Clin. Ololuryngol, 1992, 17, 449-451
RAPID COMMUNICATION
Salvage laryngectomy after radical radiotherapy for laryngeal carcinoma R.P.CRELLIN*, M . N . G A Z E * , AILEEN WHITEt, A . G . D . M A R A N f ’ & R.H.MACDOUGALL* * Department of Clinical Oncology, Western General Hospital, Edinburgh and tOtolaryngology Unit, Royal
Infirmary,
Edinburgh EH3 9EN, U K Accepted for publication 30 January 1992 CREI.LIN R . P . , G A Z E M . N . , W H I T E A . , M A R A N A . G . D . & M A C D O U G A L L R . H .
(1992) Clin. Otolaryngol. 17, 449-451
Salvage laryngectomy after radical radiotherapy for laryngeal cancer Of 376 patients who were treated by radical radiotherapy for squamous carcinoma of the larynx, 56 subsequently underwent total laryngectomy. Residual or recurrent tumour was identified in 43 of the resection specimens, and necrosis alone in 13 cases, although a positive biopsy had been obtained in 3 of these prior to salvage laryngectomy. No disease related factors such as site or stage of the original tumour, o r treatment related factors such as radiation type or dose, were found to be predictive of whether o r not tumour was present. The clinical opinion of an experienced surgeon was found to have a positive predictive value of 0.86 for the presence of tumour. The fistula rate of salvage laryngectomy, 15 out of 56, was similar to that of other series. The actuarial cause specific 5-year survival for patients with tumour was 0.589, and for patients with necrosis only was 0.923. Intercurrent, smoking related disease was the cause of death in 16 of the 33 patients who have died. Keywords
carcinoma qf the larynx
laryngectomy
Radical radiotherapy is the treatment of choice for early laryngeal carcinoma. Even in advanced disease, where the chance of successful control by irradiation is less, many clinicians still advocate primary radiotherapy to preserve speech.’ When there is residual o r recurrcnt disease after radiotherapy, surgery offers a second chance to cure the patient. Occasionally, no tumour is identified in the resected larynx. This may be anticipated if the indication for surgery was necrosis following irradiation, or a surprise finding when a recurrence was suspected. This paper is an audit of laryngectomy practice in Edinburgh following primary radical radiotherapy for laryngeal cancer. The aim was to identify the proportion of resection specimens which contained no tumour, and to relate this finding to the indications for, complications of and outcome after salvage laryngectomy. Correspondence: Dr R.P.Crellin, Department of Clinical Oncology. Western General Hospital. Edinburgh EH4 2XU, UK
morbidity
radiotherapJ.
recurrence
Patients and methods In the I I years from 1974 to 1984, 376 patients received radical radiotherapy for squamous carcinoma of the larynx at the Western General Hospital, Edinburgh.’ In the majority of patients. treatment was given by small, parallel opposed, wedged lateral fields covering only the larynx and any involved nodcs. Usually a dose of 52.5-55 Gy in 20 fractions over 4 weeks was prescribed using 4 MV X-rays. neutron^.^ Twenty-one patients were treated with Laryngectomy was subsequently performed in 56 patients (1 5%), including 6 treated with neutrons. Laryngectomy was required in 24 of 99 patients with supraglottic tumours (24%) but only 32 of 268 (12%) with glottic tumours. Laryngectomy was performed in 13 of 135 patients (10%) with T, tumours, 13 of 89 (15%) with T2 tumours, 23 of 99 (23%) with T3 tumours and 7 of 44 (16%) with T, tumours. The surgical and radiotherapeutic case notes of these 56 patients, 48 men and 8 women, were reviewed to find the indications for laryngectomy, the preoperative opinion about
449
450
R.P.Crellin et al.
the presence of tumour, the operative morbidity, the pathology of the resected larynx, and the outcome. Survival rates were calculated by the life-table method and compared using the logrank test.
Results PAI‘HOLOGY OF RESECTION SPECIMENS
Histological examination of the 56 resected larynges showed residual or recurrent tumour in 43 (77%) and no evidence of tumour in 13 (23%). Various factors have been assessed to see whether they were more likely to be related to either necrosis or tumour. The patient and disease related factors include sex, site, tumour stage and nodal status. The treatment related factors include the presence of a trachcostomy, radiation dose and radiation type. None of these factors, either disease or treatment related, was found to be related to the presence o r absence of tumour in the excised larynx. Of the 6 patients treated with neutrons, 4 had tumour in the resected larynx and two necrosis only. INVICATIOKS FOR LARYNGECTOMY
The principal symptoms prior to surgery were stridor in 12 patients, hoarseness without stridor in 37 patients, of whom 4 also had pain and 3 foetor, pain alone in 4 patients, and foetor alone in 3. There was no difference in the distribution of these symptoms between those who were found to have cancer, and those with no evidence of tumour. Biopsy was undertaken prior to laryngectomy in 46 patients. This showed the presence of residual or recurrent disease in 36 of the 43 cases found to have tumour in the resection specimen. Biopsy was negative in 5 and not performed in 2 patients in this group. The biopsy was positive for cancer in 3 of the 13 cases where subsequently no tumour was found in the resection specimen, negative in 2 and not performed in 8. The surgeon’s opinion prior to laryngectomy as to whether or not the larynx harboured residual or recurrent disease was correct in 42 of the 56 cases. In only 1 patient was tumour found in the resected larynx when necrosis alone was suspected, and in 7 patients necrosis without malignancy was found when turnour was suspected (in 3 cases because of a positive biopsy). Thus the sensitivity of the surgeon’s opinion that cancer was present was 0.98, but the specificity was only 0.46. The positive predictive value of the surgeon’s opinion was 0.86, the negative predictive value 0.86, and the likelihood ratio 1.12. OUTCOME OF SURGERY
The principal complication of surgery was fistula formation which occurred in 15 patients. The fistula was pharyngo-
cutaneous in 14 patients and pharyngo-tracheal in 1. Other complications noted were pharyngeal stricture in 6 patients, infection in 5, dehiscence in 3, arterial blowout in 2 and tracheal stenosis in one. The distribution of these complications did not differ significantly between those patients with tumour and those with necrosis only. Survival was significantly better for those who did not have cancer in the resected larynx. The actuarial cause specific survival at 5 years was 0.923 in this group, compared with 0.589 in those who had recurrent or residual malignancy. The magnitude of this difference was 0.334 (95% confidence intervals, 0.075-0.593). Among patients found to have tumour in the resected larynx there was n o difference in survival between those with residual disease and those in whom the disease recurred following a complete response to radiotherapy. Of the 43 patients operated on for recurrent or residual disease, only 12 patients are alive and well. Sixteen have died from this cancer and 2 had recurrent cancer when last seen. Thirteen have died from other causes (infection 5; myocardial infarction, 3; cerebrovascular disease, 2, and 3 other malignancy (in each case, lung cancer)). Of the 13 patients who had no evidence of tumour, 9 are alive and well. One died post-operatively, 2 died from a myocardial infarction and 1 died from carcinoma of the maxillary antrum.
Discussion Radical radiotherapy is the mainstay of treatment for laryngeal carcinoma, though it is recognized that surgical salvage of radiation failures may be required, especially if the disease is advanced at presentation. Diagnosis of residual or recurrent disease may, however, be difficult. While a biopsy may indicate recurrent disease, a negative biopsy result is not reassuring if the clinical picture is suspicious. Persistent endolaryngeal oedcma indicates recurrent tumour in about half, yet even multiple deep biopsies of oedematous laryngeal mucosa often fail to demonstrate this: and may precipitate frank necrosis. As recurrence frequently coexists with a biopsy showing necrosis only does not chondronecr~sis,~ exclude persistent malignancy. For these reasons, and because a painful, necrotic, oedematous larynx is of little value to the patient and should perhaps be removed anyway if it fails to improve with conservative therapy, a surgeon must be prepared to act on suspicion of recurrence, without absolute proof. In this series, a high index of suspicion was maintained, so that only 1 case of recurrence was thought to be necrosis only. However, the specificity of this suspicion was lower than its sensitivity, with 4 patients having cancerfree larynges removed on suspicion of recurrence, without prior histological verification. While occasionally limited surgery for recurrence may be possible,b total laryngectomy is the most frequently performed procedure. It is believed by some that a small recur-
Salvage laryngectomy
rence may be completely removed at biopsy, leading to a ‘biopsy cure’,’ as apparently happened in 3 cases in this series. Alternatively, small foci of tumour were missed by the pathologist during examination of the larynx. This is a more plausible explanation in initially advanced tumours where it would be naive to imagine that all malignancy except a small superficial area had been eradicated by irradiation. Nevertheless, patients in whom no tumour is found, even if in reality small islands of malignant cells do exist, have an excellent prognosis. The complications of salvage laryngectomy in the present series are typical, with a fistula rate of 27%, similar to the 32’- 37%’ reported by others. Fcw reported series of salvage laryngectomy acknowledge the proportion of larynges removed for necrosis in the absence of tumour, yet this is clearly important in the assessment of this procedure. Our figure of 23% compares with 12% reported from Liverpool,* and 3% from Toronto, Canada.’ Fortunately, survival for this group of patients is good, and quality of life is paradoxically often better when a painful, necrotic and barely functional larynx is removed. At 59%, the survival rate for those undergoing laryngectomy for recurrent or residual cancer is similar to the range of 39-56% reported by other^.^,^ The fact that about half the deaths in this group are not due to laryngeal cancer but t o other smoking related disease, indicates that crude rather than cause specific survival rates should be used in any costbenefit analysis of the treatment of laryngeal cancer. In summary, this audit of salvage laryngectomy after radical radiotherapy has shown that about onc-quarter of procedures are performed for necrosis alone, but that these patients have a significantly better prognosis than if cancer is present. No disease or treatment related factors appear to be predictive of whether a resected larynx will show necrosis
451
alone or tumour. The best predictor of recurrence is the experienced surgeon’s clinical opinion. The observed rate of complications and survival following salvage laryngectomy is similar to that seen in other centres.
Acknowledgements We wish to thank Miss G.R.Kerr for statistical advice.
References GAZE M.N., MACKILLOP, W.J., O’SULLIVAK B., GILBERT R., MAKI E. & LUNDCREN J. (1991) The management of laryngeal cancer: a comparison of practice in the United Kingdom and Canada. Clin. Oncol. 3, 296 (Abstract) HEYWOOD J.M., KERR G.R., ORR J.A., MACDOUCALL R.H. & MACKII.I.OP W.J. (1992) Primary radiotherapy for carcinoma of the larynx: the Edinburgh experience. Radiother. Oncol. (in press) MACDOUGALL R.H., ORRJ.A., KERRG.R. & DUNCAN W. (1 990) Fast neutron treatment for squamous cell carcinoma of the head and neck: final report of Edinburgh randomised trial. Er. Med. J . 301, 1241-1242 Fu K.K., WOODEHOUSE R.J., QUIVEYJ.M., PHILLIPS T.L. & DEW, H.H. (1982) The significance of laryngeal edema following radiotherapy of carcinoma of the vocal cord. Cancer 49, 655-658 KEBNEM., HARWDODA.R., BRYCED.P. & VAN NOSTRAND A.W.P. (1982) Histopathological study of radionecrosis in laryngeal carcinoma. Laryngoscope 92, 173-180 CROLLG.A., VAS DEK BROEKP., TIWARIR.M., MANNIJ . J . & SNOW G.B. (1985) Vertical partial laryngectomy for recurrent glottic carcinoma after irradiation. Head Neck Surg. 7 , 390-393 NICHOLSR.D., STINEP.H. & GREENAWALD K.J. (1980) Partial laryngectomy after radiation failure. Larvngoscope 90, 57 1-575 VIANIL., STELLP.M. & DALBY J.E. (1991) Recurrence after radiotherapy for glottic cancer. Cancer 67, 577-584 CROILG.A., GERRITSEN G.J., TIWARI R.M. & SNOWG.B. (1989) Primary radiotherapy with surgery in reserve for advanced laryngeal carcinoma - results and complications. Eur. J . Surg. Oncol. 15, 350-356
RAPID COMMUNICATION
Salvage laryngectomy after radical radiotherapy for laryngeal carcinoma R.P.CRELLIN*, M . N . G A Z E * , AILEEN WHITEt, A . G . D . M A R A N f ’ & R.H.MACDOUGALL* * Department of Clinical Oncology, Western General Hospital, Edinburgh and tOtolaryngology Unit, Royal
Infirmary,
Edinburgh EH3 9EN, U K Accepted for publication 30 January 1992 CREI.LIN R . P . , G A Z E M . N . , W H I T E A . , M A R A N A . G . D . & M A C D O U G A L L R . H .
(1992) Clin. Otolaryngol. 17, 449-451
Salvage laryngectomy after radical radiotherapy for laryngeal cancer Of 376 patients who were treated by radical radiotherapy for squamous carcinoma of the larynx, 56 subsequently underwent total laryngectomy. Residual or recurrent tumour was identified in 43 of the resection specimens, and necrosis alone in 13 cases, although a positive biopsy had been obtained in 3 of these prior to salvage laryngectomy. No disease related factors such as site or stage of the original tumour, o r treatment related factors such as radiation type or dose, were found to be predictive of whether o r not tumour was present. The clinical opinion of an experienced surgeon was found to have a positive predictive value of 0.86 for the presence of tumour. The fistula rate of salvage laryngectomy, 15 out of 56, was similar to that of other series. The actuarial cause specific 5-year survival for patients with tumour was 0.589, and for patients with necrosis only was 0.923. Intercurrent, smoking related disease was the cause of death in 16 of the 33 patients who have died. Keywords
carcinoma qf the larynx
laryngectomy
Radical radiotherapy is the treatment of choice for early laryngeal carcinoma. Even in advanced disease, where the chance of successful control by irradiation is less, many clinicians still advocate primary radiotherapy to preserve speech.’ When there is residual o r recurrcnt disease after radiotherapy, surgery offers a second chance to cure the patient. Occasionally, no tumour is identified in the resected larynx. This may be anticipated if the indication for surgery was necrosis following irradiation, or a surprise finding when a recurrence was suspected. This paper is an audit of laryngectomy practice in Edinburgh following primary radical radiotherapy for laryngeal cancer. The aim was to identify the proportion of resection specimens which contained no tumour, and to relate this finding to the indications for, complications of and outcome after salvage laryngectomy. Correspondence: Dr R.P.Crellin, Department of Clinical Oncology. Western General Hospital. Edinburgh EH4 2XU, UK
morbidity
radiotherapJ.
recurrence
Patients and methods In the I I years from 1974 to 1984, 376 patients received radical radiotherapy for squamous carcinoma of the larynx at the Western General Hospital, Edinburgh.’ In the majority of patients. treatment was given by small, parallel opposed, wedged lateral fields covering only the larynx and any involved nodcs. Usually a dose of 52.5-55 Gy in 20 fractions over 4 weeks was prescribed using 4 MV X-rays. neutron^.^ Twenty-one patients were treated with Laryngectomy was subsequently performed in 56 patients (1 5%), including 6 treated with neutrons. Laryngectomy was required in 24 of 99 patients with supraglottic tumours (24%) but only 32 of 268 (12%) with glottic tumours. Laryngectomy was performed in 13 of 135 patients (10%) with T, tumours, 13 of 89 (15%) with T2 tumours, 23 of 99 (23%) with T3 tumours and 7 of 44 (16%) with T, tumours. The surgical and radiotherapeutic case notes of these 56 patients, 48 men and 8 women, were reviewed to find the indications for laryngectomy, the preoperative opinion about
449
450
R.P.Crellin et al.
the presence of tumour, the operative morbidity, the pathology of the resected larynx, and the outcome. Survival rates were calculated by the life-table method and compared using the logrank test.
Results PAI‘HOLOGY OF RESECTION SPECIMENS
Histological examination of the 56 resected larynges showed residual or recurrent tumour in 43 (77%) and no evidence of tumour in 13 (23%). Various factors have been assessed to see whether they were more likely to be related to either necrosis or tumour. The patient and disease related factors include sex, site, tumour stage and nodal status. The treatment related factors include the presence of a trachcostomy, radiation dose and radiation type. None of these factors, either disease or treatment related, was found to be related to the presence o r absence of tumour in the excised larynx. Of the 6 patients treated with neutrons, 4 had tumour in the resected larynx and two necrosis only. INVICATIOKS FOR LARYNGECTOMY
The principal symptoms prior to surgery were stridor in 12 patients, hoarseness without stridor in 37 patients, of whom 4 also had pain and 3 foetor, pain alone in 4 patients, and foetor alone in 3. There was no difference in the distribution of these symptoms between those who were found to have cancer, and those with no evidence of tumour. Biopsy was undertaken prior to laryngectomy in 46 patients. This showed the presence of residual or recurrent disease in 36 of the 43 cases found to have tumour in the resection specimen. Biopsy was negative in 5 and not performed in 2 patients in this group. The biopsy was positive for cancer in 3 of the 13 cases where subsequently no tumour was found in the resection specimen, negative in 2 and not performed in 8. The surgeon’s opinion prior to laryngectomy as to whether or not the larynx harboured residual or recurrent disease was correct in 42 of the 56 cases. In only 1 patient was tumour found in the resected larynx when necrosis alone was suspected, and in 7 patients necrosis without malignancy was found when turnour was suspected (in 3 cases because of a positive biopsy). Thus the sensitivity of the surgeon’s opinion that cancer was present was 0.98, but the specificity was only 0.46. The positive predictive value of the surgeon’s opinion was 0.86, the negative predictive value 0.86, and the likelihood ratio 1.12. OUTCOME OF SURGERY
The principal complication of surgery was fistula formation which occurred in 15 patients. The fistula was pharyngo-
cutaneous in 14 patients and pharyngo-tracheal in 1. Other complications noted were pharyngeal stricture in 6 patients, infection in 5, dehiscence in 3, arterial blowout in 2 and tracheal stenosis in one. The distribution of these complications did not differ significantly between those patients with tumour and those with necrosis only. Survival was significantly better for those who did not have cancer in the resected larynx. The actuarial cause specific survival at 5 years was 0.923 in this group, compared with 0.589 in those who had recurrent or residual malignancy. The magnitude of this difference was 0.334 (95% confidence intervals, 0.075-0.593). Among patients found to have tumour in the resected larynx there was n o difference in survival between those with residual disease and those in whom the disease recurred following a complete response to radiotherapy. Of the 43 patients operated on for recurrent or residual disease, only 12 patients are alive and well. Sixteen have died from this cancer and 2 had recurrent cancer when last seen. Thirteen have died from other causes (infection 5; myocardial infarction, 3; cerebrovascular disease, 2, and 3 other malignancy (in each case, lung cancer)). Of the 13 patients who had no evidence of tumour, 9 are alive and well. One died post-operatively, 2 died from a myocardial infarction and 1 died from carcinoma of the maxillary antrum.
Discussion Radical radiotherapy is the mainstay of treatment for laryngeal carcinoma, though it is recognized that surgical salvage of radiation failures may be required, especially if the disease is advanced at presentation. Diagnosis of residual or recurrent disease may, however, be difficult. While a biopsy may indicate recurrent disease, a negative biopsy result is not reassuring if the clinical picture is suspicious. Persistent endolaryngeal oedcma indicates recurrent tumour in about half, yet even multiple deep biopsies of oedematous laryngeal mucosa often fail to demonstrate this: and may precipitate frank necrosis. As recurrence frequently coexists with a biopsy showing necrosis only does not chondronecr~sis,~ exclude persistent malignancy. For these reasons, and because a painful, necrotic, oedematous larynx is of little value to the patient and should perhaps be removed anyway if it fails to improve with conservative therapy, a surgeon must be prepared to act on suspicion of recurrence, without absolute proof. In this series, a high index of suspicion was maintained, so that only 1 case of recurrence was thought to be necrosis only. However, the specificity of this suspicion was lower than its sensitivity, with 4 patients having cancerfree larynges removed on suspicion of recurrence, without prior histological verification. While occasionally limited surgery for recurrence may be possible,b total laryngectomy is the most frequently performed procedure. It is believed by some that a small recur-
Salvage laryngectomy
rence may be completely removed at biopsy, leading to a ‘biopsy cure’,’ as apparently happened in 3 cases in this series. Alternatively, small foci of tumour were missed by the pathologist during examination of the larynx. This is a more plausible explanation in initially advanced tumours where it would be naive to imagine that all malignancy except a small superficial area had been eradicated by irradiation. Nevertheless, patients in whom no tumour is found, even if in reality small islands of malignant cells do exist, have an excellent prognosis. The complications of salvage laryngectomy in the present series are typical, with a fistula rate of 27%, similar to the 32’- 37%’ reported by others. Fcw reported series of salvage laryngectomy acknowledge the proportion of larynges removed for necrosis in the absence of tumour, yet this is clearly important in the assessment of this procedure. Our figure of 23% compares with 12% reported from Liverpool,* and 3% from Toronto, Canada.’ Fortunately, survival for this group of patients is good, and quality of life is paradoxically often better when a painful, necrotic and barely functional larynx is removed. At 59%, the survival rate for those undergoing laryngectomy for recurrent or residual cancer is similar to the range of 39-56% reported by other^.^,^ The fact that about half the deaths in this group are not due to laryngeal cancer but t o other smoking related disease, indicates that crude rather than cause specific survival rates should be used in any costbenefit analysis of the treatment of laryngeal cancer. In summary, this audit of salvage laryngectomy after radical radiotherapy has shown that about onc-quarter of procedures are performed for necrosis alone, but that these patients have a significantly better prognosis than if cancer is present. No disease or treatment related factors appear to be predictive of whether a resected larynx will show necrosis
451
alone or tumour. The best predictor of recurrence is the experienced surgeon’s clinical opinion. The observed rate of complications and survival following salvage laryngectomy is similar to that seen in other centres.
Acknowledgements We wish to thank Miss G.R.Kerr for statistical advice.
References GAZE M.N., MACKILLOP, W.J., O’SULLIVAK B., GILBERT R., MAKI E. & LUNDCREN J. (1991) The management of laryngeal cancer: a comparison of practice in the United Kingdom and Canada. Clin. Oncol. 3, 296 (Abstract) HEYWOOD J.M., KERR G.R., ORR J.A., MACDOUCALL R.H. & MACKII.I.OP W.J. (1992) Primary radiotherapy for carcinoma of the larynx: the Edinburgh experience. Radiother. Oncol. (in press) MACDOUGALL R.H., ORRJ.A., KERRG.R. & DUNCAN W. (1 990) Fast neutron treatment for squamous cell carcinoma of the head and neck: final report of Edinburgh randomised trial. Er. Med. J . 301, 1241-1242 Fu K.K., WOODEHOUSE R.J., QUIVEYJ.M., PHILLIPS T.L. & DEW, H.H. (1982) The significance of laryngeal edema following radiotherapy of carcinoma of the vocal cord. Cancer 49, 655-658 KEBNEM., HARWDODA.R., BRYCED.P. & VAN NOSTRAND A.W.P. (1982) Histopathological study of radionecrosis in laryngeal carcinoma. Laryngoscope 92, 173-180 CROLLG.A., VAS DEK BROEKP., TIWARIR.M., MANNIJ . J . & SNOW G.B. (1985) Vertical partial laryngectomy for recurrent glottic carcinoma after irradiation. Head Neck Surg. 7 , 390-393 NICHOLSR.D., STINEP.H. & GREENAWALD K.J. (1980) Partial laryngectomy after radiation failure. Larvngoscope 90, 57 1-575 VIANIL., STELLP.M. & DALBY J.E. (1991) Recurrence after radiotherapy for glottic cancer. Cancer 67, 577-584 CROILG.A., GERRITSEN G.J., TIWARI R.M. & SNOWG.B. (1989) Primary radiotherapy with surgery in reserve for advanced laryngeal carcinoma - results and complications. Eur. J . Surg. Oncol. 15, 350-356
