.

15.5.9 Target Volume Delineation

• If surgical resection was performed, radiotherapy volumes will be targeting the extent of disease prior to resection; thus, fusion of CT simulation imaging with imaging studies at presentation including a CT with IV contrast and MR T1 post-contrast-/T2-weighted imaging is imperative for the delineation of the initial gross tumor volume (GTV). Fusion may be limited by differences in neck extension.

• If the normal tissues have returned to their normal positions, the GTV excludes the volume that extends into the normal tissue or cavity. The GTV must include all infiltrative disease detected at initial presentation.

• While there is no agreed upon CTV margins available, given the benign nature of JNA, the clinical tumor volume (CTV) may be reduced to <1 cm, especially around critical structures. The CTV does not extend outside the patient and is anatomically constrained by structures that serve as a barrier to spread (i.e., uninvolved bone), and thus, smaller margins may be used in head and neck primary sites. If an operative bed is present, the CTV will include the entire operative bed.

• Elective nodal irradiation is not necessary.

• Examples of target volume delineation for a large advanced JNA are shown (Figure 15.6)

Fig. 15.6

Axial CT simulation planning images (column a) and corresponding pre-chemotherapy MR T1 post-contrast and T2 imaging (columns b and c, respectively) for a teenager with a large juvenile nasopharyngeal angiofibroma stage IIIA status post presurgical embolization and subtotal resection. Presurgical MR T1 post-contrast along with MR T2 imaging was fused with CT planning images. Note that planning accuracy can be limited by the quality of fusion between components. The initial tumor involved the nasopharynx and extended into the left nasal cavity, maxillary and sphenoid sinuses, as well as the left pterygopalatine fossa via the sphenopalatine foramen. There was an extension through the pharyngobasilar fascia to the left foramen lacerum and evidence of erosion of the anterior aspects of the base of skull with extension to the cavernous sinus and left vidian canal. Additionally, there was extension to the sphenoid sinus and floor of the sella turcica. There was post-obstructive opacification of much of the left paranasal cavities. Only subtotal resection was performed given proximity to the internal carotid artery. The extent of tumor on presurgery T1/T2 imaging is delineated to form GTV1 (light blue) and CTV1 (dark blue). The CTV1 was drawn on each slice to consider anatomic constraints, to avoid primarily inflammatory changes. CTV1 also includes the initial extension to the left pterygopalatine fossa. In this case, CTV1 was treated to a total of 41.4Gy (RBE) using passively scattered protons given the large tumor size, multiple treatment interruptions, and severity of bleeding. Given the treatment response and changes in tumor volume which can affect proton dosimetry at the soft tissue-air interface, the patient was re-simulated and replanned at several intervals during treatment. Please note that the CTV to PTV margins depend on patient- and institution-dependent factors, and therefore, recommendations for uniform expansions cannot be made. In the case of treatment with proton therapy, no PTV margin is applied. Please note that these are representative slices and not all slices have been included

15.5.10 Treatment Planning

• Please refer to Sect. 2.11 for a discussion of treatment planning.

• In the case of radiotherapy, IMRT or proton beam therapy is preferred in order to maximize target coverage while decreasing dose to surrounding normal tissue [50].

• Relevant OARs include the spinal cord, optic chiasm, optic nerves, lens/cornea, cochlea, hypothalamus, pituitary gland, temporal lobes, and oral cavity. Additional structures contoured as OARs will depend on extent of disease. Normal tissues constraints for head and neck OARs can be found by references such as QUANTEC [17].

15.5.11 Outcomes

• Patients undergoing primary surgical resection alone achieve a disease-free rate of approximately 85–90% [51, 52].

• Similar control rates have been seen in patients treated with primary radiotherapy [41, 46, 50].