39_02

For lesions involving the ethmoid sinuses or olfactory region, the CTV should also include the cribriform plate. A third CTV may be delineated to encompass the tract of cranial nerve V2 to the foramen rotundum if there is perineural invasion. For primary radiotherapy using IMRT, the CTV1, consisting of the gross tumor volume plus a margin of 1 to 2 cm, receives the full dose of 66 to 70 Gy. In patients receiving neoadjuvant chemotherapy, target volume definition is based on the extent of disease before chemotherapy.
For three-dimensional (3D) conformal radiotherapy, the initial target volume for postoperative radiotherapy consists of the surgical bed with 1- or 2-cm margins, depending on the surgical pathology findings and the proximity of critical structures. The boost volume consists of areas at greatest risk for recurrence, such as close or positive resection margins or regions of perineural invasion, with 1- to 2-cm margins.
For small anteroinferior septal lesions, brachytherapy can be accomplished by a single-plane implant of the lesion with 2-cm margins. Elective neck irradiation is not given routinely even in patients with large tumors or ENB.
Setup and Field Arrangement
For target volumes <5 cm deep, an electron technique similar to that described for nasal vestibule carcinomas is used. Treatment devices include lead skin collimation to obtain a sharp penumbra as well as bolus material in the nasal cavity, postoperative defects, and on skin scars. An intraoral stent is used to depress the tongue, provide a patent airway, and aid in immobilization. Tungsten internal eye shields may be used if the target volume approaches the orbits (Fig. 39.3).
For 3D conformal or IMRT, the patient is immobilized in a supine position with the head positioned such that the hard palate is perpendicular to the treatment couch. Scars are marked with thin radio-opaque wires, bolus and other devices are positioned, and transverse CT images are obtained from the vertex to the upper mediastinum. For IMRT, rigid immobilization is necessary, including using special head and shoulder thermoplastic masks that extend down to the upper thorax. The shoulders may be additionally depressed and fixed, for example, using wrist straps tethered to a footboard. Target volumes are delineated as previously described.
For IMRT, multiple gantry angles are used based on beam-optimization algorithms. An example of a 10-field noncoplanar arrangement with two vertex beams is shown in Figure 39.4. The beam angle selections are based on the same principles as for 3D conformal therapy:
• Preference for the shortest path to the target;
• Avoidance of direct irradiation of the critical structures (eg, avoid beam entry through the contralateral eye after ipsilateral exenteration); and
• Use of as large beam separation as possible.



Inverse planning is usually done and multiple iterations may be necessary to ensure that the following are accomplished:
• Targets are covered;
• Normal tissue constraints are respected; and
• Dose is relatively homogenous.
Dose calculations should include heterogeneity corrections because of the significant amount of air and bone of the sinuses. The radiation oncologist must work closely with the physicist and dosimetrist. It is important to realize that the criteria for accepting or rejecting the plan may not be evident from the dose-volume histogram.
For 3D conformal radiotherapy, anterior oblique wedge-pair photon fields are appropriate for lesions located in the anterior lower half of the nasal cavity. Opposed-lateral fields may be used to treat tumors at the posterior part of the nasal fossa, provided the ethmoid cells are not involved. The optic pathway can be excluded from the radiation fields with this setup. For primaries of the upper nasal cavity and ethmoidal air cells, a three-field setup allows coverage of the ethmoid cells while sparing the optic apparatus. CT-based treatment planning is necessary to select beam and wedge angles (usually 45 to 60 degrees) and the relative loading of the fields, as well as to evaluate the dose to critical structures such as brain, brainstem, and optic structures.
Dose Fractionation Schedule
The dose schedule for low dose rate brachytherapy is 60 to 65 Gy during 5 to 7 days. The external-beam regimen for primary radiotherapy is 50 Gy in 25 fractions followed by a boost of 16 to 20 Gy in 8 to 10 fractions, depending on the size of the lesion. Postoperative radiotherapy consists of 50 Gy to elective tissue, 56 Gy to the operative bed, and 60 Gy to the tumor bed, with an optional boost to close or positive surgical margins, all given at 2 Gy per fraction. The dose regimens for IMRT are summarized in Table 39.3.

Maxillary Sinuses
Target Volume
Because maxillary cancers are usually diagnosed in a locally advanced stage and surgery is the primary therapy, most patients receive postoperative radiotherapy. Delineation of target volumes is based physical examination, pretreatment imaging, intraoperative findings (tumor extension relative to critical structures such as orbital wall, cribriform plate, cranial nerve foramina, and ease of resection), and pathologic findings (such as positive margin, perineural invasion).
IMRT is the preferred treatment method as it generally yields better dose distribution in terms of tumor coverage and normal tissue-sparing than 3D conformal radiotherapy. The CTV1 consists of the primary tumor bed with 1.0- to 1.5-cm margin of normal tissue. The CTV2 encompasses the operative bed, including the bony orbit after orbital exenteration and the ethmoid, frontal, and/or sphenoid sinuses if explored during surgery. A third CTV may be delineated to encompass the tract of cranial nerve V2 to the foramen rotundum if there is perineural invasion. A CTVHR (Table 39.3) may be outlined; for example, gross macroscopic residual tumor or positive margins, to which a higher dose may be delivered. An example of an IMRT plan for postoperative radiotherapy is shown in Figure 39.5.
For primary radiotherapy using IMRT, the prescription doses are 66 to 70 Gy to the gross tumor volume (prechemotherapy for those receiving systemic treatment), plus a 1- to 1.5-cm margin of normal-appearing tissue (CTV1), 59 to 63 Gy to other secondary clinical target volumes such as the rest of the involved sinus and wider region around the primary target, and 56 to


57 Gy to the tracts of nerves if there is perineural invasion and to elective nodal regions. An example of an IMRT plan for primary definitive radiotherapy of a T4N0 squamous cell carcinoma is shown in Figure 39.6.

For postoperative radiotherapy using a 3D conformal technique, the initial target volume consists of the operative bed with 1- to 2-cm margins. The boost field consists of the primary tumor bed and areas at higher risk for recurrence such as positive resection margins or perineural invasion. Radiation is administered to the neck following node dissection if multiple nodes are involved and/or there is presence of extracapsular extension. Elective radiation of ipsilateral submandibular and subdigastric nodes is given in patients with squamous cell or poorly differentiated carcinoma.
Setup and Field Arrangement
The patient is immobilized in a supine position with the head slightly hyperextended to bring the floor of the orbit parallel to the axis of the anterior field. An intraoral stent is used to open the mouth and depress the tongue out of the radiation field. Following palatectomy, the stent can be designed to hold a water-filled balloon to obliterate the large air cavity in the surgical defect in order to improve dose homogeneity. An orbital exenteration defect can also be filled directly with a water-filled balloon to decrease the dose delivered to the temporal lobe. Marking of the lateral canthi, oral commissures, external auditory canals, and external scars facilitates target volume delineation. The planning CT scan should include the entire head in order to allow the use of vertex beams. The principles of target delineation and plan evaluation for IMRT of the maxillary sinus cancer are the same as those described for nasal cavity and ethmoid tumors.
For 3D conformal radiotherapy, a three-field technique consisting of an anterior and right and left lateral fields is used for tumors involving the suprastructure or extending to the roof of the nasal cavity and ethmoid cells. The lateral fields may have a 5-degree posterior tilt and 60-degree wedges. The relative loading varies from 1:0.15:0.15 to 1:0.07:0.07 depending on the tumor location and photon energy. For the initial target volume, the superior border of the anterior portal is above the crista galli to encompass the ethmoids and, in the absence of orbital invasion, at the lower edge of the cornea to cover the orbital floor. The inferior border is 1 cm below the floor of the sinus


and the medial border is 1 to 2 cm (or more if necessary) across the midline to cover contralateral ethmoidal extension. The lateral border is 1 cm beyond the apex of the sinus or falling off the skin. The superior border of the lateral portals follows the floor of the anterior cranial fossa, the anterior border is behind the lateral bony canthus parallel to the slope of the face, the posterior border covers the pterygoid plates, and the inferior border corresponds to that of the anterior portal. The boost volume encompasses the tumor bed while sparing the optic pathway.
Anterior and ipsilateral wedge-pair (usually 45-degree wedges) photon fields are used for tumors of the infrastructure with no extension into the orbit or ethmoids. If necessary, the lateral portal can have a 5-degree inferior tilt to avoid beam divergence into the contralateral eye. Lateral-opposed photon fields are preferred for tumors of the infrastructure spreading across midline through the hard palate. If necessary, the fields can be slightly angled (5-degree inferior tilt from the ipsilateral side and 5-degree superior tilt from the contralateral side) to avoid irradiating the contralateral eye. The use of a half-beam with the isocenter placed at the level of the orbital floor and the upper half of the fields shielded further reduces exposure of the eyes by beam divergence.
The eyes and the optic pathway are of particular concern. With 3D conformal techniques, it is generally possible to shield the cornea in patients with limited involvement of the medial or inferior orbital wall to avoid keratitis. If the tumor invades the orbital cavity without necessitating orbital exenteration, care should be taken to avoid irradiation of the lacrimal gland to prevent xerophthalmia. It is important to keep the dose to the contralateral optic nerve as well as the optic chiasm below 54 Gy in 27 fractions to prevent bilateral blindness.
Treatment of the Neck
For squamous and undifferentiated carcinoma, elective neck irradiation is recommended (13). Ipsilateral upper neck treatment is delivered using a lateral appositional electron field (usually 12 MeV). With conventional radiotherapy, careful matching is required to prevent hot or cold spots. The superior border of the field slopes up from the horizontal ramus of the mandible anteriorly to match the inferior border of the primary portal posteriorly, leaving a small triangle over the cheek untreated. The anterior border is just behind the oral commissure, the posterior border is at the mastoid process, and the inferior border is at the thyroid notch (above the arytenoids). The nodal volume can also be covered using IMRT with sparing of the parotid gland. Alternatively, the primary tumor bed and the upper neck can be treated with IMRT with the isocenter above the arytenoids and matched to a separate unmodulated lower neck field. This allows sparing of the laryngeal structures using a larynx block.
If the maxillary sinus is being treated with conventional radiotherapy (non-IMRT), the central axes of the primary (sinus) fields and the opposed-lateral upper neck fields all are placed in the plane of the inferior border of the maxillary fields (ie, usually 1 cm below the floor of the maxillary sinus). An independent collimator jaw is used to shield the caudal half of the maxillary fields and the cephalad half of the neck field. The junction between the primary and the neck fields can be moved during the course of treatment to reduce dose heterogeneity in this region. Portal reduction is made after 42 Gy and treatment to the posterior neck continues with abutting electron fields to the desired dose. The middle and lower neck is irradiated with an anterior appositional photon field matched to the inferior border of the opposed-lateral upper neck fields.
Dose Fractionation Schedule
Table 39.3 summarizes the dose regimens for IMRT. With 3D conformal techniques, the dose for postoperative radiotherapy at 2 Gy per fraction is 50 Gy for elective nodal treatment, 56 Gy to the operative bed, 60 Gy to the tumor bed if resection margins are negative, and 66 Gy if margins are positive. For primary radiotherapy, the total dose to the primary tumor at 2 Gy per fraction is 66 to 70 Gy. The contralateral optic nerve and chiasm are excluded from the field after a dose of 50 to 54 Gy. When the tumor invades structures adjacent to the optic chiasm, a dose of up to 60 Gy to the chiasm may be acceptable (potentially higher control probability with a still relatively low risk for visual impairment [44]) after clear discussion with the patient.
Follow-Up and Recurrences
Salvage is possible for some patients with persistent or recurrent lesions. In particular, recurrent cancers of the nasal vestibule remain curable with salvage surgery after primary radiotherapy or occasionally with salvage radiotherapy after primary surgery. Regional recurrences can be treated successfully with neck dissection with or without postoperative radiotherapy depending on pathologic features. Treatment options are limited for tumors that recur after combined modality therapy, although a few highly selected patients may qualify for reirradiation with curative intent. Cumulative doses of radiotherapy to neural tissues (spinal cord, brainstem, brain, optic structures) are the main limitation to reirradiation.
Most oncologists recommend a baseline physical examination together with CT or MRI for patients with nasal cavity or paranasal sinus tumors 3 months posttreatment. Common practice is to repeat clinical examination and imaging when indicated every 4 months for the first 3 years posttreatment, every 6 months for the fourth and fifth years posttreatment, and annually thereafter. In addition to evaluating for possible tumor recurrence, these follow-up visits are critical with respect to the identification and management of side effects of treatment.
Results of Treatment
The results of treatment have improved during the past 4 decades, with overall survival increasing progressively from 33% ± 18% in the 1960s, to 42% ± 15% in the 1970s, 54% ± 15% in the 1980s, and 56% ± 13% in the 1990s (p <.001) (17). In a systematic review of published series spanning 40 years, Dulguerov et al. (17) demonstrated a progressive improvement in outcome for all treatment modalities (surgery, surgery + radiotherapy, and radiotherapy).
Nasal Vestibule
Table 39.4 summarizes the results of six series of patients treated by brachytherapy, external-beam irradiation, or both, modalities (18,34,37,46,47). Either brachytherapy or external-beam radiotherapy cures up to 95% to 100% of small (up to 2 cm) tumors. When adequate doses of radiation are used, 70% to 80% of lesions >2 cm can be controlled. Although as many as 40% of patients with larger lesions who do not receive elective nodal irradiation will fail in the neck, most can be salvaged and ultimate regional control is excellent. Proper selection of radiation technique, dose, and fractionation results in a low rate of severe late complications.
An excellent analysis was conducted by the Groupe Europeen de Curietherapie (43). Of 1,676 carcinomas of the skin of the nose and nasal vestibule treated by brachytherapy or external-beam irradiation, the overall local control rate was 93%. Local control was dependent on tumor size (<2 cm, 96%; 2 to 3.9 cm, 88%; ≥4 cm, 81%), site (external surface, 94%; vestibule, 75%), as well as new versus recurrent tumors (95% vs.


88%). Local control was independent of histology for tumors <4 cm, but for those >4 cm, basal cell carcinomas were more frequently controlled than were squamous cell carcinomas. There were few complications (necrosis, 2%). The local control rate with surgery was approximately 90%.

Nasal Cavity and Ethmoid Sinuses
Table 39.5 summarizes the results of relatively large series focusing on nasal cavity and ethmoid sinus tumors. Overall local control rates range from 60% to 80% (4,5,10). Results are best for patients with lesions confined to the nasal septum that are generally small and well controlled with primary radiotherapy. Interstitial brachytherapy alone may be the treatment of choice for such patients. Regional failure rates are low. In the MDACC series, nodal recurrence was approximately 5% in patients with nasal cavity tumors who did not receive elective nodal treatment. Complications such as soft tissue necrosis, nasal stenosis, and visual impairment are seen in 5% to 11% of patients.
The data on 783 of a total of 981 patients with nasal cavity cancer included in the Surveillance, Epidemiology, and End Results (SEER) database from 1988 through 1998 were recently analyzed (9). Squamous cell carcinoma was the most common tumor type (49.3%), followed by ENB (13.2%). More than half of the cases presented with a small primary tumor (T1), and only 5% had positive nodes at diagnosis. Overall mean (median) survival was 76 months and overall 5-year survival 56.7%. On multivariate analysis, male gender, increasing age, T stage, N stage, and poorer tumor grade adversely affected survival (p <.05). Radiotherapy was administered in 50.5% of patients, and also independently predicted poorer survival (p =.03), likely due to selection of patients with poor prognostic features such as perineural invasion, positive margins, or poor performance status (medically unfit for surgery) for radiotherapy. Five-year survival by tumor type, T stage, and N stage is shown in Table 39.6 (9). Five-year survival also correlates with tumor dedifferentiation: 75.3%, 61.9%, 47.6%, and 36.8% for well-, moderately, poorly, and undifferentiated cancers, respectively.
Esthesioneuroblastoma
Among the 783 cases of nasal cavity cancer extracted from the SEER database, 103 (13.2%) were ENB. Median survival was 88 months and overall 5-year survival 63.6% (9). Tables 39.7 and 39.8 summarize the results of treatment. The prognosis of patients with stage A disease is excellent. Overall, 30% of patients with stage B tumor died of the disease. About 60% of patients with stage C tumors died of the disease, primarily because of failure to control the primary tumor. Distant metastasis is uncommon (10%) even in locoregionally advanced disease.
Spaulding et al. (64) reported results for 25 patients treated at the University of Virginia Medical Center from 1959 through 1986 and followed for 2 years after therapy. There had been a gradual evolution of treatment with progressive introduction of craniofacial resections, complex field megavoltage radiation, and, for stage C disease, the addition of chemotherapy. Therefore, patients were divided into two groups, based on treatment era, for comparative analysis. Although the series is relatively small, it revealed two interesting findings for this rare disease: (a) extensive craniofacial resection does not appear to confer a major advantage over wide local excision for patients with stage B lesions, and (b) the addition of chemotherapy to craniofacial resection and radiotherapy for patients with stage C tumors may yield a higher disease-specific survival.



An analysis of 72 sinonasal neuroendocrine tumors treated at MDACC between 1982 and 2002 included a spectrum of histologies: ENB (31 patients), sinonasal undifferentiated carcinoma (SNUC, 16 patients), neuroendocrine carcinoma (NEC, 18 patients), and small cell carcinoma (7 patients). The overall survival at 5 years was 93.1% for patients with ENB, 62.5% for SNUC, 64.2% for NEC, and 28.6% for small cell carcinoma (p = .0029; log-rank test). The local control rate at 5 years also was superior for patients with ENB (96.2%) compared with patients who had SNUC (78.6%), NEC (72.6%), or small cell carcinoma (66.7%) (p = .04). The regional failure rate at 5 years was 8.7% for patients with ENB, 15.6% for SNUC, 12.9% for NEC, and 44.4% for small cell carcinoma. The corresponding distant metastasis rates were 0% for ENB, 25.4% for SNUC, 14.1% for NEC, and 75.0% small cell carcinoma. ENB had excellent local and distant control rates with local therapy alone (58). Eight patients with ENB were treated at MDACC during the past 3.5 years with surgery and adjuvant radiotherapy using IMRT to 60 Gy. One patient had stage B disease and seven had stage C, of whom five had intracranial extension. There were no local recurrences and one nodal recurrence was salvaged surgically. All eight patients were alive with no evidence of disease at the last follow-up.

Maxillary Sinuses
For patients with carcinoma of the maxillary sinuses, the combination of surgery and radiation yields 5-year local control and survival rates of 44% to 80% (Table 39.9). These rates are better than those achieved with either surgery or radiotherapy alone.


For radiotherapy alone, the 5-year local control rate ranges from 22% to 39% and the 5-year overall survival rate is 22% to 40%.

Sequelae of Treatment
Soft Tissue and Bone
The formation of nasal cavity synechiae (fibrous mucosal bands causing airway stenosis) can be prevented by intermittent dilation of the nasal passages with a petroleum-coated cotton swab until mucositis has resolved. Dry mucous membranes can be managed symptomatically with saline nasal spray. Soft-tissue or cartilage necrosis is uncommon after therapy with an estimated incidence of 5% to 10% (19,35,52,53).

Eyes and Optic Pathway
Chronic keratitis and iritis (“dry-eye syndrome”) can develop after radiotherapy if tumor extension to the orbital cavity mandates irradiation of the lacrimal gland to doses of more than 30


to 40 Gy (53). Without lacrimal irradiation, fewer than 20% of patients treated with up to 55 Gy to the cornea develop chronic corneal injury (29). There is an approximately 5% risk (at 5 years) of cataract formation after doses of up to 10 Gy to the lenses using conventional fractionation; this risk increases to 50% at 5 years after 18 Gy (21).

Radiation retinopathy is rare after doses of less than 45 Gy, but the incidence increases to about 50% after doses of 45 to 55 Gy (52). The reported incidence of optic neuropathy is <5% after 50 to 60 Gy but increases to around 30% for doses of 61 to 78 Gy. The parameters that influence the risk of radiation-induced optic neuropathy were recently analyzed in 273 patients treated between 1964 and 2000 in whom the radiation fields included the optic nerves and/or chiasm (46). The likelihood of developing optic neuropathy was primarily influenced by the total dose, but fraction size was marginally significant. The 5-year rates of freedom from optic neuropathy were 95% for doses ≤63 Gy treated once daily, 98% for doses ≤63 Gy treated twice daily, 78% for doses >63 Gy treated once daily, and 91% for doses >63 Gy treated twice daily. On multivariate analysis, the risk of optic neuropathy was correlated with increasing total dose (p =.0047). A trend was seen with increasing patient age (p =.091), once daily versus twice-daily fractionation (p =.068), and overall treatment time (p =.097). When the target volumes include the optic pathway, special attention must be paid to hot spots and dose per fraction.