Dose Sparing of Brainstem and Spinal Cord for Re-Irradiating Recurrent Head and Neck Cancer with Intensity-Modulated Radiotherapy

Article Outline

• Abstract
• Introduction
• Methods and Materials
  • Patient characteristics
  • Patient setup
  • Target delineation and dose prescription
  • Beam arrangement
  • Dose optimization
• Results
• Discussion
• Conclusions
• References
• Copyright

Abstract 

Because of the dose limit for critical structures such as brainstem and spinal cord, administering a dose of 60 Gy to patients with recurrent head and neck cancer is challenging for those who received a previous dose of 60−70 Gy. Specifically, previously irradiated head and neck patients may have received doses close to the tolerance limit to their brainstem and spinal cord. In this study, a reproducible intensity-modulated radiation therapy (IMRT) treatment design is presented to spare the doses to brainstem and spinal cord, with no compromise of prescribed dose delivery. Between July and November 2008, 7 patients with previously irradiated, recurrent head and neck cancers were treated with IMRT. The jaws of each field were set fixed with the goal of shielding the brainstem and spinal cord at the sacrifice of partial coverage of the planning target volume (PTV) from any particular beam orientation. Beam geometry was arranged to have sufficient coverage of the PTV and ensure that the constraints of spinal cord <10 Gy and brainstem <15 Gy were met. The mean maximum dose to the brainstem was 12.1 Gy (range 6.1−17.3 Gy), and the corresponding mean maximum dose to spinal cord was 10.4 Gy (range 8.2−14.1 Gy). For most cases, 97% of the PTV volume was fully covered by the 95% isodose volume. We found empirically that if the angle of cervical spine curvature (Cobb's angle) was less than ∼30°, patients could be treated by 18 fields. Six patients met these criteria and were treated in 25 minutes per fraction. One patient exceeded a 30° Cobb's angle and was treated by 31 fields in 45 minutes per fraction. We have demonstrated a new technique for retreatment of head and neck cancers. The angle of cervical spine curvature plays an important role in the efficiency and effectiveness of our approach.

Key Words: Recurrent head-and-neck cancer, Dose sparing, IMRT

Back to Article Outline

Introduction 

Approximately 35,000 new cases of head and neck cancer are diagnosed each year in the United States.¹ The local control rate ranges from 30%−50% by surgery or radiation therapy.², ³ Failure of tumor control contributed to a 50% death rate for patients with loco-regional recurrence.⁴ For patients with recurrent, unresectable head and neck cancer, concomitant chemotherapy and radiation therapy (CCRT) is a potential option.⁵, ⁶, ⁷, ⁸ Strategies to target the radio-resistant cells include the use of radiation sensitizer with aggressive dose escalation. The intended dose prescription for re-irradiation is 60 Gy to the planning target volume (PTV) delivered by a hyper-fractioned and accelerated scheme (twice per day, 1.5 Gy per fraction)⁵ or by an alternating week schedule (2 Gy per fraction, 1 week on and 1 week off).⁸ However, the feasibility of re-irradiation without severe toxicity of normal tissue tolerance is limited. In most cases, the neck, including brainstem and spinal cord, reached the limit of full dose tolerance during the initial treatment. Risks of developing fistulae, carotid rupture, osteoradionecrosis, soft tissue necrosis, and radiation neuropathy have been reported.⁶, ⁹ Although the radiobiological research indicates that the dose for 5% myelopathy is 59.3 Gy,¹⁰ lifetime clinically acceptable dose limit to spinal cord is still taken as 50 Gy in general. For patients who previously received a prescription dose of 60−70 Gy, the residual dose to reach radiation dose tolerance for brainstem and spinal cord could be <10−15 Gy, depending on the treatment plan for the previous irradiation. Other investigators have used IMRT treatments for recurrent head and neck cancer.¹¹, ¹², ¹³ The results show that the dose coverage of PTV might be sacrificed (V_95%, percentage volume received at least 95% prescription, <95%) to stay within the tolerance limit of the brainstem and spinal cord.¹¹, ¹² An excellent dosimetric result was shown by Lee et al.¹³; however, no details for the technique used for treatment planning were provided. Our approach is specifically designed for Varian linear accelerators (Varian Medical System, Inc., Palo Alto, CA) treating with dynamic multileaf collimator (DMLC) intensity-modulated radiation therapy (IMRT). The potential application to other vendors is provided in the discussion. Because the maximum residual dose constraint of 15 Gy at any point in the brainstem and spinal cord constitutes approximately 20% of the prescription dose 60−70 Gy, a 7-field IMRT with sliding window technique shaped by a Varian Millennium 120-leaf MLC¹⁴ was used for those retreated head and neck patients in the early trials of our study. However, it is difficult, if not impossible, to meet the PTV coverage requirements while maintaining the brainstem and spinal cord constraints, because of the unavoidable radiation transmission of the Varian MLC in dynamic mode (up to 4−6%).¹⁵ Consequently, the use of secondary collimators (jaws) or central island blocks¹⁶ could be 2 approaches to reduce the transmission and to shape the fields used in the treatment of paraspinal tumors. In this study, the jaws or secondary collimators were chosen for 2 reasons. First, a single central block was not suitable for all fields, especially for shielding of the asymmetrically shaped brainstem. Moreover, field-specific central blocks are time consuming to fabricate and use. Second, the treatment planning system (Eclipse, version 8.1; Varian Medical System, Inc.) does not support the use of blocks during inverse planning.

Our planning scheme evolved from a 7-field IMRT with central block in concept and ended with 18−31 split fields. The details of the treatment planning technique and dosimetric results are presented in this study. Planning reproducibility and treatment efficiency are also discussed.

Back to Article Outline

Methods and Materials 

Patient characteristics 

Between July and November 2008, 7 patients with previous irradiated, recurrent head and neck cancers were treated with IMRT in our institute. Table 1 lists the clinical details of these 7 patients including age, sex, tumor site, treatment dose prescription, and elapsed time between the end of first treatment and recurrence.

Table 1. Patient characteristics and treatment prescriptions

Pt. No.	Age	Gender	Historical Medical		Recurrent Medical
			Site	Tx Dose (Gy)	Site	Re-Tx Dose (Gy)⁎	Elapsed Time (mo)†
1	58	M	Larynx	70	Glottis	45+25.2	8
2	73	M	Base of tongue	68.4	Base of tongue	50.4+15.6	12
3	66	M	Base of tongue	70	Glottis	45+21.6	5
4	53	F	Larynx/tongue	67.5	Superglottis	50.4+16.2	90
5	66	M	Base of tongue	64	Oropharynx	50.4+16.2	39
6	62	F	Retromolar trigone	60	Retromolar area	40.8+25.2	18
7	61	F	Larynx	66	Larynx	50.4+16.2	9

Patient setup 

Patients were immobilized with customized thermoplastic facemask (Bionix Development Corp., Toledo, OH) and shoulder pulls (WFR/Aquaplast Corp., Wyckoff, NJ). A headrest was chosen for patients to have a fixed, reproducible, and comfortable posture, and the most important, a straight line of cervical spine as possible. To quantify the curvature of cervical spine, the Cobb's angle measurement used in scoliosis diagnosis was borrowed.¹⁷ In this study, the angle was measured from the superior end plate of vertebra C3 to the inferior end plate of vertebra T2 in a sagittal view. After the posture adjustment, patients were scanned with a LightSpeed CT scanner (GE Medical Systems, Milwaukee, WI) with 2.5-mm slice thickness. Image sets were transferred to treatment planning system (Varian Eclipse, version 8.1) for contouring and planning.

Target delineation and dose prescription 

Two treatment plans used for initial PTV and boost PTV were performed for all patients. The target delineation began with the unresectable, visible tumor volume as the gross tumor volume (GTV). The initial clinical target volume (CTV) included GTV and all major head and neck lymph nodes. An initial PTV was then created from the initial CTV by adding a 3−5-mm margin to account for setup error and organ motion. The boost (CTV) included the GTV and any positive diseased lymph nodes diagnosed by MRI, PET, or any other physiological examinations. A 3−5-mm margin was added on the boost CTV to be the boost PTV. The dose was prescribed ranging from 45−50.4 Gy for PTV, and ranging from 15.6−25.2 Gy for boost PTV. The total prescription doses ranging from 66−70.2 Gy are shown in Table 1. Depending on the patient condition, doses were delivered by fractions of either regular (1.8 Gy/fraction) or BID (twice per day, 1.2 Gy/fraction).

Beam arrangement 

We established a template for developing our treatment plans through an iterative technique. As shown in Fig. 1, the design of the plan began with a pair of posterior-to-anterior (PA) fields. Jaws of the field were set fixed with the priority of shielding the brainstem and spinal cord, followed by partial coverage of PTV. The collimators were rotated to be parallel to the spinal cord. The margin between field edge and spinal cord was set to 5 mm to account for the penumbra. The margin between field edge and brainstem could be ignored because of the higher tolerance dose of the brainstem. Figure 1 displays the fields of the paired PA fields. Other fields with different gantry angles were then inserted every 20°−25°. The process was to fix one jaw to shield the cord and then iteratively adjust the gantry angles by ±5° so as to maximize the volume of the PTV in the beam's eye view (BEV).

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 1. 

  Field A and B at gantry angle equal to 180° in IEC scale. Jaws were adjusted to block brainstem and spinal cord and open to partial PTV.

We found that a total of 17−18 fields were used for each of the 6 patients. Excluding the paired PA fields (Fig. 1), the space between two gantry angles varied from 15−30°, depending on the shape of the target volumes of each patient. The remaining patient, who had an excessively large inherent cervical spine curvature, required 31 fields in the final plan. Figure 2 shows the Cobb's angle of cervical spine curvature for the different patients. For patient 7, the Cobb's angle was 30.7°. Two fields with different collimator angles at the same gantry angle were required to shield the spinal cord with the jaw. Table 2 summarizes the Cobb's angles and the number of fields used for all planned patients. The number of fields used for some boost plans was less than the corresponding initial plans because the boost PTV was smaller and sometimes grew asymmetrically around the spinal cord. Although the same beam arrangement could still be used for the boost plan, using the BEV, we removed those gantry angles that did not cover the smaller boost volume. This resulted in a one-fourth reduction in the total number of fields and consequently a reduction in treatment time for the boost phase.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 2. 

  Cobb's angle of cervical curvature (C3 to T2) measured from left sagittal view and fields at gantry angle of 270°. For patient 7, two fields were essential to cover the least PTV for the angle of cervical curvature larger of 31°.

Table 2. Cobb's angles and number of fields

Dose optimization 

No particular constraints for planning optimization were needed for the brainstem and spinal cord once they were well protected by the jaws. The dose coverage of PTV was strongly dependent on the beam arrangement. A dose constraint of 100% volume of PTV covered by 95% of the prescription dose was achievable once each part of PTV was covered by more than half the number of fields. We accepted up to a 10% gradient around the PTV prescription dose (i.e., the dose variation in PTV was allowed to vary between 95% and 105% of the prescription dose). Our gradient was evaluated using the dose-volume histograms (DVHs) of the PTV. A cold spot was allowed after it was found to be a 95% prescription dose located within the 5-mm PTV expansion margin abutting the brainstem and spinal cord. A maximum of 105% of prescription dose inside the PTV was generally achieved.

Back to Article Outline

Results 

Table 3 shows the maximum doses to brainstem and spinal cord for summation plans (initial plan plus boost plan) and Table 4 lists the dose coverages and homogeneities of PTVs in individual plans. Patient 2 is typical and representative of most patients, whereas patient 7 shows the effect of a large Cobb's angle to the planning. Sample dose distributions and DVHs for patients 2 and 7 are shown in Fig. 3, Fig. 4, Fig. 5, Fig. 6.

Table 3. Maximum doses to brainstem and spinal cord

Table 4. Dose coverages and homogeneities of the planning target volumes

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 3. 

  Dose distributions for patient 2 displayed in (A) axial, (B) frontal, and (C) sagittal planes. The color wash began at 10 Gy.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 4. 

  Dose-volume histograms of brainstem, spinal cord, PTV and boost PTV for patient 2. The dose prescriptions to PTV and boost PTV were 50.4 and 66 Gy.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 5. 

  Dose distributions for patient 7 displayed in (A) axial, (B) frontal, and (C) sagittal planes. The color wash began at 10 Gy.

• 
  • View Large Image
  • Download to PowerPoint

• Fig. 6. 

  Dose-volume histograms of brainstem, spinal cord, PTV and boost PTV for patient 7. The dose prescriptions to PTV and boost PTV were 50.4 and 66.6 Gy.

Patient 2 received the prescribed doses of 50.4 Gy for PTV and additional 16.2 Gy for boost PTV. Both the treatment plans for PTV and boost PTV were done with 18 fields. As is shown in Fig. 3, there was no isodose line >10 Gy invading the brainstem and spinal cord. The maximum doses to brainstem and spinal cord were 6.1 and 8.7 Gy, respectively (Fig. 4). As shown on the sagittal view, the brainstem was geometrically separated from the PTV. Consequently, the dose distribution was relatively easy to constrain from the brainstem and spinal cord while ensuring that PTV and boost PTV received at least 95% of the prescribed dose. The dose gradients across the PTVs are evaluated using two parameters: the maximum point dose and the dose difference between 95% and 5% volumes (D_5%-D_95%). The maximum point dose to the PTVs is nearly 109%, yet the D_5%-D_95% was less than 6% of the prescription.

Patient 7 had the same prescription as patient 2, but was a much more difficult plan. Both the treatment plans of PTV and boost PTV were designed with 31 fields. The maximum doses to brainstem and spinal cord were 14.3 and 14.1 Gy, respectively, which exceeded the criteria. Although the number of fields was almost double as a result of the curvature of the spinal cord, the dosimetric results were not as good as other patients. Figure 5 shows that the 10-Gy isodose line parallels the anterior surface of the brainstem and spinal cord, presumably because of the accumulated scatter dose from the jaws. The increase of safe margin might be helpful to decrease the undesired scatter dose, i.e., from 5 to 10 mm. However, the dose coverage would then decrease, especially for PTVs abutting the brainstem and spinal cord. For this patient, the percentage volume of the PTV for 95% prescription was 96.3%, whereas of boost PTV was only 96.2% when using a 5-mm margin.

For the 6 patients who were treated by approximately 18 fields, the total treatment time was about 25 minutes per fraction, including 5 minutes for setup and 20 minutes for dose delivery. Patient 7 had a 45-minute total treatment time.

For patients with straight spinal cord, i.e., Cobb's angle of cervical curve <30°, the treatment planning template was applicable. With the first trial of equally spaced 18 fields, the angles of collimators should always be adjusted based on the curvature of the spinal cord for each patient. Fields sizes (jaw positions) were then tuned to fit the safeguard for spinal cord, as well as cover partial PTVs. The gantry angles were not modified until the finish of initial dose calculation. For most cases, the use of the template could get passable results after the initial trial. The treatment planning could be refined within 2−3 trials, with minor adjustments of beam arrangement.

Back to Article Outline

Discussion 

Because the transmission of the jaws is very low (<1%), the brainstem and spinal cord during the retreatment were well shielded. However, the use of the jaws caused the difficulties in treatment design depending on patient posture and consequently required a large number of fields. The curvature of cervical spine in the treatment area (C3 to T2) could be crucial for the beam arrangement. It was found if the Cobb's angle of cervical spine curvature is larger than 30°, the number of fields would be doubled because 2 fields would be needed in the same gantry angle. Cervical spine curvature varies and is measured on the BEVs. In pretreatment simulation, the Cobb's angle of cervical spine curvature could be measured easily on a sagittal scout. This can be used to establish the acceptability of a particular patient setup. Moreover it could be used for the initial patient treatment and prospectively minimize cord dose to avoid difficulties for retreatment if the patient recurred.

A correction of patient posture might be taken if necessary, which should involve replacement of headrest, refabricating of the facemask, and repositioning of the shoulder pulls. The 30° angle is an empirical guide thus far. Refinement of this criterion will require more studies. In general, the spinal cord must be as straight as possible, but not necessarily parallel to the couch. In this study, all gantry angles of the split fields were tilted to have the better coverage of PTV after a few trials. Fortunately, the increase of field number and gantry angles not only compensated for the cold band behind the block, but also provided the flexibility for dose optimization. The increase in treatment time results from the 18 fields required to provide the degrees of freedom for optimization of the dose distribution.

The efficiency of this treatment scheme could be questionable because 18 fields and a 25-min-per-fraction treatment time is beyond the normal. The problem we are dealing with is specific to Varian machines, with a head design that is different from other vendors. For instance, transmissions of <1% for Elekta MLCs (Elekta Beam Modulator, Elekta Oncology Systems, Crawley, UK) and <0.5% for the binary MLCs of Tomotherapy (TomoTherapy Inc., Madison, WI) were reported.¹⁸, ¹⁹, ²⁰ The effect of the different MLCs could affect the need for this approach. Nevertheless, the use of MLCs with lower transmission might reduce the number of fields, as well as lessen the freedom of optimization along with the abundant gantry angles.

The general applicability of the new rotational therapy techniques like volumetric modulated arc therapy (VMAT) and RapicArc (Varian Medical Systems, Inc.), or even relatively well-established techniques like Hi-Art Tomotherapy (TomoTherapy Inc.)²¹, ²², ²³ to this problem have not been investigated. However, the use of 18 or 31 fields in this study suggests that rotational therapy may be successfully applied to this problem.

Back to Article Outline

Conclusions 

The justification to re-irradiate patients with recurrent head and neck cancer is truly tenable. As a result, dose sparing for brainstem and spinal cord became the keystone to administer the retreatment. A reproducible IMRT treatment technique using jaws to block brainstem and spinal cord was presented in this paper. For patients with straight spinal cords, a treatment with 18 fields in 25-min-per-fraction is a feasible approach for treatment of this patient population.

Back to Article Outline

References 

1. American Cancer Society. Cancer facts and figures. Altanta, GA: Author; 2008;
   • View In Article
2. Million R, Cassisi N, Clark J. Cancer of the head and neck. In:  DeVita VT,  Hellman S,  Rosenberg SA editor. Cancer: Principles and Practice of Oncology. Philadelphia: J.B. Lippincott; 1989;p. 488–590
   • View In Article
3. Vokes EE, Weichselbaum RR, Lippman SM, et al. Head and neck cancer. N. Engl. J. Med. 1993;328:184–194
   • View In Article
   • MEDLINE
   • CrossRef
4. Nishijima W, Takooda S, Tokita N, et al. Analyses of distant metastases in squamous cell carcinoma of the head and neck and lesions above the clavicle at autopsy. Arch. Otolaryngol. Head. Neck. Surg. 1993;119:65–68
   • View In Article
   • MEDLINE
5. Haraf DJ, Weichselbaum RR, Vokes EE. Re-irradiation with concomitant chemotherapy of unresectable recurrent head and neck cancer: A potentially curable disease. Ann. Oncol. 1996;7:913–918
   • View In Article
   • MEDLINE
6. Spencer SA, Harris J, Wheeler RH, et al. RTOG 96-10: Reirradiation with concurrent hydroxyurea and 5-fluorouracil in patients with squamous cell cancer of the head and neck. Int. J. Radiat. Oncol. Biol. Phys. 2001;51:1299–1304
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (73 KB)
   • CrossRef
7. Spencer S, Wheeler R, Peters G, et al. Phase 1 trial of combined chemotherapy and reirradiation for recurrent unresectable head and neck cancer. Head. Neck. 2003;25:118–122
   • View In Article
   • MEDLINE
   • CrossRef
8. Biagioli MC, Harvey M, Roman E, et al. Intensity-modulated radiotherapy with concurrent chemotherapy for previously irradiated, recurrent head and neck cancer. Int. J. Radiat. Oncol. Biol. Phys. 2007;69:1067–1073
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (219 KB)
   • CrossRef
9. Langer CJ, Harris J, Horwitz EM, et al. Phase II study of low-dose paclitaxel and cisplatin in combination with split-course concomitant twice-daily reirradiation in recurrent squamous cell carcinoma of the head and neck: Results of Radiation Therapy Oncology Group Protocol 9911. J. Clin. Oncol. 2007;25:4800–4805
   • View In Article
   • CrossRef
10. Schultheiss TE. The radiation dose-response of the human spinal cord. Int. J. Radiat. Oncol. Biol. Phys. 2008;71:1455–1459
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (209 KB)
    • CrossRef
11. Chen YJ, Kuo JV, Ramsinghani NS, et al. Intensity-modulated radiotherapy for previously irradiated, recurrent head-and-neck cancer. Med. Dosim. 2002;27:171–176
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (157 KB)
    • CrossRef
12. Chua DT, Sham JS, Leung LH, et al. Re-irradiation of nasopharyngeal carcinoma with intensity-modulated radiotherapy. Radiother. Oncol. 2005;77:290–294
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (83 KB)
    • CrossRef
13. Lee N, Chan K, Bekelman JE, et al. Salvage re-irradiation for recurrent head and neck cancer. Int. J. Radiat. Oncol. Biol. Phys. 2007;68:731–740
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (520 KB)
    • CrossRef
14. Hunt MA, Zelefsky MJ, Wolden S, et al. Treatment planning and delivery of intensity-modulated radiation therapy for primary nasopharynx cancer. Int. J. Radiat. Oncol. Biol. Phys. 2001;49:623–632
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (667 KB)
    • CrossRef
15. LoSasso T, Chui CS, Ling CC. Physical and dosimetric aspects of a multileaf collimation system used in the dynamic mode for implementing intensity modulated radiotherapy. Med. Phys. 1998;25:1919–1927
    • View In Article
    • MEDLINE
    • CrossRef
16. Jena R, Luhana F, Brooke SL, et al. Conformal rotation therapy with central axis beam block is a feasible alternative to intensity-modulated radiotherapy for chordomas of the cervical spine. Clin. Oncol. (R. Coll. Radiol.). 2004;16:449–456
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (632 KB)
    • CrossRef
17. Cobb JR. Outline for the study of scoliosis. In:  Edwards JW editors. Instructional Course Lectures. Ann Arbor, MI: The American Academy of Orthopaedic Surgeons; 1948;p. 261–275
    • View In Article
18. Patel I, Glendinning AG, Kirby MC. Dosimetric characteristics of the Elekta Beam Modulator. Phys. Med. Biol. 2005;50:5479–5492
    • View In Article
    • MEDLINE
    • CrossRef
19. Balog JP, Mackie TR, Wenman DL, et al. Multileaf collimator interleaf transmission. Med. Phys. 1999;26:176–186
    • View In Article
    • MEDLINE
    • CrossRef
20. Sarkar V, Lin L, Shi C, et al. Quality assurance of the multileaf collimator with helical tomotherapy: Design and implementation. Med. Phys. 2007;34:2949–2956
    • View In Article
    • CrossRef
21. Yu CX. Intensity-modulated arc therapy with dynamic multileaf collimation: An alternative to tomotherapy. Phys. Med. Biol. 1995;40:1435–1449
    • View In Article
    • MEDLINE
    • CrossRef
22. Otto K. Volumetric modulated arc therapy: IMRT in a single gantry arc. Med. Phys. 2008;35:310–317
    • View In Article
    • CrossRef
23. Mackie TR, Holmes T, Swerdloff S, et al. Tomotherapy: A new concept for the delivery of dynamic conformal radiotherapy. Med. Phys. 1993;20:1709–1719
    • View In Article
    • MEDLINE
    • CrossRef