Revisit of Primary Malignant Neoplasms of the Trachea: Clinical Characteristics and Survival Analysis

Jpn J Clin Oncol 1997;27(5)305 309 Revisit of Primary Malignant Neoplasms of the Trachea: Clinical Characteristics and Survival Analysis -, -, - - 1 Chest Department and 2 Section of Thoracic Surgery, Veterans General Hospital Taipei and School of Medicine, National Yang-Ming University, Taipei, Taiwan A clinical review with an analysis of prognostic factors, including clinical characteristics, histological classifications, presenting symptoms/signs and treatment modalities, was conducted in 67 patients with primary malignant neoplasms of the trachea who were seen at the Veterans General Hospital Taipei between 1979 and 1994. The incidence of tracheal cancer was 140 times less than lung cancer during this period. Delayed diagnosis of more than 6 months after onset of symptoms occurred in one-third of the patients owing to lack of specific symptoms/signs indicative of tracheal disease. Epidermoid carcinoma was the most frequent histological type encountered and accounted for half of the cases. Surgical resection was the first choice of treatment for all patients if the disease was locally confined, except for small cell carcinoma and lymphoma. Radiotherapy was given if the patient was not suitable for surgery. Single and multivariate analyses showed that clinical symptoms and signs were not related to prognosis, except for general malaise and acute respiratory failure. Patients suffering from adenoid cystic carcinoma and mucoepidermoid carcinoma had a better prognosis than other histological diagnoses. Whether the patient received radiotherapy or not proved to be a significant prognostic factor in the patients. Patients with tracheal cancers had a poorer prognosis than those with lung cancer. Key words: trachea neoplasm adenoid cystic carcinoma epidermoid carcinoma survival analysis INTRODUCTION Primary malignant neoplasms of the trachea are rare, accounting for 0.1% of all malignancies (1). Lung cancers were 180 times more frequently diagnosed than tracheal malignancies (2). The correct diagnosis is seldom made initially and a delay in diagnosis is common. This delay in diagnosis is most likely due to a lack of early symptoms or non-specific symptoms (1 4). The rare reports in the literature not only show the rarity of tracheal carcinoma, but also indicate the limited experience with this malignancy at any given institution. Detailed clinical manifestations and histological classifications were reported about two decades ago (1 3). The treatment modalities for tracheal tumors have also been discussed but with no consensus (5 11). Univariate survival analysis had been tried before (3,6). However, no well accepted statistical methods have been used. Also no multivariate analyses have been applied before. The present study was undertaken to investigate the Received March 3, 1997; accepted May 14, 1997 For reprints and all correspondence: Yuh-Min Chen, Chest Department, VGH-Taipei, Shih-pai road, Taipei, Taiwan clinical features, management and survival analysis of our patients using uni- and multivariate analyses. MATERIALS AND METHODS From 1979 to 1994, there were 67 cases of histocytologically proven primary malignant neoplasms of the trachea in the Veterans General Hospital Taipei, Taiwan. All tumors arising outside the trachea, including direct extension from adjacent organs or distant metastases to the trachea, were excluded. Primary bronchogenic carcinoma showing tracheal involvement by direct extension was also excluded. The clinical characteristics of the patients are shown in Table 1. There were 55 men and 12 women. The mean age was 61 years with a range between 25 and 85 years. Forty-eight of 67 patients smoked. All patients received fiber-optic bronchoscopic examinations, chest radiography and CT scans. Whole-body bone scans and brain scans were performed if clinically indicated. Surgical resection was the first choice of treatment if the disease was locally confined. Post-operative radiotherapy was optional. Radiotherapy was the primary treatment modality if the patient could not receive surgical treatment. Chemotherapy was the primary treatment modality in patients suffering from small cell carcinoma and malignant lymphoma. Chemotherapy was an optional treatment after initial treatment with surgery and/or

306 Malignant tracheal tumors radiotherapy. Drugs of choice were usually cyclophosphamide, epirubicin and cisplatin for cancers other than small cell carcinoma and lymphoma. Cisplatin and etoposide were given for patients with small cell carcinoma. Laser removal of the intratracheal tumor was the palliative treatment of choice. Mechanical ventilation was given to four patients who initially presented with acute respiratory failure. Patient survival time was calculated from the initial diagnosis of the primary disease to their death or last follow-up. Univariate survival analysis was based on the Kaplan Meier estimate and log-rank test. Multivariate analysis was performed with the Cox regression method. RESULTS CLINICAL CHARACTERISTICS Epidermoid carcinoma was the most frequent histological type of malignant tracheal tumor found and accounted for 52.2% of cases, followed by adenocarcinoma (14.9%), poorly differentiated carcinoma (10.4%), adenoid cystic carcinoma (7.5%) and others (15%). The mean age of all patients was 61 years. However, the mean age was only 29 and 41 years in mucoepidermoid carcinoma and adenoid cystic carcinoma, respectively. Adenocarcinoma and poorly differentiated carcinoma tended to occur in older patients. Delayed diagnosis of more than 6 months after symptoms onset occurred in one-third of patients. Median survival for all patients after diagnosis of malignant tracheal tumors was 6 months (Fig. 1). Median survival varied widely with different histological types of tracheal tumors. Those who suffered from adenoid cystic carcinoma or mucoepidermoid carcinoma had significantly better survival than other types of histology. Those who suffered from epidermoid carcinoma, adenocarcinoma or carcinosarcoma had worse survival than other types of cancer statistically. The median survival time was longest in the lymphoma group and shortest in the carcinosarcoma group (Table 1). The most frequent presenting symptom/sign was cough (71.6%), followed by dyspnea (65.7%), hemoptysis (38.8%), stridor (38.8%), hoarseness (31.3%) and others (Table 2). Most Figure 1. Kaplan Meier survival curve for 67 patients with malignant tracheal tumors diagnosed between 1979 and 1994. The median survival time was 6 months. of the symptoms were non-specific. Specific symptom/sign indicating structural lesions of the trachea or nearby structures included stridor (38.8%), superior vena cava (SVC) syndrome (13.4%) and dysphagia (10.4%). ANATOMICAL LOCATION AND STAGING More than half of the cases of epidermoid carcinoma, adenocarcinoma and small cell carcinoma had their primary location in the lower trachea. whereas poorly differentiated carcinoma and adenoid cystic carcinoma had their lesions mainly in the upper trachea. Tumor origin from the anterior wall and lateral wall of the trachea was noted in 40 of 67 patients (59.7%). Twenty cases had their tumor origin from the posterior wall of the trachea (29.9%). Only adenoid cystic carcinoma had no tumor origin from the posterior tracheal wall. Almost equal numbers of patients had tumor obstruction of more than 50% of the tracheal lumen. Cervical lymph node involvement occurred in 22.4% and distant metastases occurred in 14.9% of patients. Cervical lymph nodes metastases could be found in all types of tracheal tumors. Distant metastases were not found in patients suffering from adenoid cystic carcinoma (Table 3). Table 1. Clinical characteristics of 67 patients with malignant primary tracheal neoplasms Patient Mean age Sex (M/F) Smoking Delayed Median survival p number (%) (range) (years) (yes/no) diagnosis (%)* (range) (months) Epidermoid carcinoma 35 (52.2) 63 (43 79) 30/5 27/8 12 (34.3) 5 (0 165) 0.044 Adenocarcinoma 10 (14.9) 66 (58 76) 8/2 7/3 2 (20) 3.5 (1 22) 0.028 Poorly differentiated carcinoma 7 (10.4) 69 (55 80) 6/1 4/3 3 (42.9) 5.5 (0 42) 0.614 Adenoid cystic carcinoma 5 (7.5) 41 (34 49) 3/2 5/0 4 (80) 64 (40 120 ) 0.011 Small cell carcinoma 4 (6.0) 60 (37 85) 3/1 2/2 0 13.5 (3 57) 0.99 Mucoepidermoid carcinoma 3 (4.5) 29 (25 34) 2/1 1/2 2 (67) 74 (11 140 ) 0.011 Lymphoma 2 (3.0) 62 (58,66) 2/0 1/1 0 76.5 (56 97) 0.311 Carcinosarcoma 1 (1.5) 64 1/0 1/0 0 <1 0.007 Total 67 61 55/12 48/19 23 (34.3) 6 (0 140 ) *Delayed diagnosis meant onset of symptoms to correct diagnosis more than 6 months. Sixteen of 67 patients were initially mis-diagnosed as suffering from COPD or asthma. When compared with other histological types of primary malignant tracheal neoplasms. Patient was still alive.

Nucleic Acids Jpn Research, J Clin 1994, Oncol Vol. 1997;27(5) 22, No. 1 307 Table 2. Presenting symptoms/signs of 67 patients with malignant primary tracheal neoplasms Number of patients Epidermoid Adenocarcinoma Poorly differentiated Adenoid cystic Small cell Others* Total carcinoma carcinoma carcinoma carcinoma (% of total cases) Cough 26 5 5 3 3 6 48 (71.6) Dyspnea 25 4 3 5 2 5 44 (65.7) Hemoptysis 13 5 3 1 1 3 26 (38.8) Stridor 14 1 3 4 1 3 26 (38.8) Hoarseness 14 3 2 1 1 0 21 (31.3) Body weight loss 4 6 3 1 0 1 15 (22.4) Wheezing 9 2 0 2 0 0 13 (19.4) SVC syndrome 6 0 2 0 1 0 9 (13.4) Chest pain 2 2 2 1 0 1 8 (11.9) Dysphagia 6 0 1 0 0 0 7 (10.4) General malaise 3 2 1 0 0 0 6 (9.0) Pneumonia 3 1 0 0 0 1 5 (7.5) Acute respiratory failure 3 0 0 0 0 1 4 (6.0) Loss of consciousness 2 1 0 0 0 0 3 (4.5) *Including 3 mucoepidermoid carcinoma, 2 lymphoma and 1 carcinosarcoma. Table 3. Tumor location and staging Number of patients Epidermoid Adenocarcinoma Poorly differentiated Adenoid cystic Small cell Others* Total carcinoma carcinoma carcinoma carcinoma (% of total cases) Longitudinal aspect Upper trachea 16 3 4 3 0 3 29 (43.3) Lower trachea 19 7 3 2 4 3 38 (56.7) Coronal aspect Anterior wall 4 2 3 2 1 0 12 (17.9) Lateral wall 14 5 3 3 2 1 28 (41.8) Posterior wall 12 3 0 0 0 5 28 (41.8) Circumference 5 0 1 0 1 0 7 (10.4) Tracheal obstruction 50% 16 7 4 2 3 2 34 (50.7) >50% 19 3 3 3 1 4 33 (49.3) Cervical lymph node involvement 7 2 2 1 2 1 15 (22.4) Distant metastases 4 4 1 0 1 0 10 (14.9) *Including 3 mucoepidermoid carcinoma, 2 lymphoma and 1 carcinosarcoma. Table 4. Treatment modalities of 67 patients with malignant primary tracheal neoplasms* Surgery Radiotherapy Chemotherapy S/T + R/T R/T +C/T S/T + R/T + C/T Supportive care (S/T) (R/T) (C/T) alone Epidermoid carcinoma 1 19 0 3 8 0 4 Adenocarcinoma 0 6 1 0 3 0 0 Poorly differentiated carcinoma 0 6 0 0 1 0 0 Adenoid cystic carcinoma 1 1 0 3 0 0 0 Small cell carcinoma 0 1 1 0 2 0 0 Mucoepidermoid carcinoma 0 0 0 2 0 1 0 Lymphoma 0 0 0 0 2 0 0 Carcinosarcoma 1 0 0 0 0 0 0 Total (%) 3 (4.5) 33 (49.3) 2 (3.0) 8 (11.9) 16 (23.9) 1 (1.5) 4 (6.0) *Ten of 67 patients received laser palliative treatment during their disease course. Patients disease was too terminal to receive specific treatment.

308 Malignant tracheal tumors TREATMENT MODALITIES Radiotherapy was the main treatment modality in our patients. Radiotherapy alone was given to 49.3% patients. Radiotherapy in combination with surgery and/or chemotherapy was given to another 37.3% patients. Only 13.5% of patients did not receive radiotherapy. Chemotherapy in combination with radiotherapy was mainly used in patients with small cell carcinoma or malignant lymphoma. However, there were 8 out of 35 patients of epidermoid carcinoma and 3 out of 10 patients of adenocarcinoma receiving chemotherapy after completion of radiotherapy. Four patients received no specific anti-tumor treatment because their disease was found too late to receive specific treatment (Table 4). SURVIVAL ANALYSIS Univariate and multivariate analyses were performed. Factors analyzed included clinical characteristics (age, sex, smoking, delayed diagnosis), presenting symptoms and signs (cough, dyspnea, hemoptysis, stridor, hoarseness, body weight loss, wheezing, SVC syndrome, chest pain, dysphagia, general malaise, pneumonia, acute respiratory failure, loss of consciousness), tumor location and staging and treatment modality. The significant positive predictor included age 65, histology, delayed diagnosis, surgery and radiotherapy in univariate analysis. The significant negative predictor included distant metastases, loss of consciousness, general malaise and acute respiratory failure in univariate analysis. Multivariate analysis showed that age, dysphagia, general malaise, trachea obstruction, histology, distant metastases, radiotherapy and acute respiratory failure were significant prognostic factors for survival. We then compared the survival times with these tracheal cancers with those with lung cancers. During the same time period, there were 9389 lung cancers diagnosed in our hospital. The median survival time with lung cancer was longer than that with tracheal tumors (7 vs 6 months, p = 0.009). When we compared tracheal tumors without distant metastases (57 cases) with stage IIIb lung cancer (792 cases), the median survival time was longer with tracheal tumor (8 months) than with stage IIIb lung cancer (7 months) (p = 0.0001, Fig. 2). However, when cases with adenoid cystic carcinoma, small cell carcinoma, mucoepidermoid carcinoma and lymphoma were excluded from the analysis, the median survival time was only 5 months with tracheal tumors (p = 0.14). When we compared tracheal tumors with distant metastases (10 cases) with stage IV lung cancer (3329 cases), the median survival time was longer with lung cancer (3 vs 5 months, p = 0.002, Fig. 3). DISCUSSION Epidermoid carcinoma was the most common type of malignant tracheal tumor, representing more than half of all cases in our patients. Most cases of epidermoid carcinoma occurred in smokers over 40 years of age. Depending on institutions, either epidermoid carcinoma or adenoid cystic carcinoma was the predominant type of malignant tracheal tumor (2,3,5,9,11,12). Histological classification was a significant prognostic factor in Figure 2. Kaplan Meier survival curves for tracheal tumors without distant metastases (57 cases, median survival 8 months) and stage IIIb lung cancer (792 cases, median survival 7 months). p = 0.0001 in log-rank test. However, when adenoid cystic carcinoma, small cell carcinoma, mucoepidermoid carcinoma and lymphoma were excluded from the analysis, the median survival time was only 5 months in patients with tracheal tumors (p = 0.14 in log-rank test). Figure 3. Kaplan Meier survival curves for tracheal tumors with distant metastases (10 cases, median survival 3 months) and stage IV lung cancer (3329 cases, median survival 5 months). p = 0.002 in log-rank test. both univariate and multivariate analyses. Mucoepidermoid carcinoma and adenoid cystic carcinoma had a statistically better prognosis than other types of tracheal cancer. Other types of cancer which also had a longer median survival included lymphoma and small cell carcinoma. However, it was not statistically significant. From our data, the age at diagnosis of adenoid cystic carcinoma and mucoepidermoid carcinoma was younger than for epidermoid carcinoma, adenocarcinoma and poorly differentiated carcinoma. Those tracheal tumors which had a better prognosis tended to occur in younger patients (Table 1). Delayed diagnosis was also most frequently found in younger patients, such as adenoid cystic carcinoma and mucoepidermoid carcinoma. The initial impression in this age group was usually adult onset asthma, when their chest roentgenographic finding was normal

Nucleic Acids Jpn Research, J Clin 1994, Oncol Vol. 1997;27(5) 22, No. 1 309 and the possibility of cancer had not been considered. For those patients who had delayed diagnosis, most of them suffered from indolent cancers rather than more aggressive cancers. This could partially explain why those patients who had a delayed diagnosis had a better prognosis than those who did not have delayed diagnosis in univariate analysis. Pulmonary symptoms, such as cough, dyspnea and hemoptysis, were not related to prognosis. SVC syndrome and hoarseness both meant that patients disease was at least far advanced and inoperable. However, they were not significant in relation to survival. Loss of consciousness was a significant prognostic factor in univariate analysis and dysphagia was a significant factor in multivariate analysis. Symptoms of general malaise usually meant that the patients tracheal lumen was obstructed to such a significant extent that patients felt discomfort and dyspnea when performing their usual daily activities. The worst condition of tracheal obstruction was acute respiratory failure and this reflected a poor prognosis of patients both in uni- and multivariate analyses. Locations of tumor, longitudinal or coronal aspect, was not related to survival. Cervical lymph node involvement occurred in 22.4% of patients and it was not a significant prognostic factor. Distant metastases occurred in 14.9% of patients and was a poor prognostic factor. Radiotherapy was the main treatment modality in our patients, reflecting that the majority of patients suffered from locally far-advanced disease or distant metastases beyond the feasibility of surgical curative treatment. Regarding treatment modalities, patients receiving surgical treatment or radiotherapy had longer survival times in univariate analysis. However, it was only radiotherapy that was significant in multivariate analysis. When comparing malignant tracheal tumors with lung cancer during the same time period, we found that lung cancer was 140 times more common than tracheal cancer (9389 vs 67). This was slightly less than in previous reports (1). The survival was poor in tracheal cancers compared with lung cancers, however, when we considered that tracheal involvement was considered to be T4 disease (at least stage IIIb) in lung cancer. We had to compare the survival of patients with tracheal tumors (without distant metastases) with those with stage IIIb lung cancer and patients with tracheal tumors (with distant metastases) with those with stage IV lung cancer. Then, survival was better for tracheal tumors without distant metastases than for stage IIIb lung cancer. The survival time was still worse with tracheal tumors with distant metastases than with stage IV lung cancers. In conclusion, delayed diagnosis frequently occurred in patients with tracheal cancers. Patients suffering from adenoid cystic carcinoma and mucoepidermoid carcinoma had good prognosis. Patients with epidermoid carcinoma and adenocarcinoma had poor prognosis. Clinical symptoms and signs were not related to prognosis, except general malaise and acute respiratory failure. Whether or not radiotherapy was given proved a significant prognostic factor in our patients. Patients with tracheal cancers had a poorer prognosis compared with lung cancer. Acknowledgment We thank Dr Jacqueline Ming Liu for correcting the English. References 1. Ranke EJ, Presley SS, Holinger PH. Tracheogenic carcinoma. J Am Med Assoc 1962;182:121 4. 2. Houston HE, Payne WS, Harrison EG Jr, Olsen AM. Primary cancers of the trachea. Arch Surg 1969;99:132 40. 3. Hajdu SI, Huvos AG, Goodner JT, Foote FW Jr, Beattie EJ Jr. Carcinoma of the trachea clinicopathologic study of 41 cases. Cancer 1970;25:1448 56. 4. Adachi MM, Pamies RJ. Carcinoma of the trachea: a hidden tumor. Hosp Pract (Off Ed) 1993;28:81 4. 5. Grillo HC, Mathisen W. Primary tracheal tumors: treatment and results. Ann Thorac Surg 1990;49:69 77. 6. Grillo HC, Mathisen DJ, Wain JC. Management of tumors of the trachea. Oncology 1992;6:61 7. 7. Emami B. The Grillo, et al. article reviewed. Oncology 1992;6:68 70. 8. Roth JA. The Grillo, et al. article reviewed. Oncology 1992;6:72. 9. Fields JN, Rigaud G, Emami BN. Primary tumors of the trachea: results of radiation therapy. Cancer 1989;63:2429 33. 10. Cheung AYC. Radiotherapy for primary carcinoma of the trachea. Radiother Oncol 1989;14:279 85. 11. Mathisen DJ. Surgical management of tracheobronchial disease. Clin Chest Med 1992;13:151 71. 12. Zimmer W, Deluca SA. Primary tracheal neoplasm: recognition, diagnosis and evaluation. Am Fam Physician 1992;45:2651 7.