- Email: [email protected]
Treatment of stage I nasopharyngeal carcinoma: Analysis of the patterns of relapse and the results of withholding elective neck irradiation
hr. J. Radrorron Oncology Em/. Php.,Vol.17, pp.II83-1I90 F'rinkd inthe U.S.A. All rights rererved.
Copyright
0360-3016/89 $3.00 +.fXl 0 1989Pergamon Press plc
0 Original Contribution TREATMENT OF STAGE I NASOPHARYNGEAL CARCINOMA: ANALYSIS THE PATTERNS OF RELAPSE AND THE RESULTS OF WITHHOLDING ELECTIVE NECK IRRADIATION
OF
ANNE W. M. LEE, F.R.C.R., JONATHAN S. T. SHAM, F.R.C.R., Y. F. POON, F.R.C.R. AND JOHN H. C. Ho, M.D., D.Sc., F.R.C.P., F.R.C.R. (D&T),
F.R.C.R.A.,
F.A.C.R.
Institute of Radiology and Oncology, Medical and Health Department, Queen Elizabeth Hospital, Wylie Road, Kowloon, Hong Kong This is a retrospective analysis of 196 patients with nasopharyngeal carcinoma Stage I (Ho’s classification) treated by megavoltage radiation during 1980-1984. The primary target volume included all potential sites of local invasion and the first station lymph nodes at retropharyngeal spaces. Two different dose schedules were used, both gave a total tumor dose biologically equivalent to 65 Gy by conventional fractionation, and both achieved a S-year actuarial local-recurrence-free survival of 88%. Elective neck irradiation was withheld in all except seven patients. The overall 7-year actuarial survival was 85%, but the relapse-free survival was only 62%. The patterns of relapse, prognostic factors, and treatment complications were analyzed. Eighteen patients (9%) recurred locally. Radical retreatment with radiation achieved complete remission in seven out of fifteen cases. Distant failure occurred in 17 patients (9%). Although 57 (30%) of the 189 patients without elective neck irradiation subsequently showed lymph node involvement, none of the seven regionally-treated patients relapsed. The successful regional salvage rate was 81% overall (46 out of 57 patients), but 90% (44 of 49) for those properly treated with whole neck irradiation. However, the 7-year actuarial survival was lower in patients with nodal relapse than those without (70% versus 87%) because of the associated higher incidence of hematogenous dissemination. The various aspects of treatment, the value of elective neck irradiation in particular, are discussed. Nasopharyngeal carcinoma, Stage I (Ho’s classification), Megavoltage radiation therapy, No elective whole neck irradiation, Patterns of relapse, Prognostic factors, Treatment complications.
and End Results Reporting (AJC) (1) subdivide tumors confined to the nasopharynx into Tl and T2 according to whether one or more walls are involved. This has been strongly criticized because this has no proven prognostic significance, NPC often spreads submucosally, and it is not always possible to define with accuracy the limits of tumor within the small and irregularly-shaped nasopharyngeal space (6, 14, 19, 23, 25, 30, 32). Thus, the Ho’s classification (19), which groups all tumors confined to the nasopharynx irrespective of the number of walls involved as Stage I (T 1NOMO), is used throughout the subsequent discussion. Table 1 summarizes the available data on Stage I squamous-cell carcinoma of the nasopharynx. Clinically staged T 1NOM0 cases constituted only 7 to 25% (median = 13%) of the published series, and their 5-year survival rate ranged from 35 to 83% (median = 58%). There is no dispute that radiation therapy (RT) is the
INTRODUCTION Nasopharyngeal carcinoma (NPC)is notorious for its high propensity for extensive local infiltration, lymphatic spread, and even hematogenous dissemination. Early detection is by no means easy because of the inaccessibility to direct visualization. It is not surprising, therefore, that only a minority of patients present with disease still confined to the nasopharynx. Specific data on this subgroup is scanty and their patterns of failure have not been separately analyzed. However, such information will be crucial for further attempts to optimize treatment. A review of the literature on NPC is confusing because of the inclusion of non-epithelial malignancies in many reported series and the different staging systems used by different institutes. Both the classifications recommended by the International Union Against Cancer (UICC) (40) and the American Joint Committee for Cancer Staging
Acknowledgements-The
authors wish to thank all the staff of the Institute for their contribution in the management of this series of patients, and MS G. Lau for her secretarial assistance. Accepted for publication 15 June 1989.
Presented in part at the E.O.R.T.C. Symposium on Recent Advances in Cancer Management, in Hong Kong, 20-23 September, 1987. Reprint requests to: Anne W. M. Lee, F.R.C.R. 1183
1184
I. J. Radiation Oncology 0 Biology 0 Physics Table 1. Summary
of data on squamous-cell
carcinoma
Time period
Authors
1952-74 1955-76 1958-69 1933-56 1959-78 1971-80 1958-83 1969-74 1956-73 1954-77 1940-68 1970-76 1970-79 1961-76 1940-7 I 1960-83
Baker (4) Bedwinek
et al. (5) Bertelsen et a/. (6) Bohorquez (7) Cellai et al. (8) Dickson and Flares (12) Haghbin et al. (15)
Ho et al. (21) Hoppe et al. (22) Mesic et al. (28) Moench and Phillips (30) Payne (3 I ) Rahima et al. (35) Urdaneta et al. (4 1) Wang (43) Yamashita et al. (46)
AND MATERIALS
Patient characteristics There was a total of 199 patients. Three cases who defaulted before completion of treatment were excluded from analysis. Of the 196 evaluable patients, 142 were males and 54 females. Their age ranged from 22 to 79 Table 2. Updated
5-year result of Ho’s randomized
of the nasopharynx
As % of series
5-year survival %
20
20 13 7 9 8 17 10 9 21 10 16 8 20 18 25 16
35 46 50 67 80 58 46 83 63
years (median = 47 years). Fifty-two (27%) of them were under 40 years old at the time of diagnosis. All had histologically proven undifferentiated or poorly differentiated squamous-cell carcinoma of the nasopharynx (lymphoepithelioma is considered as a variant of undifferentiated carcinoma). The observation period ranged from 14 to 87 months (median = 53 months). Clinicul staging All patients were staged by thorough physical examination, anterior rhinoscopy, indirect +_direct nasopharyngoscopy, blood biochemistries, chest radiograph, and five X ray views of the nasopharyngeal region as recommended by Ho (17). Fifty-three patients had additional investigations with conventional tomography for confirmation of the extent of disease. Computerized tomography was not available for routine staging until late 1982, hence its findings were ignored in the present analysis. In 62 patients (32%), the tumor appeared to involve only one wall within the nasopharynx, whereas in the rest, two or more walls had been infiltrated, that is, 68% of our cases would have been Staged as II (T2NOMO) instead of I (TlNOMO) by the UICC and the AJC systems. Radiation therapy The treatment techniques and dosage schemes have been described in detail in Ho’s previous publications (18, 20).
study ( 18) on the value of ENI for Stage I nasopharyngeal
Survival rates (%) ENI
No. of cases
Overall
Yes No
75 77
71 76
ENI = Elective neck irradiation.
Stage I (Ho’s classification)
No. of cases
14 6 13 II 23 8 175 17 24 24 II 18 7 62 16
treatment of choice, but the optimal technique and dosage have yet to be properly explored. Among the controversies is the need for routine prophylactic treatment of the whole regional lymphatics. This is being strongly advocated by many authorities (2, 5, 7,9,22,26-30, 32, 34, 37, 39,42, 43,46). Although Ho (18), based on information from a randomized study carried out in the early 1970’s, suggested that neck irradiation can be postponed until the first clinical evidence of lymph node (LN) involvement without any significant jeopardy on both the overall survival and the disease-free survival. Table 2 shows the updated result of that trial. Because of this early experience, our treatment policy is to omit elective whole neck irradiation (ENI) unless patient cannot be closely observed, but the first station LN at the retropharyngeal spaces together with all potential sites of local invasion are routinely included within the primary target volume. This is a retrospective analysis of the patterns of failure in the Stage I cases treated during 1980-1984. METHODS
December 1989, Volume 17, Number 6
Disease-free 65 71
carcinoma
Relapse rates (%) Local
Regional
28 25
5 25
Distant 15 5
Stage I nasopharyngeal carcinoma 0 A. W. M. LEE et al.
1185
Seven patients had EN1 via an anterior field with median shield for larynx and spinal cord. Using 4.5 or 6 MeV photons, a cumulative dose of 39.2 Gy at 90% isodose level was given in 7 weekly fractions. This was equivalent to 59.4 Gy by conventional fractionation (NSD = 1629 ret).
Fig. 1. The composite dose distribution
at the principal plane.
All patients were treated by 4.5 or 6 MeV photons using two lateral parallel opposed fields supplemented by an
anterior field. Besides the nasopharynx, the target volume included the posterior parts of both nasal fossae and maxillary sinuses, posterior ethmoidal cells, sphenoid sinuses, inferior orbital fissures, petro-occipital fissures, medial parts of petrous bones, and the retropharyngeal spaces. The composite dose distribution at the principal plane is illustrated in Figure 1. Because of a serious shortage of treatment machines, the 40 patients treated before March 1981 were given a mean tumor dose (MTD) of 4.2 Gy per fraction, twice weekly, for a total of 12 fractions. The nominal standard dose (NSD) (13) was 1861 ret, and this was biologically equivalent to 65 Gy by conventional fractionation of 2 Gy daily. With increasing concern about late radiation damages and the addition of a linear accelerator, the dose per fraction was reduced in the subsequent 156 patients. The schedule consisted of 2.5 Gy MTD per fraction, 4 times weekly, for 24 fractions. This gave a similar NSD of 1870 ret, but a lower dose (1286 versus 1477 btu) in terms of brain tolerance units (btu) (33). Eleven patients (6%) with residual growths were given additional boosts 2 to 12 weeks after the scheduled course. Seven of them were treated by intracavitary caesium (ICC) for another lo-40 Gy at 5-10 mm depth, one was given an additional 5 Gy via reduced external fields, and in three the boost consisted of lo-20 Gy by ICC plus 5-10 Gy by external irradiation.
Salvage treatment Among the patients who subsequently developed cervical lymph-adenopathy, five were treated palliatively with involved field RT, and three were left untreated (because of uncontrolled local recurrence in two and distant metastases in one patient). Forty-nine patients were treated radically with whole neck irradiation as described, using dose schedules of 39.2 Gy/7 fractions/42 days (NSD = 1629 ret) in 28 patients, 45.6 Gy/12 fractions/38 days (NSD = 1683 ret) in 8, and 48 Gy/12 fractions/38 days (NSD = 1772 ret) in 13. These were biologically equivalent to 59.4 Gy, 59.1 Gy, and 64.5 Gy, respectively. Seven of them had additional local boosts for residual nodal diseases. Fifteen patients with local recurrence were re-treated radically with external irradiation using various techniques and dose schedules. The total MTD given ranged from 50 Gy to 66 Gy at conventional fractionation. Chemotherapy was used in 12 patients-six for locoregional recurrences and six for distant metastases. Various combinations had been used, including the use of Cisplatin regimens in five cases.
RESULTS Survival Figure 2 shows the survival curves of the whole series of Stage I patients. The 7-year actuarial survival rate
\\
0.6.
\ \ \
\a_
a 2 2
0.7
1-m
‘d
0.5.
\ -----1 -\______2!_
0.6
.
x 5 z
0.4
.
p” I 0.3
,
a 0.2
0.1
1 01 0
6
12
16
24
30
30
42
46
54
60
66
7176
64
Months
Fig. 2. The actuarial survival (A) and actuarial relapse-free survival (B) curves of 196 patients with stage I nasopharyngeal carcinoma.
1186
1. J. Radiation Oncology 0 Biology 0 Physics
Fig. 3. The actuarial local-recurrence-free survival achieved by the two different fractionation schemes.
(ASR) was 85%, but the relapse-free only 62%.
survival
curves
(RFS) was
Response to primary treatment All patients achieved primary complete remission (CR)-185 (94%) after the scheduled course of RT, whereas the remaining eleven after booster doses. Both fractionation schedules achieved similar local control (Fig. 3) and the 5-year actuarial local-recurrence-free survival was 88%.
Patterns of‘ relapse and response to salvage treatment Sixty-eight patients (35%) subsequently relapsed-the first site of failure being local in 1 1, regional in 53, and distant in 4 cases. The patterns of failure are shown in Table 3. Local. Eighteen patients (9%) relapsed locally 7 to 59 months (median = 25 months) after completion of primary treatment. The tumor recurred in the nasal fossa in one case, remained confined within the nasopharynx in eight, but extended cranially to cause erosion of the base of skull +- cranial nerve palsies in nine patients.
December 1989, Volume 17, Number 6
Radical re-treatment by external irradiation achieved CR for more than 6 months in 7 (47%) out of 15 patients. The response rate was only l/8 (13%) in patients with cranial extension, but 6/7 (86%) in those without. Among the seven complete responders, four remained diseasefree for over 18 to 30 months after salvage treatment, one recurred locally at 12 months, whereas two others died of hematogenous dissemination at 12 months. Patients with local recurrence were associated with a higher risk of failure at other sites. Altogether, four (22%) had subsequently developed distant metastases, and eight others (44%) had regional relapse documented at the time of local recurrence or thereafter. Regional. None of the seven patients with ENI relapsed, whereas 57 (30%) of the 189 patients without ENI developed lymphatic metastases 3 to 72 months (median = 16 . months) after initial staging. Fifty of them manifested at the upper, four mid, and three lower cervical level as defined in Ho’s classification (19). Forty-four (90%) of 49 patients with therapeutic irradiation to the whole neck achieved complete nodal control for more than 6 months, whereas two of five patients treated with involved field RT remained free of nodal disease. Thus, the overall regional salvage rate was 81% (46 out of 57). Only two patients free from failure at other sites suffered from persistent regional involvement. Despite the high salvage rate, the 7-year ASR of the 57 patients with regional relapse was lower than those without-70% versus 87% (Fig. 4). The difference could have been greater as there were three defaulted cases, six deaths from intercurrent diseases, and two deaths from treatment complications in the latter group, but no similar losses in the former. The difference in outcome with respect to the presence or absence of regional and/or local relapses is shown in Table 4. Patients with regional involvement were associated with a higher risk of hematogenous dissemination. Excluding
0.6
16)
\
\-__ \
0.6
.
L_\ ---
Table 3. Patterns of relapse in Stage I nasopharyngeal carcinoma
a :
Regional*
0.6
-
0.5
-
2
0.4
-
2 ;
0.3
.
0.2
.
0.1
,
(A)
2
Sites of relapse Local
I I---,--,
0.7
‘d
Distant
Number of patients 4
37 4 10 3 9
1
A
2
o\ 0
6
12
18
24
30
33
42
46
64
60
66
72
78
64
Mosrhs
Total = 68 * All occurred
in patients
without
elective neck irradiation.
Fig. 4. The actuarial survival curves of the 57 patients gional relapse (A) and the I39 patients without (B).
with re-
Stage I nasopharyngeal carcinoma 0 A. W. M. LEE et al.
Table 4. Outcome by type of loco-regional relapse
Table 5. Analysis of possible prognostic factors
Cases with loco-regional relapse Outcome
Local
Regional
Alive without disease Alive with L-R disease Alive with D disease Died of L-R disease Died of D disease Died of intercurrent cause Died of RT complication Untraced
1 1 0 2 3 -
35 2 1 0 8 -
7
46
Both
None
3 6 0 1 -
117 0 I 0 3 6 2 3
11
132
7-year results (%) Factor
Total
1
1187
L-R = Local and/or regional; D = Distant; RT = Radiation therapy.
ASR
RFS
Age
~40 yr. 240 yr.
89 82
64 59
Sex
Female Male
89 82
69 57
Extent
1 wall >1 wall
85 87
69 59
Relapse (%) Local
Regional
Distant
10 9
24 31
8 9
ASR = Overall actuarial survival rate; RFS = Actuarial relapse-free survival rate.
reported (24). Among the 22 patients affected in the present series, hyperprolactinemia was found in 17 patients, secondary hypothyroidism in three, and pan-hypopituitarism in two. Cranial neuropathy involved the second nerve in four patients, the sixth in one, and the last four in three cases. Thus far, two patients have died-one of brainstem and the other of temporal lobe necrosis. The overall incidence rate was 28% (48/ 170) for patients who had only one basic course of RT, 55% (6/l 1) for those with additional boosts, and 40% (6/l 5) for cases who received two courses of external irradiation, respectively. The differences between the three groups are not statistically significant. Further analysis of patients who had only one basic course of RT using the same technique shows that the overall incidence rate was 54% (19/35) for those treated with the 4.2 Gy per fraction scheme, but 2 1% (29/ 135) for those with the 2.5 Gy per fraction scheme. The difference is significant (p = 0.0 12). However, if only 4-year cumulated incidences are compared, the difference-29%
been previously patients who also suffered from local recurrence, the metastatic rate was 9/46 (20%) in cases with regional disease, but 4/132 (3%) in those without. Six patients with persistent loco-regional disease were treated with various combination chemotherapy, but all failed to respond. Distant. Hematogenous metastases were detected in 17 patients (9%) 11 to 61 months (median = 25 months)
after the initial staging. They were documented after local and/or regional relapses in all except four cases. The commonest sites included bones ( 13 patients), lungs ( 13) and liver (7). The median survival following manifestation of metastases was 12 months. Of the six cases treated with chemotherapy, only one patient with lung secondaries showed partial response to Cisplatin lasting for 14 months. Analysis of possible prognostic factors Table 5 shows the 7-year ASR and RFS when analyzed by age, sex, and the extent of primary tumor infiltration. Although females and patients below 40-years old did slightly better than males and older patients, respectively, the differences are not statistically significant. When patients are subclassified according to whether one or more walls within the nasopharynx are involved, there is no significant difference between the two groups both in the survival results and the patterns of relapse. Intercurrent diseases Six relapse-free patients died from causes unrelated to NPC. Two of them succumbed to a second malignancy in the bronchus. Treatment complications Sixty patients (3 1%) developed one or more late irradiation sequelae besides varying degrees of dryness of mouth. Table 6 lists the damages and their incidences. The latent interval for the neurological complications ranged from 14 to 70 months (median = 44 months). Symptomatic hypothalamic-pituitary dysfunction has
Table 6. Late irradiation sequelae Radiation therapy: no. of courses One
No. of patients Latency <4 years Latency >4 years Incidence at sites Brainstem Cranial nerve Temporal lobe Hypo-pit. axis TM joint Ear
4.2 Gy
2.5 Gy
One + boost
Two
Total
10 9
26 3
5 I
5 1
46 14
-
-
1 4 4
I 8 6 22 12 29
4 3 7 53 10
3 1 9 14
I I 2 5 1
Hypo-pit. axis = Hypothalamic-pituitary = Temporo-mandibular joint.
-
axis; TM joint
1188
1. J. Radiation Oncology 0 Biology 0 Physics
(10/35) versus 19% (26/l 35) fails to reach statistical nificance (p = 0.46).
sig-
DISCUSSION Because of the difficulty, if not impossibility, of resecting the base of skull or the retropharyngeal spaces, surgery has little role in the primary treatment of NPC, even for Stage I cases. RT is the modality of choice as both the primary tumor and its lymphatic metastases are radiosensitive. The anatomical proximity to critical structures, however, presents special limitation. To achieve the highest cure rate with minimal morbidity remains one of the greatest challenges for radiation oncologists. The retrospective analysis of Hoppe et ul. (22) demonstrated that although nearly one-third (6/17) of their patients with local recurrence had initial T 1 growths, there had been no such failures in the 7 years following the adoption of larger treatment volumes. This suggests that local involvement is often more extensive than clinically evident, and hence the field coverage is a significant factor in determining the chance of success. This well-recognized characteristic of extensive tumor infiltration and difficulty of accurately defining the extent of microscopic involvement dictate that all potential sites must be included within the target volume, even for apparently Tl growths (18, 20, 22, 46). Advances in radiologic technology undoubtedly allow better delineation of the macroscopic extent. With the introduction of computerized tomography and magnetic resonance imaging, a significant number of Stage I cases would have been upstaged and the incidence of geographic miss might have been reduced. But whether such additions, with their present degree of accuracy, can allow safe reduction in field size has yet to be cautiously tested. The dose-response relationship for Tl growths is still poorly defined. Although Chang et al. (9) and Bedwinek et al. (5) suggested that 60 Gy (at conventional fractionation of 2 Gy daily) may be adequate, Moench and Phillips (30) Mesic et al. (28), and Vikram t’t al. (42) demonstrated that better control can be achieved with higher doses. Like most other centers, we advocate a total dose biologically equivalent to 65 Gy. Primary CR lasting more than 6 months was achieved in 185/ 196 (94%) of our Stage I patients, and only 17 (9%) of them subsequently recurred. With this high local control rate, routine boosting by intracavitary caesium insertion is not warranted. However. careful re-assessment after the completion of RT is obviously important, as patients with residual tumors may benefit from additional doses. With appropriate boosting, all of our eleven partial-responders eventually attained CR, and only one subsequently relapsed. Of the patients with local recurrence, 7/l 5 (47%) responded completely to aggressive re-treatment with external irradiation. Similar to the experience of Wang (44) the response rate was high for patients with recurrence still confined within the nasopharynx, but extremely poor
December 1989, Volume 17, Number 6
for those with cranial extension (6/7 versus l/8). Thus far, only one of the seven complete-responders recurred again, but longer observation is required to assess the longterm control and morbidity rates of the second course of RT. Besides the problem of local failure, these relapsed patients were associated with a greater risk of developing lymphatic and/or hematogenous metastases, the incidence being 44% and 22%, respectively. Additional treatment may be needed-EN1 (if not previously given) should at least be considered. The need for including the whole regional lymphatics in the primary treatment of Stage I patients remains controversial. ENI is strongly advocated by many authors (2, 5, 7, 9, 22, 26-30, 32, 34, 37, 39, 42, 43, 46) because of the high incidence of overt LN involvement even for Tl tumors (ranging from 62% to 86% in the literature), the universal experience of very low regional relapse rates following EN1 in node-negative cases, and the low morbidity of the treatment. However, its exact impact on the final outcome has never been extensively tested by controlled trials, and conflicting results showing a lack of significant benefit have been published (6, 10, 18,25). There are only two studies with specific data for T 1NOM0 patients. Chang et ul. (9) reported that patients with ENI had a lower incidence of developing LN metastases (6% versus 16%). Ho ( 18) also found a similar trend (5% versus 25%). However, he showed that, with prompt salvage by therapeutic irradiation, the 5-year ASR and disease-free survival rates in the group without initial ENI were by no means inferior (Table 2). Hence, it was suggested that EN1 can safely be withheld if the patient could be closely observed. The present series confirms Ho’s previous finding that 70%) of Stage I patients do not require prophylactic treatment to the whole neck, provided that the first station LN at the retropharyngeal spaces are adequately covered within the primary target volume. Radical therapeutic irradiation of the whole neck successfully salvaged 44/49 (90%) of the relapsed cases, and only two patients suffered from uncontrolled regional disease alone. However, the present analysis indicates that there is a difference in the 7-year ASR when patients with nodal relapse are compared to those without (70% versus 87%). Of the 178 local-recurrence-free patients, 46 (26%) developed cervical LN metastases, and they were associated with a higher risk of hematogenous dissemination (9/46 versus 4/ 132, that is, 20% versus 3%). It is impossible to tell from this retrospective analysis whether such distant metastases arose from the primary growths or occult foci in the untreated regional LN, but we share the concern of Mendenhall et al. (27) that a significant proportion of them might have originated from the latter. Although this question about the value of EN1 on the ultimate survival can only be answered by a prospective randomized trial with an adequate number of patients, there is little doubt that EN1 can at least improve the relapse-free rate. Our present 7-year RFS of 62% is far from being ideal; regional relapse, though largely salvage-
Stage 1 nasopharyngeal carcinoma 0 A. W. M. LEE et al.
able, is psychologically distressing. Hence, if treatment machine availability allows, EN1 ought to be considered, especially for patients who cannot be closely observed. No significant prognostic factors can be identified in our present analysis. Similar to the studies by Chatani et al. (10) Hoppe et al. (22), and Yamashita et al. (46), our data support the criticisms (6, 14, 19, 23, 25, 30, 32) that subclassification according to whether one or more walls within the nasopharynx are involved, as suggested by the AJC and the UICC systems, has no value in determining the treatment strategy or predicting prognosis. The relatively high incidence of late treatment complications is one of our major concerns. With increasing evidence that nervous tissues are much less tolerant of large fractional doses than connective tissues, and hence, Ellis’ (13) NSD formula (TD = NSD X f124 X TO.“) has to be modified for more accurate reflection of tolerance (3, 11, 16, 33, 38, 45), our dose per fraction was reduced from 4.2 Gy to 2.5 Gy, and the brain tolerance unit formula (TD = btu X No.45X T”.03)suggested by Pezner and Archambeau (33) is also applied in the estimation of doses. To achieve a homogeneous dose of 65 Gy to all potential sites of local invasion, the dose to brain tissues included within the target volume (NSD = 1859 ret, btu = 1212) will exceed the usually quoted tolerance limit of 1700 ret (36) or 1050 btu (33), even if treatment is given at the
1189
conventional fractionation rate of 2 Gy daily. A definite risk of morbidity seems inevitable for the best chance of survival. The total doses given by our two fractionation schemes were 186 1 ret and 1870 ret in terms of NSD, but 1477 and 1286 in terms of btu. Although the overall incidence of damages by the former scheme of 4.2 Gy per fraction is indeed significantly higher (54% versus 2 1%, p = 0.012), this should not be taken as valid proof of the greater damaging effect by the larger fractional dose, because of the longer observation period for patients treated by the former scheme and the long latent interval before development of sequelae (median = 44 months). As the shortest follow-up for survivors by the latter fractionation schedule was 4 years, comparison of relative risk should best be based on the 4-year cumulated data for patients who had only one basic course of RT using the same technique. With the same survival rates (90%) by both schemes, the cumulated incidence of late complications was higher with the former-29% versus 19%. However, this difference is not statistically significant (p = 0.46). Longer follow-up is necessary to assess the ultimate result. Further reduction of fractional dose to 1.8-2 Gy is probably required for significant improvement, but major break-through may have to await the development of effective and non-toxic radioprotectors.
REFERENCES 1. American Joint Committee for Cancer Staging and End Results Reporting: Manual for staging of cancer 1978. Chicago: American Joint Committee; 1978:38. 2. Andel, J. G. van; Hop, W. C. J. Carcinoma of the nasopharynx: a review of 86 cases. Clin. Radiol. 33:95-99; 1982. 3. Aristizabal, S. A.; Caldwell, W. L.; Avila, J. The relationship of time-dose fractionation factors to complications in the treatment of pituitary tumors by irradiation. Int. J. Radiat. Oncol. Biol. Phys. 2:667-673; 1977. 4. Baker, S. R. Nasopharyngeal carcinoma: clinical course and results of therapy. Head Neck Surg. 3:8- 14; 1980. 5. Bedwinek, J. M.; Perez, C. A.; Keys, D. J. Analysis of failures after definitive irradiation for epidermoid carcinoma of the nasopharynx. Cancer 4512725-2729; 1980. 6. Bertelsen, K.; Andersen, A. P.; Elbrand, 0.; Lund C. Malignant tumors of the nasopharynx. Acta Radiol. Ther. Phys. Biol. 14:177-186; 1975. 7. Bohorquez, J. Factors that modify the radio-response of cancer of the nasopharynx. Am. J. Roentgenol. 126:863876; 1976. 8. Cellai, E.; Chiavacci, A.; Olmi, P.; Carcangiu, M. L. Carcinoma of the nasopharynx: results of radiation therapy. Acta Radio]. Oncol. 21 (Fast. 2):87-95; 1982. 9. Chang, C.; Liu, T.; Chang, Y.; Cao, S. Radiation therapy of nasopharyngeal carcinoma. Acta Radiol. Oncol. 19(Fasc. 6):433-438; 1980. IO. Chatani, M.; Teshima, T.; Inoue, T.; Azuma, I.; Yoshimura, H.; Oshitani, T.; Hashiba, M.; Nishiyama, K.; Tsutsui, K.; Fujimura, T.; Araki, Y.; Hishikawa, Y.; Arita, S.; Yamada, C.; Kondo, K.; Kagemoto, M.; Tanaka, M.; Imajo, Y. Radiation therapy for nasopharyngeal carcinoma: retrospective review of 105 patients based on a survey of Kansai Cancer Therapist Group. Cancer 57:2267-2271; 1986. 11. Cohen, L.; Creditor, M. Iso-effect tables for tolerance of
12.
13. 14.
15.
16.
17. 18.
19.
20.
21.
irradiated normal human tissues. lnt. J. Radiat. Oncol. Biol. Phys. 9:233-241; 1983. Dickson, R. I.; Flores, A. D. Nasopharyngeal carcinoma: An evaluation of 134 patients treated between 197 1-l 980. Laryngoscope 95:276-283; 1985. Ellis, F. Dose, time and fractionation: a clinical hypothesis. Clin. Radiol. 20: l-7; 1969. Fu, K. K. Prognostic factors of carcinoma of the nasopharynx (editorial). Int. J. Radiat. Oncol. Biol. Phys. 6:523-526; 1980. Haghbin, M.; Kramer, S.; Patchefsky, A. S.; Prestipino, A. J.; Diener-West, M. D. Carcinoma of the nasopharynx: a 25-year study. Am. J. Clin. Oncol. (CCT) 8:384-392; 1985. Harris, J. R.; Levene, M. B. Visual complications following irradiation for pituitary adenomas and craniopharyngiomas. Radiology 120:167-171; 1976. Ho, J. H. C. Radiologic diagnosis of nasopharyngeal cancer. J.A.M.A. 220:396-398; 1972. Ho, J. H. C. An epidemiologic and clinical study of nasopharyngeal carcinoma. Int. J. Radiat. Oncol. Biol. Phys. 4: 183-198; 1978. Ho, J. H. C. Stage classification of nasopharyngeal carcinoma: A review. In: De The, G., Eto, Y., eds. Nasopharyngeal carcinoma: etiology and control, IARC Scientific Publications No. 20. Lyon: International Agency for Research on Cancer; 1978:99- 113. Ho, J. H. C. Nasopharynx. In: Halnan, K. E., Boak, J. L., Crowther, D., von Essen, C. E., Orr, J. S., Peckham, M. J., eds. Treatment of cancer. London: Chapman and Hall; 1982:249-267. Ho, J. H. C.; Lau, W. H.; Fong, M.; Chan, C. L.; Au, G. K. H. Treatment of nasopharyngeal carcinoma (NPC). In: Grundmann, E., Krueger, G. R. F., Ablashi, D. V., eds. Nasopharyngeal carcinoma, cancer campaign, Vol. 5.
1190
22.
23.
24.
25.
26. 27.
28.
29.
30.
3 1. 32.
33.
1. J. Radiation Oncology 0 Biology 0 Physics
Stuttgart and New York: Gustav Fischer Verlag; 198 1:279285. Hoppe, R. T.; Goffinet, D. R.; Bagshaw, M. A. Carcinoma of the nasopharynx: eighteen years’ experience with megavoltage radiation therapy. Cancer 37:2605-26 12; 1976. Huang, S.-C. Nasopharyngeal cancer: a review of 1605 patients treated radically with cobalt 60. Int. J. Radiat. Oncol. Biol. Phys. 6:401-407; 1980. Lam, K. S. L.; Ho, J. H. C.; Lee, A. W. M.; Tse, V. K. C.; Chan, P. K.; Wang, C.; Ma, J. T. C.; Yeung, R. T. T. Symptomatic hypothalamic-pituitary dysfunction in nasopharyngeal carcinoma patients following radiation therapy: a retrospective study. Int. J. Radiat. Oncol. Biol. Phys. 13: 1343-1350; 1987. Larsson, L.-G.; Seelig, I. Malignant nasopharyngeal tumors: result of radiation therapy. Acta Radio]. Ther. Phys. Biol. 15:209-218; 1976. Lederman, M. Cancer of the nasopharynx: its natural history and treatment. Illinois: Springfield; 196 1. Mendenhall, W. M.; Million, R. R.; Cassisi, N. J. Elective neck irradiation in squamous-cell carcinoma of the head and neck. Head Neck Surg. 3: 15-20; 1980. Mesic, J. B.; Fletcher, G. H.; Goepfert, H. Megavoltage irradiation of epithelial tumors of the nasopharynx. Int. J. Radiat. Oncol. Biol. Phys. 7:447-453: 198 I. Million, R. R.; Fletcher, G. H.: Jesse, R. H. Evaluation of elective irradiation of the neck for squamous-cell carcinoma of the nasopharynx, tonsillar fossa and base of tongue. Am. J. Roentgenol. 80:973-988; 1963. Moench, H. C.; Phillips, T. L. Carcinoma of the nasopharynx: review of 146 patients with emphasis on radiation dose and time factors. Am. J. Surg. 124:5 15-5 18; 1972. Payne, D. G. Carcinoma of the nasopharynx. J. Otolaryngol. 12:197-202; 1983. Perez, C. A.; Ackerman, L. V.; Mill, W. B.; Ogura, J. H.; Powers, W. E. Cancer of the nasopharynx: factors influencing prognosis. Cancer 24: I - 17; 1969. Pezner, R. D.; Archambeau, J. 0. Brain tolerance unit: A method to estimate the risk of radiation brain injury for various dose schedules. Int. J. Radiat. Oncol. Biol. Phys. 7: 397-402; 1981.
December 1989, Volume 17, Number 6 34. Qin, D.; Hu, Y.; Yan, J.; Xu, G.; Cai, W.; Wu, X.; Cao, D.; Gu, X. Analysis of I379 patients with nasopharyngeal carcinoma treated by radiation. Cancer 6 1: 1117- 1124; 1988. 35. Rahima, M.; Rakowsky, E.; Barzilay, J.; Sidi J. Carcinoma of the nasopharynx: an analysis of 9 1 cases and a comparison ofdiffering treatment approaches. Cancer 58:843-849; 1986. 36. Rottenberg, D. A.; Chemik, N. L.; Deck, M. D. F.; Ellis, F.; Posner, J. B. Cerebral necrosis following radiotherapy of extracranial neoplasms. Ann. Neurol. 1:339-357; 1977. 37. Scanlon, P. W.; Rhodes, R. E.; Woolner, L. B.; Devine, K. D.; McBean, J. B. Cancer of the nasopharynx: 142 patients treated in the 1I year period 1950-1960. Am. J. Roentgenol. 99:3 13-325; 1967. 38. Scheline, G. E.; Wara, W. M.; Smith, V. Therapeutic irradiation and brain injury. Int. J. Radiat. Oncol. Biol. Phys. 6:1215-1228; 1980. 39. Thompson, R. W.; Doggett, R. L. S.; Bagshaw, M. A. Tenyear experience with linear accelerator irradiation of cancer of the nasopharynx. Radiology 97: 149-155; 1970. 40. Union Internationale Contre le Cancer (UICC). TNM classification of malignant tumors, 3rd edition. Geneva: UICC; 1978:27-32. 4 1. Urdaneta, N.; Fischer, J. J.; Vera, R.; Gutierrez, E. Cancer of the nasopharynx: review of 43 cases treated with supervoltage radiation therapy. Cancer 37: 1707- 17 12; 1976. 42. Vikram, B.; Mishra, U. B.; Strong, E. W.; Manolatos, S. Patterns of failure in carcinoma of the nasopharynx: 1. Failure at the primary site. Int. J. Radiat. Oncol. Biol. Phys. 11: 1455-1459; 1985. 43. Wang, C. C. Treatment of malignant tumors of the nasopharynx. Otolaryngol. Clin. N. Amer. 13:477-48 1; 1980. 44. Wang, C. C. Re-irradiation of recurrent nasopharyngeal carcinoma-treatment techniques and results. Int. J. Radiat. Oncol. Biol. Phys. 13:953-956; 1987. 45. Wigg, D. R.; Koschel, K.; Hodgson, G. S. Tolerance of the mature human central nervous system to photon irradiation. Br. J. Radiol. 54:787-798; 1981. 46. Yamashita, S.; Kondo, M.; Inuyama, Y.; Hashimoto, S. Improved survival of patients with nasopharyngeal squamous cell carcinoma. Int. J. Radiat. Oncol. Biol. Phys. 12: 307-3 12: 1986.
Copyright
0360-3016/89 $3.00 +.fXl 0 1989Pergamon Press plc
0 Original Contribution TREATMENT OF STAGE I NASOPHARYNGEAL CARCINOMA: ANALYSIS THE PATTERNS OF RELAPSE AND THE RESULTS OF WITHHOLDING ELECTIVE NECK IRRADIATION
OF
ANNE W. M. LEE, F.R.C.R., JONATHAN S. T. SHAM, F.R.C.R., Y. F. POON, F.R.C.R. AND JOHN H. C. Ho, M.D., D.Sc., F.R.C.P., F.R.C.R. (D&T),
F.R.C.R.A.,
F.A.C.R.
Institute of Radiology and Oncology, Medical and Health Department, Queen Elizabeth Hospital, Wylie Road, Kowloon, Hong Kong This is a retrospective analysis of 196 patients with nasopharyngeal carcinoma Stage I (Ho’s classification) treated by megavoltage radiation during 1980-1984. The primary target volume included all potential sites of local invasion and the first station lymph nodes at retropharyngeal spaces. Two different dose schedules were used, both gave a total tumor dose biologically equivalent to 65 Gy by conventional fractionation, and both achieved a S-year actuarial local-recurrence-free survival of 88%. Elective neck irradiation was withheld in all except seven patients. The overall 7-year actuarial survival was 85%, but the relapse-free survival was only 62%. The patterns of relapse, prognostic factors, and treatment complications were analyzed. Eighteen patients (9%) recurred locally. Radical retreatment with radiation achieved complete remission in seven out of fifteen cases. Distant failure occurred in 17 patients (9%). Although 57 (30%) of the 189 patients without elective neck irradiation subsequently showed lymph node involvement, none of the seven regionally-treated patients relapsed. The successful regional salvage rate was 81% overall (46 out of 57 patients), but 90% (44 of 49) for those properly treated with whole neck irradiation. However, the 7-year actuarial survival was lower in patients with nodal relapse than those without (70% versus 87%) because of the associated higher incidence of hematogenous dissemination. The various aspects of treatment, the value of elective neck irradiation in particular, are discussed. Nasopharyngeal carcinoma, Stage I (Ho’s classification), Megavoltage radiation therapy, No elective whole neck irradiation, Patterns of relapse, Prognostic factors, Treatment complications.
and End Results Reporting (AJC) (1) subdivide tumors confined to the nasopharynx into Tl and T2 according to whether one or more walls are involved. This has been strongly criticized because this has no proven prognostic significance, NPC often spreads submucosally, and it is not always possible to define with accuracy the limits of tumor within the small and irregularly-shaped nasopharyngeal space (6, 14, 19, 23, 25, 30, 32). Thus, the Ho’s classification (19), which groups all tumors confined to the nasopharynx irrespective of the number of walls involved as Stage I (T 1NOMO), is used throughout the subsequent discussion. Table 1 summarizes the available data on Stage I squamous-cell carcinoma of the nasopharynx. Clinically staged T 1NOM0 cases constituted only 7 to 25% (median = 13%) of the published series, and their 5-year survival rate ranged from 35 to 83% (median = 58%). There is no dispute that radiation therapy (RT) is the
INTRODUCTION Nasopharyngeal carcinoma (NPC)is notorious for its high propensity for extensive local infiltration, lymphatic spread, and even hematogenous dissemination. Early detection is by no means easy because of the inaccessibility to direct visualization. It is not surprising, therefore, that only a minority of patients present with disease still confined to the nasopharynx. Specific data on this subgroup is scanty and their patterns of failure have not been separately analyzed. However, such information will be crucial for further attempts to optimize treatment. A review of the literature on NPC is confusing because of the inclusion of non-epithelial malignancies in many reported series and the different staging systems used by different institutes. Both the classifications recommended by the International Union Against Cancer (UICC) (40) and the American Joint Committee for Cancer Staging
Acknowledgements-The
authors wish to thank all the staff of the Institute for their contribution in the management of this series of patients, and MS G. Lau for her secretarial assistance. Accepted for publication 15 June 1989.
Presented in part at the E.O.R.T.C. Symposium on Recent Advances in Cancer Management, in Hong Kong, 20-23 September, 1987. Reprint requests to: Anne W. M. Lee, F.R.C.R. 1183
1184
I. J. Radiation Oncology 0 Biology 0 Physics Table 1. Summary
of data on squamous-cell
carcinoma
Time period
Authors
1952-74 1955-76 1958-69 1933-56 1959-78 1971-80 1958-83 1969-74 1956-73 1954-77 1940-68 1970-76 1970-79 1961-76 1940-7 I 1960-83
Baker (4) Bedwinek
et al. (5) Bertelsen et a/. (6) Bohorquez (7) Cellai et al. (8) Dickson and Flares (12) Haghbin et al. (15)
Ho et al. (21) Hoppe et al. (22) Mesic et al. (28) Moench and Phillips (30) Payne (3 I ) Rahima et al. (35) Urdaneta et al. (4 1) Wang (43) Yamashita et al. (46)
AND MATERIALS
Patient characteristics There was a total of 199 patients. Three cases who defaulted before completion of treatment were excluded from analysis. Of the 196 evaluable patients, 142 were males and 54 females. Their age ranged from 22 to 79 Table 2. Updated
5-year result of Ho’s randomized
of the nasopharynx
As % of series
5-year survival %
20
20 13 7 9 8 17 10 9 21 10 16 8 20 18 25 16
35 46 50 67 80 58 46 83 63
years (median = 47 years). Fifty-two (27%) of them were under 40 years old at the time of diagnosis. All had histologically proven undifferentiated or poorly differentiated squamous-cell carcinoma of the nasopharynx (lymphoepithelioma is considered as a variant of undifferentiated carcinoma). The observation period ranged from 14 to 87 months (median = 53 months). Clinicul staging All patients were staged by thorough physical examination, anterior rhinoscopy, indirect +_direct nasopharyngoscopy, blood biochemistries, chest radiograph, and five X ray views of the nasopharyngeal region as recommended by Ho (17). Fifty-three patients had additional investigations with conventional tomography for confirmation of the extent of disease. Computerized tomography was not available for routine staging until late 1982, hence its findings were ignored in the present analysis. In 62 patients (32%), the tumor appeared to involve only one wall within the nasopharynx, whereas in the rest, two or more walls had been infiltrated, that is, 68% of our cases would have been Staged as II (T2NOMO) instead of I (TlNOMO) by the UICC and the AJC systems. Radiation therapy The treatment techniques and dosage schemes have been described in detail in Ho’s previous publications (18, 20).
study ( 18) on the value of ENI for Stage I nasopharyngeal
Survival rates (%) ENI
No. of cases
Overall
Yes No
75 77
71 76
ENI = Elective neck irradiation.
Stage I (Ho’s classification)
No. of cases
14 6 13 II 23 8 175 17 24 24 II 18 7 62 16
treatment of choice, but the optimal technique and dosage have yet to be properly explored. Among the controversies is the need for routine prophylactic treatment of the whole regional lymphatics. This is being strongly advocated by many authorities (2, 5, 7,9,22,26-30, 32, 34, 37, 39,42, 43,46). Although Ho (18), based on information from a randomized study carried out in the early 1970’s, suggested that neck irradiation can be postponed until the first clinical evidence of lymph node (LN) involvement without any significant jeopardy on both the overall survival and the disease-free survival. Table 2 shows the updated result of that trial. Because of this early experience, our treatment policy is to omit elective whole neck irradiation (ENI) unless patient cannot be closely observed, but the first station LN at the retropharyngeal spaces together with all potential sites of local invasion are routinely included within the primary target volume. This is a retrospective analysis of the patterns of failure in the Stage I cases treated during 1980-1984. METHODS
December 1989, Volume 17, Number 6
Disease-free 65 71
carcinoma
Relapse rates (%) Local
Regional
28 25
5 25
Distant 15 5
Stage I nasopharyngeal carcinoma 0 A. W. M. LEE et al.
1185
Seven patients had EN1 via an anterior field with median shield for larynx and spinal cord. Using 4.5 or 6 MeV photons, a cumulative dose of 39.2 Gy at 90% isodose level was given in 7 weekly fractions. This was equivalent to 59.4 Gy by conventional fractionation (NSD = 1629 ret).
Fig. 1. The composite dose distribution
at the principal plane.
All patients were treated by 4.5 or 6 MeV photons using two lateral parallel opposed fields supplemented by an
anterior field. Besides the nasopharynx, the target volume included the posterior parts of both nasal fossae and maxillary sinuses, posterior ethmoidal cells, sphenoid sinuses, inferior orbital fissures, petro-occipital fissures, medial parts of petrous bones, and the retropharyngeal spaces. The composite dose distribution at the principal plane is illustrated in Figure 1. Because of a serious shortage of treatment machines, the 40 patients treated before March 1981 were given a mean tumor dose (MTD) of 4.2 Gy per fraction, twice weekly, for a total of 12 fractions. The nominal standard dose (NSD) (13) was 1861 ret, and this was biologically equivalent to 65 Gy by conventional fractionation of 2 Gy daily. With increasing concern about late radiation damages and the addition of a linear accelerator, the dose per fraction was reduced in the subsequent 156 patients. The schedule consisted of 2.5 Gy MTD per fraction, 4 times weekly, for 24 fractions. This gave a similar NSD of 1870 ret, but a lower dose (1286 versus 1477 btu) in terms of brain tolerance units (btu) (33). Eleven patients (6%) with residual growths were given additional boosts 2 to 12 weeks after the scheduled course. Seven of them were treated by intracavitary caesium (ICC) for another lo-40 Gy at 5-10 mm depth, one was given an additional 5 Gy via reduced external fields, and in three the boost consisted of lo-20 Gy by ICC plus 5-10 Gy by external irradiation.
Salvage treatment Among the patients who subsequently developed cervical lymph-adenopathy, five were treated palliatively with involved field RT, and three were left untreated (because of uncontrolled local recurrence in two and distant metastases in one patient). Forty-nine patients were treated radically with whole neck irradiation as described, using dose schedules of 39.2 Gy/7 fractions/42 days (NSD = 1629 ret) in 28 patients, 45.6 Gy/12 fractions/38 days (NSD = 1683 ret) in 8, and 48 Gy/12 fractions/38 days (NSD = 1772 ret) in 13. These were biologically equivalent to 59.4 Gy, 59.1 Gy, and 64.5 Gy, respectively. Seven of them had additional local boosts for residual nodal diseases. Fifteen patients with local recurrence were re-treated radically with external irradiation using various techniques and dose schedules. The total MTD given ranged from 50 Gy to 66 Gy at conventional fractionation. Chemotherapy was used in 12 patients-six for locoregional recurrences and six for distant metastases. Various combinations had been used, including the use of Cisplatin regimens in five cases.
RESULTS Survival Figure 2 shows the survival curves of the whole series of Stage I patients. The 7-year actuarial survival rate
\\
0.6.
\ \ \
\a_
a 2 2
0.7
1-m
‘d
0.5.
\ -----1 -\______2!_
0.6
.
x 5 z
0.4
.
p” I 0.3
,
a 0.2
0.1
1 01 0
6
12
16
24
30
30
42
46
54
60
66
7176
64
Months
Fig. 2. The actuarial survival (A) and actuarial relapse-free survival (B) curves of 196 patients with stage I nasopharyngeal carcinoma.
1186
1. J. Radiation Oncology 0 Biology 0 Physics
Fig. 3. The actuarial local-recurrence-free survival achieved by the two different fractionation schemes.
(ASR) was 85%, but the relapse-free only 62%.
survival
curves
(RFS) was
Response to primary treatment All patients achieved primary complete remission (CR)-185 (94%) after the scheduled course of RT, whereas the remaining eleven after booster doses. Both fractionation schedules achieved similar local control (Fig. 3) and the 5-year actuarial local-recurrence-free survival was 88%.
Patterns of‘ relapse and response to salvage treatment Sixty-eight patients (35%) subsequently relapsed-the first site of failure being local in 1 1, regional in 53, and distant in 4 cases. The patterns of failure are shown in Table 3. Local. Eighteen patients (9%) relapsed locally 7 to 59 months (median = 25 months) after completion of primary treatment. The tumor recurred in the nasal fossa in one case, remained confined within the nasopharynx in eight, but extended cranially to cause erosion of the base of skull +- cranial nerve palsies in nine patients.
December 1989, Volume 17, Number 6
Radical re-treatment by external irradiation achieved CR for more than 6 months in 7 (47%) out of 15 patients. The response rate was only l/8 (13%) in patients with cranial extension, but 6/7 (86%) in those without. Among the seven complete responders, four remained diseasefree for over 18 to 30 months after salvage treatment, one recurred locally at 12 months, whereas two others died of hematogenous dissemination at 12 months. Patients with local recurrence were associated with a higher risk of failure at other sites. Altogether, four (22%) had subsequently developed distant metastases, and eight others (44%) had regional relapse documented at the time of local recurrence or thereafter. Regional. None of the seven patients with ENI relapsed, whereas 57 (30%) of the 189 patients without ENI developed lymphatic metastases 3 to 72 months (median = 16 . months) after initial staging. Fifty of them manifested at the upper, four mid, and three lower cervical level as defined in Ho’s classification (19). Forty-four (90%) of 49 patients with therapeutic irradiation to the whole neck achieved complete nodal control for more than 6 months, whereas two of five patients treated with involved field RT remained free of nodal disease. Thus, the overall regional salvage rate was 81% (46 out of 57). Only two patients free from failure at other sites suffered from persistent regional involvement. Despite the high salvage rate, the 7-year ASR of the 57 patients with regional relapse was lower than those without-70% versus 87% (Fig. 4). The difference could have been greater as there were three defaulted cases, six deaths from intercurrent diseases, and two deaths from treatment complications in the latter group, but no similar losses in the former. The difference in outcome with respect to the presence or absence of regional and/or local relapses is shown in Table 4. Patients with regional involvement were associated with a higher risk of hematogenous dissemination. Excluding
0.6
16)
\
\-__ \
0.6
.
L_\ ---
Table 3. Patterns of relapse in Stage I nasopharyngeal carcinoma
a :
Regional*
0.6
-
0.5
-
2
0.4
-
2 ;
0.3
.
0.2
.
0.1
,
(A)
2
Sites of relapse Local
I I---,--,
0.7
‘d
Distant
Number of patients 4
37 4 10 3 9
1
A
2
o\ 0
6
12
18
24
30
33
42
46
64
60
66
72
78
64
Mosrhs
Total = 68 * All occurred
in patients
without
elective neck irradiation.
Fig. 4. The actuarial survival curves of the 57 patients gional relapse (A) and the I39 patients without (B).
with re-
Stage I nasopharyngeal carcinoma 0 A. W. M. LEE et al.
Table 4. Outcome by type of loco-regional relapse
Table 5. Analysis of possible prognostic factors
Cases with loco-regional relapse Outcome
Local
Regional
Alive without disease Alive with L-R disease Alive with D disease Died of L-R disease Died of D disease Died of intercurrent cause Died of RT complication Untraced
1 1 0 2 3 -
35 2 1 0 8 -
7
46
Both
None
3 6 0 1 -
117 0 I 0 3 6 2 3
11
132
7-year results (%) Factor
Total
1
1187
L-R = Local and/or regional; D = Distant; RT = Radiation therapy.
ASR
RFS
Age
~40 yr. 240 yr.
89 82
64 59
Sex
Female Male
89 82
69 57
Extent
1 wall >1 wall
85 87
69 59
Relapse (%) Local
Regional
Distant
10 9
24 31
8 9
ASR = Overall actuarial survival rate; RFS = Actuarial relapse-free survival rate.
reported (24). Among the 22 patients affected in the present series, hyperprolactinemia was found in 17 patients, secondary hypothyroidism in three, and pan-hypopituitarism in two. Cranial neuropathy involved the second nerve in four patients, the sixth in one, and the last four in three cases. Thus far, two patients have died-one of brainstem and the other of temporal lobe necrosis. The overall incidence rate was 28% (48/ 170) for patients who had only one basic course of RT, 55% (6/l 1) for those with additional boosts, and 40% (6/l 5) for cases who received two courses of external irradiation, respectively. The differences between the three groups are not statistically significant. Further analysis of patients who had only one basic course of RT using the same technique shows that the overall incidence rate was 54% (19/35) for those treated with the 4.2 Gy per fraction scheme, but 2 1% (29/ 135) for those with the 2.5 Gy per fraction scheme. The difference is significant (p = 0.0 12). However, if only 4-year cumulated incidences are compared, the difference-29%
been previously patients who also suffered from local recurrence, the metastatic rate was 9/46 (20%) in cases with regional disease, but 4/132 (3%) in those without. Six patients with persistent loco-regional disease were treated with various combination chemotherapy, but all failed to respond. Distant. Hematogenous metastases were detected in 17 patients (9%) 11 to 61 months (median = 25 months)
after the initial staging. They were documented after local and/or regional relapses in all except four cases. The commonest sites included bones ( 13 patients), lungs ( 13) and liver (7). The median survival following manifestation of metastases was 12 months. Of the six cases treated with chemotherapy, only one patient with lung secondaries showed partial response to Cisplatin lasting for 14 months. Analysis of possible prognostic factors Table 5 shows the 7-year ASR and RFS when analyzed by age, sex, and the extent of primary tumor infiltration. Although females and patients below 40-years old did slightly better than males and older patients, respectively, the differences are not statistically significant. When patients are subclassified according to whether one or more walls within the nasopharynx are involved, there is no significant difference between the two groups both in the survival results and the patterns of relapse. Intercurrent diseases Six relapse-free patients died from causes unrelated to NPC. Two of them succumbed to a second malignancy in the bronchus. Treatment complications Sixty patients (3 1%) developed one or more late irradiation sequelae besides varying degrees of dryness of mouth. Table 6 lists the damages and their incidences. The latent interval for the neurological complications ranged from 14 to 70 months (median = 44 months). Symptomatic hypothalamic-pituitary dysfunction has
Table 6. Late irradiation sequelae Radiation therapy: no. of courses One
No. of patients Latency <4 years Latency >4 years Incidence at sites Brainstem Cranial nerve Temporal lobe Hypo-pit. axis TM joint Ear
4.2 Gy
2.5 Gy
One + boost
Two
Total
10 9
26 3
5 I
5 1
46 14
-
-
1 4 4
I 8 6 22 12 29
4 3 7 53 10
3 1 9 14
I I 2 5 1
Hypo-pit. axis = Hypothalamic-pituitary = Temporo-mandibular joint.
-
axis; TM joint
1188
1. J. Radiation Oncology 0 Biology 0 Physics
(10/35) versus 19% (26/l 35) fails to reach statistical nificance (p = 0.46).
sig-
DISCUSSION Because of the difficulty, if not impossibility, of resecting the base of skull or the retropharyngeal spaces, surgery has little role in the primary treatment of NPC, even for Stage I cases. RT is the modality of choice as both the primary tumor and its lymphatic metastases are radiosensitive. The anatomical proximity to critical structures, however, presents special limitation. To achieve the highest cure rate with minimal morbidity remains one of the greatest challenges for radiation oncologists. The retrospective analysis of Hoppe et ul. (22) demonstrated that although nearly one-third (6/17) of their patients with local recurrence had initial T 1 growths, there had been no such failures in the 7 years following the adoption of larger treatment volumes. This suggests that local involvement is often more extensive than clinically evident, and hence the field coverage is a significant factor in determining the chance of success. This well-recognized characteristic of extensive tumor infiltration and difficulty of accurately defining the extent of microscopic involvement dictate that all potential sites must be included within the target volume, even for apparently Tl growths (18, 20, 22, 46). Advances in radiologic technology undoubtedly allow better delineation of the macroscopic extent. With the introduction of computerized tomography and magnetic resonance imaging, a significant number of Stage I cases would have been upstaged and the incidence of geographic miss might have been reduced. But whether such additions, with their present degree of accuracy, can allow safe reduction in field size has yet to be cautiously tested. The dose-response relationship for Tl growths is still poorly defined. Although Chang et al. (9) and Bedwinek et al. (5) suggested that 60 Gy (at conventional fractionation of 2 Gy daily) may be adequate, Moench and Phillips (30) Mesic et al. (28), and Vikram t’t al. (42) demonstrated that better control can be achieved with higher doses. Like most other centers, we advocate a total dose biologically equivalent to 65 Gy. Primary CR lasting more than 6 months was achieved in 185/ 196 (94%) of our Stage I patients, and only 17 (9%) of them subsequently recurred. With this high local control rate, routine boosting by intracavitary caesium insertion is not warranted. However. careful re-assessment after the completion of RT is obviously important, as patients with residual tumors may benefit from additional doses. With appropriate boosting, all of our eleven partial-responders eventually attained CR, and only one subsequently relapsed. Of the patients with local recurrence, 7/l 5 (47%) responded completely to aggressive re-treatment with external irradiation. Similar to the experience of Wang (44) the response rate was high for patients with recurrence still confined within the nasopharynx, but extremely poor
December 1989, Volume 17, Number 6
for those with cranial extension (6/7 versus l/8). Thus far, only one of the seven complete-responders recurred again, but longer observation is required to assess the longterm control and morbidity rates of the second course of RT. Besides the problem of local failure, these relapsed patients were associated with a greater risk of developing lymphatic and/or hematogenous metastases, the incidence being 44% and 22%, respectively. Additional treatment may be needed-EN1 (if not previously given) should at least be considered. The need for including the whole regional lymphatics in the primary treatment of Stage I patients remains controversial. ENI is strongly advocated by many authors (2, 5, 7, 9, 22, 26-30, 32, 34, 37, 39, 42, 43, 46) because of the high incidence of overt LN involvement even for Tl tumors (ranging from 62% to 86% in the literature), the universal experience of very low regional relapse rates following EN1 in node-negative cases, and the low morbidity of the treatment. However, its exact impact on the final outcome has never been extensively tested by controlled trials, and conflicting results showing a lack of significant benefit have been published (6, 10, 18,25). There are only two studies with specific data for T 1NOM0 patients. Chang et ul. (9) reported that patients with ENI had a lower incidence of developing LN metastases (6% versus 16%). Ho ( 18) also found a similar trend (5% versus 25%). However, he showed that, with prompt salvage by therapeutic irradiation, the 5-year ASR and disease-free survival rates in the group without initial ENI were by no means inferior (Table 2). Hence, it was suggested that EN1 can safely be withheld if the patient could be closely observed. The present series confirms Ho’s previous finding that 70%) of Stage I patients do not require prophylactic treatment to the whole neck, provided that the first station LN at the retropharyngeal spaces are adequately covered within the primary target volume. Radical therapeutic irradiation of the whole neck successfully salvaged 44/49 (90%) of the relapsed cases, and only two patients suffered from uncontrolled regional disease alone. However, the present analysis indicates that there is a difference in the 7-year ASR when patients with nodal relapse are compared to those without (70% versus 87%). Of the 178 local-recurrence-free patients, 46 (26%) developed cervical LN metastases, and they were associated with a higher risk of hematogenous dissemination (9/46 versus 4/ 132, that is, 20% versus 3%). It is impossible to tell from this retrospective analysis whether such distant metastases arose from the primary growths or occult foci in the untreated regional LN, but we share the concern of Mendenhall et al. (27) that a significant proportion of them might have originated from the latter. Although this question about the value of EN1 on the ultimate survival can only be answered by a prospective randomized trial with an adequate number of patients, there is little doubt that EN1 can at least improve the relapse-free rate. Our present 7-year RFS of 62% is far from being ideal; regional relapse, though largely salvage-
Stage 1 nasopharyngeal carcinoma 0 A. W. M. LEE et al.
able, is psychologically distressing. Hence, if treatment machine availability allows, EN1 ought to be considered, especially for patients who cannot be closely observed. No significant prognostic factors can be identified in our present analysis. Similar to the studies by Chatani et al. (10) Hoppe et al. (22), and Yamashita et al. (46), our data support the criticisms (6, 14, 19, 23, 25, 30, 32) that subclassification according to whether one or more walls within the nasopharynx are involved, as suggested by the AJC and the UICC systems, has no value in determining the treatment strategy or predicting prognosis. The relatively high incidence of late treatment complications is one of our major concerns. With increasing evidence that nervous tissues are much less tolerant of large fractional doses than connective tissues, and hence, Ellis’ (13) NSD formula (TD = NSD X f124 X TO.“) has to be modified for more accurate reflection of tolerance (3, 11, 16, 33, 38, 45), our dose per fraction was reduced from 4.2 Gy to 2.5 Gy, and the brain tolerance unit formula (TD = btu X No.45X T”.03)suggested by Pezner and Archambeau (33) is also applied in the estimation of doses. To achieve a homogeneous dose of 65 Gy to all potential sites of local invasion, the dose to brain tissues included within the target volume (NSD = 1859 ret, btu = 1212) will exceed the usually quoted tolerance limit of 1700 ret (36) or 1050 btu (33), even if treatment is given at the
1189
conventional fractionation rate of 2 Gy daily. A definite risk of morbidity seems inevitable for the best chance of survival. The total doses given by our two fractionation schemes were 186 1 ret and 1870 ret in terms of NSD, but 1477 and 1286 in terms of btu. Although the overall incidence of damages by the former scheme of 4.2 Gy per fraction is indeed significantly higher (54% versus 2 1%, p = 0.012), this should not be taken as valid proof of the greater damaging effect by the larger fractional dose, because of the longer observation period for patients treated by the former scheme and the long latent interval before development of sequelae (median = 44 months). As the shortest follow-up for survivors by the latter fractionation schedule was 4 years, comparison of relative risk should best be based on the 4-year cumulated data for patients who had only one basic course of RT using the same technique. With the same survival rates (90%) by both schemes, the cumulated incidence of late complications was higher with the former-29% versus 19%. However, this difference is not statistically significant (p = 0.46). Longer follow-up is necessary to assess the ultimate result. Further reduction of fractional dose to 1.8-2 Gy is probably required for significant improvement, but major break-through may have to await the development of effective and non-toxic radioprotectors.
REFERENCES 1. American Joint Committee for Cancer Staging and End Results Reporting: Manual for staging of cancer 1978. Chicago: American Joint Committee; 1978:38. 2. Andel, J. G. van; Hop, W. C. J. Carcinoma of the nasopharynx: a review of 86 cases. Clin. Radiol. 33:95-99; 1982. 3. Aristizabal, S. A.; Caldwell, W. L.; Avila, J. The relationship of time-dose fractionation factors to complications in the treatment of pituitary tumors by irradiation. Int. J. Radiat. Oncol. Biol. Phys. 2:667-673; 1977. 4. Baker, S. R. Nasopharyngeal carcinoma: clinical course and results of therapy. Head Neck Surg. 3:8- 14; 1980. 5. Bedwinek, J. M.; Perez, C. A.; Keys, D. J. Analysis of failures after definitive irradiation for epidermoid carcinoma of the nasopharynx. Cancer 4512725-2729; 1980. 6. Bertelsen, K.; Andersen, A. P.; Elbrand, 0.; Lund C. Malignant tumors of the nasopharynx. Acta Radiol. Ther. Phys. Biol. 14:177-186; 1975. 7. Bohorquez, J. Factors that modify the radio-response of cancer of the nasopharynx. Am. J. Roentgenol. 126:863876; 1976. 8. Cellai, E.; Chiavacci, A.; Olmi, P.; Carcangiu, M. L. Carcinoma of the nasopharynx: results of radiation therapy. Acta Radio]. Oncol. 21 (Fast. 2):87-95; 1982. 9. Chang, C.; Liu, T.; Chang, Y.; Cao, S. Radiation therapy of nasopharyngeal carcinoma. Acta Radiol. Oncol. 19(Fasc. 6):433-438; 1980. IO. Chatani, M.; Teshima, T.; Inoue, T.; Azuma, I.; Yoshimura, H.; Oshitani, T.; Hashiba, M.; Nishiyama, K.; Tsutsui, K.; Fujimura, T.; Araki, Y.; Hishikawa, Y.; Arita, S.; Yamada, C.; Kondo, K.; Kagemoto, M.; Tanaka, M.; Imajo, Y. Radiation therapy for nasopharyngeal carcinoma: retrospective review of 105 patients based on a survey of Kansai Cancer Therapist Group. Cancer 57:2267-2271; 1986. 11. Cohen, L.; Creditor, M. Iso-effect tables for tolerance of
12.
13. 14.
15.
16.
17. 18.
19.
20.
21.
irradiated normal human tissues. lnt. J. Radiat. Oncol. Biol. Phys. 9:233-241; 1983. Dickson, R. I.; Flores, A. D. Nasopharyngeal carcinoma: An evaluation of 134 patients treated between 197 1-l 980. Laryngoscope 95:276-283; 1985. Ellis, F. Dose, time and fractionation: a clinical hypothesis. Clin. Radiol. 20: l-7; 1969. Fu, K. K. Prognostic factors of carcinoma of the nasopharynx (editorial). Int. J. Radiat. Oncol. Biol. Phys. 6:523-526; 1980. Haghbin, M.; Kramer, S.; Patchefsky, A. S.; Prestipino, A. J.; Diener-West, M. D. Carcinoma of the nasopharynx: a 25-year study. Am. J. Clin. Oncol. (CCT) 8:384-392; 1985. Harris, J. R.; Levene, M. B. Visual complications following irradiation for pituitary adenomas and craniopharyngiomas. Radiology 120:167-171; 1976. Ho, J. H. C. Radiologic diagnosis of nasopharyngeal cancer. J.A.M.A. 220:396-398; 1972. Ho, J. H. C. An epidemiologic and clinical study of nasopharyngeal carcinoma. Int. J. Radiat. Oncol. Biol. Phys. 4: 183-198; 1978. Ho, J. H. C. Stage classification of nasopharyngeal carcinoma: A review. In: De The, G., Eto, Y., eds. Nasopharyngeal carcinoma: etiology and control, IARC Scientific Publications No. 20. Lyon: International Agency for Research on Cancer; 1978:99- 113. Ho, J. H. C. Nasopharynx. In: Halnan, K. E., Boak, J. L., Crowther, D., von Essen, C. E., Orr, J. S., Peckham, M. J., eds. Treatment of cancer. London: Chapman and Hall; 1982:249-267. Ho, J. H. C.; Lau, W. H.; Fong, M.; Chan, C. L.; Au, G. K. H. Treatment of nasopharyngeal carcinoma (NPC). In: Grundmann, E., Krueger, G. R. F., Ablashi, D. V., eds. Nasopharyngeal carcinoma, cancer campaign, Vol. 5.
1190
22.
23.
24.
25.
26. 27.
28.
29.
30.
3 1. 32.
33.
1. J. Radiation Oncology 0 Biology 0 Physics
Stuttgart and New York: Gustav Fischer Verlag; 198 1:279285. Hoppe, R. T.; Goffinet, D. R.; Bagshaw, M. A. Carcinoma of the nasopharynx: eighteen years’ experience with megavoltage radiation therapy. Cancer 37:2605-26 12; 1976. Huang, S.-C. Nasopharyngeal cancer: a review of 1605 patients treated radically with cobalt 60. Int. J. Radiat. Oncol. Biol. Phys. 6:401-407; 1980. Lam, K. S. L.; Ho, J. H. C.; Lee, A. W. M.; Tse, V. K. C.; Chan, P. K.; Wang, C.; Ma, J. T. C.; Yeung, R. T. T. Symptomatic hypothalamic-pituitary dysfunction in nasopharyngeal carcinoma patients following radiation therapy: a retrospective study. Int. J. Radiat. Oncol. Biol. Phys. 13: 1343-1350; 1987. Larsson, L.-G.; Seelig, I. Malignant nasopharyngeal tumors: result of radiation therapy. Acta Radio]. Ther. Phys. Biol. 15:209-218; 1976. Lederman, M. Cancer of the nasopharynx: its natural history and treatment. Illinois: Springfield; 196 1. Mendenhall, W. M.; Million, R. R.; Cassisi, N. J. Elective neck irradiation in squamous-cell carcinoma of the head and neck. Head Neck Surg. 3: 15-20; 1980. Mesic, J. B.; Fletcher, G. H.; Goepfert, H. Megavoltage irradiation of epithelial tumors of the nasopharynx. Int. J. Radiat. Oncol. Biol. Phys. 7:447-453: 198 I. Million, R. R.; Fletcher, G. H.: Jesse, R. H. Evaluation of elective irradiation of the neck for squamous-cell carcinoma of the nasopharynx, tonsillar fossa and base of tongue. Am. J. Roentgenol. 80:973-988; 1963. Moench, H. C.; Phillips, T. L. Carcinoma of the nasopharynx: review of 146 patients with emphasis on radiation dose and time factors. Am. J. Surg. 124:5 15-5 18; 1972. Payne, D. G. Carcinoma of the nasopharynx. J. Otolaryngol. 12:197-202; 1983. Perez, C. A.; Ackerman, L. V.; Mill, W. B.; Ogura, J. H.; Powers, W. E. Cancer of the nasopharynx: factors influencing prognosis. Cancer 24: I - 17; 1969. Pezner, R. D.; Archambeau, J. 0. Brain tolerance unit: A method to estimate the risk of radiation brain injury for various dose schedules. Int. J. Radiat. Oncol. Biol. Phys. 7: 397-402; 1981.
December 1989, Volume 17, Number 6 34. Qin, D.; Hu, Y.; Yan, J.; Xu, G.; Cai, W.; Wu, X.; Cao, D.; Gu, X. Analysis of I379 patients with nasopharyngeal carcinoma treated by radiation. Cancer 6 1: 1117- 1124; 1988. 35. Rahima, M.; Rakowsky, E.; Barzilay, J.; Sidi J. Carcinoma of the nasopharynx: an analysis of 9 1 cases and a comparison ofdiffering treatment approaches. Cancer 58:843-849; 1986. 36. Rottenberg, D. A.; Chemik, N. L.; Deck, M. D. F.; Ellis, F.; Posner, J. B. Cerebral necrosis following radiotherapy of extracranial neoplasms. Ann. Neurol. 1:339-357; 1977. 37. Scanlon, P. W.; Rhodes, R. E.; Woolner, L. B.; Devine, K. D.; McBean, J. B. Cancer of the nasopharynx: 142 patients treated in the 1I year period 1950-1960. Am. J. Roentgenol. 99:3 13-325; 1967. 38. Scheline, G. E.; Wara, W. M.; Smith, V. Therapeutic irradiation and brain injury. Int. J. Radiat. Oncol. Biol. Phys. 6:1215-1228; 1980. 39. Thompson, R. W.; Doggett, R. L. S.; Bagshaw, M. A. Tenyear experience with linear accelerator irradiation of cancer of the nasopharynx. Radiology 97: 149-155; 1970. 40. Union Internationale Contre le Cancer (UICC). TNM classification of malignant tumors, 3rd edition. Geneva: UICC; 1978:27-32. 4 1. Urdaneta, N.; Fischer, J. J.; Vera, R.; Gutierrez, E. Cancer of the nasopharynx: review of 43 cases treated with supervoltage radiation therapy. Cancer 37: 1707- 17 12; 1976. 42. Vikram, B.; Mishra, U. B.; Strong, E. W.; Manolatos, S. Patterns of failure in carcinoma of the nasopharynx: 1. Failure at the primary site. Int. J. Radiat. Oncol. Biol. Phys. 11: 1455-1459; 1985. 43. Wang, C. C. Treatment of malignant tumors of the nasopharynx. Otolaryngol. Clin. N. Amer. 13:477-48 1; 1980. 44. Wang, C. C. Re-irradiation of recurrent nasopharyngeal carcinoma-treatment techniques and results. Int. J. Radiat. Oncol. Biol. Phys. 13:953-956; 1987. 45. Wigg, D. R.; Koschel, K.; Hodgson, G. S. Tolerance of the mature human central nervous system to photon irradiation. Br. J. Radiol. 54:787-798; 1981. 46. Yamashita, S.; Kondo, M.; Inuyama, Y.; Hashimoto, S. Improved survival of patients with nasopharyngeal squamous cell carcinoma. Int. J. Radiat. Oncol. Biol. Phys. 12: 307-3 12: 1986.