- Email: [email protected]
Effect of tumor bulk on local control and survival of patients with T1 glottic cancer
Radiotherapy and Oncology 47 (1997) 161–166
Effect of tumor bulk on local control and survival of patients with T1 glottic cancer Sarada P. Reddy*, Najeeb Mohideen, Silvio Marra, James E. Marks Loyola-Hines Department of Radiotherapy, Loyola University Chicago, 2160 South First Avenue, Maywood, IL 60153, USA Received 11 April 1997; revised version received 17 September 1997; accepted 2 October 1997
Abstract Purpose: To evaluate the effect of tumor bulk in relation to various tumor-related prognostic factors and treatment-related variables on local control and survival of patients with T1 N0 M0 squamous cell carcinoma of the glottis. Materials and methods: In 114 patients with T1 squamous cell carcinoma of the glottic larynx who were irradiated with curative intent, we determined the effect of tumor bulk in relation to mucosal extent (stage and anterior commissure involvement), histologic differentiation and various radiation factors, especially overall treatment time on local control and survival. Tumors were classified retrospectively as small surface lesions or bulky tumors. Seventy-seven patients had small lesions and 37 had bulky tumors. The anterior commissure was involved with cancer in 43 patients. The overall duration of irradiation ranged from 39 to 64 days. The median follow-up time was 6 years (range 5–24 years). Results: The 5-year actuarial local control rate for all patients was 82% after radiotherapy and 92% after salvage laryngectomy. On univariate analysis, bulky tumors and tumors involving the anterior commissure showed an adverse effect on local control, whereas the overall duration of irradiation had a borderline significance. The actuarial local control rate was 91% for small tumors and 58% for bulky tumors (P = 0.0002), 88% when the anterior commissure was not involved and 67% when the anterior commissure was involved (P = 0.01) and 89% when radiation was given in less than 50 days and 73% when irradiation exceeded 50 days (P = 0.06). On multivariate analysis, tumor bulk was the only significant factor that affected local control (P = 0.02). The 5-year actuarial survival for all patients was 73% and the disease-free survival was 92%. Conclusion: This study shows that tumor bulk has a highly significant effect on the radiation control of T1 glottic cancer. Patients who had bulky tumors had lower local control and disease-free survival rates than those patients who had small tumors. 1998 Elsevier Science Ireland Ltd. All rights reserved Keywords: T1 glottic cancer; Tumor bulk; Local control; Ultimate control; Voice preservation; Survival; Complications
1. Introduction Irradiation of early glottic cancer yields 5-year local control rates of 80–95% and ultimate control rates of 90–100% after salvage surgery of recurrences [2–4,9,14–20]. In the quest to improve these results, radiation oncologists continue to search for possible prognostic factors that might significantly affect the local control of early glottic cancer. Since the early 1960s, no one has been able to establish a clear relationship between tumor control and radiation dose for T1 glottic cancer [2,5,6,8,12,13,17–19]. During the 1980s, improved local control rates were reported when * Corresponding author.
radiation was delivered in a shorter overall time, with higher dose fractionation schedules [4,11,15,17,18]. During the same era, a decrease in the local control rate was reported as the size of the lesion increased for patients with T1a glottic cancer [15] and for patients with bulky T1 and T2 lesions [3]. The effect of tumor bulk on local control of T1 glottic cancer, regardless of the number of true vocal cords involved, has not been reported in the literature. The purpose of this study is to determine the effect of tumor bulk on local control, survival and voice preservation in patients with T1 glottic cancer and to evaluate its influence on various tumor-related prognostic factors and treatment-related variables. The prognostic factors studied include the tumor stage, anterior commissure involvement,
0167-8140/98/$19.00 1998 Elsevier Science Ireland Ltd. All rights reserved PII S0167-8140 (97 )0 0196-5
162
S.P. Reddy et al. / Radiotherapy and Oncology 47 (1997) 161–166
histologic differentiation, radiation beam energy, field size, total dose, fraction size and total duration of irradiation.
2. Materials and methods 2.1. Patient population From 1976 to 1990, 126 patients with histologically confirmed T1 squamous cell carcinomas of the true vocal cords were referred to our department for curative irradiation. Radiotherapy charts and medical records were reviewed. Based on the findings recorded at indirect and direct laryngoscopy and on radiographic findings, all tumors were restaged according to the 1992 UICC staging system [1]. Patients irradiated from 1976 to 1992 were chosen for this analysis because we found meticulously drawn tumor diagrams, carefully documented tumor descriptions in the medical records and laryngeal tomograms and CT scans from 1976 onwards. Twelve of the 126 patients were excluded from this analysis; four patients died within 1 year and eight patients were lost to follow-up in less than 2 years after irradiation. All 12 patients who were excluded from this analysis had no local recurrence when they were last seen. The remaining 114 patients who form the basis for this study were followed for a minimum of 5 years or until death and for a maximum of 24 years (median 6 years). Tumor characteristics are listed in Table 1. There were 111 male and three female patients, a finding
Table 1 Tumor characteristics and treatment results No. of patients
Local control % (95% CI)a
Size Small 77 91 Bulky 37 58 Stage T1a 94 80 T1b 20 90 AC involvement No 71 88 Fraction size (Gy) 1.8 46 73 2 68 85 Total dose (Gy) 60–65 28 87 66 36 89 67–70 50 71 Total days irradiation ≤50 54 89 >50 60 73
Log-rank Survival % (95% CI)b
Log-rank
(84–98) (40–76)
0.0002
97 (95–99) 84 (62–92)
0.0017
(71–89) (80–100)
0.09
89 (82–96) 100
0.21
(80–96)
0.01
95 (90–100)
0.06
(60–86) (76–94)
0.09
86 (75–97) 94 (88–100)
0.09
(74–100) (78–100) (58–84)
0.09
91 (79–103) 95 (86–104) 87 (77–97)
0.3
(81–98) (61–85)
0.06
96 (92–100) 86 (76–96)
0.1
AC, anterior commissure; CI, confidence interval. Five-year local control. Five-year disease-free survival.
a
b
Fig. 1. Classification of T1 glottic cancer according to tumor bulk. (A,B) Small tumors. (C,D) Bulky tumors.
that was not surprising given the large number of veterans in the study. Eighty-one patients were white, 32 were black and one was oriental. The patients’ ages ranged from 44 to 79 years, with a median of 60 years. 2.2. Tumor classification Since the present UICC staging system does not consider tumor bulk in the substaging of T1 glottic cancer, we retrospectively classified all T1 glottic tumors that we had irradiated either as small or bulky lesions, regardless of whether one or both true vocal cords were involved. Our classification was based on the description of the tumors, on the diagrams drawn in the medical records at the time of diagnostic endoscopy and on the radiographic tests, which were laryngeal tomograms before the availability of computerized tomography, and CT scans of the larynx later on. Surface lesions not involving an entire true vocal cord and those involving less than the anterior one-third of both true vocal cords (horseshoe-shaped lesions) were classified as small tumors. Large exophytic and/or infiltrative lesions involving an entire true vocal cord and horseshoe-shaped lesions involving more than the anterior one-third of both true vocal cords were classified as bulky tumors (Fig. 1). So that bias could be avoided, each lesion was classified separately as small or bulky by a panel of three physicians, consisting of two radiation oncologists and an otolaryngologist who reached a consensus on tumor bulk. Seventyseven tumors were classified as small and 37 were classified as bulky tumors. Anterior commissure was involved in 14 patients with small tumors and in 29 patients with bulky tumors.
163
S.P. Reddy et al. / Radiotherapy and Oncology 47 (1997) 161–166
2.3. Radiotherapy The majority of patients were irradiated through opposed lateral portals and a few were irradiated through right and left anterior oblique portals. We used cobalt-60 g-rays to irradiate 90 patients, whereas 24 were irradiated with 4 MV X-rays. Total dose ranging from 60 to 70 Gy (median 66 Gy) was delivered as a continuous course prescribed at the intersection of the two beams at midplane. Wedge filters were used when appropriate for optimizing the dose distribution to achieve a homogeneity of ±5% of the prescribed dose. Both fields were irradiated daily, 5 days a week. Doses per fraction ranged from 1.8 to 2.0 Gy. The total duration of irradiation lasted from 42 to 60 days. Treatment parameters are listed in Table 1. The end-points analyzed in this study were local control after primary radiotherapy, ultimate control after salvage surgery for radiotherapy failures, voice quality and voice preservation, overall and disease-free survival and complications. Survival was measured from the date of pathologic diagnosis to the date of last follow-up or death. Local control was measured from the last day of irradiation until the date of biopsy-proven recurrence. Biopsies were performed on all patients with recurrence. 2.4. Statistical analysis Local control, overall survival and disease-free survival probabilities for each possible prognostic factor were calculated by the Kaplan–Meier method [10]. The significance of differences between the paired local control curves was compared by both the log-rank test and the Wilcoxon method. In addition, the influence of tumor bulk on local control when the cancer extended onto the anterior commissure, as well as when the treatment time exceeded 50 days, was assessed by the x2-test. Multivariate analysis of factors that appeared to affect local control and survival was performed with Cox’s proportional hazards model. Statistical analysis was done with the SAS program (SAS Institute, Cary, NC).
to have a significant effect on radiation control of early glottic cancer, whereas the effect of total duration of irradiation appeared to be of borderline significance (Table 1). The local control rate for small tumors was 91%, compared to 58% for bulky tumors (P = 0.0002). Eighty-eight percent of patients without anterior commissure involvement and 67% with anterior commissure involvement had local control (P = 0.01). The local control rate was 89% for patients whose irradiation was completed in 50 days or less and 73% for those whose irradiation lasted for more than 50 days (P = 0.06). Tumor stage, histologic differentiation, radiation beam energy, field size, total dose and fraction size of irradiation showed no significant effect on local control of T1 glottic cancer. When the analysis was controlled for tumor bulk, anterior commissure involvement and total duration of irradiation beyond 50 days did not show an adverse effect on local control (Table 2). The local control rate for small tumors was 86% when the anterior commissure was involved and 92% when the anterior commissure was not involved (NS) and for bulky tumors the control rates were 62 and 63%, respectively (NS). However, anterior commissure involvement was noted more frequently in patients with bulky tumors than in patients with small tumors (78 versus 18%), a finding which relates the lower local control seen in patients with anterior commissure involvement with tumor bulk rather than its location. Likewise, the local control rate was 95% for patients with small tumors irradiated within 50 days and 86% for those irradiated over more than 50 days (NS). For patients with bulky tumors, the rates were 67 and 60%, respectively (NS). Therefore, it appears that tumor bulk is the only significant factor that influences the local control of T1 glottic cancer, irrespective of anterior commissure involvement and total duration of irradiation. Multivariate analysis of tumor bulk, anterior commissure involvement, total duration of irradiation and other tumorTable 2 Influence of tumor bulk, anterior commissure (AC) involvement and total days of irradiation on local control Local control (no. controlled /total no.) (%)
P-value (x2)
58/63 (92) 12/14 (86)
NS
40/42 (95) 30/35 (86)
NS
5/8 (63) 18/29 (62)
NS
8/12 (67) 15/25 (60)
NS
3. Results 3.1. Local control The local control rate at 5 years for the entire group of patients after primary radiotherapy was 82% (93/114). Twenty-one patients developed local recurrences. Sixteen (76%) of these 21 recurrences occurred within 2 years and 20 (95%) recurred by 4 years. The twenty-first patient had a recurrence shortly after the fourth year and no recurrences developed thereafter. Univariate analysis of all prognostic factors studied showed tumor bulk and anterior commissure involvement
Small tumors AC involvement No Yes Treatment days ≤50 >50 Bulky tumors AC involvement No Yes Treatment days ≤50 >50
164
S.P. Reddy et al. / Radiotherapy and Oncology 47 (1997) 161–166
and treatment-related variables studied showed only tumor bulk to have a significant effect on local control of T1 glottic cancer (P = 0.017). 3.2. Ultimate local control after surgical salvage Salvage surgery for 19 of the 21 patients with recurrences consisted of 13 total laryngectomies and six hemilaryngectomies. After successful surgical salvage of 12 patients (seven total laryngectomies and five hemilaryngectomies), the ultimate local control was 92% for the entire group of 114 patients (Table 3). When analyzed according to tumor size, five (71%) of the seven recurrences in patients with small tumors and seven (50%) of the 14 recurrences in patients with bulky tumors were salvaged with surgery, bringing the ultimate local control to 97% for small tumors and 81% for bulky tumors (Table 3). Seven bulky recurrent tumors could not be salvaged, despite six of these patients undergoing a total laryngectomy. Salvage surgery was not attempted for two patients because of medical contraindications. These two patients and the seven who failed salvage laryngectomy succumbed to local tumor progression. 3.3. Voice quality and preservation Although voice quality was not measured objectively, it was improved subjectively in 94 (82%) patients, with the remaining 20 (18%) patients experiencing no change. No patient complained that hoarseness was worse after irradiation. Hoarseness of voice gradually returned or became worse during follow-up for 23 patients which was due to local recurrence in 21 patients and due to severe laryngeal edema in two patients. Five of the 21 recurrences were salvaged with hemilaryngectomy, bringing the ultimate voice preservation to 86% (98/114) (Table 3). Ninety-two percent of patients with small tumors retained their voices, compared to 73% of those with bulky tumors. 3.4. Survival The overall survival and disease-free survival rates at 5 years for all patients with T1 glottic cancer after primary Table 3 Ultimate local control and voice preservation after primary radiotherapy and salvage surgery Tumor size
Small Bulky Total
No. of patients
77 37 114
No. with recurrence
7 14 21
No. salvaged TL
HL
4 3 7
1 4 5
Ultimate control (n) (%)
Voice preservation (n) (%)
75 (97) 30 (81) 105 (92)
71 (92) 27 (73) 98 (86)a
TL, total laryngectomy; HL, hemilaryngectomy. a One patient who had TL for chondronecrosis is not included.
radiotherapy and salvage laryngectomy for post-irradiation recurrences were 73 and 92%, respectively. On univariate and multivariate analysis only tumor bulk significantly influenced disease-free survival. Disease-free survival at 5 years for small and bulky tumors was 97 and 84%, respectively (P = 0.0017) (Table 1) and 95% when the anterior commissure was not involved with cancer and 83% when it was involved (P = 0.06). 3.5. Complications Severe complications were noted in 1.7% (two of 114) of patients. These included severe glottic edema requiring permanent tracheostomy in one patient and chondroradionecrosis requiring total laryngectomy in another. Both patients had small tumors and received 70 Gy, with less than 2 Gy per fraction.
4. Discussion According to the current UICC staging system, T1 glottic cancers are divided into two sub-groups, i.e. tumors involving one true vocal cord (T1a) and tumors involving both true vocal cords (T1b). The tumor volume or bulk is not taken into consideration in this substaging. Thus, a small surface lesion measuring a few millimeters as well as a bulky infiltrating lesion involving an entire true vocal cord will both be staged as T1a, whereas a small surface lesion involving the anterior one-third or less of both true vocal cords as well as a bulky infiltrating tumor involving both true vocal cords will be staged as T1b. This heterogeneity of tumor size and mucosal extent within the T1 category of glottic cancers might be expected to influence the results of radiotherapy. The influence of tumor stage on the radiation control of T1 glottic cancer has been studied by several authors who found no difference in local control whether one or both true vocal cords were involved [2,6,7,11,14,15,17,19]. Our study showed no difference between the radiation control rates of T1a lesions and T1b lesions. Analysis of the risk of local failure after radiotherapy for early glottic cancer showed the probability of success to be closely related to the volume or bulk of the lesion. Dickens et al. [3] analyzed the rates of tumor control according to tumor bulk in a group of patients with T1 and T2 glottic cancer irradiated with curative intent. They found no failures for lesions measuring less than 5 mm, failure rates of 4% for tumors measuring 5–15 mm and failure rates of 26% for lesions larger than 15 mm. Mendenhall et al. [15] noted a decrease in tumor control for T1a glottic cancer as the size of the lesion increased. We found significantly higher rates of local control and disease-free survival for patients with small T1 glottic tumors as compared to bulky lesions. The effect of anterior commissure involvement on local control of T1 glottic cancer has been studied by several
S.P. Reddy et al. / Radiotherapy and Oncology 47 (1997) 161–166
authors and most studies failed to show a relationship between anterior commissure involvement and local control [5,8,14,15,17,18,20]. Although anterior commissure involvement showed an adverse effect on the local control of T1 glottic cancer in our univariate analysis, on multivariate analysis we found that the effect was of no significance. Recent studies showed improved local control for patients with T1 glottic cancer when the total radiation dose was delivered in a shorter overall treatment time with a higher dose fractionation schedule. Rudoltz et al. [17] reported 100% local control when irradiation was completed within 42 days, in contrast to 50% when treatment lasted from 55 to 66 days. Schwaibold et al. [18] reported a lower local control rate for patients with T1 glottic cancer irradiated with 180 Gy per fraction as compared to those irradiated with 200 cGy per fraction. They also studied local control as a function of treatment duration and reported no failures in patients whose irradiation was completed in 51 days. For a combined group of patients with T1 and T2 stage glottic cancers, Fein et al. [4] reported a lower local control rate at 2 years when radiation therapy was extended over more than 50 days, as compared to less than 50 days. However, they found no significant difference in local control when T1 tumors were analyzed separately. The abovequoted studies did not analyze the significance of tumor bulk as a prognostic factor for patients with T1 glottic cancers irradiated with curative intent. In our study, on univariate analysis, the effect of overall duration of irradiation beyond 50 days on local control of T1 glottic cancer appeared to be of borderline significance. However, on multivariate analysis, we found that the effect of total treatment time was of no significance. Most authors have failed to demonstrate a clear-cut dose– response effect for early carcinoma of the glottic larynx [2, 5,6,8,12,13,17–19]. It is possible that, in most series, the doses used for irradiation of T1 glottic tumors, which are of relatively small volume, are already on the plateau of the sigmoid dose–response curve, thus obscuring any dose– response relationship that may exist. In our study, the median total dose was 66 Gy and the difference in local control above and below this dose was not significant. However, in T2 vocal cord cancer, reports of higher local control using hyper fractionated RT (total dose 74.4–76.8 Gy) have been reported [22] and we would recommend using at least 70 Gy by conventional fractionation or .74.2 Gy by hyperfractionation in the bulky T1 patients. We would also recommend closer follow-up for patients with bulky disease so as to improve salvage rates as seven of 14 patients with initial bulky disease who recurred could not be salvaged despite six of the seven undergoing a total laryngectomy. Several reports, as well as ours, show that histologic differentiation [8,17,18], radiation field size [2,4,9,12,17,18, 21] and beam energy [4,17] have no influence on local tumor control. Tumor bulk was the only significant factor in this study that influenced the disease-free survival at 5 years. Rates of
165
ultimate local control and voice preservation after salvage surgery were higher for patients with small tumors than for patients with bulky lesions.
5. Conclusion Tumor bulk was the only significant factor in multivariate and univariate analysis influencing the local control and disease-free survival of patients with T1 glottic cancer after external beam radiotherapy. Other factors which were significant on univariate analysis included anterior commissure involvement for local control. Based on these findings, we suggest that the present UICC staging for glottic cancer be modified to incorporate tumor bulk in the substaging of these cancers.
Acknowledgements The authors wish to thank Preston Bricker for his help with statistics and Bonnie Kalemba for typing the manuscript. References [1] UICC (International Union Against Cancer). TNM Classification of Malignant Tumors. Larynx – Classification. Springer, Berlin, 1987. [2] Akine, Y., Tokita, N., Ogino, T., Tsukiyama, I., Egawa, S., Saikawa, M. et al. Radiotherapy of T1 glottic cancer with 6 MeV X-rays. Int. J. Radiat. Oncol. Biol. Phys. 20: 1215–1218, 1991. [3] Dickens, W.J., Cassisi, N.J., Million, R.R. and Bova, F.J. Treatment of early vocal cord carcinoma: a comparison of apples and apples. Laryngoscope 93: 216–219, 1983. [4] Fein, D.A., Lee, W.R., Hanlon, A.L., Ridge, J.A., Curran, W.J. and Coia, L.R. Do overall treatment time, field size, and treatment energy influence local control of T1-T2 squamous cell carcinomas of glottic larynx. Int. J. Radiat. Oncol. Biol. Phys. 34: 823–831, 1996. [5] Fletcher, G.H. and Klein, R. Dose-time-volume relationship in squamous-cell carcinoma of the larynx. Radiology 82: 1032–1041, 1964. [6] Harwood, A.R., Hawkins, N.V., Rider, W.D. and Bryce, D.P. Radiotherapy of early glottic cancer-I. Int. J. Radiat. Oncol. Biol. Phys. 5: 473–476, 1979. [7] Harwood, A.R. and Tierie, A. Radiotherapy of early glottic cancer-II. Int. J. Radiat. Oncol. Biol. Phys. 5: 477–482, 1979. [8] Horiot, J.C., Fletcher, G.H., Ballantyne, A.J. and Lindberg, R.D. Analysis of failures in early vocal-cord cancer. Radiology 103: 663–665, 1972. [9] Inoue, T., Inoue, T., Ikeda, H., Teshima, T. and Murayama, S. Prognostic factor of telecobalt therapy for early glottic carcinoma. Cancer 70: 2797–2801, 1992. [10] Kaplan, E.L. and Meier, P. Nonparametric estimation from incomplete observations. J. Am. Stat. Assoc. 53: 457–481, 1958. [11] Kim, R.Y., Marks, M.E. and Salter, M.M. Early-stage glottic cancer: importance of dose fractionation in radiation therapy. Radiology 182: 273–275, 1992. [12] Lustig, R.A., MacLean, C.J., Hanks, G.E. and Kramer, S. The patterns of care outcome studies: results of the national practice in carcinoma of the larynx. Int. J. Radiat. Oncol. Biol. Phys. 10: 2357–2362, 1984. [13] Marks, J.E., Lowry, L.D., Lerch, I. and Griem, M.L. Glottic cancer:
166
[14]
[15]
[16]
[17]
S.P. Reddy et al. / Radiotherapy and Oncology 47 (1997) 161–166 an analysis of recurrence as related to dose, time, and fractionation. Am. J. Radiol. 117: 540–547, 1973. Mendenhall, W.M., Parsons, J.T., Stringer, S.P., Cassisi, N.J. and Million, R.R. T1-T2 vocal cord carcinoma: a basis for comparing the results of radiotherapy and surgery. Head Neck Surg. 10: 373– 377, 1988. Mendenhall, W.M., Parsons, J.T., Million, R.R. and Fletcher, G.H. T1-T2 squamous cell carcinoma of the glottic larynx treated with radiation therapy: relationship of dose-fractionation factors to local control and complications. Int. J. Radiat. Oncol. Biol. Phys. 15: 1267–1273, 1988. Pellitteri, P.K., Kennedy, T.L., Vrabec, D.P., Beiler, D. and Hellstrom, M. Radiotherapy: the mainstay in the treatment of early glottic carcinoma. Arch. Otolaryngol. Head Neck Surg. 117: 297–301, 1991. Rudoltz, M.S., Benammar, A. and Mohiuddin, M. Prognostic factors for local control and survival in T1 squamous cell carcinoma of the glottis. Int. J. Radiat. Oncol. Biol. Phys. 26: 767–772, 1993.
[18] Schwaibold, F., Scariato, A., Nunno, M., Wallner, P.E., Lustig, R.A., Rouby, E. et al. The effect of fraction size on control of early glottic cancer. Int. J. Radiat. Oncol. Biol. Phys. 14: 451–454, 1988. [19] Small, W., Mittal, B.B., Brand, W.N., Shetty, R.M., Rademaker, A.W., Beck, G.G. et al. Results of radiation therapy in early glottic carcinoma: multivariate analysis of prognostic and radiation therapy variables. Radiology 183: 789–794, 1992. [20] Stevenson, J.M., Juillard, G.Y. and Selch, M.T. Stages 1 and 2 epidermoid carcinoma of the glottic larynx: involvement of the anterior commissure. Radiology 182: 797–799, 1992. [21] Teshima, T., Chatani, M. and Inoue, T. Radiation therapy for early glottic cancer (T1N0M0): II. Prospective randomized study concerning radiation field. Int. J. Radiat. Oncol. Biol. Phys. 18: 119–123, 1990. [22] Parsons, J.T., Mendenhall, W.M., Stringer, S.P., Cassis, N.J. and Million, R.R. Twice-a-day radiotherapy for squamous cell carcinoma of the head and neck. Head Neck 15: 87–96, 1993.
Effect of tumor bulk on local control and survival of patients with T1 glottic cancer Sarada P. Reddy*, Najeeb Mohideen, Silvio Marra, James E. Marks Loyola-Hines Department of Radiotherapy, Loyola University Chicago, 2160 South First Avenue, Maywood, IL 60153, USA Received 11 April 1997; revised version received 17 September 1997; accepted 2 October 1997
Abstract Purpose: To evaluate the effect of tumor bulk in relation to various tumor-related prognostic factors and treatment-related variables on local control and survival of patients with T1 N0 M0 squamous cell carcinoma of the glottis. Materials and methods: In 114 patients with T1 squamous cell carcinoma of the glottic larynx who were irradiated with curative intent, we determined the effect of tumor bulk in relation to mucosal extent (stage and anterior commissure involvement), histologic differentiation and various radiation factors, especially overall treatment time on local control and survival. Tumors were classified retrospectively as small surface lesions or bulky tumors. Seventy-seven patients had small lesions and 37 had bulky tumors. The anterior commissure was involved with cancer in 43 patients. The overall duration of irradiation ranged from 39 to 64 days. The median follow-up time was 6 years (range 5–24 years). Results: The 5-year actuarial local control rate for all patients was 82% after radiotherapy and 92% after salvage laryngectomy. On univariate analysis, bulky tumors and tumors involving the anterior commissure showed an adverse effect on local control, whereas the overall duration of irradiation had a borderline significance. The actuarial local control rate was 91% for small tumors and 58% for bulky tumors (P = 0.0002), 88% when the anterior commissure was not involved and 67% when the anterior commissure was involved (P = 0.01) and 89% when radiation was given in less than 50 days and 73% when irradiation exceeded 50 days (P = 0.06). On multivariate analysis, tumor bulk was the only significant factor that affected local control (P = 0.02). The 5-year actuarial survival for all patients was 73% and the disease-free survival was 92%. Conclusion: This study shows that tumor bulk has a highly significant effect on the radiation control of T1 glottic cancer. Patients who had bulky tumors had lower local control and disease-free survival rates than those patients who had small tumors. 1998 Elsevier Science Ireland Ltd. All rights reserved Keywords: T1 glottic cancer; Tumor bulk; Local control; Ultimate control; Voice preservation; Survival; Complications
1. Introduction Irradiation of early glottic cancer yields 5-year local control rates of 80–95% and ultimate control rates of 90–100% after salvage surgery of recurrences [2–4,9,14–20]. In the quest to improve these results, radiation oncologists continue to search for possible prognostic factors that might significantly affect the local control of early glottic cancer. Since the early 1960s, no one has been able to establish a clear relationship between tumor control and radiation dose for T1 glottic cancer [2,5,6,8,12,13,17–19]. During the 1980s, improved local control rates were reported when * Corresponding author.
radiation was delivered in a shorter overall time, with higher dose fractionation schedules [4,11,15,17,18]. During the same era, a decrease in the local control rate was reported as the size of the lesion increased for patients with T1a glottic cancer [15] and for patients with bulky T1 and T2 lesions [3]. The effect of tumor bulk on local control of T1 glottic cancer, regardless of the number of true vocal cords involved, has not been reported in the literature. The purpose of this study is to determine the effect of tumor bulk on local control, survival and voice preservation in patients with T1 glottic cancer and to evaluate its influence on various tumor-related prognostic factors and treatment-related variables. The prognostic factors studied include the tumor stage, anterior commissure involvement,
0167-8140/98/$19.00 1998 Elsevier Science Ireland Ltd. All rights reserved PII S0167-8140 (97 )0 0196-5
162
S.P. Reddy et al. / Radiotherapy and Oncology 47 (1997) 161–166
histologic differentiation, radiation beam energy, field size, total dose, fraction size and total duration of irradiation.
2. Materials and methods 2.1. Patient population From 1976 to 1990, 126 patients with histologically confirmed T1 squamous cell carcinomas of the true vocal cords were referred to our department for curative irradiation. Radiotherapy charts and medical records were reviewed. Based on the findings recorded at indirect and direct laryngoscopy and on radiographic findings, all tumors were restaged according to the 1992 UICC staging system [1]. Patients irradiated from 1976 to 1992 were chosen for this analysis because we found meticulously drawn tumor diagrams, carefully documented tumor descriptions in the medical records and laryngeal tomograms and CT scans from 1976 onwards. Twelve of the 126 patients were excluded from this analysis; four patients died within 1 year and eight patients were lost to follow-up in less than 2 years after irradiation. All 12 patients who were excluded from this analysis had no local recurrence when they were last seen. The remaining 114 patients who form the basis for this study were followed for a minimum of 5 years or until death and for a maximum of 24 years (median 6 years). Tumor characteristics are listed in Table 1. There were 111 male and three female patients, a finding
Table 1 Tumor characteristics and treatment results No. of patients
Local control % (95% CI)a
Size Small 77 91 Bulky 37 58 Stage T1a 94 80 T1b 20 90 AC involvement No 71 88 Fraction size (Gy) 1.8 46 73 2 68 85 Total dose (Gy) 60–65 28 87 66 36 89 67–70 50 71 Total days irradiation ≤50 54 89 >50 60 73
Log-rank Survival % (95% CI)b
Log-rank
(84–98) (40–76)
0.0002
97 (95–99) 84 (62–92)
0.0017
(71–89) (80–100)
0.09
89 (82–96) 100
0.21
(80–96)
0.01
95 (90–100)
0.06
(60–86) (76–94)
0.09
86 (75–97) 94 (88–100)
0.09
(74–100) (78–100) (58–84)
0.09
91 (79–103) 95 (86–104) 87 (77–97)
0.3
(81–98) (61–85)
0.06
96 (92–100) 86 (76–96)
0.1
AC, anterior commissure; CI, confidence interval. Five-year local control. Five-year disease-free survival.
a
b
Fig. 1. Classification of T1 glottic cancer according to tumor bulk. (A,B) Small tumors. (C,D) Bulky tumors.
that was not surprising given the large number of veterans in the study. Eighty-one patients were white, 32 were black and one was oriental. The patients’ ages ranged from 44 to 79 years, with a median of 60 years. 2.2. Tumor classification Since the present UICC staging system does not consider tumor bulk in the substaging of T1 glottic cancer, we retrospectively classified all T1 glottic tumors that we had irradiated either as small or bulky lesions, regardless of whether one or both true vocal cords were involved. Our classification was based on the description of the tumors, on the diagrams drawn in the medical records at the time of diagnostic endoscopy and on the radiographic tests, which were laryngeal tomograms before the availability of computerized tomography, and CT scans of the larynx later on. Surface lesions not involving an entire true vocal cord and those involving less than the anterior one-third of both true vocal cords (horseshoe-shaped lesions) were classified as small tumors. Large exophytic and/or infiltrative lesions involving an entire true vocal cord and horseshoe-shaped lesions involving more than the anterior one-third of both true vocal cords were classified as bulky tumors (Fig. 1). So that bias could be avoided, each lesion was classified separately as small or bulky by a panel of three physicians, consisting of two radiation oncologists and an otolaryngologist who reached a consensus on tumor bulk. Seventyseven tumors were classified as small and 37 were classified as bulky tumors. Anterior commissure was involved in 14 patients with small tumors and in 29 patients with bulky tumors.
163
S.P. Reddy et al. / Radiotherapy and Oncology 47 (1997) 161–166
2.3. Radiotherapy The majority of patients were irradiated through opposed lateral portals and a few were irradiated through right and left anterior oblique portals. We used cobalt-60 g-rays to irradiate 90 patients, whereas 24 were irradiated with 4 MV X-rays. Total dose ranging from 60 to 70 Gy (median 66 Gy) was delivered as a continuous course prescribed at the intersection of the two beams at midplane. Wedge filters were used when appropriate for optimizing the dose distribution to achieve a homogeneity of ±5% of the prescribed dose. Both fields were irradiated daily, 5 days a week. Doses per fraction ranged from 1.8 to 2.0 Gy. The total duration of irradiation lasted from 42 to 60 days. Treatment parameters are listed in Table 1. The end-points analyzed in this study were local control after primary radiotherapy, ultimate control after salvage surgery for radiotherapy failures, voice quality and voice preservation, overall and disease-free survival and complications. Survival was measured from the date of pathologic diagnosis to the date of last follow-up or death. Local control was measured from the last day of irradiation until the date of biopsy-proven recurrence. Biopsies were performed on all patients with recurrence. 2.4. Statistical analysis Local control, overall survival and disease-free survival probabilities for each possible prognostic factor were calculated by the Kaplan–Meier method [10]. The significance of differences between the paired local control curves was compared by both the log-rank test and the Wilcoxon method. In addition, the influence of tumor bulk on local control when the cancer extended onto the anterior commissure, as well as when the treatment time exceeded 50 days, was assessed by the x2-test. Multivariate analysis of factors that appeared to affect local control and survival was performed with Cox’s proportional hazards model. Statistical analysis was done with the SAS program (SAS Institute, Cary, NC).
to have a significant effect on radiation control of early glottic cancer, whereas the effect of total duration of irradiation appeared to be of borderline significance (Table 1). The local control rate for small tumors was 91%, compared to 58% for bulky tumors (P = 0.0002). Eighty-eight percent of patients without anterior commissure involvement and 67% with anterior commissure involvement had local control (P = 0.01). The local control rate was 89% for patients whose irradiation was completed in 50 days or less and 73% for those whose irradiation lasted for more than 50 days (P = 0.06). Tumor stage, histologic differentiation, radiation beam energy, field size, total dose and fraction size of irradiation showed no significant effect on local control of T1 glottic cancer. When the analysis was controlled for tumor bulk, anterior commissure involvement and total duration of irradiation beyond 50 days did not show an adverse effect on local control (Table 2). The local control rate for small tumors was 86% when the anterior commissure was involved and 92% when the anterior commissure was not involved (NS) and for bulky tumors the control rates were 62 and 63%, respectively (NS). However, anterior commissure involvement was noted more frequently in patients with bulky tumors than in patients with small tumors (78 versus 18%), a finding which relates the lower local control seen in patients with anterior commissure involvement with tumor bulk rather than its location. Likewise, the local control rate was 95% for patients with small tumors irradiated within 50 days and 86% for those irradiated over more than 50 days (NS). For patients with bulky tumors, the rates were 67 and 60%, respectively (NS). Therefore, it appears that tumor bulk is the only significant factor that influences the local control of T1 glottic cancer, irrespective of anterior commissure involvement and total duration of irradiation. Multivariate analysis of tumor bulk, anterior commissure involvement, total duration of irradiation and other tumorTable 2 Influence of tumor bulk, anterior commissure (AC) involvement and total days of irradiation on local control Local control (no. controlled /total no.) (%)
P-value (x2)
58/63 (92) 12/14 (86)
NS
40/42 (95) 30/35 (86)
NS
5/8 (63) 18/29 (62)
NS
8/12 (67) 15/25 (60)
NS
3. Results 3.1. Local control The local control rate at 5 years for the entire group of patients after primary radiotherapy was 82% (93/114). Twenty-one patients developed local recurrences. Sixteen (76%) of these 21 recurrences occurred within 2 years and 20 (95%) recurred by 4 years. The twenty-first patient had a recurrence shortly after the fourth year and no recurrences developed thereafter. Univariate analysis of all prognostic factors studied showed tumor bulk and anterior commissure involvement
Small tumors AC involvement No Yes Treatment days ≤50 >50 Bulky tumors AC involvement No Yes Treatment days ≤50 >50
164
S.P. Reddy et al. / Radiotherapy and Oncology 47 (1997) 161–166
and treatment-related variables studied showed only tumor bulk to have a significant effect on local control of T1 glottic cancer (P = 0.017). 3.2. Ultimate local control after surgical salvage Salvage surgery for 19 of the 21 patients with recurrences consisted of 13 total laryngectomies and six hemilaryngectomies. After successful surgical salvage of 12 patients (seven total laryngectomies and five hemilaryngectomies), the ultimate local control was 92% for the entire group of 114 patients (Table 3). When analyzed according to tumor size, five (71%) of the seven recurrences in patients with small tumors and seven (50%) of the 14 recurrences in patients with bulky tumors were salvaged with surgery, bringing the ultimate local control to 97% for small tumors and 81% for bulky tumors (Table 3). Seven bulky recurrent tumors could not be salvaged, despite six of these patients undergoing a total laryngectomy. Salvage surgery was not attempted for two patients because of medical contraindications. These two patients and the seven who failed salvage laryngectomy succumbed to local tumor progression. 3.3. Voice quality and preservation Although voice quality was not measured objectively, it was improved subjectively in 94 (82%) patients, with the remaining 20 (18%) patients experiencing no change. No patient complained that hoarseness was worse after irradiation. Hoarseness of voice gradually returned or became worse during follow-up for 23 patients which was due to local recurrence in 21 patients and due to severe laryngeal edema in two patients. Five of the 21 recurrences were salvaged with hemilaryngectomy, bringing the ultimate voice preservation to 86% (98/114) (Table 3). Ninety-two percent of patients with small tumors retained their voices, compared to 73% of those with bulky tumors. 3.4. Survival The overall survival and disease-free survival rates at 5 years for all patients with T1 glottic cancer after primary Table 3 Ultimate local control and voice preservation after primary radiotherapy and salvage surgery Tumor size
Small Bulky Total
No. of patients
77 37 114
No. with recurrence
7 14 21
No. salvaged TL
HL
4 3 7
1 4 5
Ultimate control (n) (%)
Voice preservation (n) (%)
75 (97) 30 (81) 105 (92)
71 (92) 27 (73) 98 (86)a
TL, total laryngectomy; HL, hemilaryngectomy. a One patient who had TL for chondronecrosis is not included.
radiotherapy and salvage laryngectomy for post-irradiation recurrences were 73 and 92%, respectively. On univariate and multivariate analysis only tumor bulk significantly influenced disease-free survival. Disease-free survival at 5 years for small and bulky tumors was 97 and 84%, respectively (P = 0.0017) (Table 1) and 95% when the anterior commissure was not involved with cancer and 83% when it was involved (P = 0.06). 3.5. Complications Severe complications were noted in 1.7% (two of 114) of patients. These included severe glottic edema requiring permanent tracheostomy in one patient and chondroradionecrosis requiring total laryngectomy in another. Both patients had small tumors and received 70 Gy, with less than 2 Gy per fraction.
4. Discussion According to the current UICC staging system, T1 glottic cancers are divided into two sub-groups, i.e. tumors involving one true vocal cord (T1a) and tumors involving both true vocal cords (T1b). The tumor volume or bulk is not taken into consideration in this substaging. Thus, a small surface lesion measuring a few millimeters as well as a bulky infiltrating lesion involving an entire true vocal cord will both be staged as T1a, whereas a small surface lesion involving the anterior one-third or less of both true vocal cords as well as a bulky infiltrating tumor involving both true vocal cords will be staged as T1b. This heterogeneity of tumor size and mucosal extent within the T1 category of glottic cancers might be expected to influence the results of radiotherapy. The influence of tumor stage on the radiation control of T1 glottic cancer has been studied by several authors who found no difference in local control whether one or both true vocal cords were involved [2,6,7,11,14,15,17,19]. Our study showed no difference between the radiation control rates of T1a lesions and T1b lesions. Analysis of the risk of local failure after radiotherapy for early glottic cancer showed the probability of success to be closely related to the volume or bulk of the lesion. Dickens et al. [3] analyzed the rates of tumor control according to tumor bulk in a group of patients with T1 and T2 glottic cancer irradiated with curative intent. They found no failures for lesions measuring less than 5 mm, failure rates of 4% for tumors measuring 5–15 mm and failure rates of 26% for lesions larger than 15 mm. Mendenhall et al. [15] noted a decrease in tumor control for T1a glottic cancer as the size of the lesion increased. We found significantly higher rates of local control and disease-free survival for patients with small T1 glottic tumors as compared to bulky lesions. The effect of anterior commissure involvement on local control of T1 glottic cancer has been studied by several
S.P. Reddy et al. / Radiotherapy and Oncology 47 (1997) 161–166
authors and most studies failed to show a relationship between anterior commissure involvement and local control [5,8,14,15,17,18,20]. Although anterior commissure involvement showed an adverse effect on the local control of T1 glottic cancer in our univariate analysis, on multivariate analysis we found that the effect was of no significance. Recent studies showed improved local control for patients with T1 glottic cancer when the total radiation dose was delivered in a shorter overall treatment time with a higher dose fractionation schedule. Rudoltz et al. [17] reported 100% local control when irradiation was completed within 42 days, in contrast to 50% when treatment lasted from 55 to 66 days. Schwaibold et al. [18] reported a lower local control rate for patients with T1 glottic cancer irradiated with 180 Gy per fraction as compared to those irradiated with 200 cGy per fraction. They also studied local control as a function of treatment duration and reported no failures in patients whose irradiation was completed in 51 days. For a combined group of patients with T1 and T2 stage glottic cancers, Fein et al. [4] reported a lower local control rate at 2 years when radiation therapy was extended over more than 50 days, as compared to less than 50 days. However, they found no significant difference in local control when T1 tumors were analyzed separately. The abovequoted studies did not analyze the significance of tumor bulk as a prognostic factor for patients with T1 glottic cancers irradiated with curative intent. In our study, on univariate analysis, the effect of overall duration of irradiation beyond 50 days on local control of T1 glottic cancer appeared to be of borderline significance. However, on multivariate analysis, we found that the effect of total treatment time was of no significance. Most authors have failed to demonstrate a clear-cut dose– response effect for early carcinoma of the glottic larynx [2, 5,6,8,12,13,17–19]. It is possible that, in most series, the doses used for irradiation of T1 glottic tumors, which are of relatively small volume, are already on the plateau of the sigmoid dose–response curve, thus obscuring any dose– response relationship that may exist. In our study, the median total dose was 66 Gy and the difference in local control above and below this dose was not significant. However, in T2 vocal cord cancer, reports of higher local control using hyper fractionated RT (total dose 74.4–76.8 Gy) have been reported [22] and we would recommend using at least 70 Gy by conventional fractionation or .74.2 Gy by hyperfractionation in the bulky T1 patients. We would also recommend closer follow-up for patients with bulky disease so as to improve salvage rates as seven of 14 patients with initial bulky disease who recurred could not be salvaged despite six of the seven undergoing a total laryngectomy. Several reports, as well as ours, show that histologic differentiation [8,17,18], radiation field size [2,4,9,12,17,18, 21] and beam energy [4,17] have no influence on local tumor control. Tumor bulk was the only significant factor in this study that influenced the disease-free survival at 5 years. Rates of
165
ultimate local control and voice preservation after salvage surgery were higher for patients with small tumors than for patients with bulky lesions.
5. Conclusion Tumor bulk was the only significant factor in multivariate and univariate analysis influencing the local control and disease-free survival of patients with T1 glottic cancer after external beam radiotherapy. Other factors which were significant on univariate analysis included anterior commissure involvement for local control. Based on these findings, we suggest that the present UICC staging for glottic cancer be modified to incorporate tumor bulk in the substaging of these cancers.
Acknowledgements The authors wish to thank Preston Bricker for his help with statistics and Bonnie Kalemba for typing the manuscript. References [1] UICC (International Union Against Cancer). TNM Classification of Malignant Tumors. Larynx – Classification. Springer, Berlin, 1987. [2] Akine, Y., Tokita, N., Ogino, T., Tsukiyama, I., Egawa, S., Saikawa, M. et al. Radiotherapy of T1 glottic cancer with 6 MeV X-rays. Int. J. Radiat. Oncol. Biol. Phys. 20: 1215–1218, 1991. [3] Dickens, W.J., Cassisi, N.J., Million, R.R. and Bova, F.J. Treatment of early vocal cord carcinoma: a comparison of apples and apples. Laryngoscope 93: 216–219, 1983. [4] Fein, D.A., Lee, W.R., Hanlon, A.L., Ridge, J.A., Curran, W.J. and Coia, L.R. Do overall treatment time, field size, and treatment energy influence local control of T1-T2 squamous cell carcinomas of glottic larynx. Int. J. Radiat. Oncol. Biol. Phys. 34: 823–831, 1996. [5] Fletcher, G.H. and Klein, R. Dose-time-volume relationship in squamous-cell carcinoma of the larynx. Radiology 82: 1032–1041, 1964. [6] Harwood, A.R., Hawkins, N.V., Rider, W.D. and Bryce, D.P. Radiotherapy of early glottic cancer-I. Int. J. Radiat. Oncol. Biol. Phys. 5: 473–476, 1979. [7] Harwood, A.R. and Tierie, A. Radiotherapy of early glottic cancer-II. Int. J. Radiat. Oncol. Biol. Phys. 5: 477–482, 1979. [8] Horiot, J.C., Fletcher, G.H., Ballantyne, A.J. and Lindberg, R.D. Analysis of failures in early vocal-cord cancer. Radiology 103: 663–665, 1972. [9] Inoue, T., Inoue, T., Ikeda, H., Teshima, T. and Murayama, S. Prognostic factor of telecobalt therapy for early glottic carcinoma. Cancer 70: 2797–2801, 1992. [10] Kaplan, E.L. and Meier, P. Nonparametric estimation from incomplete observations. J. Am. Stat. Assoc. 53: 457–481, 1958. [11] Kim, R.Y., Marks, M.E. and Salter, M.M. Early-stage glottic cancer: importance of dose fractionation in radiation therapy. Radiology 182: 273–275, 1992. [12] Lustig, R.A., MacLean, C.J., Hanks, G.E. and Kramer, S. The patterns of care outcome studies: results of the national practice in carcinoma of the larynx. Int. J. Radiat. Oncol. Biol. Phys. 10: 2357–2362, 1984. [13] Marks, J.E., Lowry, L.D., Lerch, I. and Griem, M.L. Glottic cancer:
166
[14]
[15]
[16]
[17]
S.P. Reddy et al. / Radiotherapy and Oncology 47 (1997) 161–166 an analysis of recurrence as related to dose, time, and fractionation. Am. J. Radiol. 117: 540–547, 1973. Mendenhall, W.M., Parsons, J.T., Stringer, S.P., Cassisi, N.J. and Million, R.R. T1-T2 vocal cord carcinoma: a basis for comparing the results of radiotherapy and surgery. Head Neck Surg. 10: 373– 377, 1988. Mendenhall, W.M., Parsons, J.T., Million, R.R. and Fletcher, G.H. T1-T2 squamous cell carcinoma of the glottic larynx treated with radiation therapy: relationship of dose-fractionation factors to local control and complications. Int. J. Radiat. Oncol. Biol. Phys. 15: 1267–1273, 1988. Pellitteri, P.K., Kennedy, T.L., Vrabec, D.P., Beiler, D. and Hellstrom, M. Radiotherapy: the mainstay in the treatment of early glottic carcinoma. Arch. Otolaryngol. Head Neck Surg. 117: 297–301, 1991. Rudoltz, M.S., Benammar, A. and Mohiuddin, M. Prognostic factors for local control and survival in T1 squamous cell carcinoma of the glottis. Int. J. Radiat. Oncol. Biol. Phys. 26: 767–772, 1993.
[18] Schwaibold, F., Scariato, A., Nunno, M., Wallner, P.E., Lustig, R.A., Rouby, E. et al. The effect of fraction size on control of early glottic cancer. Int. J. Radiat. Oncol. Biol. Phys. 14: 451–454, 1988. [19] Small, W., Mittal, B.B., Brand, W.N., Shetty, R.M., Rademaker, A.W., Beck, G.G. et al. Results of radiation therapy in early glottic carcinoma: multivariate analysis of prognostic and radiation therapy variables. Radiology 183: 789–794, 1992. [20] Stevenson, J.M., Juillard, G.Y. and Selch, M.T. Stages 1 and 2 epidermoid carcinoma of the glottic larynx: involvement of the anterior commissure. Radiology 182: 797–799, 1992. [21] Teshima, T., Chatani, M. and Inoue, T. Radiation therapy for early glottic cancer (T1N0M0): II. Prospective randomized study concerning radiation field. Int. J. Radiat. Oncol. Biol. Phys. 18: 119–123, 1990. [22] Parsons, J.T., Mendenhall, W.M., Stringer, S.P., Cassis, N.J. and Million, R.R. Twice-a-day radiotherapy for squamous cell carcinoma of the head and neck. Head Neck 15: 87–96, 1993.