- Email: [email protected]
A forward-planned treatment technique using multisegments in the treatment of head-and-neck cancer
Int. J. Radiation Oncology Biol. Phys., Vol. 59, No. 2, pp. 584 –594, 2004 Copyright © 2004 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/04/$–see front matter
doi:10.1016/j.ijrobp.2004.02.005
PHYSICS CONTRIBUTION
A FORWARD-PLANNED TREATMENT TECHNIQUE USING MULTISEGMENTS IN THE TREATMENT OF HEAD-AND-NECK CANCER NANCY LEE, M.D., CLAYTON AKAZAWA, C.M.D., PAM AKAZAWA, C.M.D., JEANNE M. QUIVEY, M.D., CHRIS TANG, LYNN J. VERHEY, PH.D., AND PING XIA, PH.D. Department of Radiation Oncology, University of California-San Francisco Medical Center, San Francisco, CA Purpose: To describe in detail a forward-planned multisegment technique (FPMS) as an alternative treatment method for patients who are not suitable for inverse-planned intensity-modulated radiation therapy (IP-IMRT), or for situations where IP-IMRT is not available in a medical clinic. Methods and Materials: Between April 1995 and February 2002, 38 primary head-and-neck patients were treated using the FPMS technique, which has evolved over the past 7 years at our medical center. In the most recent version of the FPMS technique, which includes 5 patients examined in this analysis, the primary tumor and the upper neck nodes were treated with 7 gantry angles, including an anterior, 2 lateral, 2 anterior oblique, and 2 posterior oblique beams with a total of 13 beam shapes formed by multileaf collimators (MLC), called MLC segments. The shape of each MLC segment was carefully designed, and the associated weights were optimized through manual iterations. The lower neck nodes and the supraclavicular nodes were treated with a split-beam anterior field, matched to the inferior border of the FPMS plan at the isocenter. With an autosequencing delivery system, all fields, including dynamic wedges, can be automatically treated. The dosimetric accuracy of this technique was verified with a phantom plan and measured with an ionization chamber, as well as film dosimetry. A sample FPMS plan is described in detail, and the average results for the 5 patients treated with FPMS are retrospectively compared to results for similar patients treated with IP-IMRT. Results: The gross tumor volume was prescribed to 70 Gy (2.12 Gy/fraction) at the 88% isodose line, whereas the clinical target volume received a dose of 59.4 Gy (1.8 Gy/fraction) at the 75% isodose line. The maximum dose to the brainstem and spinal cord was below 54 and 45 Gy, respectively, comparable to IP-IMRT. The mean dose to the parotid glands was 32 Gy with FPMS vs. 26 Gy with IP-IMRT. Average delivery time was shorter for FPMS (15 min) than IP-IMRT (30 min), whereas the planning time depended on the expertise of the planner. Dosimetric accuracy for FPMS and IP-IMRT plans using phantom measurements was similar, within 1% of the phantom plan. With a median follow-up of 31 months, there was no local-regional recurrence, and the incidence of xerostomia is reduced compared to conventional techniques. Conclusion: FPMS achieved plans comparable to those for IP-IMRT and is an ideal alternative treatment technique for a center without the capabilities of IP-IMRT or for a patient who is not a suitable candidate, because of prolonged treatment time. The treatment outcomes from our clinical experience indicate that FPMS can achieve excellent local freedom from progression rates without causing excessive toxicity. Lastly, IP-IMRT plans should be comparable to, if not better than, FPMS plans in the treatment of head-and-neck cancer. © 2004 Elsevier Inc. IMRT, Head and neck, Forward planning, Multisegment.
INTRODUCTION
IMRT (inverse-planned IMRT) is the most ideal treatment technique for head-and-neck cancer patients, this technology is not readily available in every radiotherapy clinic throughout the world. Several factors contribute to this, including economic and technical constraints. First, because IP-IMRT requires a new treatment planning software program as well as new hardware capability to deliver the radiation, it may be difficult for some radiotherapy clinics to shift from their existing treatment techniques to IP-IMRT.
Each year, there are increasing numbers of head-and-neck cancer patients around the world being treated with IMRT (intensity-modulated radiation therapy) (1– 6). IMRT is a refinement of three-dimensional conformal radiotherapy (3D-CRT), which allows the modulation of radiation beams so that a high dose can be delivered to the tumor target while the dose to the surrounding normal tissues is significantly reduced (7–9). Although computer-optimized or IPReprint requests to: Ping Xia, Ph.D., Department of Radiation Oncology, 1600 Divisadero St. H-1031, UCSF Comprehensive Cancer Center, San Francisco, CA 94143-1708. Tel: (415) 3537194; Fax: (415) 353-9883; E-mail: [email protected] Poster session at the American Society of Therapeutic Radiol-
ogy and Oncology, 44th Annual Meeting, October 2002, New Orleans, LA. Received Aug 26, 2003, and in revised form Feb 2, 2004. Accepted for publication Feb 6, 2004. 584
A forward planning technique
● N. LEE et al.
585
Fig. 1. Lateral simulation film of a patient’s hyperextended neck.
Second, IP-IMRT for head-and-neck cancer involves complex planning and treatment delivery, requiring a dedicated team effort of physicists and dosimetrists. Third, the longer daily treatment time associated with IP-IMRT may be a hindrance for a busy clinic trying to accommodate all patients requesting IP-IMRT. Lastly, some patients are simply not ideal candidates for IP-IMRT, because of their inability to remain immobilized for a prolonged time during the course of treatment. FPMS (forward-planned multisegmented technique)— some would call this technology “simple” or forwardplanned IMRT— can improve dose distributions over conventional opposed-lateral fields (10, 11). The drawback with FPMS is that plan optimization is done using manual iteration by the planning dosimetrist or physicist and is heavily dependent on the experience of the planner. However, with proper beam placement and beam weighting, the number of iterations can be significantly reduced. Since 1995, we have treated many of our patients using FPMS, and this technique has evolved over the past 7 years. Currently, the radiation treatments are delivered using computer-controlled static multileaf collimators (MLC) with autosequencing software (Siemens Medical System, Concord, CA). However, some
patients were treated using partial-transmission blocks (1995–1997). Local-regional control rates have been excellent, and the patients had preservation of the salivary function (10). The purpose of this article is to describe in detail our most recent version of the FPMS technique in the treatment of head-and-neck cancer. METHODS AND MATERIALS Between April 1995 and February 2002, 38 primary head-and-neck patients were treated using FPMS techniques. In our initial experience of 33 patients, 24 patients were treated with a 5-gantry angle, 10-field technique (field within a field technique) using partial-transmission cerrobend blocks. The partial-transmission blocks were cut manually and required manual insertion for each field. The time for block manufacturing ranged between 4 and 6 hours, and the treatment time ranged between 20 and 35 min. With computer-controlled multileaf collimator (Siemens Medical System, Concord, CA) and autosequencing delivery capability (Primview, Siemens), the daily treatment time decreased significantly to approximately 15 min for 9 patients.
586
I. J. Radiation Oncology
● Biology ● Physics
Volume 59, Number 2, 2004
Target and sensitive structure delineation Treatment planning was done using the Pinnacle system (ADAC, Phillips Medical System). The gross tumor volume (GTV), clinical target volume (CTV), and the normal critical tissues were outlined on each axial CT slice. The GTV was defined as the gross extent of the tumor seen on imaging and/or determined based on physical examination. The CTV included the GTV and potential direct routes of microscopic spread, as well as margin to account for patient motion and setup errors (10). The normal structures included and evaluated were the brainstem, spinal cord, parotid glands, and mandible. Beam configuration Figure 2 shows typical beam angle arrangements using International Electrotechnical Commission convention. Four of the 7 beam directions contained multiple apertures (or segments). Depending on the particular case, the segment(s) used in a given angle can be tailored to maximize the coverage of the target while minimizing the normal tissue exposure. One can design, depending on the case, up to 3 segments at a given angle. The numbers of segments in each beam including or excluding the CTV, spinal cord, and brainstem in each segment are listed in Table 1. For example, 3 segments are used at the left anterior oblique beam, at a gantry angle of 60° (fields 7, 8, and 13). The first segment was constructed to treat both the GTV and CTV with a margin of 5 mm to account for penumbra, along with a spinal cord block. The second segment was constructed to treat both the GTV and CTV at the same angle, without shielding the spinal cord. The third segment was constructed to treat only the GTV with a 5-mm margin to account for penumbra. For this GTV boost segment, both the spinal cord and the brainstem are shielded. Similar principles were applied to the other beam directions. Table 1 lists sample beam weightings and wedge angles used for a particular plan. All beams are split-beam with the couch angle at 0° for purposes of matching with a split-beam anterior field used to treat the lower neck and the supraclavicular nodes. A single isocenter was used for all beams, including the low neck and supraclavicular field. Figures 3a and 3b show the different segment shapes superimposed on their associated digitally reconstructed
Fig. 2. Typical beam arrangements used in FPMS technique.
The most recent version of the FPMS technique consists of 7 gantry angles with 13 total field shapes. We describe the most updated FPMS technique in detail below and include the dose–volume histograms of the first 5 patients that were treated using the newer version.
Simulation and immobilization For each patient, an initial simulation was done to establish the position, isocenter, and immobilization of the patient (See Fig. 1). The patient’s head was hyperextended with the tip of the uvula and base of skull parallel to the beam axis in the x direction or extended if the patient could not tolerate hyperextension. Patients were then immobilized using a head or head-and-shoulder thermoplastic mask supported on a Timo neck support (Uni-Frame system, MED-TEC). The cervical spinal cord and brainstem were aligned as straight as possible and parallel to the couch, so that the primary tumor could be separated from the neck nodes, resulting in optimal MLC shielding. CT scans in serial 3-mm axial slices were obtained for treatment planning (10).
Table 1. Typical planning parameters used in the forward-planned multisegment technique Field
1
Name Gantry angle Wedge Weight Cord block Brainstem block Beam energy
RPO1 210° 60° 6% Yes Yes 18 Mv
2
3
4
5
6
7
8
9
10
11*
12*
13*
Rt LAT RAO1 270° 300° 45° 45° 5% 7% Yes Yes
RAO2 300° – 11% No
AP1 0° – 7% Yes
AP2 0° 45° 20% No
LAO1 60° 45° 7% Yes
LAO2 60° – 11% No
Lt LAT 90° – 5% Yes
LPO 150° 60° 6% Yes
RPO2 210° – 5% Yes
RPO3 210° – 5% Yes
LAO3 60° – 5% Yes
Yes 18 Mv
No 18 Mv
No 18 Mv
No 6 Mv
No 6 Mv
No 18 Mv
Yes 18 Mv
Yes 18 Mv
Yes 18 Mv
Yes 6 Mv
Yes 18 Mv
No 6 Mv
* Fields 11–13 are to boost the GTV only.
A forward planning technique
● N. LEE et al.
587
Table 2. End point doses to selected structures, total monitor units, and estimated planning time for FPMS and IP-IMRT plans Tumor targets
FPMS IP-IMRT
D99-GTV
D99-CTV
D95-GTV
D95-CTV
V93-GTV
V93-CTV
68.9 Gy 69.3 Gy
50.23 Gy 54.3 Gy
70.2 Gy 71.2 Gy
58.9 Gy 60.6 Gy
0.2 cc 0.1 cc
15.0 cc 9.0 cc
Total monitor units 500 1200
Planning time 8–12 h 8–12 h
Sensitive structures
FPMS IP-IMRT
Spinal cordmax 44 Gy 43 Gy
Brainstemmax 54 Gy 45 Gy
Mandiblemax 73.3 Gy 71.6 Gy
Parotidmean 32.0 Gy 26.1 Gy
Abbreviations: FPMS ⫽ forward-planned multisegment technique; IP-IMRT ⫽ inverse-planned intensity-modulated radiotherapy; GTV ⫽ gross tumor volume; CTV ⫽ clinical target volume. D99-GTV, D99-CTV ⫽ Dose encompassing 99% of the GTV, or CTV, respectively. D95-GTV D95-CTV ⫽ Dose encompassing 95% of the GTV, or CTV, respectively. V93-GTV, V93-CTV ⫽ Volumes (cc) received less than 93% of the prescribed doses to the GTV and CTV, respectively. Spinal cord-max ⫽ Maximum dose to the spinal cord. Brainstem-max ⫽ Maximum dose to the brainstem. Mandible-max ⫽ Maximum dose to the mandible. Parotid-mean ⫽ Mean dose to the parotid glands.
radiographs for a T3N1 supraglottic carcinoma case. Notice that for this particular case, segment RAO1 treated the GTV and CTV without shielding the spinal cord or the brainstem. A spinal cord block was placed for the RAO2 segment. Finally, in segment RAO3, only the GTV was treated, whereas both the spinal cord and the brainstem were blocked. Figure 3c displays the axial, sagittal, and coronal slices of the same patient, along with the isodose distributions. The light blue 70 Gy line nicely encompasses the GTV, whereas the yellow 60 Gy line conforms to the CTV shown in the red colorwash. Once the typical beam arrangements were set, numerous adjustments and fine tuning were done to increase dose homogeneity and to decrease the dose delivered to the normal tissues. The shape of each MLC was carefully designed, along with the associated weights. Manual iterations were done to optimize the plan. The oblique field angles were rotated slightly to feather out the hot spots. The thick end of wedges (heel end) was placed anteriorly. Collimator angles were rotated 90° to effectively shield cord and brainstem. The average time spent in treatment planning—including contouring, defining beams, calculating dose, optimizing by iteration, evaluating, and, finally, accepting the plan—was approximately 8 to 12 hours and, in very experienced hands, 4 to 8 hours. Delivery and quality assurance At the start of radiotherapy, all 13 films were taken to verify the treatment fields. Once all portal images matched exactly with the digitally reconstructed radiographs, weekly orthogonal portal films were taken to verify the treatment isocenter. Necessary shifts in the isocenter were made during treatment at the discretion of the treating physician. For a selected FPMS plan, a phantom plan was created
with the same beam arrangements and intensity distributions as used in the treatment plan. The dose distribution was recalculated in the phantom based on its geometry. The phantom consisted of a stack of square solid water slabs with the approximate dimensions of a human head. The dosimetric accuracy of this technique was verified with a phantom plan and measured with an ionization chamber, as well as MOSFET detectors (12). The dosimetric accuracy was within 1% of the phantom plan.
RESULTS Forward-planned multisegment treatment planning and delivery All 38 patients completed the intended treatment using FPMS technique. Because our FPMS technique has evolved over the past 7 years, 5 patients treated with the most recent version of the FPMS method will be analyzed. Using an autosequencing delivery system, all fields, including dynamic wedges, were treated automatically and in succession. The actual treatment delivery time was 15 min for the FPMS plans. The total number of monitor units used for the FPMS plans was on average around 500, considerably less than the number used for the IP-IMRT, as shown in Table 2. Plan evaluation of target volume coverage and normal tissue sparing Figure 4 shows the isodose distributions on axial, sagittal, and coronal images for the oropharyngeal study case generated with the FPMS plan. The red colorwash indicates the GTV, whereas the light green colorwash indicates the CTV. The isodose lines are displayed on an absolute dose scale
588
I. J. Radiation Oncology
● Biology ● Physics
Fig. 3.
Volume 59, Number 2, 2004
A forward planning technique
● N. LEE et al.
589
Fig. 3. (Cont’d). Fig. 3. (a, b) Beam’s-eye views of 13 digitally reconstructed radiographs for a T3N1 supraglottic larynx carcinoma patient. Notice the different MLC shapes in each segment of a particular angle used in this plan. (c) Axial, sagittal, and coronal dose distributions of the same patient. The light blue isodose line corresponds to the 70 Gy line that conforms nicely to the gross tumor volume outlined by the dark blue lines. The yellow isodose lines conform to the CTV in the red colorwash.
ranging from 30 to 70 Gy. The contours of the GTV and the CTV are outlined in red and green, respectively. Notice that the GTV is completely encompassed by the 70 Gy isodose line. The high-risk subclinical disease area is also encompassed by the 59.40 Gy isodose line. No unexpected hot spots were found outside these target volumes. Notice the nice conformality of the isodose curves to the targets, whereas the parotid glands were outside the 45 Gy line, and the majority were within the 30 Gy line. Figure 5 displays the dose–volume histograms of 5 patients treated with the 13-field FPMS technique. All 5 patients had similar coverage of the target volumes, GTV, and CTV. Depending on the respective case, the dose to the spinal cord differed. The average median dose for the parotid glands for all 5 patients was 32.2 Gy. In general, FPMS plans achieved the delivery of the desired prescription dose while minimizing the dose delivered to the normal tissue. Compared to traditional treatment planning, FPMS allows the sparing of parotid glands, which ultimately should impact on the salivary function, as well as the quality of life of patients undergoing radiotherapy.
DISCUSSION Advancements in technology have made it possible to treat many complicated and irregularly shaped head-andneck tumors (1–11, 13). Further advances in 3D-CRT, such as IMRT, allow even better coverage of the tumor target while minimizing the dose delivered to the normal tissues (1–10). IMRT can be divided into two broad categories, forward planning vs. inverse planning. In forward planning, the planning dosimetrist or physicist selects the number, energy, weighting, and angle of beams. The computer then calculates the dose distribution and generates beam’s-eye views, along with dose–volume histograms. The plan is optimized by manual iteration or trial and error. This is in contrast to inverse planning, where one begins by defining the desired dose to the target and the normal tissues, and the computer with its optimization program will go through multiple iterations seeking to find the best beam parameters that will yield the desired dose distribution. Recently, several articles have addressed the physical as well as clinical advantage of IP-IMRT over forward-
590
I. J. Radiation Oncology
● Biology ● Physics
Volume 59, Number 2, 2004
Fig. 4. Axial, sagittal, and coronal dose distributions of a patient with oropharyngeal carcinoma treated with FPMS plan. The red colorwash represents the gross tumor volume, whereas the green colorwash represents the clinical target volume. The dark blue and the light blue lines are the 70 Gy and the 60 Gy lines, respectively.
planned IMRT or conventional radiotherapy (9, 13–16). Since the introduction of IP-IMRT, difficult and advanced head-and-neck tumors that are impossible to cover adequately without exceeding the tolerance of the respective normal tissues are now possible to cover adequately with IP-IMRT. This is true for sites including nasopharynx, oropharynx, and larynx. The conclusion has always been similar, that IP-IMRT achieved better coverage, though more inhomogeneous, when compared to conventional treatment techniques. In addition, there is always more sparing of surrounding normal tissues. However, not every center has a dedicated team of physicists, dosimetrists, and therapists that can handle large volumes of IP-IMRT patients. Without this dedicated team approach, IP-IMRT for complex head-andneck treatment planning may be nearly impossible. In addition, the prolonged treatment time for each IP-IMRT may not be feasible in many busy radiotherapy clinics. Patients may not be able to hold still in the same position throughout the entire IMRT treatment course. Therefore, given the above factors, and because an excellent dose distribution covering the tumor target while sparing the
nearby normal tissues can be achieved with the FPMS technique (10, 11, 13, 17), it may be more efficient to treat patients with FPMS until IP-IMRT becomes feasible in a busy clinic. FPMS can be especially helpful when there is popular demand for IMRT. Instead of having to inform the patients and the referring physicians that one does not have the capability of treating patients using IP-IMRT, one can use the FPMS technique and achieve a similar treatment plan, as shown in Table 2. Please note that IP-IMRT is still superior when compared to the 13-field FPMS with regard to target coverage and normal tissue protection. The FPMS technique is merely an alternative approach to treating head-and-neck cancer patients for the reasons stated above. The initial IP-IMRT experience at Memorial SloanKettering Cancer Center was slightly different from that at the University of California-San Francisco (UCSF) Medical Center (1, 9). Both centers have shown that, given proper dose constraints, the improvement of IPIMRT was due to good coverage in the retropharynx, base of skull, and the medial aspects of the nodal volumes. In addition, IP-IMRT delivered lower doses to the
A forward planning technique
● N. LEE et al.
Fig. 5. Dose–volume histograms for 5 patients treated with forward-planned multisegmented technique (FPMS): (a) gross target volume, (b) clinical target volume, (c) spinal cord, (d) brainstem, and (e) parotid gland. Figure continues on pages 592 and 593.
591
592
I. J. Radiation Oncology
● Biology ● Physics
Volume 59, Number 2, 2004
Fig. 5. (Cont’d).
A forward planning technique
● N. LEE et al.
593
Fig. 5. (Cont’d).
spinal cord, brainstem, mandible, temporal lobes, and parotid glands when compared with 3D-CRT. No attempts were made to spare the parotid glands at Memorial (with a mean dose of 60 Gy), whereas parotid sparing was a goal at UCSF (with a typical mean dose around 26 Gy). An in-house treatment planning system was used at Memorial vs. a commercial system at UCSF. Treatment planning times at Memorial, including contouring, beam definition, optimization, evaluation, and documentation, were approximately 8 hours for IP-IMRT vs. 12 hours for the 3D plan. The daily patient setup and treatment time on average was 15–20 min shorter for IP-IMRT vs. 3D-CRT. In contrast, treatment planning times at UCSF were similar for both IP-IMRT vs. FPMS techniques, in general 8 to 12 hours. Delivery was also shorter with FPMS, 15 min vs. 40 min for IP-IMRT per patient. Perhaps the differences in treatment planning and delivery systems could explain the differences between the two centers. Both centers, however, did conclude that a proper set of dose constraints was essential in IP-IMRT. IP-IMRT can be dangerous when dose specifications are set inappropriately and improperly. Although it is facilitated by computer optimization, IP-IMRT plans are not always optimal, especially in the setting of wrongly specified dose constraints. IP-IMRT may even achieve an inferior plan for a given patient when compared to conventional treatment techniques. An appropriate specification of dose constraints is therefore essential to obtain an optimal result in inverse planning. Otherwise, even in the setting of multiple itera-
tions, the computer may end with the same solution. This can be frustrating to both the physician and the physicists. Therefore, the question is: What is considered the optimal IP-IMRT plan for a given patient? Because the forwardplanning multisegment technique can achieve a very good plan, we believe that all IP-IMRT plans for head-and-neck cancer patients should be just as good as, if not better than, the FPMS plans. This will give the treating physicians or the physicists a confidence interval with which he or she can choose or reject a given plan. Therefore, it is in the authors’ opinion that all IP-IMRT plans should not be inferior to what can be achieved with FPMS, which is in a way a very fancy 3D conformal plan. One final important issue must not be overlooked when treating head-and-neck patients using highly conformal techniques (2, 3). Because both FPMS and IMRT have very sharp dose falloff gradients between the target and surrounding normal tissue, adequate target volume delineation is absolutely essential. Therefore, no matter what technology or excellent supporting staff that one center has, precise target volume delineation is absolutely critical. The treatment planning system will not treat areas not drawn on the CT slices, and the algorithm will even “work hard” to spare regions that are not contoured. Given these factors, at the University of California San Francisco Medical Center, target volume delineation is done with a multidisciplinary team approach consisting of the radiation oncologist, neuroradiologist, and, in the postoperative setting, head-andneck surgeon. The target volumes of each case are carefully
594
I. J. Radiation Oncology
● Biology ● Physics
and accurately defined jointly by the multidisciplinary team. Marginal misses of tumor have been avoided in our patients. CONCLUSION The forward-planned multisegment technique allows the delivery of a high dose to the target coverage while minimizing
Volume 59, Number 2, 2004
the normal tissue toxicity when compared to conventional techniques. In a center without the capabilities of IPIMRT— or in situations where patients are not ideal candidates for that technique—FPMS can be useful and can achieve a dose distribution similar to that with IP-IMRT. Lastly, all IP-IMRT plans should be at least comparable to, if not better than, FPMS plans in the treatment of head-and-neck cancer.
REFERENCES 1. Lee N, Xia P, Quivey JM, et al. Intensity-modulated radiotherapy in the treatment of nasopharyngeal carcinoma: An update of the UCSF experience. Int J Radiat Oncol Biol Phys 2002;53:12–21. 2. Chao C, Ozyigit G, Tran BH, et al. Patterns of failure in patients receiving definitive and postoperative IMRT for headand-neck cancer. Int J Radiat Oncol Biol Phys 2003;55:312– 321. 3. Lee N, Xia P, Fischbein NJ, et al. Intensity-modulated radiation therapy for head and neck cancer: The UCSF experience focusing on target volume delineation. Int J Radiat Oncol Biol Phys 2003;57:49–60. 4. Claus F, Boterberg T, Ost P, et al. Short term toxicity profile for 32 sinonasal cancer patients treated IMRT. Can we avoid dry eye syndrome? Radiother Oncol 2002;64:205–208. 5. Butler EB, The BS, Grant WH, III, et al. SMART (simultaneous modulated accelerated radiation therapy) boost: A new accelerated fractionation schedule for the treatment of head and neck cancer with intensity modulated radiotherapy. Int J Radiat Oncol Biol Phys 1999;45:21–32. 6. Dawson LA, Anzai Y, Marsh L, et al. Patterns of localregional recurrence following parotid sparing conformal and segmental intensity modulated radiotherapy for head and neck cancer. Int J Radiat Oncol Biol Phys 2000;46:1117–1126. 7. Nutting C, Dearnaley DP, Webb S. Intensity-modulated radiation therapy: A clinical review. Br J Radiol 2000;73:459– 469. 8. Xia P, Fu KK, Wong GW, et al. Comparison of treatment plans involving intensity modulated radiotherapy for nasopharyngeal carcinoma. Int J Radiat Oncol Biol Phys 2000;48: 329–337. 9. 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.
10. Sultanem K, Shu KK, Xia P, et al. Three-dimensional intensity-modulated radiotherapy in the treatment of nasopharyngeal carcinoma: The University of California-San Francisco Experience. Int J Radiat Oncol Biol Phys 2000;48:711–722. 11. van Dieren EB, Nowak PJCM, Wijers O, et al. Beam intensity modulation using tissue compensators or dynamic multileaf collimation in three-dimensional conformal radiotherapy of primary cancers of the oropharynx and larynx, including the elective neck. Int J Radiat Oncol Biol Phys 2000;47:1299– 1309. 12. Chuang CF, Verhey LJ, Xia P. Investigation of the use of MOSFET for clinical IMRT dosimetric verification. Med Phys 2002;29:1109–1115. 13. Zabel A, Thilmann C, Zuna I, et al. Comparison of forward planned conformal radiation therapy and inverse planned intensity modulated radiation therapy for esthesioneuroblastoma. Br J Radiol 2002;75:356–361. 14. Kam MKM, Cha RMC, Suen J, et al. Intensity-modulated radiotherapy in nasopharyngeal carcinoma: Dosimetric advantage over conventional plans and feasibility of dose escalation. Int J Radiat Oncol Biol Phys 2003;56:145–157. 15. Hsiung CY, Yorke ED, Chui CS, et al. Intensity-modulated radiotherapy versus conventional three-dimensional conformal radiotherapy for boost or salvage treatment of nasopharyngeal carcinoma. Int J Radiat Oncol Biol Phys 2002;53: 638–647. 16. Cheng JCH, Chao KSC, Low DA. Comparison of intensity modulated radiation therapy (IMRT) techniques for nasopharyngeal carcinoma. Int J Cancer 2001;96:126–131. 17. Xia P, Pickett B, Vigneault E, et al. Forward or inversely planned segmental multileaf collimator IMRT and sequential tomotherapy to treat multiple dominant intraprostatic lesions of prostate cancer to 90 Gy. Int J Radiat Oncol Biol Phys 2001;51:244–254.
doi:10.1016/j.ijrobp.2004.02.005
PHYSICS CONTRIBUTION
A FORWARD-PLANNED TREATMENT TECHNIQUE USING MULTISEGMENTS IN THE TREATMENT OF HEAD-AND-NECK CANCER NANCY LEE, M.D., CLAYTON AKAZAWA, C.M.D., PAM AKAZAWA, C.M.D., JEANNE M. QUIVEY, M.D., CHRIS TANG, LYNN J. VERHEY, PH.D., AND PING XIA, PH.D. Department of Radiation Oncology, University of California-San Francisco Medical Center, San Francisco, CA Purpose: To describe in detail a forward-planned multisegment technique (FPMS) as an alternative treatment method for patients who are not suitable for inverse-planned intensity-modulated radiation therapy (IP-IMRT), or for situations where IP-IMRT is not available in a medical clinic. Methods and Materials: Between April 1995 and February 2002, 38 primary head-and-neck patients were treated using the FPMS technique, which has evolved over the past 7 years at our medical center. In the most recent version of the FPMS technique, which includes 5 patients examined in this analysis, the primary tumor and the upper neck nodes were treated with 7 gantry angles, including an anterior, 2 lateral, 2 anterior oblique, and 2 posterior oblique beams with a total of 13 beam shapes formed by multileaf collimators (MLC), called MLC segments. The shape of each MLC segment was carefully designed, and the associated weights were optimized through manual iterations. The lower neck nodes and the supraclavicular nodes were treated with a split-beam anterior field, matched to the inferior border of the FPMS plan at the isocenter. With an autosequencing delivery system, all fields, including dynamic wedges, can be automatically treated. The dosimetric accuracy of this technique was verified with a phantom plan and measured with an ionization chamber, as well as film dosimetry. A sample FPMS plan is described in detail, and the average results for the 5 patients treated with FPMS are retrospectively compared to results for similar patients treated with IP-IMRT. Results: The gross tumor volume was prescribed to 70 Gy (2.12 Gy/fraction) at the 88% isodose line, whereas the clinical target volume received a dose of 59.4 Gy (1.8 Gy/fraction) at the 75% isodose line. The maximum dose to the brainstem and spinal cord was below 54 and 45 Gy, respectively, comparable to IP-IMRT. The mean dose to the parotid glands was 32 Gy with FPMS vs. 26 Gy with IP-IMRT. Average delivery time was shorter for FPMS (15 min) than IP-IMRT (30 min), whereas the planning time depended on the expertise of the planner. Dosimetric accuracy for FPMS and IP-IMRT plans using phantom measurements was similar, within 1% of the phantom plan. With a median follow-up of 31 months, there was no local-regional recurrence, and the incidence of xerostomia is reduced compared to conventional techniques. Conclusion: FPMS achieved plans comparable to those for IP-IMRT and is an ideal alternative treatment technique for a center without the capabilities of IP-IMRT or for a patient who is not a suitable candidate, because of prolonged treatment time. The treatment outcomes from our clinical experience indicate that FPMS can achieve excellent local freedom from progression rates without causing excessive toxicity. Lastly, IP-IMRT plans should be comparable to, if not better than, FPMS plans in the treatment of head-and-neck cancer. © 2004 Elsevier Inc. IMRT, Head and neck, Forward planning, Multisegment.
INTRODUCTION
IMRT (inverse-planned IMRT) is the most ideal treatment technique for head-and-neck cancer patients, this technology is not readily available in every radiotherapy clinic throughout the world. Several factors contribute to this, including economic and technical constraints. First, because IP-IMRT requires a new treatment planning software program as well as new hardware capability to deliver the radiation, it may be difficult for some radiotherapy clinics to shift from their existing treatment techniques to IP-IMRT.
Each year, there are increasing numbers of head-and-neck cancer patients around the world being treated with IMRT (intensity-modulated radiation therapy) (1– 6). IMRT is a refinement of three-dimensional conformal radiotherapy (3D-CRT), which allows the modulation of radiation beams so that a high dose can be delivered to the tumor target while the dose to the surrounding normal tissues is significantly reduced (7–9). Although computer-optimized or IPReprint requests to: Ping Xia, Ph.D., Department of Radiation Oncology, 1600 Divisadero St. H-1031, UCSF Comprehensive Cancer Center, San Francisco, CA 94143-1708. Tel: (415) 3537194; Fax: (415) 353-9883; E-mail: [email protected] Poster session at the American Society of Therapeutic Radiol-
ogy and Oncology, 44th Annual Meeting, October 2002, New Orleans, LA. Received Aug 26, 2003, and in revised form Feb 2, 2004. Accepted for publication Feb 6, 2004. 584
A forward planning technique
● N. LEE et al.
585
Fig. 1. Lateral simulation film of a patient’s hyperextended neck.
Second, IP-IMRT for head-and-neck cancer involves complex planning and treatment delivery, requiring a dedicated team effort of physicists and dosimetrists. Third, the longer daily treatment time associated with IP-IMRT may be a hindrance for a busy clinic trying to accommodate all patients requesting IP-IMRT. Lastly, some patients are simply not ideal candidates for IP-IMRT, because of their inability to remain immobilized for a prolonged time during the course of treatment. FPMS (forward-planned multisegmented technique)— some would call this technology “simple” or forwardplanned IMRT— can improve dose distributions over conventional opposed-lateral fields (10, 11). The drawback with FPMS is that plan optimization is done using manual iteration by the planning dosimetrist or physicist and is heavily dependent on the experience of the planner. However, with proper beam placement and beam weighting, the number of iterations can be significantly reduced. Since 1995, we have treated many of our patients using FPMS, and this technique has evolved over the past 7 years. Currently, the radiation treatments are delivered using computer-controlled static multileaf collimators (MLC) with autosequencing software (Siemens Medical System, Concord, CA). However, some
patients were treated using partial-transmission blocks (1995–1997). Local-regional control rates have been excellent, and the patients had preservation of the salivary function (10). The purpose of this article is to describe in detail our most recent version of the FPMS technique in the treatment of head-and-neck cancer. METHODS AND MATERIALS Between April 1995 and February 2002, 38 primary head-and-neck patients were treated using FPMS techniques. In our initial experience of 33 patients, 24 patients were treated with a 5-gantry angle, 10-field technique (field within a field technique) using partial-transmission cerrobend blocks. The partial-transmission blocks were cut manually and required manual insertion for each field. The time for block manufacturing ranged between 4 and 6 hours, and the treatment time ranged between 20 and 35 min. With computer-controlled multileaf collimator (Siemens Medical System, Concord, CA) and autosequencing delivery capability (Primview, Siemens), the daily treatment time decreased significantly to approximately 15 min for 9 patients.
586
I. J. Radiation Oncology
● Biology ● Physics
Volume 59, Number 2, 2004
Target and sensitive structure delineation Treatment planning was done using the Pinnacle system (ADAC, Phillips Medical System). The gross tumor volume (GTV), clinical target volume (CTV), and the normal critical tissues were outlined on each axial CT slice. The GTV was defined as the gross extent of the tumor seen on imaging and/or determined based on physical examination. The CTV included the GTV and potential direct routes of microscopic spread, as well as margin to account for patient motion and setup errors (10). The normal structures included and evaluated were the brainstem, spinal cord, parotid glands, and mandible. Beam configuration Figure 2 shows typical beam angle arrangements using International Electrotechnical Commission convention. Four of the 7 beam directions contained multiple apertures (or segments). Depending on the particular case, the segment(s) used in a given angle can be tailored to maximize the coverage of the target while minimizing the normal tissue exposure. One can design, depending on the case, up to 3 segments at a given angle. The numbers of segments in each beam including or excluding the CTV, spinal cord, and brainstem in each segment are listed in Table 1. For example, 3 segments are used at the left anterior oblique beam, at a gantry angle of 60° (fields 7, 8, and 13). The first segment was constructed to treat both the GTV and CTV with a margin of 5 mm to account for penumbra, along with a spinal cord block. The second segment was constructed to treat both the GTV and CTV at the same angle, without shielding the spinal cord. The third segment was constructed to treat only the GTV with a 5-mm margin to account for penumbra. For this GTV boost segment, both the spinal cord and the brainstem are shielded. Similar principles were applied to the other beam directions. Table 1 lists sample beam weightings and wedge angles used for a particular plan. All beams are split-beam with the couch angle at 0° for purposes of matching with a split-beam anterior field used to treat the lower neck and the supraclavicular nodes. A single isocenter was used for all beams, including the low neck and supraclavicular field. Figures 3a and 3b show the different segment shapes superimposed on their associated digitally reconstructed
Fig. 2. Typical beam arrangements used in FPMS technique.
The most recent version of the FPMS technique consists of 7 gantry angles with 13 total field shapes. We describe the most updated FPMS technique in detail below and include the dose–volume histograms of the first 5 patients that were treated using the newer version.
Simulation and immobilization For each patient, an initial simulation was done to establish the position, isocenter, and immobilization of the patient (See Fig. 1). The patient’s head was hyperextended with the tip of the uvula and base of skull parallel to the beam axis in the x direction or extended if the patient could not tolerate hyperextension. Patients were then immobilized using a head or head-and-shoulder thermoplastic mask supported on a Timo neck support (Uni-Frame system, MED-TEC). The cervical spinal cord and brainstem were aligned as straight as possible and parallel to the couch, so that the primary tumor could be separated from the neck nodes, resulting in optimal MLC shielding. CT scans in serial 3-mm axial slices were obtained for treatment planning (10).
Table 1. Typical planning parameters used in the forward-planned multisegment technique Field
1
Name Gantry angle Wedge Weight Cord block Brainstem block Beam energy
RPO1 210° 60° 6% Yes Yes 18 Mv
2
3
4
5
6
7
8
9
10
11*
12*
13*
Rt LAT RAO1 270° 300° 45° 45° 5% 7% Yes Yes
RAO2 300° – 11% No
AP1 0° – 7% Yes
AP2 0° 45° 20% No
LAO1 60° 45° 7% Yes
LAO2 60° – 11% No
Lt LAT 90° – 5% Yes
LPO 150° 60° 6% Yes
RPO2 210° – 5% Yes
RPO3 210° – 5% Yes
LAO3 60° – 5% Yes
Yes 18 Mv
No 18 Mv
No 18 Mv
No 6 Mv
No 6 Mv
No 18 Mv
Yes 18 Mv
Yes 18 Mv
Yes 18 Mv
Yes 6 Mv
Yes 18 Mv
No 6 Mv
* Fields 11–13 are to boost the GTV only.
A forward planning technique
● N. LEE et al.
587
Table 2. End point doses to selected structures, total monitor units, and estimated planning time for FPMS and IP-IMRT plans Tumor targets
FPMS IP-IMRT
D99-GTV
D99-CTV
D95-GTV
D95-CTV
V93-GTV
V93-CTV
68.9 Gy 69.3 Gy
50.23 Gy 54.3 Gy
70.2 Gy 71.2 Gy
58.9 Gy 60.6 Gy
0.2 cc 0.1 cc
15.0 cc 9.0 cc
Total monitor units 500 1200
Planning time 8–12 h 8–12 h
Sensitive structures
FPMS IP-IMRT
Spinal cordmax 44 Gy 43 Gy
Brainstemmax 54 Gy 45 Gy
Mandiblemax 73.3 Gy 71.6 Gy
Parotidmean 32.0 Gy 26.1 Gy
Abbreviations: FPMS ⫽ forward-planned multisegment technique; IP-IMRT ⫽ inverse-planned intensity-modulated radiotherapy; GTV ⫽ gross tumor volume; CTV ⫽ clinical target volume. D99-GTV, D99-CTV ⫽ Dose encompassing 99% of the GTV, or CTV, respectively. D95-GTV D95-CTV ⫽ Dose encompassing 95% of the GTV, or CTV, respectively. V93-GTV, V93-CTV ⫽ Volumes (cc) received less than 93% of the prescribed doses to the GTV and CTV, respectively. Spinal cord-max ⫽ Maximum dose to the spinal cord. Brainstem-max ⫽ Maximum dose to the brainstem. Mandible-max ⫽ Maximum dose to the mandible. Parotid-mean ⫽ Mean dose to the parotid glands.
radiographs for a T3N1 supraglottic carcinoma case. Notice that for this particular case, segment RAO1 treated the GTV and CTV without shielding the spinal cord or the brainstem. A spinal cord block was placed for the RAO2 segment. Finally, in segment RAO3, only the GTV was treated, whereas both the spinal cord and the brainstem were blocked. Figure 3c displays the axial, sagittal, and coronal slices of the same patient, along with the isodose distributions. The light blue 70 Gy line nicely encompasses the GTV, whereas the yellow 60 Gy line conforms to the CTV shown in the red colorwash. Once the typical beam arrangements were set, numerous adjustments and fine tuning were done to increase dose homogeneity and to decrease the dose delivered to the normal tissues. The shape of each MLC was carefully designed, along with the associated weights. Manual iterations were done to optimize the plan. The oblique field angles were rotated slightly to feather out the hot spots. The thick end of wedges (heel end) was placed anteriorly. Collimator angles were rotated 90° to effectively shield cord and brainstem. The average time spent in treatment planning—including contouring, defining beams, calculating dose, optimizing by iteration, evaluating, and, finally, accepting the plan—was approximately 8 to 12 hours and, in very experienced hands, 4 to 8 hours. Delivery and quality assurance At the start of radiotherapy, all 13 films were taken to verify the treatment fields. Once all portal images matched exactly with the digitally reconstructed radiographs, weekly orthogonal portal films were taken to verify the treatment isocenter. Necessary shifts in the isocenter were made during treatment at the discretion of the treating physician. For a selected FPMS plan, a phantom plan was created
with the same beam arrangements and intensity distributions as used in the treatment plan. The dose distribution was recalculated in the phantom based on its geometry. The phantom consisted of a stack of square solid water slabs with the approximate dimensions of a human head. The dosimetric accuracy of this technique was verified with a phantom plan and measured with an ionization chamber, as well as MOSFET detectors (12). The dosimetric accuracy was within 1% of the phantom plan.
RESULTS Forward-planned multisegment treatment planning and delivery All 38 patients completed the intended treatment using FPMS technique. Because our FPMS technique has evolved over the past 7 years, 5 patients treated with the most recent version of the FPMS method will be analyzed. Using an autosequencing delivery system, all fields, including dynamic wedges, were treated automatically and in succession. The actual treatment delivery time was 15 min for the FPMS plans. The total number of monitor units used for the FPMS plans was on average around 500, considerably less than the number used for the IP-IMRT, as shown in Table 2. Plan evaluation of target volume coverage and normal tissue sparing Figure 4 shows the isodose distributions on axial, sagittal, and coronal images for the oropharyngeal study case generated with the FPMS plan. The red colorwash indicates the GTV, whereas the light green colorwash indicates the CTV. The isodose lines are displayed on an absolute dose scale
588
I. J. Radiation Oncology
● Biology ● Physics
Fig. 3.
Volume 59, Number 2, 2004
A forward planning technique
● N. LEE et al.
589
Fig. 3. (Cont’d). Fig. 3. (a, b) Beam’s-eye views of 13 digitally reconstructed radiographs for a T3N1 supraglottic larynx carcinoma patient. Notice the different MLC shapes in each segment of a particular angle used in this plan. (c) Axial, sagittal, and coronal dose distributions of the same patient. The light blue isodose line corresponds to the 70 Gy line that conforms nicely to the gross tumor volume outlined by the dark blue lines. The yellow isodose lines conform to the CTV in the red colorwash.
ranging from 30 to 70 Gy. The contours of the GTV and the CTV are outlined in red and green, respectively. Notice that the GTV is completely encompassed by the 70 Gy isodose line. The high-risk subclinical disease area is also encompassed by the 59.40 Gy isodose line. No unexpected hot spots were found outside these target volumes. Notice the nice conformality of the isodose curves to the targets, whereas the parotid glands were outside the 45 Gy line, and the majority were within the 30 Gy line. Figure 5 displays the dose–volume histograms of 5 patients treated with the 13-field FPMS technique. All 5 patients had similar coverage of the target volumes, GTV, and CTV. Depending on the respective case, the dose to the spinal cord differed. The average median dose for the parotid glands for all 5 patients was 32.2 Gy. In general, FPMS plans achieved the delivery of the desired prescription dose while minimizing the dose delivered to the normal tissue. Compared to traditional treatment planning, FPMS allows the sparing of parotid glands, which ultimately should impact on the salivary function, as well as the quality of life of patients undergoing radiotherapy.
DISCUSSION Advancements in technology have made it possible to treat many complicated and irregularly shaped head-andneck tumors (1–11, 13). Further advances in 3D-CRT, such as IMRT, allow even better coverage of the tumor target while minimizing the dose delivered to the normal tissues (1–10). IMRT can be divided into two broad categories, forward planning vs. inverse planning. In forward planning, the planning dosimetrist or physicist selects the number, energy, weighting, and angle of beams. The computer then calculates the dose distribution and generates beam’s-eye views, along with dose–volume histograms. The plan is optimized by manual iteration or trial and error. This is in contrast to inverse planning, where one begins by defining the desired dose to the target and the normal tissues, and the computer with its optimization program will go through multiple iterations seeking to find the best beam parameters that will yield the desired dose distribution. Recently, several articles have addressed the physical as well as clinical advantage of IP-IMRT over forward-
590
I. J. Radiation Oncology
● Biology ● Physics
Volume 59, Number 2, 2004
Fig. 4. Axial, sagittal, and coronal dose distributions of a patient with oropharyngeal carcinoma treated with FPMS plan. The red colorwash represents the gross tumor volume, whereas the green colorwash represents the clinical target volume. The dark blue and the light blue lines are the 70 Gy and the 60 Gy lines, respectively.
planned IMRT or conventional radiotherapy (9, 13–16). Since the introduction of IP-IMRT, difficult and advanced head-and-neck tumors that are impossible to cover adequately without exceeding the tolerance of the respective normal tissues are now possible to cover adequately with IP-IMRT. This is true for sites including nasopharynx, oropharynx, and larynx. The conclusion has always been similar, that IP-IMRT achieved better coverage, though more inhomogeneous, when compared to conventional treatment techniques. In addition, there is always more sparing of surrounding normal tissues. However, not every center has a dedicated team of physicists, dosimetrists, and therapists that can handle large volumes of IP-IMRT patients. Without this dedicated team approach, IP-IMRT for complex head-andneck treatment planning may be nearly impossible. In addition, the prolonged treatment time for each IP-IMRT may not be feasible in many busy radiotherapy clinics. Patients may not be able to hold still in the same position throughout the entire IMRT treatment course. Therefore, given the above factors, and because an excellent dose distribution covering the tumor target while sparing the
nearby normal tissues can be achieved with the FPMS technique (10, 11, 13, 17), it may be more efficient to treat patients with FPMS until IP-IMRT becomes feasible in a busy clinic. FPMS can be especially helpful when there is popular demand for IMRT. Instead of having to inform the patients and the referring physicians that one does not have the capability of treating patients using IP-IMRT, one can use the FPMS technique and achieve a similar treatment plan, as shown in Table 2. Please note that IP-IMRT is still superior when compared to the 13-field FPMS with regard to target coverage and normal tissue protection. The FPMS technique is merely an alternative approach to treating head-and-neck cancer patients for the reasons stated above. The initial IP-IMRT experience at Memorial SloanKettering Cancer Center was slightly different from that at the University of California-San Francisco (UCSF) Medical Center (1, 9). Both centers have shown that, given proper dose constraints, the improvement of IPIMRT was due to good coverage in the retropharynx, base of skull, and the medial aspects of the nodal volumes. In addition, IP-IMRT delivered lower doses to the
A forward planning technique
● N. LEE et al.
Fig. 5. Dose–volume histograms for 5 patients treated with forward-planned multisegmented technique (FPMS): (a) gross target volume, (b) clinical target volume, (c) spinal cord, (d) brainstem, and (e) parotid gland. Figure continues on pages 592 and 593.
591
592
I. J. Radiation Oncology
● Biology ● Physics
Volume 59, Number 2, 2004
Fig. 5. (Cont’d).
A forward planning technique
● N. LEE et al.
593
Fig. 5. (Cont’d).
spinal cord, brainstem, mandible, temporal lobes, and parotid glands when compared with 3D-CRT. No attempts were made to spare the parotid glands at Memorial (with a mean dose of 60 Gy), whereas parotid sparing was a goal at UCSF (with a typical mean dose around 26 Gy). An in-house treatment planning system was used at Memorial vs. a commercial system at UCSF. Treatment planning times at Memorial, including contouring, beam definition, optimization, evaluation, and documentation, were approximately 8 hours for IP-IMRT vs. 12 hours for the 3D plan. The daily patient setup and treatment time on average was 15–20 min shorter for IP-IMRT vs. 3D-CRT. In contrast, treatment planning times at UCSF were similar for both IP-IMRT vs. FPMS techniques, in general 8 to 12 hours. Delivery was also shorter with FPMS, 15 min vs. 40 min for IP-IMRT per patient. Perhaps the differences in treatment planning and delivery systems could explain the differences between the two centers. Both centers, however, did conclude that a proper set of dose constraints was essential in IP-IMRT. IP-IMRT can be dangerous when dose specifications are set inappropriately and improperly. Although it is facilitated by computer optimization, IP-IMRT plans are not always optimal, especially in the setting of wrongly specified dose constraints. IP-IMRT may even achieve an inferior plan for a given patient when compared to conventional treatment techniques. An appropriate specification of dose constraints is therefore essential to obtain an optimal result in inverse planning. Otherwise, even in the setting of multiple itera-
tions, the computer may end with the same solution. This can be frustrating to both the physician and the physicists. Therefore, the question is: What is considered the optimal IP-IMRT plan for a given patient? Because the forwardplanning multisegment technique can achieve a very good plan, we believe that all IP-IMRT plans for head-and-neck cancer patients should be just as good as, if not better than, the FPMS plans. This will give the treating physicians or the physicists a confidence interval with which he or she can choose or reject a given plan. Therefore, it is in the authors’ opinion that all IP-IMRT plans should not be inferior to what can be achieved with FPMS, which is in a way a very fancy 3D conformal plan. One final important issue must not be overlooked when treating head-and-neck patients using highly conformal techniques (2, 3). Because both FPMS and IMRT have very sharp dose falloff gradients between the target and surrounding normal tissue, adequate target volume delineation is absolutely essential. Therefore, no matter what technology or excellent supporting staff that one center has, precise target volume delineation is absolutely critical. The treatment planning system will not treat areas not drawn on the CT slices, and the algorithm will even “work hard” to spare regions that are not contoured. Given these factors, at the University of California San Francisco Medical Center, target volume delineation is done with a multidisciplinary team approach consisting of the radiation oncologist, neuroradiologist, and, in the postoperative setting, head-andneck surgeon. The target volumes of each case are carefully
594
I. J. Radiation Oncology
● Biology ● Physics
and accurately defined jointly by the multidisciplinary team. Marginal misses of tumor have been avoided in our patients. CONCLUSION The forward-planned multisegment technique allows the delivery of a high dose to the target coverage while minimizing
Volume 59, Number 2, 2004
the normal tissue toxicity when compared to conventional techniques. In a center without the capabilities of IPIMRT— or in situations where patients are not ideal candidates for that technique—FPMS can be useful and can achieve a dose distribution similar to that with IP-IMRT. Lastly, all IP-IMRT plans should be at least comparable to, if not better than, FPMS plans in the treatment of head-and-neck cancer.
REFERENCES 1. Lee N, Xia P, Quivey JM, et al. Intensity-modulated radiotherapy in the treatment of nasopharyngeal carcinoma: An update of the UCSF experience. Int J Radiat Oncol Biol Phys 2002;53:12–21. 2. Chao C, Ozyigit G, Tran BH, et al. Patterns of failure in patients receiving definitive and postoperative IMRT for headand-neck cancer. Int J Radiat Oncol Biol Phys 2003;55:312– 321. 3. Lee N, Xia P, Fischbein NJ, et al. Intensity-modulated radiation therapy for head and neck cancer: The UCSF experience focusing on target volume delineation. Int J Radiat Oncol Biol Phys 2003;57:49–60. 4. Claus F, Boterberg T, Ost P, et al. Short term toxicity profile for 32 sinonasal cancer patients treated IMRT. Can we avoid dry eye syndrome? Radiother Oncol 2002;64:205–208. 5. Butler EB, The BS, Grant WH, III, et al. SMART (simultaneous modulated accelerated radiation therapy) boost: A new accelerated fractionation schedule for the treatment of head and neck cancer with intensity modulated radiotherapy. Int J Radiat Oncol Biol Phys 1999;45:21–32. 6. Dawson LA, Anzai Y, Marsh L, et al. Patterns of localregional recurrence following parotid sparing conformal and segmental intensity modulated radiotherapy for head and neck cancer. Int J Radiat Oncol Biol Phys 2000;46:1117–1126. 7. Nutting C, Dearnaley DP, Webb S. Intensity-modulated radiation therapy: A clinical review. Br J Radiol 2000;73:459– 469. 8. Xia P, Fu KK, Wong GW, et al. Comparison of treatment plans involving intensity modulated radiotherapy for nasopharyngeal carcinoma. Int J Radiat Oncol Biol Phys 2000;48: 329–337. 9. 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.
10. Sultanem K, Shu KK, Xia P, et al. Three-dimensional intensity-modulated radiotherapy in the treatment of nasopharyngeal carcinoma: The University of California-San Francisco Experience. Int J Radiat Oncol Biol Phys 2000;48:711–722. 11. van Dieren EB, Nowak PJCM, Wijers O, et al. Beam intensity modulation using tissue compensators or dynamic multileaf collimation in three-dimensional conformal radiotherapy of primary cancers of the oropharynx and larynx, including the elective neck. Int J Radiat Oncol Biol Phys 2000;47:1299– 1309. 12. Chuang CF, Verhey LJ, Xia P. Investigation of the use of MOSFET for clinical IMRT dosimetric verification. Med Phys 2002;29:1109–1115. 13. Zabel A, Thilmann C, Zuna I, et al. Comparison of forward planned conformal radiation therapy and inverse planned intensity modulated radiation therapy for esthesioneuroblastoma. Br J Radiol 2002;75:356–361. 14. Kam MKM, Cha RMC, Suen J, et al. Intensity-modulated radiotherapy in nasopharyngeal carcinoma: Dosimetric advantage over conventional plans and feasibility of dose escalation. Int J Radiat Oncol Biol Phys 2003;56:145–157. 15. Hsiung CY, Yorke ED, Chui CS, et al. Intensity-modulated radiotherapy versus conventional three-dimensional conformal radiotherapy for boost or salvage treatment of nasopharyngeal carcinoma. Int J Radiat Oncol Biol Phys 2002;53: 638–647. 16. Cheng JCH, Chao KSC, Low DA. Comparison of intensity modulated radiation therapy (IMRT) techniques for nasopharyngeal carcinoma. Int J Cancer 2001;96:126–131. 17. Xia P, Pickett B, Vigneault E, et al. Forward or inversely planned segmental multileaf collimator IMRT and sequential tomotherapy to treat multiple dominant intraprostatic lesions of prostate cancer to 90 Gy. Int J Radiat Oncol Biol Phys 2001;51:244–254.