Laryngeal cancer in Colorado

LARYNGEAL CANCER IN COLORADO James A. Harris, MD, MBA, Arlen D. Meyers, MD, MBA, and Chris Smith, MD

Previous studies of laryngeal cancer have shown a decrease in the male-to-female ratio and that the sex ratio in glottic tumors is higher than that for other laryngeal sites. The purpose of this study was to characterize and identify changes in the demographics, laryngeal site predilection, geographic distribution, trends in tumor stage at diagnosis, and surgical caseload distribution in Colorado. The Colorado Central Cancer Registry for the years 1979 to 1990 was reviewed for cases of laryngeal cancer. The data were analyzed using chi-square and gradientin-proportions methods. One-thousand two-hundred sixty-five (1265) cases were identified with a male-to-female ratio of 4.3: 1. There was a significant trend of an increasing proportion of cases attributable to women. Glottic carcinoma accounted for 50.1% of cases, with a male-to-female ratio of 7.6:l. The Denver metropolitan area accounted for 57% of all cases, which was not disproportionate to the population. From 1981 to 1990 there was a significant decrease in the proportion of stage I cases and an increase in stage II cases. The teaching hospitals associated with the University of Colorado were responsible for performing 44.5% of all surgery for laryngeal cancer. The results indicate that laryngeal cancer in Colorado shares similar epidemiologic characteristics to those reported in other studies. Of most concern is the increasing proportion of women with laryngeal cancer and the decrease in cases diagnosed at stage I. HEAD & NECK 1993;15:398-404 0 1993 John Wiley & Sons, Inc.

From the Department of OtolaryngologyiHead and Neck Surgery, University of Colorado Health Sciences Center, Denver, Colorado. Acknowledgments: The authors thank Jack L. Finch, MS, Dennis L. Lezotte. PhD. and Anna E. Bar6n PhD for their assistance in this project Address reprint requests to Dr. Meyers, MD, Department of Otolaryngology1Head and Neck Surgery, Box 6-210, UCHSC, 4200 East Ninth Avenue, Denver, CO 80262. CCC 0148-64031931050398-07 0 1993 John Wiley & Sons, Inc

398

Laryngeal Cancer in Colorado

Laryngeal cancer accounts for approximately 1.3% of all new reported cancer cases nationwide.' Previous studies have shown that the main risk factors for the development of laryngeal cancer are tobacco and Other factors that may increase risk include wood dust, metal dust, hair dyes, asbestos, and A recent study by DeRienzo et a1.,6 on the epidemiologic aspects of laryngeal cancer, has suggested that over the past 30 years the observed increased incidence of laryngeal cancer in women in the United States is most likely secondary to an increase in the number of women smokers. Another study has shown that the majority of laryngeal cancers are located in the glottis and that the male-to-female ratio of glottic tumors is significantly higher than the ratio at other sites. The purpose of this study was to present the data contained in the Colorado Central Cancer Registry for laryngeal cancer cases for the years 1979 to 1990, and to compare the data to those reported in the studies mentioned above. Additionally, the cases were analyzed for geographic distribution within the state to identify any abnormally high incidence areas, and to determine the trends, if any, in the stage of the tumor at the time of diagnosis. As the primary objective of this study is to present the epidemiology of laryngeal cancer in Colorado, no data will be presented on survival and how this is impacted by different treatment modalities. This is the topic of a study in progress. As more attention is directed toward containment of medical costs, the issue of centralization

HEAD & NECK

Septernber/October 1993

the ratio was not statistically significant ( p > .05). However, there was a statistically significant ( p < .05) trend of a n increasing proportion of women diagnosed with laryngeal cancer (Figure 2). The Registry contains complete data for the years 1979 to 1990 for the Denver metropolitan area (Adams, Arapahoe, Boulder, Denver, Jefferson counties); therefore, only these counties were used to calculate the yearly incidence for the study period, as presented in Figure 3. The Cancer Registry was felt to contain complete statewide data for the years 1988 and 1989. The incidence of laryngeal cancer statewide for each of those years was 3.2 and 3.6 per 100,000 population, respectively. The Denver metropolitan area accounted for 54% to 55% of the Colorado population during the study period. The total number of patients diagnosed with laryngeal cancer from the metropolitan area represented 57.3% of the entire study population. This did not represent a significant difference (p > .05) from what would be expected based on the population in the Denver metropolitan area. Ten counties, including five in the Denver metropolitan area, which had more than 40 cases over the years 1979 to 1990, accounted for 1077 cases or 85% of the laryngeal cancer cases identified. Sixty percent of all the cases occurred in subjects greater than 60 years old. Table 1 illustrates the number of cases in each age category, as well as the male-to-female ratios. Chi-square analysis revealed that the proportion of males greater than 60 years was significantly different

of complex or highly specialized procedures is raised. Centralization of certain services could potentially offer benefits of decreased cost, improved ancillary services, and, of course, expertise by virtue of training or experience. In an effort to determine the extent of of centralization, which already exists for major laryngeal oncologic surgery, an evaluation of the types, number, and locations of surgical treatments for laryngeal cancer was performed. MATERIALS AND METHODS

The Colorado Central Cancer Registry Data Base for the years 1979 to 1990 was reviewed for cases of laryngeal cancer. Data concerning subject age, sex, place of residence, tumor histology, laryngeal site of the tumor, American Joint Committee on Cancer (AJC) stage a t diagnosis, and surgery performed were reviewed and analyzed. Differences in proportions and trends were analyzed using chi-square and gradient-inproportions methods, respectively. RESULTS

There were 1265 cases of laryngeal cancer identified for the years 1979 to 1990. The total was comprised of 1025 men and 238 women; two cases were not specified for gender. Figure 1 illustrates the number of cases of laryngeal cancer by year and sex, showing the gradual increase in the number of females over the study period. The overall male-to-female ratio was 4.3 : 1. An arbitrary division of the data into two 6-year periods yielded male-to-female ratios of 4.7: 1 and 3 9 : 1, respectively. The difference in

I

'

Limb. I

2o 0

I

/

I

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

Y m

FIGURE 1. Total number of cases of laryngeal cancer each year in Colorado, broken down by sex.

Laryngeal Cancer in Colorado

HEAD & NECK

September/October 1993

399

30 25 . 20

.

Percent 15 -

10 5 1 0 1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

Yeu

FIGURE 2. Percentage of all cases of laryngeal cancer attributable to females each year.

( p < .05) from the proportion in the other age categories. The histologic diagnosis was squamous cell carcinoma in 93.8%of all cases. Verrucous carcinoma accounted for 1.2% of the cases, while cases of adenocarcinoma, mucoepidermoid,adenosquamous, chondroma, lymphoma, and spindle cell carcinoma comprised 2.1%. No histologic type was given for 2.9%of the cases. Examination of laryngeal cancer sites showed a predilection for the glottis, which accounted for one half of all the cases. Table 2 provides the frequency of cases at each laryngeal site and the male-to-female ratios. Trend analysis showed no significant change in the proportion of glottic carcinoma or the ratio of glottic-to-supraglottic carcinoma over the study period. However, the analysis of sex proportions revealed a statistically significant ( p < .05) increase in the proportion of glottic cancers attributable to women. Figure 4 illustrates this trend.

When the laryngeal sites were examined by the age categories, “40 years or younger” and “greater than 40 years,” there did not appear to be any major difference in site distribution. In patients less than 40 years of age, 48.3% of the cancer cases were glottic, as compared with 49.3% in the other age category. The former age group had 33.33% of the cases in the supraglottis, while the latter had 32%. When examining the glottic-to-supraglottic ratio in each of the age categories presented in Table 1, there were no statistical differences ( p > .05) between age groups. The AJC stage at diagnosis was recorded in 439 of the 1078 total cases from 1981 to 1990. The staging was determined clinically, surgically, or pathologically. The percentage of cases diagnosed in each stage was as follows: 42.1% stage I, 19.4% stage 11, 19.4% stage 111, 18.7% stage IV, and 0.4% as carcinoma in situ. Figure 5 illustrates the trend in percentage of cases diag-

4.5 4 3.5

3 Number 2.5 2 1.5 1

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

Year

FIGURE 3. Overall incidence of laryngeal cancer in the Denver metropolitan area per 100,000 population.

400

Laryngeal Cancer in Colorado

HEAD & NECK

SeptemberiOctober 1993

Table 2. Larvnaeal sites-case

Table 1. Age categorization of all cases of laryngeal cancer No of cases

540 41 --50 51 - 6 0 61 - 70 71 - 80 81 -90 291 To!a!

total

Male:female

24 104 378 459 232 64 4

1.9 8.2 29.9 36.3 18.3 5.1 0. 3

1265

100.0

Glottis Supraglottis Subglottis Cartilage Other

3.8: 1 3.2: 1 3.4: 1 4.6: 1 73:l 4.8: 1 3.0: 1

Total

. _

30

'

0

.

1979

.

---

1980

-

.

.

.- - ..

1982

634 41 3 17 7 194

50 1 32 7 1.3 06 15 3

1265

1co.c

Male: female

76:l 25:l 32:l 08.1

4 1:l

DISCUSSION

The Colorado Central Cancer Registry is the central source for data on cancer in Colorado and has been accumulating data since 1979. Since that time, approximately 90% of all cancer cases diagnosed in Colorado and 100% of the cases in the Denver metropolitan area have been recorded. In 1989, laryngeal cancer was estimated to have been responsible for the loss of 283 years of potential life in Colorado.' This study repre-

._

__

-.

.

. ...

.

..

.

.

+ - . - - . +

1981

% of total

laryngectomy, total laryngectomy without lymph node dissection, total laryngectomy with lymph node dissection, laryngectomy not specified, and surgery of regional or distant sites- nodes only. The four teaching hospitals associated with the University of Colorado Health Sciences Center (Denver General Hospital, Denver Veterans Administration Hospital, University Hospital, and Fitzsimmons Army Medical Center) accounted for 445, or 44.596, of the surgeries performed. Table 3 provides the number of each type of surgery performed.

nosed as stage I and stage I1 by year. Trend analysis revealed a significant ( p < .05) decrease in the proportion of sitage I and a n increase in stage I1 cancers during those years. There was no significant change in the proportion of cases diagnosed in stage I11 or stage IV from 1981 to 1990. Of the 297 cases which were staged clinically, there were 47.8% stage I, 20.2% stage 11, 17.2% stage 111, 13.1% stage IV, and 1.6%carcinoma in situ cases. Trend analysis revealed a statistically significant ( p < .05) increase in the proportion of stage I1 cases identified clinically over the study period. The 49 Colorado hospitals which report to the registry were responsible for 999 surgeries performed for laryngeal cancer during the years 1980 to 1990. These surgeries were performed as primary therapy, in conjunction with other treatment modalities, or for salvage. These cancerdirected surgeries included local resection or laser ablation with pathologic specimen, partial

25

No. of cases

Site

% of

distribution

--

1983

+-

/ - -

1984

1985

4-

1986

---I

1 -

1987

1988

1989

1990

Year

FIGURE 4. Percentage of cases of glottic cancer attributable to women each year.

Laryngeal Canc:er in C:olorado

HEAD & NECK

SeptemberiOctober 1993

401

70

I

I

60 50

-

40

-Stage

Percent

1 %

Stage 2 %

30

20 10

0 1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

Year

FIGURE 5. Percentage of cases diagnosed as AJC stages I and II by year.

sents the first review of this type on laryngeal cancer from data contained in the Colorado Central Cancer Registry. The incidence of laryngeal cancer in Colorado was found to be approximately four cases per 100,000 population. This is similar to that reported by the National Cancer Institute, 4.5 cases per 100,000 p o p ~ l a t i o n .DeRienzo ~ et al. examined the incidence of laryngeal cancer in Texas women over two 15-year periods and noted a decline in the male-to-female ratio of 5.6: 1 to 4.5: 1, indicating an increased incidence among women. Although our study did not examine 30 years of data, two 6-year periods were examined which revealed male-to-female ratios of 4.7 : 1 and 3.9:1, respectively. The difference was not statistically significant; however, trend analysis confirmed a statistically significant increase in the percentage of women diagnosed with laryngeal cancer. The incidence data for the Denver Metropolitan area, from the Colorado Department of Health,' presented in Figure 6, shows a

Table 3. Cases performed by teaching hospitals for the years 1980 to 1990. Type of surgery

No of cases

No by teaching

% By teaching

26 8 36 2 65 5 61 5 37 7 37 5 44.5

LSEilaser

231

PL 2 LND

152

TL - LND TL + LND Laryngectomy NS Regional or distal LND

55 338 191 32

62 55 36 208 72 12

Total

999

445

LSE, local surgical excision, PL, partial laryngectomy, LND, lymph node dissection, JL, total laryngectomy, NS, not specihed

402

Laryngeal Cancer in Colorado

gradual increase from the years 1979 to 1988. Over the study period, there has been essentially no change in the sex ratio for the entire population of Colorado. It is felt that the increased incidence of laryngeal cancer among women is either due to the increased proportion of women who smoke or a difference in rates of decline in smoking between men and women.6 Laryngeal cancer typically affects individuals in the sixth or seventh decade of life. This study was in agreement with a previous report which suggested that more than 60% of the cases occur in patients greater than 60 years of age." There was also a statistically significant increase in the male-to-female ratio in the group greater than 60 years old. The possibility that the increase in this ratio was due to an increase in the glottic-tosupraglottic ratio (glottic cancer is associated with a significantly higher male-to-female ratio than supraglottic cancer; see Table 2) in patients over the age of 60 was not supported by statistical analysis. Approximately 2% of the 1265 cases w x e in individuals 40 years or younger, similar to the finding of Shvero et al. The AJC stage at diagnosis, recorded in eight individuals in this age category, was stage I for four cases. Because of the small number of cases in this group, no meaningful statement about trends in stage at diagnosis for this age category can be made. Although studies are not in total agreement, most have found that the glottic region was the most frequently involved site and had a significant male predominan~e.~.l'l3 The present investigation resulted in similar findings, with glottic cancer accounting for 50.1% of the cases and a male-to-female ratio of 7.6:l. This ratio was significantly different from other laryngeal

HEAD & NECK

SeptemberiOctober 1993

-

09

__

~

___

-

-

07 -

t

-

~

~

----+--.+--

sites which had ratios ranging from 0.75: 1 to 4.1 : 1. Yang et al.7 proposed that the differences in site predilection might be due to factors related to differences in “anatomy and/or physiology of the vocal cords (glottis) in the two sexes that either enhances accumulation of the carcinogens in those of males, or facilitates their removal in females.” They did not find significant differences in alcohol consumption or tobacco use that could account for the difference. A study performed by the Alcohol and Drug Abuse Division of the Office of Data Analysis and Evaluation of Colorado14 showed that, in 1985, 68% of the state population used alcohol and 9.7% were “problem” drinkers (based on use pattern, physiologic and psychologic dependence, and social disruption). Of these problem drinkers, 80.2% were men and 19.8% were women. Perhaps differences in alcohol use may account for some of the discrepancy in sex proportions in laryngeal cancer. The statewide distribution of laryngeal cancer did not appear to have any geographic predilection that was out of proportion to regional population differences. Accordingly, the Denver metropolitan area, which comprises 55% of the state population, accounted for 57% of all cases reported. There is little probability that natural, industrial, or individual factors are significantly influencing regional incidence rates. Because surviival rates are dependent upon early detection iind treatment, the significant trend of a decrease in stage I cancers and an increase in stage [I is concerning. However, because only 41%o f the cases had a stage recorded, the possibility of a recording bias toward more

Laryngeal Cancer in Colorado

-

_

-

-

advanced stages is present; therefore, inferences from the data may be less reliable. In addition, the AJC staging system is subject to errors in interpretation and/or clinical judgment.13 In this study, the AJC stage, as reported in the data base, was from clinical, pathologic, or surgical staging procedures. This can result in a comparison of dissimilar entities and can introduce errors into the analyses. However, even when the cases that were staged clinically were analyzed separately from those staged surgically and pathologically, there was still a significant trend of an increasing proportion of cases diagnosed at stage 11. This trend may be due to delays in patients seeking medical attention or delays in physician referrals to appropriate specialists. Another possibility is that improved examination techniques are allowing better evaluation of tumor site involvement and, therefore, more accurate staging. It is doubtful that the trend could be due to shifts in proportions of glottic and supraglottic carcinomas, as there has been no significant changes or trends in their relative proportions for the study period. It was not surprising to find that the four hospitals associated with the University of Colorado performed almost half of the surgeries for laryngeal cancer. Except for Denver General Hospital, the other three are major state referral centers, and would be expected to perform a significant portion of the surgery performed in the state, as well as a higher proportion of more advanced procedures versus less complex ones. Otolaryngologists associated with the University of Colorado teaching hospitals have comprised approximately 10% of the state’s total number of

HEAD & NECK

SeptemberiOctober 1993

403

otolaryngologists since the early 1980s. Although otolaryngologists are clearly not the only physicians who perform head and neck oncologic surgery, it might be reasonable to estimate that approximately 10% of the state’s head and neck surgeons are performing almost 45% of the surgery for laryngeal cancer. When analyzing by case type, as described in Table 3, the average number of total and partial laryngectomies performed per head and neck surgeon associated with the teaching hospitals is approximately 37 over the years 1980 to 1990. In contrast, the surgeon in the state who is not associated with the University of Colorado Health Sciences Center would have performed approximately eight total and partial laryngectomies in a similar period. In Colorado, the characteristics of laryngeal cancer, such as incidence rates, sex distribution, laryngeal site predilection, and histologic types of tumors, are similar to those found in other studies. The most significant findings were the increase in the proportion of women diagnosed with laryngeal cancer and the decrease in the percent of patients diagnosed with stage I disease. The actual significance of the latter remains to be evaluated in terms of treatment rendered and survival. Also of significance was the proportion of surgery for laryngeal cancer being performed at the major teaching hospitals. The more extensive experience at the teaching hospitals implies some degree of expertise and supports the role of the regional referral center. This study represents only the initial examination of the information on laryngeal cancer contained in the data base of the Colorado Central Cancer Registry. Efforts are currently being made to examine the data base for information on different treatment modalities and survival statistics which, in turn, may affect the approach to laryngeal cancer in Colorado.

404

Laryngeal Cancer in Colorado

REFERENCES

1. Young J, Asire A, Polltell E. SEER Program: cancer incidence and mortality in the United States 1973-1976. BHEW Publication #(NIH)78-1837. Washington, DC: US.Government Printing Office, 1978. 2. Moore C. Cigarette smoking and cancer of the mouth, pharynx, and larynx. JAMA 1971;218:553-558. 3. Schottenfeld D, Gantt R, Wynder E. The role of alcohol and tobacco in multiple primary cancers of the upper digestive system, larynx, and lung: a prospective study. Prev Med 1974;3:277-293. 4. Burch J, Howe G, Miller A, Semenciu R. Tobacco, alcohol, asbestos, and nickel in the etiology of cancer of the larynx: a case control study. J Natl Cancer Inst 1981;67:1219- 1224. 5. Brown L, Mason T, Pickle L, et al. Occupational risk factors for laryngeal cancer on the Texas Gulf Coast. Cancer Res 1988;48:1960-1964. 6. DeRienzo D, Greenberg D, Fraire A. Carcinoma of the larynx: changing incidence in women. Arch Otolaryngol Head Neck Surg 1991;117:681-684. 7. Yang P, Thomas D, Daling J, et al. Differences in the sex ratio of laryngeal cancer incidence rates by anatomic subsite. J Clin Epidemiol 1989;42:755-758. 8. Karp S , Finch J, Bott R, et al. Cancer in Colorado 1979 to 1988: prevention, incidence, survival, and mortality. Denver: American Cancer Society, Colorado Department of Health, and Colorado Central Cancer Registry, 1991. 9. Ries L, Hankey B, Edwards B. Cancer statistics reuiew1973-1987. NIH Publication #90-2789. Bethesda, MD: National Cancer Institute, 1990. 10. Shvero J, Hadar T, Segal et al. Laryngeal carcinoma in patients 40 years of age and older. Cancer 1987;60:30923096. 11. Levi F, Vecchia C. Sex ratio of laryngeal cancer by anatomical site (letter). J Clin Epidemiol 1990;43:729- 731. 12. Mendez P, Maves M, Panje W. Squamous cell carcinoma of the head and neck in patients under 40 years of age. Arch Otolaryngol 1985;111:762-764. 13. Harris H, Watson F, Spratt J. Carcinoma of the larynx: a retrospective study of 144 cases. A m J Surg 1969;118:676- 684. 14. Mendelson B. The alcohol and drug problem in Colorado: characteristics and trends from 1980 to 1990. Colorado Alcohol and Drug Abuse Division, Office of Data Analysis and Evaluation, 1990.

HEAD & NECK

Septernber/October 1993