Ini J. Radimon Oncology Bml Php.. Vol. 19. pp. I 177-1182 Printed in the U.S.A. All rights reserved.

0360-3016/90 $3.00 + .oO Copyright 0 1990 Pergamon Press plc

??Original Contribution

COMPUTED

TOMOGRAPHY IN NASOPHARYNGEAL PART II: IMPACT ON SURVIVAL

CARCINOMA:

E. CELLAI, M.D.,* P. OLMI, M.D.,* A. CHIAVACCI, M.D.,* G. GIANNARDI, M.D.,? R. FARGNOLI, M.D.,? N. VILLARI, M.D.+ AND C. FALLAI, M.D.* Universityand Hospital, Florence Two hundredand seventeen consecutive patients affected by nasopharyngeal carcinoma (NBC) were treated with radiotherapy alone, with curative intent, from 1970 to 1985 at the Radiotherapy Unit of the University and Hospital of Florence. A group (A) of 111 patients staged with conventional clinical and radiological method was compared to a second group (B) of 106 patients who underwent CT staging before treatment. Group B showed better 5-year NED survival and local control; only the differences in local control were significant (p < 0.01). As to primary control statistically significant differences were observed in T2 and T4 cases. We feel that CT could have contributed to the improvement, probably through a more reliable display of the primary extent and a more adequately planned radiotherapeutic treatment. With CT staging we could not increase our skills in prognostically separating stages according to UICC criteria (1978); in Group B only T2 patients presented significant differences in primary control when compared to T3 and T4 patients. However, a multivariate analysis of prognostic factors showed that nodal involvement, primarily, and histology, secondarily, were the most important factors; T stage showed a minor influence on prognosis. Nasopharynx,

Neoplasms,

CT-staging, Radiotherapy.

and prognosis. In our paper we will test the hypothesis that a better correlation exists between T stages determined by CT and prognosis. Finally, some prognostic factors will be re-evaluated through multivariate analysis to establish whether the advent of CT has modified their individual significance.

INTRODUCTION

The capability of CT in diagnosing the extent of primary tumor and nodal disease in nasopharyngeal carcinoma is commonly agreed upon. Information provided by CT can be incorporated in treatment planning; furthermore, CT can replace the traditional simulation procedure (2,6, 19) and help in the choice of the most adequate treatment technique (14). CT has advantages over the politomography such as better displaying bones and soft tissue changes contemporaneously. Thus, it has been reported that CT has provided changes in T stage in a high percentage of cases (13, 16, 19). The frequent T stage conversion with CT-staging (1, 26-28) has been confirmed by Part I of the present series; the advantages of CT did not appear as important in staging nodal disease, which explains why we are focusing particularly on T-stage variables in the present study. In the first part of this series an older series staged with conventional clinical and radiological method was compared to a more recent group of patients who underwent CT before treatment, with regard to survival and local control, for the whole series and according to T extension. A lack of correlation has been reported by several authors (3, 7, 13, 16) between T stage

METHODS

MATERIALS

Two hundred and seventeen consecutive patients affected by nasopharyngeal carcinoma (NPC) were treated with radiotherapy alone, with curative intent, from 1970 to 1985; the minimum follow-up was 3 years. One hundred and eleven cases, irradiated before December 1978, were staged with clinical examination and polytomography (group A); the remaining 106 patients were accrued from December 1978 and staged with the help of CT (Group B). When the clinical situation required more complicated techniques than two lateral opposed fields (e.g. 3-field technique for nasal spread), treatment planning was carried out with the CT as well. The male to female ratio was 2.96 in group B and, respectively, 2.93 in the group A; the mean age was 5 1.2 years and 52.3 years, respectively. Distribution of patients according to

timento di Fisiopatologia

* Departmentof Radiotherapy. + Imaging Department. Reprint requests to: E. Cellai, Unita’ di Radioterapia,

AND

Clinica, Universita’ Morgagni, Care&, 50 134 Firenze, Italy. Accepted for publication 24 May 1990.

Dipar1177

di Firenze,

v.le

I.

1178

J. Radiation Oncology 0 Biology 0 Physics

Table 1. Distribution Micheau’s

of patients according classification

ALL

CASESW

(217)

GROUP

AH

(1111

GROUP

BO-O

(106)

No. of

No. of Undifferentiated carcinoma of nasopharyngeal type Squamous cell carcinoma Low grade Middle grade High grade Unclassified carcinoma Total

100

to

Group B

Group A Pts

November 1990, Volume 19, Number 5

(S)

Pts

(%)

55

(50)

50

(47)

49 21 14 8 7 111

(44)

50 25 24

(48) I

I

I

I

I

I

1

119456789

(6)

:, 106

(5)

Micheau’s classification (17) is reported in Table 1; the undifferentiated carcinoma of nasopharyngeal type (U.C.N.T.) showed the highest incidence in both groups: 50% in group A and 47% in group B, respectively. The distribution by T and N stage according to the UICC Staging System (24) (1978) has been previously reported (Tables 1 and 2, Part I). The incidence of nodal involvement appeared similar; T 1 cases are definitely less frequent in group B, whereas T3 showed a higher incidence in group A. Treatment techniques are depicted in Figure 1. The therapeutic approach always consisted of irradiation of large volumes including the primary tumor, whatever its extension, and of cervical areas, in most cases even if a nodal involvement was not clinically evident. Fractionation schedules and total tumor dose (T.T.D.) are shown in Figure 1. If clinically uninvolved, supraclavicular areas were given 5000 cGy and cervical nodes 6000 cGy. A 6oCo source was used in all cases. After administering 4400 cGy with lateral opposing portals, the posterior nodal areas were treated with 10 MeV electron beams to spare un-

1

10 YEARS

Fig. 2. IO-year NED survival for the overall series, group A and group B.

derlying spinal cord. Furthermore, electron beams were used to boost residual adenopathies. A 31 MeV photonbeam from a betatron was used to supplement the dose at the primary site, delivering 1OOO- 1500 cGy up to a total dose of 7000 cGy. Survival and local control were estimated by the life-table method of Berkson-Gage (4) and the obtained percentage rates were compared with the Greenwood formula (11). A multivariate analysis of the prognostic factors of survival was carried out according to the Cox regression model (5). Several variables were examined in a preliminary analysis in accordance with a step-down procedure. Only variables which had displayed a preliminary significance level were submitted to further analysis in conformity with the Cox model. Among initially studied factors were: age, sex, histology (lymphoepithelioma, low, middle and high-grade carcinoma), T and N stage and staging with and without CT. RESULTS

Actuarial NED survival results are reported in Figure 2 for the overall series and group A and B. A more favorable outcome appeared for group A, 47% at 10 years ALL CASES

W

(217 1

GROUP

AO-Ol111)

GROUP

9 O-011061

. . . 1;: 3~~-~~~;~~wek XX......

7r

. .. . . . . . . . . .

‘ii5””.............“...

Fig. 1. Distribution of patients according to treatment technique and treatment volume, fractionation schedule, and mean tumor dose.(*) In 25 out of 4 1 cases the anterior field was used to boost the primary, even without anterior spread.(**) The supraclavicular area was not irradiated in 12 cases of group A.

YEARS

Fig. 3. IO-year local control group B.

for the overall series, group A and

CT and NCP: survival 0 E. CELLAIet al. Table 2. Cumulative

5-year NED survival at the primary Group

Total cases

Tl T2 T3 T4 Total Significant

1179

site according

to T stages Group

A

B

No. of cases

% survival

+SD

No. of cases

% survival

+SD

No. of cases

% survival

+SD

31 69 46 65 217

69 IO 63 43 60

9 6 8 I 4

30 35 14 32 111

68 57 68 30 53

10 9 13 9 5

7 34 32 33 106

71 84 61 56 67

24 7 9 9 5

differences

between

categories: Group A c p cc 0.05 + Group B; T2 -

versus 36% for the older cases (group B); however, the difference was not significant (p = 0.08). Local diseasefree survival curves (Fig. 3) showed a difference (59% vs 40% at 10 years) which reached a significant level (p < 0.01). When disease-free survival (at the primary site only) according to T extension was analyzed, results, on the whole, were better in group B (67% vs 53%; p < 0.05); in particular T2 and T4 cases showed a significantly more favorable outcome in group B versus A (p -K0.05). As far as interstage comparison is concerned, Tl to T3 lesions had better 5-year results than T4 (whole series, group A). In group B only T2 patients revealed significant differences when compared to more advanced neoplasms (T3-T4) (Table 2). NED survival and local control of group A and B were compared according to T and N stages, clustered into composite categories. Due to the prognostic relevance of N factor, it was considered interesting to determine whether a better correlation between T stage and prognosis could be found through comparison of different stages with similar nodal disease. Although NED survival showed a general trend towards better rates, in no subset of group A, compared to equivalent subsets of group B,

Table 3. Cumulative

No. of cases

% survival

was a statistical significance reached. Limited primary T l2 NO- 1 did better than advanced lesions T4 NO- 1 (whole series, group A) (Table 3). In relation to the local-regional control (Table 4) a definite improvement was observed in all categories with limited adenopathy or without lymph node involvement, but only in the T4-NO- 1 subset was the difference between A and B cases large enough to reach significance (p < 0.0 1). When the interstage comparison was carried out T 1-2NO-1 lesions had a significantly higher percentage of 5-year local control than T4NO- 1 (whole series, group A); better local control was also observed in Tl-2 N2-3 and T3 N2-3 cases versus T4 N2-3 (whole series, group B). In the multivariate analysis the prognostic determinants of NED survival were nodal disease at presentation and histology, both in the whole series and group A and B. Additional factors reaching a significant level were T extent (whole series and group B) (Table 5) and fractionation (borderline significance in the whole series and group A). The multivariate analysis of local control (Table 6) confirmed nodal involvement as the most important factor (whole series, group A and B). In the overall series,

5-year NED survival by composite

Total cases

Group

SSD

No. of cases

p < 0.05 -+ T2: T4 + P < 0.05 -+ T4.

subsets according

to T and N stages

A

% survival

Group

+SD

No. of cases

B

% survival

+SD

Tl-2 NO- 1

52

68

-I

34

67

8

18

72

11

T3 NO- 1

16

58

13

5

60

22

11

51

16

T4 NO-l

32

44

9

11

27

13

21

55

11

Tl-2 N2-3

54

32

7

31

25

8

23

43

I1

T3 N2-3

30

36

10

9

22

14

21

42

12

T4 N2-3

33

26

8

21

21

9

12

32

14

Total

217

44

4

111

38

5

106

50

5

1180

1. J. Radiation Oncology 0 Biology0 Physics Table 4. Cumulative

5-year local-regional control by composite subsets, according to T and N stages

Total cases No. of cases

% survival

Tl-2 NO- 1

52

70

T3 NO-1

16

T4 NO- 1

November 1990, Volume 19, Number 5

Group A

Group B

No. of cases

% survival

6

34

64

8

18

83

9

54

13

5

40

22

11

60

15

32

46

9

11

18

12

21

65

11

Tl-2 N2-3

54

46

7

31

35

9

23

62

11

T3 N2-3

30

52

10

9

50

18

21

53

12

T4 N2-3

33

2.5

8

21

28

10

12

21

13

Total

217

50

4

111

42

5

106

59

5

+SD

Significant or borderline differences between categoreis; T4 NO- 1 + p < .Ol

staging with CT achieved a significant level as did T extension. The stage of the primary tumor was also statistically significant in the CT-staged group. Fractionation did not reach a significant level in any group.

DISCUSSION With regard to the comparison between non CT-staged and CT-staged patients, the second group showed better 5-year NED survival and local control. The difference in local control was highly significant (a < 0.01). When a direct comparison was performed according to T extension, CT-staged patients with a relatively limited T2 lesion exhibited better control of the primary than T2 patients in group A. A statistically significant difference was observed also between T4 cases. Table 5. Results of multivariate Variables Total cases N Histology T Fraction. CT Group A N Histology Fraction. T Group B N Histology T

+SD

+

No. of cases

% survival

+SD

T4 NO-l; Tl-2 N2-3 + p = .05 + Tl-2 N2-3.

There is only partial agreement between our data and the results of Yamashita et al. (26) who described a more favorable trend in the more advanced neoplasms only (T3-4). We feel that the significance levels obtained by comparing B versus A cases, particularly in terms of local control, suggest a substantial CT contribution to the improvement through more reliable staging and more accurate treatment planning. As previously reported (see Part I) CT-staging determined a T stage conversion in 23 out of 97 cases; in 14 cases the primary volume was more adequately irradiated and in 9 cases a different technique was adopted. A clear-cut prognostic influence on control of the primary was shown in less extended lesions (Tl and/or T2) versus the far advanced cases (T4) in the whole series and in both group A and B. It was also possible to single out

analysis (Cox regression model) of the prognostic factors in relation to 5-year NED survival

Regression coefficient

Relative risk

95% confidence interval

Chi-square

p-value

0.4016 0.2932 0.2399 0.4407 0.3209

1.4942 1.3407 1.2712 1.5538 0.7255

1.2726-1.7362 1.1691-1.5375 1.0743-l .5043 1.0007-2.4 124 0.4935-I .0667

25.3858 17.606 1 7.8093 3.8547 2.6620

0.0000 0.0000

0.0052 0.0496 0.1028

0.4019 0.2935 0.4579 0.1901

1.4948 1.3411 1.5809 1.2093

1.2019-1.8589 1.1186-1.6078 1.0161-2.4595 0.9851-1.4846

15.0585 10.0565 4.1247 3.3016

0.0003 0.0015 0.0423 0.0692

0.4294 0.3137 0.3378

1.5363 1.3685 1.4591

1.2041-1.9601 1.0949- 1.7106 1.0362-2.0547

11.9344 7.599 1 4.6810

0.0006 0.0058 0.0305

CT and NCP: survival 0 E.

Table 6. Results of multivariate

Variables Total cases N Histology T CT Group A N Fraction. Histology T Group B N T Histology

CELLAI rf ul.

1181

analysis of the prognostic factors in relation to S-year local-regional control

Regression coefficient

Relative risk

95% confidence interval

0.2902 0.2035 0.2469 0.4453

1.3367 1.2256 1.2800 0.6406

0.5694 0.3612 0.1253 0.1048 0.4275 0.3901 0.1628

Chi-square

p-value

1.1310-1.5798 1.0532-l .4262 1.0644-1.5392 0.4 183-0.98 13

1 I .5922 6.9206 6.8863 4.1897

0.0007 0.0085 0.0087 0.0407

1.7672 1.4351 1.1335 1.1105

1.2464-2.5054 0.7654-2.6906 0.8652-1.4850 0.8257- 1.4936

10.2196 1.2686 0.8274 0.4804

0.0014 0.2600 0.3630 0.4882

1.5334 1.4771 1.1768

1.1665-2.0157 1.0085-2.1635 0.9082-l .5248

9.3878 4.0141 1.5169

0.0002 0.045 1 0.2181

differences that were statistically significant by comparing contiguous T stages (T2 vs T3: group B; T3 vs T4: whole series and A group). Conventionally staged patients with limited lesions (T l2) with NO or Nl (UICC 1978) (24) showed, as expected, a better trend in survival and local control versus T4NO1 cases. Among patients with extensive nodal disease (N23) T l-2 and T3 lesions had higher local control than T4

cases, with a stronger correlation between prognosis and T extension in CT-staged patients. In conclusion, the prognostic significance of T extension (according to UICC categories) was not always clear-cut, in our experience. In general, as expected, limited lesions showed significantly different results in terms of NED survival and local control, but it was difficult to separate prognostically contiguous stages (ex. T2 vs T3). Such a result, on the whole, was reached more often in the CTstaged group and could also be obtained occasionally in the non CT-staged group when the primary control was studied according to T extent. Nevertheless, the data seem to only partially support the criteria adopted for T stage definition. Staging systems with more detailed subclassifications of T stage (12) might reveal a closer correlation with prognosis. The results of the multivariate analysis stress the importance of nodal disease as a leading prognostic factor in NED survival and local control for the whole series and both non-CT-staged and CT-staged patients. CT can detect nodal disease beyond the range of physical examination (deep cervical, retropharyngeal nodes) and normal-sized cavitated nodes (15, 22). In spite of this refinement of staging, we believe that impact of CT on survival, due to detection of clinically unsuspected nodal disease, is questionable. In our series 3 out of 30 clinically NO patients were classified N 1 after CT staging. Due to their anatomical location, they were included in the high dose volume with the primary; dose and volume of treatment was not modified. It must be pointed out that, according to our policy, 6000 cGy are delivered to even clinically negative cervical nodes.

Isolated nodal relapse was uncommon. In group A two NO patients had marginal relapse (one patient was salvaged with radical neck dissection). No patient of group B with apparently negative nodes at presentation relapsed in the neck alone. Histologic type is the only other factor that was found to be of relevant strength for the overall series in disease-free survival and local control.

CONCLUSIONS CT still plays a major role in the study of NPC (Table 7). CT scans can occasionally detect primary lesions not accessible to nasopharyngoscopy (28) define extent of disease, and facilitate correct tumor staging (18, 25). CTstaged patients have a significantly higher local control than patients staged with conventional tomography. lmpact on survival is rather limited because T stage shows Table 7. Nasopharyngeal

carcinoma: role of CT

Detection Staging T extent Nodal enlargement Nodal necrosis Extranodal disease Treatment planning Positioning of portals Dose calculations Beam configurations Compensation of tissue inhomogeneity Early Response Evaluation --t Boost volume selection Follow-up First scan at 2 months Further CT scans every 6 months for 2-2.5 years Yearly CT scans + Detection of possible local relapse -+ Restaging

1182

1. J. Radiation Oncology 0 Biology 0 Physics

only a minor prognostic influence on this parameter. The relative importance of each independent prognostic factor is not modified by CT-staging. The relative role of CT and magnetic resonance imaging (MRI) is under investigation (8- 10). CT plays a vital role in treatment planning as well; data can be transferred to therapy planning computers and used for dosimetry calculations and positioning of portals. So far, information from MRI cannot be used to perform inhomogeneity dose corrections in the same way as CT numbers (23). CT can be used in the followup of irradiated patients with NPC. It is our policy to perform a baseline CT scan 2 to 3 months after the end

November 1990, Volume 19, Number 5

of the treatment. This interval allows regression of postradiotherapy oedema (20-2 1). Further CT scans are performed every 6 months for 2-2.5 years and yearly afterwards because 80% of the patients relapsing at the primary site do so within 2 years (but late relapse are occasionally encountered). Timely diagnosis of residual or recurrent disease can be rewarding for isolated local relapse can be occasionally salvaged. Intracranial recurrence and postirradiation encephalopathy can be differentiated (28). Some uncertain findings on CT (e.g. distinction between postirradiation flogosis and tumor recurrence) can be clarified with MRI (8, 10).

REFERENCES 1. Abayomi, 0. K. Computerized tomographic evaluation of tumors of the nasopharynx and maxillary sinus: potential impact on results of treatment. J. Natl. Med. Assoc. 78: 31 l-316; 1986. 2. Adam, E. J.; Berry, R. J.; Clitherow, S.; Bedford, A. Evaluation of the role of computed tomography in radiotherapy treatment planning. Clin. Radio]. 35: 147-150; 1984. 3. Bedwineck, J. M.; Perez, C. A.; Keys, D. J. Analysis of failures after definitive irradiation for epidermoid carcinoma of the nasopharynx. Cancer 4512725-2729; 1980. 4. Bergson-Gage, R. P. Calculations of survival rates for cancer. Proc. Staff Meet. Mayo Clin. 25:270-286; 1950. 5. Cox, D. R. Regression models and life tables. J. R. Stat. Sot. B 34:187-220; 1972. 6. Dobbs, H. J.; Parker, R. P.; Hodson, N. J.; Hobday, P.; Husband, J. E. The use of CT in radiotherapy treatment planning. Radiother. Oncol. 1: 133-141; 1983. 7. Fu, K. K. Prognostic factors of carcinoma of the nasopharynx. Int. J. Radiat. Oncol. Biol. Phys. 6:523-526; 1980. 8. Giannardi, G.; Cellai, E.; Fargnoli, R.; Olmi, P.; Simoncini, R.; Villari, N. La risonanza magnetica nella diagnostica delle neoplasie rinofaringee. Diagn. Immagini 8:25-36; 1988. 9. Giannardi, G.; Dal Pozzo, G. C.; Fargnoli, R.; Villari, N. La risonanza magnetica nello studio delle adenolinfopatie metastatiche de1 cello. Diagn. Immagini 6: 153- 164; 1986. 10. Giannardi, G.; Fargnoli, R.; Villari, N. Diagnostica per immagini. In: Alajmo, E., ed. I1 carcinoma rinofaringeo. Pisa: Pacini Editore; 1989:75-105. M. The errors of sampling of the survivorship 1 I. Greenwood, tables (Appendix 1). In: Report on public health and statistical subjects. London: His Majesty Stationery Office; 33: 1926. 12. Ho, J. H. C. Stage classification of nasopharyngeal carcinoma: a review. In: De the, G., Ito, Y., eds. Nasopharyngeal carcinoma: etiology and control. Lyon: International Agency for Research on Cancer; 1978:99-l 13. 13. 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. 14. Huang, Y. R. Carcinoma of nasopharynx: CT scans improve radiotherapy technique. Comput. Radiol. 8:263-269; 1984. 15. Mancuso, A.; Harnsberger, H. R.; Muraki, A. S.; Stevens, M. Computed tomography of cervical and retropharyngeal lymph nodes: Normal anatomy, variants of normal, and applications in staging head and neck cancer: I. normal anatomy. II. pathology. Radiology 148:709-723; 1983. 16. Mesic, J. B.; Fletcher, G. H.; Goepfert, H. Megavoltage ir-

17.

18.

19.

20.

2 I.

22. 23.

24.

25.

26.

27.

28.

radiation of epithelial tumors of the nasopharynx. Int. J. Radiat. Oncol. Biol. Phys. 7:447-453; 1981. Micheau, C.; Rilke, F.; Pilotti, S. Proposal for a new histopathological classification of the carcinoma of the nasopharynx. Tumori 64:5 13-5 18; 1978. Olmi, P.; Casamassima, F.; Cellai, E.; Chiavacci, A.; Dini, C.; Fargnoli, R. Ruolo della tomografia assiale computerizzata nello studio delle neoplasie rinofaringee. In: Radiologia, Vol. III. Bologna: Monduzzi, 1984:399-403. Silvestre. M. E.; Medina, E.; Andrea, M.; Monteiro-Grilo, I. Direct multiplanar computed tomography in planning radiotherapy for malignant tumors ofthe nasopharynx. Clin. Radio]. 37: I5- 18; 1986. Schaefer, S. D.; Maravilla, K. R.; Suss, R. A.; Burns, D. K.; Merkel, A.: Close, G. L. Magnetic resonance imaging vs computed tomography. Comparison in imaging oral cavity 111:730and pharyngeal carcinomas. Arch. Otolaryngol. 734; 1985. Schaefer, S. D.; Merkel, M.; Bums, D. K.; Close, L. G. Computed tomography of upper aerodigestive tract squamous cell carcinoma. Arch. Otolaryngol. 110:236-240; 1984. Som, P. Lymph nodes of the neck. Radiology 165:593-600; 1987. Sutchffe, S. B.; Henkelman, R. M. The potential application of magnetic resonance imaging in radiation oncology. In: Withers, H. R., Peters, L. J., eds. Innovations in radiation oncology. New York: Springer-Verlag; 1988:201-208. Union lnternationale Contre le Cancer (UICC). TNM: classification des tumeurs malignes. Geneve: UICC; 1978:2732. Yamashita, S.; Kondo, M.; Hashimoto, S. Conversion of T-stages of nasopharyngeai carcinoma by computed tomography. Int. J. Radiat. Oncol. Biol. Phys. 11:1017-1021; 1985. Yamashita, S.; Kondo, M.; Inuyama, Y.; Hashimoto, S. Improved survival of patients with nasopharyngeal squamous cell carcinoma. Int. J. Radiat. Oncol. Biol. Phys. 12: 307-312; 1986. Yu, Z.; Xu, G.; Huang, Y.; Su, X.; Gu, X. Value of computed tomography in staging the primary lesion (T-staging) of nasopharyngeal carcinoma (NPC): an analysis of 54 patients with special reference to parapharyngeal space. Int. J. Radiat. Oncol. Biol. Phys. 11:2143-2147; 1985. Zheng, G. V.; Zeng, Q. X.; Wu, P. H.; Yuan, C. Computed tomography in the management of nasopharyngeal carcinoma. Clin. Radio]. 40:25-29; 1989.

Computed tomography in nasopharyngeal carcinoma: Part II: Impact on survival.

Two hundred and seventeen consecutive patients affected by nasopharyngeal carcinoma (NPC) were treated with radiotherapy alone, with curative intent, ...
697KB Sizes 0 Downloads 0 Views