Intensity-modulated radiotherapy for early-stage nasopharyngeal carcinoma

A prospective study on disease control and preservation of salivary function

Authors

  • Dora L. W. Kwong M.B.B.S.,

    Corresponding author
    1. Department of Clinical Oncology, the University of Hong Kong, Queen Mary Hospital, Hong Kong, SAR, People's Republic of China
    • Department of Clinical Oncology, 1/F, Professorial Block, Queen Mary Hospital, Pokfulam Road, Hong Kong, SAR, People's Republic of China
    Search for more papers by this author
    • Fax: (011)852 28726426

  • Edmond H. N. Pow M.D.S.,

    1. Oral Rehabilitation, Faculty of Dentistry, The University of Hong Kong, Prince Philip Dental Hospital, Hong Kong, SAR, People's Republic of China
    Search for more papers by this author
  • Jonathan S. T. Sham M.D.,

    1. Department of Clinical Oncology, the University of Hong Kong, Queen Mary Hospital, Hong Kong, SAR, People's Republic of China
    Search for more papers by this author
  • Anne S. McMillan Ph.D.,

    1. Oral Rehabilitation, Faculty of Dentistry, The University of Hong Kong, Prince Philip Dental Hospital, Hong Kong, SAR, People's Republic of China
    Search for more papers by this author
  • Lucullus H. T. Leung Ph.D.,

    1. Department of Clinical Oncology, the University of Hong Kong, Queen Mary Hospital, Hong Kong, SAR, People's Republic of China
    Search for more papers by this author
  • W. Keung Leung Ph.D.,

    1. Periodontology, Faculty of Dentistry, The University of Hong Kong, Prince Philip Dental Hospital, Hong Kong, SAR, People's Republic of China
    Search for more papers by this author
  • Daniel T. T. Chua M.B.Ch.B.,

    1. Department of Clinical Oncology, the University of Hong Kong, Queen Mary Hospital, Hong Kong, SAR, People's Republic of China
    Search for more papers by this author
  • Ashley C. K. Cheng M.B.B.S.,

    1. Department of Clinical Oncology, the University of Hong Kong, Queen Mary Hospital, Hong Kong, SAR, People's Republic of China
    Search for more papers by this author
  • Po M. Wu Ph.D.,

    1. Department of Clinical Oncology, the University of Hong Kong, Queen Mary Hospital, Hong Kong, SAR, People's Republic of China
    Search for more papers by this author
  • Gordon K. H. Au M.B.B.S.

    1. Department of Clinical Oncology, the University of Hong Kong, Queen Mary Hospital, Hong Kong, SAR, People's Republic of China
    Search for more papers by this author

Abstract

BACKGROUND

Xerostomia is a uniform complication after radiotherapy (RT) for nasopharyngeal carcinoma (NPC). Dosimetric studies suggested that intensity-modulated RT (IMRT) can spare part of the parotid glands from high-dose radiation. Disease control and salivary function after IMRT for early-stage NPC was studied prospectively.

METHODS

Thirty-three patients with T1,N0–N1,M0 NPC were treated with IMRT from 2000 to 2002. The prescribed dose was 68–70 grays (Gy) in 34 fractions to gross tumor volume, 64–68 Gy to the planning target volume, and 70 Gy to enlarged cervical lymph nodes. Nineteen patients had stimulated whole salivary (SWS) flow assessment and stimulated parotid salivary (SPS) flow assessment at baseline and at 2 months, 6 months, 12 months, 18 months, and 24 months after the completion of IMRT.

RESULTS

At a median follow-up of 2 years, only 1 neck failure was observed. The 2-year and 3-year local control, distant metastases-free, and overall survival rates all were 100%. The lymph node control and progression-free survival rates were 100% at 2 years and 92.3% at 3 years, respectively. The average mean dose to the parotid gland was 38.8 Gy. The SWS and SPS flow showed continuous recovery: 60% and 47.1% of patients recovered at least 25% of their baseline SPS flow and SWS flow, respectively, at 1 year after completion of IMRT, and the proportions rose to 85.7% and 71.4%, respectively, by 2 years. The pH and buffering capacity of saliva also improved with time.

CONCLUSIONS

Parotid-sparing IMRT achieved good locoregional control, and there was continuous recovery of salivary flow, pH, and buffering capacity in the first 2 years after IMRT in patients with NPC. Cancer 2004. © 2004 American Cancer Society.

Nasopharyngeal carcinoma (NPC) is endemic in Southern China and in South East Asia. In Hong Kong, NPC is the most common head and neck malignancy.1 Compared with other malignancies in the head and neck region, patients with NPC have a younger age distribution. The peak incidence occurs at ages 40–50 years. Unlike other head and neck malignancies, NPC is not related to smoking or alcoholism. Thus, patients who are affected are relatively young and often otherwise healthy, with no comorbid illness. Radiotherapy (RT) is the primary treatment. For patients with early-stage NPC, 60–80% long-term survival after RT alone can be expected.2, 3 Because these patients are likely to survive for decades after treatment, it is important to minimize normal tissue complications due to radiation and to optimize the quality of life of survivors.

The nasopharynx (NP) is surrounded anatomically by an array of radiosensitive structures, like the parotid glands, spinal cord, brainstem, temporal lobes, optic pathways, and auditory pathways. Conventional RT for patients with NPC depends heavily on lateral opposing fields.4 Both parotid glands are irradiated entirely to near tumor dose, not because they are involved by tumor but because the parotids happen to lie in the path of radiation. Among the major salivary glands, the parotid glands are the largest and produce ≈ 60–65% of the oral salivary output.5 The submandibular glands and sublingual glands contribute ≈ 20–30% and 2–5% of salivary output, respectively. The minor salivary glands are distributed throughout the oral cavity and pharynx but are variable from patient to patient. In addition to the parotids, the submandibular, sublingual glands, and some of the minor salivary glands often are irradiated in the large fields used for conventional RT. Thus, xerostomia becomes a uniform and irreversible complication. Not only is the salivary flow reduced, but the quality of saliva also is affected. The post-RT saliva has a lower pH and buffering capacity.6 Thus, because of the change in salivary function, the oral health and quality of life of survivors are impaired.7 Tooth decay, periodontal disease, and oral candidiasis are common problems despite frequent dental care.8 Patients complain of dry mouth and may have problems with speech and swallowing.9

Intensity-modulated RT (IMRT) is an advanced form of conformal RT that conforms high-dose RT to the tumor while conforming low-dose RT to normal tissues. In addition to employing multiple beams that conform to the shape of the target, IMRT also allows for fine modulation of radiation intensity within each radiation beam. Thus, in effect, there are thousands of beamlets, each with calculated intensity to deposit a defined dose at each specific point. A good therapeutic ratio can be achieved by giving a high dose to the tumor to achieve a high probability of local control while minimizing normal tissue complications by limiting the radiation dose to normal tissues. A number of dosimetric studies have shown the potential of IMRT for sparing the parotids from excessive radiation in patients with NPC.10, 11 Early-stage disease, without parapharyngeal involvement or bulky upper cervical lymph nodes encroaching on the parotid glands, is eminently suitable for parotid-sparing IMRT. It also has been shown that the use of IMRT in patients with other head and neck malignancies preserves salivary function after treatment.12, 13 In this study, we report the early results of disease control and preservation of salivary function with IMRT for patients with early-stage NPC.

MATERIALS AND METHODS

Since the year 2000, patients with newly diagnosed, T1,N0–N1,M0 NPC (according to the 1997 American Joint Committee on Cancer staging system14) were treated with a full course of IMRT in our center. All patients were treated with RT alone. The pretreatment evaluation included a complete history and physical examination, nasopharyngoscopy, chest X-ray, complete blood count, liver and renal biochemistry, and computer tomography (CT) scans of the NP and neck. Baseline magnetic resonance imaging (MRI) scans of the NP also were obtained. However, because of limited resources, most patients had MRI scans only just before or after the start of RT. Thus, CT scans were used for staging, treatment decisions, and IMRT planning. Treatment plans were reviewed when MRI results were available. Two patients had disease up-staged from T1 to T2 and from T1 to T3 based on MRI finding. However, a review of the IMRT plans showed adequate coverage of disease, and IMRT was continued as planned originally for these two patients. Both patients had small-volume disease, and chemotherapy was not used.

Treatment Planning and Delivery

Immobilization was achieved with a custom thermoplastic cast from head to shoulders with the patient resting comfortably on a neck support and with a mouth bite. A high-resolution planning CT scan with contrast was taken from the vertex down to below the clavicles with the cast on and in the treatment position. From 1 cm above pituitary fossa to the angle of the jaw, 2.5-mm cuts were taken; and, outside this volume, 5.0-mm cuts were taken. The CT scan was used for localization. Treatment targets and organs at risk were localized on each slice of the CT scan. With a CT scan alone, it often is difficult to distinguish clearly the extent of small NP tumor from the neighboring, uninvolved NP mucosa. It also is difficult to distinguish the NP mucosa, submucosa, or retropharyngeal lymph nodes from the prevertebral muscles, because these structures have similar soft-tissue density. To avoid missing the tumor, relatively generous target volumes were used. For the NP gross tumor volume (GTV), any macroscopic tumor and the whole NP, including the Eustachian cushions on both sides and the prevertebral muscles, were delineated. The clinical target volume (CTV) was modified from Ho's technique,4 which covered the potential sites of local infiltration and with at least a 1-cm margin around the GTV. The CTV included the sphenoid sinus caudal to the base of the pituitary fossa, cavernous sinuses on both sides, the base of the skull (including the medial one-third of the petrous temporal bones and excluding the internal auditory canals and cochleae), inferior orbital fissures, foramen ovale and foramen spinosum, the anterior one-half of the clivus, the posterior one-third of the nasal cavity and antrum, the medial pterygoid muscles and parapharyngeal space up to the styloid process, and the anterior one-half of the arch of cervical vertebrae 1 (C1) and prevertebral muscles inferior to C1. Enlarged cervical lymph nodes were localized as separate lymph node GTV. Potential sites of spread in the upper jugular and accessory chain lymphatics (levels II, III, and VA)15 on both sides of the neck were included in the CTV continuous with that of the NP, and the most cranial jugular lymphatics were outlined up to the base of the skull in the poststyloid area.16 The planning target volume (PTV) was obtained with an additional 2-mm margin beyond the CTV, and the margin was included in manual contouring of the targets. The posterior parts of the submandibular glands often were included in the PTV, because they are adjacent to the upper jugulodigastric lymph node groups. The neck caudal to the chin or caudal to the most distal enlarged lymph node, whichever was more distal, was treated separately by a matching lower-anterior cervical field with midline shield to 60–66 grays (Gy) in 2-Gy daily fractions. Organs at risk that were localized included the lenses; eyes; optic nerves; optic chiasm; pituitary; temporal lobes and brain; brainstem; the inner ear, middle ear, and external auditory meatus on both sides; temporomandibular joints; parotid glands; and spinal cord.

Inverse planning was performed using the Corvus system (version 3.0; NOMOS Corporation) with simulated annealing.17 The prescribed dose to GTV was 68–70 Gy to at least 95% of the GTV and 70 Gy to macroscopically enlarged lymph nodes in 34 fractions. The fractional dose to GTV was 2.0–2.06 Gy. The prescribed dose to PTV was 66–68 Gy to at least 95% of PTV. One patient who was treated early in this series received a prescribed PTV dose of 64 Gy. The fractional dose to PTV was 1.9–2.0 Gy. Table 1 shows the prescription parameters used for inverse planning and selection of the plan. The isodose chosen for prescription usually was at not < 80% of the maximum dose. The deep lobe of the parotid glands often was included in the PTV. For parotid sparing, the objective was to limit the dose to at least one-half of the superficial parotid glands on both sides to < 20 Gy.

Table 1. Dose Constraints for Inverse Planning
VariableValue
  • Gy: grays.

  • a

    No more than 0.01 cc of spinal cord receives > 48 Gy.

  • b

    No more than 0.01 cc of brainstem receives > 58 Gy.

Gross tumor volume 
 Prescribed dose range68–70 Gy
 Minimum dose65 Gy
 Maximum dose77 Gy
 Percentage volume under dosed allowed5%
Planning target volume 
 Prescribed dose range64–68 Gy
 Minimum dose61 Gy
 Maximum dose75 Gy
 Percentage volume under dosed allowed5%
Organs at risk 
 Spinal cord 
  Mean dose40 Gy
  Maximum dose46 Gy
  Percentage volume overdosed allowed5%a
 Eyes, optic nerves, optic chiasm, temporal lobes, brain 
  Mean dose50 Gy
  Maximum dose58 Gy
  Percentage volume overdosed allowed5%
 Brainstem 
  Mean dose50 Gy
  Maximum dose56 Gy
  Percentage volume overdosed allowed5%b
 Parotid glands 
  Mean dose20 Gy
  Maximum dose66 Gy
  Percentage volume overdosed allowed50%
 Pituitary gland 
  Mean dose25 Gy
  Maximum dose45 Gy
  Percentage volume overdosed allowed5%
 Inner ears, middle ears and external auditory meatus, temporomandibular joints
  Mean dose50 Gy
  Maximum58 Gy
  Percentage volume overdosed allowed5%

From April, 2000, to September, 2001, IMRT was delivered by a dynamic, multileaf, intensity-modulating collimator called the MIMiC (NOMOS Corporation) using a slice-by-slice arc-rotation approach.18 The intensity pattern of the beam was modulated every 5° of rotation by opening or closing the leaves in the MIMiC. Treatment usually consisted of 9 300°-arc slices (gantry rotation, from 210° to 150°), using an indexing width of 1.65 cm at the isocenter. After September, 2001, IMRT was delivered with a 4-MV or 6-MV multileaf collimator (MLC) system (Varian Medical Systems, Palo Alto, CA) with step-and-shoot techniques. With the MLC, 9 coplanar, equally spaced beam angles at 0°, 40°, 80°, 120°, 160°, 200°, 240°, 280°, and 320° were used. Each treatment port was divided into multiple, 1 cm × 1 cm beamlets. Verification of the IMRT plan and dose was performed before treatment delivery. The location of the planning origin for MIMiC treatment and the position of the isocenter for MLC treatment were checked weekly with machine check films or electronic portal imaging to ensure the accuracy of the set-up during the course of treatment.

Follow-Up and Assessment

Patients were seen weekly during RT. Radiation toxicity was graded according to the Radiation Therapy Oncology Group radiation morbidity scoring criteria.19 At 6 weeks and 8 weeks after the completion of RT, nasopharyngoscopy and multiple biopsies were performed to assess disease remission in the NP. After the completion of treatment, patients were followed every month during the first year, every 2 months in the second year, and every 3–6 months thereafter. The first posttreatment CT scan was obtained 3–6 months after the completion of RT; thereafter, regular CT or MRI studies were obtained every 6 months or when they were indicated clinically.

Salivary Function Assessment

Patients on IMRT were recruited for this prospective salivary function study. The study was approved by the Institutional Review Board, and informed consent was obtained from all patients. Salivary function was assessed by dental surgeons (E.H.N.P. and A.S.M.) at a dental work-up before treatment and then at 2 months, 6 months, 12 months, 18 months, and 24 months after the completion of RT. Stimulated parotid salivary (SPS) flow and stimulated whole salivary (SWS) flow were measured. Whole saliva was collected in a sterile cup over a 5-minute period while the patient chewed on a piece of sterile rubber tubing. Saliva from the parotid gland was collected using a Lashley cup attached over either the right or left parotid duct.20 Parotid saliva was collected over a period of 15 minutes while the patient chewed on sterile rubber tubing. The device was held in place by a suction effect and is a standard, nontraumatic technique for parotid saliva collection. Each patient had only one parotid gland measured, and the same side was assessed at subsequent visits. Collection occurred at the same time of day on each occasion. Saliva volume was measured, and pH and buffering capacity were assessed immediately after collection. The amount of saliva obtained was divided by the time period of collection to obtain the flow rate in mL/minute. Buffering capacity was reported as nil, low, medium, or high.

Dose-Volume Analysis of the Targets and Parotid Glands

Dose-volume histograms of all IMRT plans were reviewed for quantitative analysis. For GTV and PTV, the minimum dose received by 95% of the target (D95) was used to reflect target coverage, and the maximum dose received by 5% of the target volume (D5) was used to reflect dose inhomogeneity within target. The prescription isodose and the mean, minimum, and maximum doses to targets also were reported.

Because the Corvus system did not allow for overlapping structures in localization, parts of the deep lobe of the parotid glands that were included in the PTV were excluded from dose-volume histograms of the parotids in the original treatment plan (Fig. 1). These would lead to an erroneously low reported dose to the parotids. To obtain the true dose-volume histogram of the parotid glands for correlation with salivary function, the original planning CT scan for individual patients was recontoured for the parotid volumes without the treatment targets. The actual treatment plan was then superimposed onto the recontoured CT scan to analyze the dose-volume histogram of the parotids. Dose-volume histogram statistics of the parotid gland chosen for salivary flow measurement were analyzed. The mean dose, the dose to 50% of the parotid gland (D50), and the percentage volume of the parotid that received > 25 Gy were reported. In addition to the dose to the parotid gland, doses to the submandibular glands and the oral cavity also were obtained and correlated with SWS flow.

Figure 1.

Isodose distribution with intensity-modulated radiotherapy (IMRT) planning at the level of the parotid glands. 1: gross target volume; 2: planning target volume; 3: right parotid; 4: left parotid; 5: spinal cord. The prescribed dose was 68 grays (Gy) at the 83%-isodose level. Most of the superficial part of the parotid glands received < 24.6 Gy (< the 30%-isodose level).

Statistical Analysis

Follow-up was counted from the date of completion of RT. The NP and cervical lymph node recurrence-free survival, distant metastases free survival, progression-free survival, and overall survival rates were estimated using the Kaplan–Meier method. The percentage reductions in SPS flow and SWS flow at different time points after the completion of RT were compared with baseline salivary flow before the start of RT. Correlations between SPS and SWS flow at 12 months after completion of RT and between the percentage reduction in SPS and SWS flow and the mean dose to parotid glands were performed using Pearson correlations. The probability (P) value was 2-sided, and P values < 0.05 were considered significant.

RESULTS

From April, 2000, to December, 2002, 33 patients with T1,N0–N1,M0 NPC were treated with IMRT. Twenty-seven patients had Stage I NPC, and 6 patients had Stage II NPC, as determined on CT scans and clinical staging. With additional information from MRI studies, 1 patient was upstaged to T2 disease, and another patient was upstaged to T3 disease. All patients had undifferentiated or poorly differentiated carcinoma. The median age at diagnosis was 43 years (range, 29–74 years). Table 2 summarizes the patient characteristics and treatment parameters. On average, the prescription isodose was 81.8% of the maximum dose and ranged from 79.1% to 86%. Table 3 summarizes the dose-volume histogram statistics for the targets. The average mean dose achieved was 71.3 Gy to the NP, 72.1 Gy to cervical lymph nodes, and 70.1 Gy to the PTV. Ninety-five percent of the NP GTV (D95) received at least 68.9 Gy, and the corresponding doses to cervical lymph node GTV and PTV were 70.0 Gy and 64.7 Gy, respectively. All patients completed IMRT. Maximum Grade 2 and 3 mucositis occurred in 48.5% and 33.3% of patients, respectively, during the course of IMRT. Maximum Grade 2 and 3 skin reactions occurred in 12.1% and 26% of patients, respectively. One patient required admission for hydration and skin care. The median overall treatment time was 45 days (range, 44–52 days), and 97% of patients completed 34 fractions within 7 weeks.

Table 2. Patient Characteristics and Treatment Parameters
Characteristics and treatment parametersNo. of patients (%)
  1. CT: computed tomography; MRI: magnetic resonance imaging; MIMiC: multileaf intensity-modulating collimator.

Gender 
 Male16 (48.5)
 Female17 (51.5)
Tumor classification 
 By CT 
  T133 (100)
 By MRI 
  T131 (94)
  T21 (3)
  T31 (3)
Lymph node status 
 N027 (81.8)
 N16 (18.2)
Prescribed dose 
Gross tumor volume 
 68 Gy20 (60.6)
 70 Gy13 (39.4)
Planning target volume 
 64 Gy1 (3.0)
 66 Gy22 (66.7)
 68 Gy10 (30.3)
Treatment delivery 
 MIMiC15 (45.5)
 Multileaf collimator18 (54.5)
Table 3. Target Dose-Volume Histogram Statistics
VariableAverage (range)
Gross tumor volumePlanning target volume
NasopharynxNeck lymph node
  1. Gy: grays.

Volume (cc)29.7 (9.7–80.8)9.3 (2.9–17.0)374.1 (246.7–564.4)
Mean dose (Gy)71.3 (69.5–74.1)72.1 (71.8–72.5)70.1 (66.4–71.8)
Minimum dose (Gy)65.6 (59.3–67.9)67.3 (66.7–67.8)23.9 (13.0–34.1)
Maximum dose (Gy)78.9 (76.1–84.9)78.1 (76.6–80.3)84.1 (79.1–88.1)
D95: 95% of target volume received dose > (Gy)68.9 (66.7–71.3)70.0 (69.7–70.4)64.7 (59.5–67.8)
D5: 5% of target volume received dose > (Gy)74.5 (71.7–78.5)74.8 (74.4–75.3)76.3 (73.2–80.0)

The median follow-up for this group was 24 months (range, 11–42 months). All but one patient achieved complete remission in the NP and cervical lymph nodes after the completion of IMRT. One patient had persistent disease in the NP and was salvaged with additional intracavitary brachytherapy. This patient had T1 disease, as assessed by CT and MRI studies. Another patient had a recurrence in a cervical lymph node at 27 months after RT. Review of the CT scan and the IMRT plan at the time of recurrence showed that the lymph node recurrence occurred at a cervical lymph node that measured < 1 cm and that was not contoured separately as lymph node GTV in IMRT planning but was included in the PTV. This patient received a prescribed dose of 64.0 Gy to the PTV, and the mean dose to the cervical lymph node was estimated at 64.2 Gy. This was the only patient in this series who had received a lower PTV dose of 64 Gy. Thus, an inadequate dose to the involved lymph node may have accounted for the lymph node failure in this patient, who underwent salvage with neck dissection. There was no NP recurrence or distant metastases. At the time of analysis, the 2-year and 3-year NP recurrence-free survival, distant metastases-free survival, and overall survival rates were 100%. The 2-year and 3-year lymph node recurrence-free survival rates were 100% and 92.3%, respectively. Figure 2 shows the progression-free survival rate for the whole group (100% at 2 years, 92.3% at 3 years).

Figure 2.

Progression-free survival of the whole group. RT: radiotherapy.

Salivary Function after Treatment

Nineteen patients from this series participated in a prospective study of salivary function assessment. All 19 patients had pre-RT assessment, 17 patients had assessment up to 12 months post-RT, and 7 patients had assessment up to 24 months post-RT. Table 4 summarizes the dose-volume histogram statistics of the parotid gland, submandibular glands, and oral cavity for these 19 patients. The average mean dose to the parotid glands was 38.8 Gy. On average, 50% of the parotid glands received > 34.6 Gy, and 66.1% of the parotid volume received > 25 Gy. The average mean dose to the submandibular glands and the oral cavity were greater than the dose to the parotid glands, because no specific attempts were made during IMRT planning to spare these structures. Figures 3 and 4 show the time course of recovery of SWS and SPS flow after the completion of RT. There was an initial reduction in SWS and SPS flow after RT, followed by a gradual, continuous recovery in the first 2 years. Table 5 summarizes the results of salivary function assessment. At 2 months post-RT, the mean SWS flow was reduced by 88.5% compared with baseline, and the mean SPS flow was reduced by 68.9%. By 1 year post-RT, the mean percentage reduction of SWS flow improved to 72.2% and further improved to 58.7% by 2 years. The recovery of SPS flow was even better: The mean percentage reduction compared with baseline improved to 14.1% by 1 year and to 12.4% by 2 years. The SPS flow at 12 months post-RT varied from no recovery to 4.8 times the baseline SPS flow; and, for 4 patients, the SPS flow at 12 months post-RT was higher than the baseline measurement. The high level of SPS flow in these 4 patients tends to improve the mean percentage reduction in SPS flow for the whole group. Another way to gauge the recovery of salivary flow is to measure the proportion of patients who did not have severe xerostomia at different time points after RT. Eisbruch et al. defined severe xerostomia as post-RT stimulated salivary flow that was < 25% of the pre-RT flow.21 At 1 year after the completion of RT, 60.0% and 47.1% of patients recovered at least 25% of their baseline SPS flow and SWS flow, respectively. By 2 years after the completion of RT, 85.7% and 71.4% of patients had recovered at least 25% of their baseline SPS and SWS flow, respectively. The recovery of SPS appeared to be more significant than SWS recovery, and there was a positive correlation between SPS flow and SWS flow at 12 months post-RT (Pearson correlation, 0.59; P = 0.019).

Table 4. Dose-Volume Histogram Statistics for Salivary Glands (n = 19 patients)
VariableAverage (range)
  1. Gy: grays.

Volume of parotid gland (cc)22.4 (11.2–45.8)
Mean dose to parotid gland (Gy)38.8 (32.0–46.1)
V25: % volume of parotid gland that received > 25 Gy66.1 (51.3–82.1)
D50: 50% volume of parotid gland that received a dose > (Gy)34.6 (25.8–44.6)
Mean dose to right submandibular gland (Gy)64.9 (57.4–70.0)
Mean dose to left submandibular gland (Gy)63.9 (54.1–68)
Mean dose to oral cavity (Gy)45.1 (38.2–49.2)
Figure 3.

Changes in stimulated whole salivary flow with time after radiotherapy (RT). Pre-RT baseline at 0 months. N: the number of patients assessed.

Figure 4.

Changes in stimulated parotid salivary flow with time after radiotherapy (RT). Pre-RT baseline at 0 months. N: the number of patients assessed.

Table 5. Recovery of Salivary Function with Time after Radiotherapy
MosNo. of patientsStimulated whole salivaStimulated parotid salivaWhole saliva mean pHWhole saliva buffering capacity: No. of patients (%)
Mean flow (mL/min)Mean % dropMean flow (mL/min)Mean % dropNil/lowMedium/high
0194.780.927.21 (5.3)18 (94.7)
2190.4788.50.1668.96.413 (68.4)6 (31.6)
6190.9280.50.21806.511 (57.9)8 (42.2)
12171.3372.20.5914.16.95 (29.4)12 (70.6)
18131.4271.60.6237.16.92 (15.4)11 (84.6)
2472.7358.70.6912.47.21 (14.3)6 (85.7)

Table 5 also shows that there was an improvement in the quality of the saliva with time after RT. The baseline saliva was alkaline, with a mean pH of 7.2, and 94.7% of patients had saliva with medium-to-high buffering capacity. Immediately after RT, the saliva was acidic, with low buffering capacity. The saliva secreted also became less acidic and had improved buffering capacity as salivary flow gradually increased with time. The mean saliva pH dropped to 6.4 at 2 months post-RT and improved to 6.9 at 1 year. By 2 years, the baseline alkalinity was regained, with the mean pH returned to 7.2. The proportion of patients who had saliva with medium-to-high buffering capacity dropped to 31.6% at 2 months post-RT but increased to 70.6% at 1 year and to 85.7% at 2 years.

The correlation between the percentage change in SPS flow at 12 months post-RT and the mean dose to the parotid was significant (Pearson correlation, −0.54; P = 0.037) (Fig. 5). The correlation between the percentage reduction in SWS flow at 12 months post-RT and the mean dose to the parotid was not significant (Pearson correlation, − 0.39; P = 0.12). There also was no significant correlation between SWS flow and mean dose to the submandibular glands or the oral cavity.

Figure 5.

This scatter plot shows the percentage change in stimulated parotid salivary (SPS) flow at 12 months postradiotherapy in relation to the mean dose to the parotid gland.

DISCUSSION

Early results showed excellent disease control with IMRT. All patients had disease controlled in the NP. The prescribed dose to GTV was 68–70 Gy, which was not different from the dose used for conventional RT. However, the actual mean dose achieved by IMRT was higher, with an average of 71.3 Gy to the NP and 72.1 Gy to cervical lymph nodes. The higher dose achieved in GTV may have contributed to better locoregional control. The only recurrence observed in this study occurred in a cervical lymph node that was not enlarged significantly but probably was involved and received about 64 Gy. Although 50–56 Gy have been used for a clinically negative neck,16 a higher dose probably is required to control more than microscopic disease, especially if RT is used alone. Relative under dosing in the involved lymph node may have contributed to lymph node failure in this patient. The experience with IMRT for NPC reported by Lee et al. from the University of California–San Francisco (UCSF) also showed excellent locoregional control, with 4-year local and locoregional control estimates of 97% and 98%, respectively.22 However, the UCSF report included patients with all stages of disease. Most patients had Stage III/IV disease and had received chemoradiotherapy. Patients from USCF received 65–70 Gy to GTV and an additional boost of 5–7 Gy with intracavitary brachytherapy after external beam irradiation. In the current series, the dose to GTV was 68–70 Gy, and no additional boost was given after primary IMRT. Only one patient received intracavitary brachytherapy for persistent NP disease. The dose to PTV in this series was 64–68 Gy, which was higher than the 60-Gy dose employed at UCSF. The use of concurrent chemoradiotherapy probably allows for a reduction in the RT dose to the PTV and a clinically negative neck. However, the patient who had a cervical lymph node recurrence after receiving 64 Gy in the current series indicates a need for caution in reducing the dose to the PTV, because there may be unrecognized subclinical disease other than microscopic disease in the PTV. Preliminary results from the current series indicate that our target coverage and prescribed doses for IMRT are adequate for patients with early-stage disease. RT alone appeared to be sufficient for treating early-stage and low-volume disease.

In this series, CT scans were used for localization and IMRT planning. However, most IMRT planning systems now allow for image fusion. Thus, additional information from MRI, positron-emission tomography, or other biologic imaging studies may help in localizing the extent of tumor. With further development and general availability of advanced imaging, the margin required for CTV may be reduced further, and the dose to uninvolved areas may be lowered.

Nineteen patients had prospective salivary function assessment. Stimulated whole salivary flow reflects salivary output from all salivary tissues, whereas SPS flow reflects output from parotid glands. Although there was a severe reduction in salivary flow in the immediate post-RT period, there was continuous recovery of both SPS and SWS flow in the first 2 years after IMRT. IMRT allows for parotid sparing, and the improvement of SPS flow after RT was more obvious than the improvement of SWS flow. In addition to the parotid glands, the submandibular glands and minor salivary glands in the oral cavity and pharynx also are irradiated in treatment for NPC. Thus, sparing the parotid glands alone will not preserve SWS flow completely. However, a statistically significant, positive correlation between SPS and SWS flow was found in this series. For some patients, the SPS flow after RT was higher than the baseline level. Thus, the increased salivary output from residual functioning parotid tissue may compensate for reduced output from other damaged salivary tissues and may contribute to the recovery of SWS flow. In addition to the recovery of salivary flow, the quality of saliva also improved with time. After RT, the saliva turned acidic and had low buffering capacity. However, within 24 months, pH and buffering capacity of saliva returned to relatively normal values. Preservation of salivary flow, pH, and buffering capacity is important for oral and dental health. Such improvements have never been observed after conventional RT.6 Xerostomia used to be considered an irreversible complication.5

Eisbruch et al. reported a correlation between parotid salivary flow recovery and mean dose received by the parotid gland.21 A mean dose threshold of 24 Gy and 26 Gy were required for unstimulated and stimulated parotid salivary flow, respectively, to recover to at least 25% of baseline flow at 12 months after completion of RT. With different modeling, Chao et al. predicted a higher threshold dose of 32 Gy for stimulated whole salivary flow.23 With our current target volume coverage and dose prescription, the average mean parotid dose received was 38.8 Gy (range, 32.0–46.1 Gy). Although the mean dose was higher than the reported threshold, recovery of both SPS and SWS were observed in the current series, with 60% and 47.1% of patients recovering at least 25% of their baseline SPS flow and SWS flow, respectively, at 12 months after completion of RT, and the proportions rose to 85.7% and 71.4%, respectively, by 24 months. A significant correlation between the mean parotid dose and the percentage change of SPS flow at 12 months post-RT also was found. Thus, although a very low whole parotid mean dose may not be achievable, sparing part of the parotid gland still appears to be useful for preservation of salivary function. Further follow-up will be required to determine the impact of parotid-sparing IMRT on the quality of life of survivors.

Ancillary