- Email: [email protected]
Squamous cell carcinoma of the skin – histopathological features and their significance for the clinical outcome
Journal of the European Academy of Dermatology and Venereology 11 (1998) 37–44
Squamous cell carcinoma of the skin – histopathological features and their significance for the clinical outcome G. Petter*, U.-F. Haustein Department of Dermatology, University of Leipzig, Liebigstrasse 21, 04103 Leipzig, Germany
Abstract Background Owing to the rising incidence of tumours, the question of reliable risk classification is becoming increasingly significant. Objective Participation in the multicentre carcinoma registry maintained by the Association of Operative and Oncological Dermatology of the German Dermatological Society requires, in addition to the parameters of clinical staging and grading already established by the International Union against Cancer, description of other histopathological criteria related to prognosis, with special attention to microstaging. Methods One hundred and eighty-four patients with squamous cell carcinoma of the skin were examined. The histological parameters, carcinoma type, Breslow index, invasion level, growth form, grading and mitotic index were recorded and classified, and clinical staging was performed. Results It was found that the clinical criterion of tumour diameter (T-category) determines the further course of the disease. The other parameters taken into account, namely pathohistological microstaging and grading, can increase the accuracy of prognosis evaluation, and in particular enable carcinomas of the T1-category to be classified as either high malignant or low malignant tumours. The desmoplastic squamous cell carcinoma subtype is a potential risk tumour. Endophytic tumours are more malignant in their development than exophytically-growing carcinomas, and the probability of recurrence and metastatic tumour spread further increases when ulceration can be detected by microscope. In order to distinguish metastatic and recurring carcinomas, in addition to determining the invasion levels after Clark, measuring the Breslow index proved to be an important criterion. Quoting the mitotic index augmented the grading system, which further improved the reliability of malignancy assessment. Conclusions The TNM categories currently applied to squamous cell carcinomas of the skin must be supplemented by additional histological parameters which, according to our findings, permit more accurate classification of high and low malignant tumours. 1998 Elsevier Science B.V. All rights reserved Keywords: Cutaneous squamous cell carcinoma; Histopathology; Prognostic factors; Microstaging
1. Introduction After basal cell carcinoma, squamous cell carcinomas (SCC) represent the second most frequent type of malignant tumours in the skin among whites [1]. The * Corresponding author: Universita¨t Leipzig, Klinik und Poliklinik fu¨r Hautkrankheiten, Liebigstrasse 21, 04103 Leipzig, Germany; Tel.: +49 341 9718600; fax: +49 341 9718609.
occurrence of new carcinomas is on the increase throughout the world, due both to increased UV exposure, resulting from the destruction of the ozone layer caused by atmospheric pollution, and people’s growing life expectancy, and thus chronic exposure to carcinogenic substances [2–4]. The annual rate of incidence is nowadays about 25 squamous cell carcinomas per 100 000 people, with the occurrence being highest among the elderly (espe-
0926-9959/98/$19.00 1998 Elsevier Science B.V. All rights reserved PII S0926-9959 (98 )0 0049-X
38
G. Petter, U.-F. Haustein / J. Eur. Acad. Dermatol. Venereol. 11 (1998) 37–44
cially people aged 60–70 years); moreover, males are significantly more likely to be affected than females [5]. In the USA there are currently about 40 new occurrences per 100 000 people every year. Sunny countries have much higher rates of incidence, especially Australia, where the annual rate is about 201 carcinomas per 100 000 inhabitants [2,6,7]. The aggressiveness and metastatic spreading ability of tumours are undisputed. This, together with their rising incidence, explains why questions of establishing uniform parameters relevant for prognosis is becoming increasingly important in oncological research. So far, tumour classification has been carried out by the Union Internationale Contre le Cancer (UICC) using the TNM system [8,9]. However, this system has the disadvantage that the biologically-different carcinomas are summarised in one group. As carcinomas of the T1 category, in particular, are also capable of metastatic spread and recurrence, it should be investigated whether prognosis accuracy can be increased by taking other pathohistological tumour criteria into account.
2. Materials and methods A total of 184 patients treated at the Dermatological Clinic at Leipzig University with SCC of the skin were included in the study. The average follow-up period was 3 years. 2.1. Histological data The following parameters were recorded and classified after examination of haematoxylin–eosin and van Gieson stained section preparations.
This system was also extended for invasion depths down to the musculature (level VI) and cartilage (level VII). Growth pattern – microscopic classification into exophytic, exophytic-ulcerating, endophytic-ulcerating and endophytic tumours. 2.1.3. Grading Cell polymorphism – classification into low polymorphic (well differentiated), moderate or high polymorphic (poorly differentiated). Mitosis rate – counting of all typical and atypical mitoses in three high-power fields (HPF) 400 × on the tumour growth front. 2.2. Clinical staging Recording of the clinical tumour size (T-classification). 2.3. Statistical methods Statistical data evaluation was carried out using survival functions in accordance with the Kaplan– Meier method using the logrank test. Cox regression analysis was performed using the likelihood-ratio test [12]. Results were deemed to be statistically significant when P , 0.05.
3. Results In 17 cases (9%) tumour progression occurred (locally recurring and/or satellite, in transit and regional lymph node metastatic spread). Eight of the carcinomas displayed recurrence and four tumours metastatic spread, whereas in another five cases, both recurrence and metastatic spread took place. The metastatic spread rate was thus 5% (n = 9) while the recurrence rate was 7% (n = 13).
2.1.1. Typing Carcinoma type – whether a common SCC was present or whether one of the rarer histological varieties had occurred.
3.1. Histological findings
2.1.2. Microstaging Breslow index – measurement of tumour thickness with an accuracy up to 0.1 mm. Clark level – determination of the invasion level I–V analogous to malignant melanoma [10,11].
3.1.1. Typing Most tumours were common SCC. High-infiltration (desmoplastic) carcinomas which spread in small tumour complexes were rare (Fig. 1). Also occasionally encountered were acantholytic and Bowen carci-
39
G. Petter, U.-F. Haustein / J. Eur. Acad. Dermatol. Venereol. 11 (1998) 37–44
Fig. 1. Squamous cell carcinoma with desmoplastic tumour branch formation (haematoxylin–eosin, 112 × ).
nomas and in two cases epithelioma metatypique intermediaire Darier (basosquamous carcinoma). When comparing tumours with progression to carcinomas without progression, the proportion with desmoplastic growth type was greater in the first group. One acantholytic SCC and one epithelioma metatypique intermediaire Darier displayed local recurrence (Table 1). In the case of desmoplastic carcinomas, the probability of renewed progression is significantly higher (P = 0.0001) than for the remaining carcinoma types. At 3 years, cumulative metastasis/recurrencefree survival (probability of freedom from renewed tumour spread) was 93% for tumours without desmoplastic type and 53% for patients presenting a desmoplastic SCC.
3.1.2. Pathohistological microstaging The distribution of tumour thickness is shown in Table 2. Most tumours with progression were thicker than 3.5 mm. Our thinnest metastasising carcinoma was 2.4 mm and the thinnest recurring SCC showed a Breslow index of 1 mm. The metastatic spread and recurrence risk of carcinomas significantly increased (P = 0.012) as the Breslow index rose (Fig. 2). Cumulative metastasis/recurrence-free survival at 3 years was 98% for tumours of ≤3.5 mm and 84% for SCC thicker than 3.5 mm. In five cases Breslow index, invasion depth, growth form and mitosis rate could not be exactly evaluated histologically. Therefore, the total number of tumours with and without progression was 16 and 163, respectively.
Table 1 SCC subtypes Tumour type
Common SCC Desmoplastic SCC Acantholytic SCC Bowen carcinoma Epithelioma metatypique intermediaire Darier Total
SCC without progression
SCC with progression
No.
%
No.
%
139 4 17 6 1 167
83 2 10 4 1
11 4 1 – 1 17
65 23 6 – 6
40
G. Petter, U.-F. Haustein / J. Eur. Acad. Dermatol. Venereol. 11 (1998) 37–44
Table 2 Tumour thickness and invasion depth SCC without progression
SCC with progression
No.
%
No.
%
Breslow index ≤3.5 mm >3.5 mm Total
86 77 163
53 47
3 13 16
19 81
Clark level I II III IV V VI VII Total
– 4 16 59 53 29 2 163
– 2 10 36 33 18 1
– – – 3 11 2 – 16
– – – 19 69 12 –
Only with invasion depths of Clark level IV or more did recurrence take place and metastatic spreading only occurred from Clark level V (Table 2). As the invasion level increased (Clark level ≥V), the prognosis with respect to metastasis/recurrence significantly deteriorated (P = 0.033). Cumulative metastasis/recurrence-free survival at 3 years was 97% for tumours with a Clark level ,V and 85% for SCC with a Clark level ≥V.
In a total of 11 cases (69%), ulceration occurred more frequently in progressive tumours than in the comparative group of squamous cell carcinomas without further progression. Here the proportion of ulcerating carcinomas was much lower, 33% (n = 53), whereas exophytic growth forms were more abundant (Table 3). The influence of growth forms on prognosis (P = 0.0017) is shown in Fig. 3. At 3 years, the cumulative metastasis/recurrence-free survival was 97% for exophytic, 94% for endophytic, 82% for exophytic-ulcerating and 77% for endophytic-ulcerating SCC. Comparing the tumour groups with and without the appearance of an ulceration, the cumulative metastasis/recurrence-free survival at 3 years was 80% and 96%, respectively (P = 0.013). It is apparent that the microscopic detection of ulceration in particular argues for higher malignancy. The particularly aggressive nature of ulcerating tumours can be explained by their significantly higher mitosis frequency (P = 0.001) and a larger degree of cell polymorphism (P = 0.001). There was also positive correlation with tumour thickness, although this was not statistically significant (P = 0.094). 3.1.3. Grading In order to appraise malignancy, the degree of cell polymorphism constituted another important parameter (Table 4). The probability of tumour progres-
Fig. 2. Probability of freedom from recurrence/metastasis by tumour thickness.
G. Petter, U.-F. Haustein / J. Eur. Acad. Dermatol. Venereol. 11 (1998) 37–44 Table 3 SCC growth forms Growth pattern
SCC without progression
SCC with progression
No.
%
No.
%
Endophytic-ulcerating 21 Exophytic 75 Endophytic 35 Exophytic-ulcerating 32 Total 163 Total SCC with ulceration 53
13 46 21 20
6 2 3 5 16 11
37.5 12.5 19 31
33
69
sion was significantly higher the more polymorphic the squamous cell carcinomas (P = 0.028). Cumulative metastasis/recurrence-free survival at 3 years was 96% for tumours with a low degree of cell polymorphism, 91% for moderate polymorphic tumours and 77% for carcinomas with a high degree of cell polymorphism. The proportion of thick carcinomas rose with decreasing differentiation. However, this observation was not statistically significant (P = 0.058). Most SCC without progression (66%) had ≤10 mitoses/3 HPF 400× on the tumour growth front, whereas SCC with progression were more often associated with a higher mitosis rate (Table 4). The strong mitotic activity of a tumour represents a poor prognostic sign. There was a significant difference
41
(P = 0.006) in metastasis/recurrence-free survival for lesions with a mitosis rate greater than 10/3 HPF 400× (3-year survival rate of 80%) compared to SCC whose mitotic index was ≤10/3 HPF 400× (3-year survival rate of 97%). 3.2. Clinical staging (T-classification) Among the non-progressive SCC (n = 167), T1class tumours were most frequent, at 87% (n = 145). Class-T2 carcinomas were rarer (13%; n = 22), while no T3-class squamous cell carcinomas were found. By contrast, distribution among progressive carcinomas (n = 17) was overwhelmingly weighted towards class-T2 tumours (53%; n = 9) and there was a lower proportion (41%; n = 7) of carcinomas smaller than 2 cm. Only one (6%) recurring squamous cell carcinoma was larger than 5 cm. As tumour-diameter increased, the probability of recurring/metastasising tumour growth also rose significantly (P = 0.0001). At 3 years cumulative metastasis/recurrence-free survival was 95% for SCC with a diameter ≤2 cm and 70% for carcinomas with a diameter .2 cm.
4. Discussion The relatively low tumour progression rate (local recurrence rate of 7%, metastatic spreading rate of
Fig. 3. Probability of freedom from recurrence/metastasis by tumour growth pattern.
42
G. Petter, U.-F. Haustein / J. Eur. Acad. Dermatol. Venereol. 11 (1998) 37–44
Table 4 Differentiation and mitotic index
Degree of differentiation Well Moderate Poor Total Mitosis rate ≤10/3 HPF 400× >10/3 HPF 400× Total
SCC without progression
SCC with progression
No
%
No.
%
– 55 93 19
33 56 11
3 9 5
18 53 29
167 108 55 163
17 66 34
5 11
31 69
16
5%) is also quoted by other authors [1,13–15]. However, since with metastatic tumours prognosis quod vitam increasingly deteriorates, correct risk classification is very important both pre- and postoperatively. Nowadays, rarer subtypes such as acantholytic, verrucous and Bowen carcinomas can be histomorphologically distinguished from ordinary squamous cell carcinomas [16,17]. Recent publications have also referred to a desmoplastic carcinoma type which harbours an increased recurrence and metastatic spreading potential, which can be explained by the occurrence of a high proportion of subclinical tumour branch formation surrounded by a usually pronounced stroma reaction [5,18]. In our group the probability of tumour progression was also significantly increased for desmoplastic carcinomas. Bowen carcinomas and acantholytic tumours were not associated with a higher progression risk, as has been postulated in the literature [19,20]. As details of the dignity of carcinoma subtypes have so far only been scanty and partly contradictory, further investigations are required with large groups of patients. Only a few authors have so far examined tumourthickness determination as a method of evaluation prognosis for cutaneous squamous cell carcinoma [13,21,22] and found a correlation between tumour thickness and malignancy. In our investigation also, Breslow index and invasion level proved to be two important indicators of further pathological process.
The thinnest recurring carcinoma had a Breslow index of 1 mm and a Clark level of IV and the thinnest metastatic tumour had a Breslow index of 2.4 mm and a Clark level of V. These findings coincide with those of other authors [13,21]. The histological classification of growth forms with respect to tumour progression behaviour on the squamous cell carcinoma of the skin has, as far as we are aware, not yet been studied. The findings show a significant connection between the occurrence of microscopically-detectable ulceration and tumour malignancy. The metastasis/recurrence-free survival rate dropped significantly when ulceration was present. Moreover, exophytic tumours had lower malignancy compared to the endophytic squamous cell carcinomas. The degree of differentiation was stated by many authors to be a major prognosis criterion [1,9,14]. The metastasis/recurrence-free survival rates in our group were lower for moderate and high polymorphic carcinomas than for low polymorphic tumours. Moreover, there was a positive correlation between tumour thickness and the degree of cell polymorphism. However, it should also be noted that 18% of our tumours with progression only displayed low polymorphism (well differentiated). The mitotic rate in three high-power fields on the tumour growth front proved to be an indicator of the spreading ability of the carcinomas. We found only two publications referring to the mitotic rate. The authors confirmed the prognostic significance [21] and believed mitotic activity to be unsuitable as a prognostic criterion [23]. We found that tumours with mitotic indices of .10/3 HPF 400× tended more frequently to progression than carcinomas with ≤10 mitoses/3 HPF 400×. Clinical size based on T-classification is an established prognosis method. The view contained in the literature [6,15,22,24] that with increasing tumour diameter the probability of metastatic spread and recurrence increases coincide with our findings. In our opinion, however, in addition to implementation of the TNM system, pathohistological microstaging should be introduced, as 41% of progressive carcinomas in our group belonged to stage T1 and thus were appraised only as slightly malignant. Cox regression analysis was carried out to determine the most important prognostic parameters of
G. Petter, U.-F. Haustein / J. Eur. Acad. Dermatol. Venereol. 11 (1998) 37–44
our group of SCC, which were tumour diameter, ulceration and desmoplastic carcinoma subtype. However, these findings are preliminary and may change in a larger group of patients. Therefore all data are entered in the multicentre carcinoma registry of the Association of Operative and Oncological Dermatology of the German Dermatological Society, which is designed to allow multivariate analysis by using a large case series with the goal of defining generally applicable prognosis factors, as has already been done for malignant melanoma [10,25–27].
5. Summary Carcinomas with a diameter of .2 cm are significantly more frequently linked with recurrence and metastasis. However since tumours in stage T1 are also metastatic and recurrent, the T-system within clinical staging is an unreliable factor for determining future course. Other microscopic criteria were therefore analysed, with a view to obtaining more accurate malignancy evaluation. Histologically-poor prognostic signs included a desmoplastic growth type, tumour thickness of .3.5 mm, Clark-Level ≥V, ulceration, poor differentiation, and over ten mitoses/3 HPF 400× on the tumour growth front. The grading system can be augmented by quoting the mitosis rate, which further increases the reliability of malignancy assessment. In order to create a uniformly-valid system for prognosis estimation, multivariate analysis of large case-series patient collectives are essential. This could be achieved by the multicentre carcinoma registry. References [1] Johnson TM, Rowe DE, Nelson BR, Swanson NA. Squamous cell carcinoma of the skin (excluding lip and oral mucosa). J Am Acad Dermatol 1992;26:467–484. [2] Gloster HM, Brodland DG. The epidemiology of skin cancer. Dermatol Surg 1996;22:217–226. [3] Gallagher RP, Ma B, McLean DI, Yang CP, Ho V, Carruthers JA, Warshawski LM. Trends in basal cell carcinoma, squamous cell carcinoma, and melanoma of the skin from 1973 through 1987. J Am Acad Dermatol 1990;23:413–421. [4] Marks R. An overview of skin cancers. Incidence and causation. Cancer 1995;75:607–612.
43
[5] Breuninger H, Dummer R, Garbe C, Hauschild A, Kaufmann R, Panizzon R, Scho¨fer H, Sebastian G, Soyer HP, Sterry W, Tilgen W. Zur Diskussion gestellt: Entwurf zu den Leitlinien zur Qualita¨tssicherung in der dermatologischen Onkologie. Derm, Dermatologische Onkologie. Omnimed Verlag 1996;2:8–13. [6] Haydon RC. Cutaneous squamous carcinoma and related lesions. Otolaryngol Clin North Am 1993;26:57–71. [7] Marks R. The epidemiology of non-melanoma skin cancer: who, why and what can we do about it. J Dermatol 1995;22:853–857. [8] Hermanek P, Sobin LH. TNM classification of malignant tumours. Springer-Verlag Berlin, Heidelberg, New York 1987:89–98. [9] Spiessl B, Beahrs OH, Hermanek P, Hutter RVP, Scheibe O, Sobin LH, Wagner G. TNM-Atlas. Illustrierter Leitfaden zur TNM/pTNM-Klassifikation maligner Tumoren. 3. Auflage. Springer-Verlag Berlin, Heidelberg, New York 1993:171– 174. [10] Clark WH, Elder DE, Horn MV. The biologic forms of malignant melanoma. Hum Pathol 1986;17:443–450. [11] Mihm MC, Clark WH, From L. The clinical diagnosis, classification and histogenetic concepts of the early stages of cutaneous malignant melanomas. New Engl J Med 1971;284:1078–1082. [12] Cox DR, Oakes D. Analysis of survival data. London: Chapman and Hall, 1984. [13] Friedman HI, Cooper PH, Wanebo HJ. Prognostic and therapeutic use of microstaging of cutaneous squamous cell carcinoma of the trunk and extremities. Cancer 1985;56:1099– 1105. [14] Frankel DH. Squamous cell carcinoma of the skin. Hosp Pract 1992;27:99–106. [15] Yeatman J, Marks R. Non-melanoma skin cancer. Med J Australia 1996;164:492–496. [16] Lever WF, Schaumburg-Lever G. Histopathology of the skin, seventh ed. Philadelphia: J.B. Lippincott, 1990:542–563. [17] McKee P. Pathology of the skin with clinical correlations. Philadelphia: J.B. Lippincott, 1989:14.12–14.20. [18] Haneke E. Histologische Varianten des Plattenepithelkarzinoms der Haut und ihre Dignita¨t. In: Breuninger H, Rassner G, editors. Operationsplanung und Erfolgskontrolle. Berlin, Heidelberg, New York: Springer-Verlag, 1989:79–85. [19] Bernstein SC, Lim KK, Brodland DG, Heidelberg KA. The many faces of squamous cell carcinoma. Dermatol Surg 1996;22:243–254. [20] Wick MR, Pettinato G, Nappi O. Adenoid (acantholytic) squamous carcinoma of the skin. J Cutan Pathol 1988;15: 351. [21] Breuninger H, Hawlitschek E. Das Mikrostaging des Plattenepithelkarzinoms der Haut und Lippen. Lichtmikroskopisch erfaßbare Prognosefaktoren. In: Tilgen W, Petzoldt D. Operative und konservative Dermato-Onkologie. Fortschritte der operativen und onkologischen Dermatologie. Band 10. Berlin, Heidelberg, New York: Springer-Verlag 1995:110–115. [22] Rowe DE, Carroll RJ, Day CL. Prognostic factors for local recurrence, metastasis, and survival rates in squamous cell
44
G. Petter, U.-F. Haustein / J. Eur. Acad. Dermatol. Venereol. 11 (1998) 37–44
carcinoma of the skin, ear, and lip: implications for treatment modality selection. J Am Acad Dermatol 1992;26:976– 990. [23] Frierson HF, Cooper PH. Prognostic factors in squamous cell carcinoma of the lower lip. Hum Pathol 1986;17:346–354. [24] Hacker SM, Flowers FP. Squamous cell carcinoma of the skin: will heightened awareness of risk factors slow its increase? Postgrad Med 1993;93:115–126. [25] Schmoeckel C, Bockelbrink A, Bockelbrink H, Koutsis J,
Braun-Falco O. Low- and high-risk malignant melanoma-I. Evaluation of clinical and histological prognosticators in 585 cases. Eur J Cancer Clin Oncol 1983;19:227–235. [26] McGovern VJ, Shaw H.M, Milton GW, McCarthy WH. Ulceration and prognosis in cutaneous malignant melanoma. Histopathology 1982;6:399–407. [27] Breslow A. Thickness, cross-sectional areas and depth of invasion in the prognosis of cutaneous melanoma. Ann Surg 1970;172:902–908.
Squamous cell carcinoma of the skin – histopathological features and their significance for the clinical outcome G. Petter*, U.-F. Haustein Department of Dermatology, University of Leipzig, Liebigstrasse 21, 04103 Leipzig, Germany
Abstract Background Owing to the rising incidence of tumours, the question of reliable risk classification is becoming increasingly significant. Objective Participation in the multicentre carcinoma registry maintained by the Association of Operative and Oncological Dermatology of the German Dermatological Society requires, in addition to the parameters of clinical staging and grading already established by the International Union against Cancer, description of other histopathological criteria related to prognosis, with special attention to microstaging. Methods One hundred and eighty-four patients with squamous cell carcinoma of the skin were examined. The histological parameters, carcinoma type, Breslow index, invasion level, growth form, grading and mitotic index were recorded and classified, and clinical staging was performed. Results It was found that the clinical criterion of tumour diameter (T-category) determines the further course of the disease. The other parameters taken into account, namely pathohistological microstaging and grading, can increase the accuracy of prognosis evaluation, and in particular enable carcinomas of the T1-category to be classified as either high malignant or low malignant tumours. The desmoplastic squamous cell carcinoma subtype is a potential risk tumour. Endophytic tumours are more malignant in their development than exophytically-growing carcinomas, and the probability of recurrence and metastatic tumour spread further increases when ulceration can be detected by microscope. In order to distinguish metastatic and recurring carcinomas, in addition to determining the invasion levels after Clark, measuring the Breslow index proved to be an important criterion. Quoting the mitotic index augmented the grading system, which further improved the reliability of malignancy assessment. Conclusions The TNM categories currently applied to squamous cell carcinomas of the skin must be supplemented by additional histological parameters which, according to our findings, permit more accurate classification of high and low malignant tumours. 1998 Elsevier Science B.V. All rights reserved Keywords: Cutaneous squamous cell carcinoma; Histopathology; Prognostic factors; Microstaging
1. Introduction After basal cell carcinoma, squamous cell carcinomas (SCC) represent the second most frequent type of malignant tumours in the skin among whites [1]. The * Corresponding author: Universita¨t Leipzig, Klinik und Poliklinik fu¨r Hautkrankheiten, Liebigstrasse 21, 04103 Leipzig, Germany; Tel.: +49 341 9718600; fax: +49 341 9718609.
occurrence of new carcinomas is on the increase throughout the world, due both to increased UV exposure, resulting from the destruction of the ozone layer caused by atmospheric pollution, and people’s growing life expectancy, and thus chronic exposure to carcinogenic substances [2–4]. The annual rate of incidence is nowadays about 25 squamous cell carcinomas per 100 000 people, with the occurrence being highest among the elderly (espe-
0926-9959/98/$19.00 1998 Elsevier Science B.V. All rights reserved PII S0926-9959 (98 )0 0049-X
38
G. Petter, U.-F. Haustein / J. Eur. Acad. Dermatol. Venereol. 11 (1998) 37–44
cially people aged 60–70 years); moreover, males are significantly more likely to be affected than females [5]. In the USA there are currently about 40 new occurrences per 100 000 people every year. Sunny countries have much higher rates of incidence, especially Australia, where the annual rate is about 201 carcinomas per 100 000 inhabitants [2,6,7]. The aggressiveness and metastatic spreading ability of tumours are undisputed. This, together with their rising incidence, explains why questions of establishing uniform parameters relevant for prognosis is becoming increasingly important in oncological research. So far, tumour classification has been carried out by the Union Internationale Contre le Cancer (UICC) using the TNM system [8,9]. However, this system has the disadvantage that the biologically-different carcinomas are summarised in one group. As carcinomas of the T1 category, in particular, are also capable of metastatic spread and recurrence, it should be investigated whether prognosis accuracy can be increased by taking other pathohistological tumour criteria into account.
2. Materials and methods A total of 184 patients treated at the Dermatological Clinic at Leipzig University with SCC of the skin were included in the study. The average follow-up period was 3 years. 2.1. Histological data The following parameters were recorded and classified after examination of haematoxylin–eosin and van Gieson stained section preparations.
This system was also extended for invasion depths down to the musculature (level VI) and cartilage (level VII). Growth pattern – microscopic classification into exophytic, exophytic-ulcerating, endophytic-ulcerating and endophytic tumours. 2.1.3. Grading Cell polymorphism – classification into low polymorphic (well differentiated), moderate or high polymorphic (poorly differentiated). Mitosis rate – counting of all typical and atypical mitoses in three high-power fields (HPF) 400 × on the tumour growth front. 2.2. Clinical staging Recording of the clinical tumour size (T-classification). 2.3. Statistical methods Statistical data evaluation was carried out using survival functions in accordance with the Kaplan– Meier method using the logrank test. Cox regression analysis was performed using the likelihood-ratio test [12]. Results were deemed to be statistically significant when P , 0.05.
3. Results In 17 cases (9%) tumour progression occurred (locally recurring and/or satellite, in transit and regional lymph node metastatic spread). Eight of the carcinomas displayed recurrence and four tumours metastatic spread, whereas in another five cases, both recurrence and metastatic spread took place. The metastatic spread rate was thus 5% (n = 9) while the recurrence rate was 7% (n = 13).
2.1.1. Typing Carcinoma type – whether a common SCC was present or whether one of the rarer histological varieties had occurred.
3.1. Histological findings
2.1.2. Microstaging Breslow index – measurement of tumour thickness with an accuracy up to 0.1 mm. Clark level – determination of the invasion level I–V analogous to malignant melanoma [10,11].
3.1.1. Typing Most tumours were common SCC. High-infiltration (desmoplastic) carcinomas which spread in small tumour complexes were rare (Fig. 1). Also occasionally encountered were acantholytic and Bowen carci-
39
G. Petter, U.-F. Haustein / J. Eur. Acad. Dermatol. Venereol. 11 (1998) 37–44
Fig. 1. Squamous cell carcinoma with desmoplastic tumour branch formation (haematoxylin–eosin, 112 × ).
nomas and in two cases epithelioma metatypique intermediaire Darier (basosquamous carcinoma). When comparing tumours with progression to carcinomas without progression, the proportion with desmoplastic growth type was greater in the first group. One acantholytic SCC and one epithelioma metatypique intermediaire Darier displayed local recurrence (Table 1). In the case of desmoplastic carcinomas, the probability of renewed progression is significantly higher (P = 0.0001) than for the remaining carcinoma types. At 3 years, cumulative metastasis/recurrencefree survival (probability of freedom from renewed tumour spread) was 93% for tumours without desmoplastic type and 53% for patients presenting a desmoplastic SCC.
3.1.2. Pathohistological microstaging The distribution of tumour thickness is shown in Table 2. Most tumours with progression were thicker than 3.5 mm. Our thinnest metastasising carcinoma was 2.4 mm and the thinnest recurring SCC showed a Breslow index of 1 mm. The metastatic spread and recurrence risk of carcinomas significantly increased (P = 0.012) as the Breslow index rose (Fig. 2). Cumulative metastasis/recurrence-free survival at 3 years was 98% for tumours of ≤3.5 mm and 84% for SCC thicker than 3.5 mm. In five cases Breslow index, invasion depth, growth form and mitosis rate could not be exactly evaluated histologically. Therefore, the total number of tumours with and without progression was 16 and 163, respectively.
Table 1 SCC subtypes Tumour type
Common SCC Desmoplastic SCC Acantholytic SCC Bowen carcinoma Epithelioma metatypique intermediaire Darier Total
SCC without progression
SCC with progression
No.
%
No.
%
139 4 17 6 1 167
83 2 10 4 1
11 4 1 – 1 17
65 23 6 – 6
40
G. Petter, U.-F. Haustein / J. Eur. Acad. Dermatol. Venereol. 11 (1998) 37–44
Table 2 Tumour thickness and invasion depth SCC without progression
SCC with progression
No.
%
No.
%
Breslow index ≤3.5 mm >3.5 mm Total
86 77 163
53 47
3 13 16
19 81
Clark level I II III IV V VI VII Total
– 4 16 59 53 29 2 163
– 2 10 36 33 18 1
– – – 3 11 2 – 16
– – – 19 69 12 –
Only with invasion depths of Clark level IV or more did recurrence take place and metastatic spreading only occurred from Clark level V (Table 2). As the invasion level increased (Clark level ≥V), the prognosis with respect to metastasis/recurrence significantly deteriorated (P = 0.033). Cumulative metastasis/recurrence-free survival at 3 years was 97% for tumours with a Clark level ,V and 85% for SCC with a Clark level ≥V.
In a total of 11 cases (69%), ulceration occurred more frequently in progressive tumours than in the comparative group of squamous cell carcinomas without further progression. Here the proportion of ulcerating carcinomas was much lower, 33% (n = 53), whereas exophytic growth forms were more abundant (Table 3). The influence of growth forms on prognosis (P = 0.0017) is shown in Fig. 3. At 3 years, the cumulative metastasis/recurrence-free survival was 97% for exophytic, 94% for endophytic, 82% for exophytic-ulcerating and 77% for endophytic-ulcerating SCC. Comparing the tumour groups with and without the appearance of an ulceration, the cumulative metastasis/recurrence-free survival at 3 years was 80% and 96%, respectively (P = 0.013). It is apparent that the microscopic detection of ulceration in particular argues for higher malignancy. The particularly aggressive nature of ulcerating tumours can be explained by their significantly higher mitosis frequency (P = 0.001) and a larger degree of cell polymorphism (P = 0.001). There was also positive correlation with tumour thickness, although this was not statistically significant (P = 0.094). 3.1.3. Grading In order to appraise malignancy, the degree of cell polymorphism constituted another important parameter (Table 4). The probability of tumour progres-
Fig. 2. Probability of freedom from recurrence/metastasis by tumour thickness.
G. Petter, U.-F. Haustein / J. Eur. Acad. Dermatol. Venereol. 11 (1998) 37–44 Table 3 SCC growth forms Growth pattern
SCC without progression
SCC with progression
No.
%
No.
%
Endophytic-ulcerating 21 Exophytic 75 Endophytic 35 Exophytic-ulcerating 32 Total 163 Total SCC with ulceration 53
13 46 21 20
6 2 3 5 16 11
37.5 12.5 19 31
33
69
sion was significantly higher the more polymorphic the squamous cell carcinomas (P = 0.028). Cumulative metastasis/recurrence-free survival at 3 years was 96% for tumours with a low degree of cell polymorphism, 91% for moderate polymorphic tumours and 77% for carcinomas with a high degree of cell polymorphism. The proportion of thick carcinomas rose with decreasing differentiation. However, this observation was not statistically significant (P = 0.058). Most SCC without progression (66%) had ≤10 mitoses/3 HPF 400× on the tumour growth front, whereas SCC with progression were more often associated with a higher mitosis rate (Table 4). The strong mitotic activity of a tumour represents a poor prognostic sign. There was a significant difference
41
(P = 0.006) in metastasis/recurrence-free survival for lesions with a mitosis rate greater than 10/3 HPF 400× (3-year survival rate of 80%) compared to SCC whose mitotic index was ≤10/3 HPF 400× (3-year survival rate of 97%). 3.2. Clinical staging (T-classification) Among the non-progressive SCC (n = 167), T1class tumours were most frequent, at 87% (n = 145). Class-T2 carcinomas were rarer (13%; n = 22), while no T3-class squamous cell carcinomas were found. By contrast, distribution among progressive carcinomas (n = 17) was overwhelmingly weighted towards class-T2 tumours (53%; n = 9) and there was a lower proportion (41%; n = 7) of carcinomas smaller than 2 cm. Only one (6%) recurring squamous cell carcinoma was larger than 5 cm. As tumour-diameter increased, the probability of recurring/metastasising tumour growth also rose significantly (P = 0.0001). At 3 years cumulative metastasis/recurrence-free survival was 95% for SCC with a diameter ≤2 cm and 70% for carcinomas with a diameter .2 cm.
4. Discussion The relatively low tumour progression rate (local recurrence rate of 7%, metastatic spreading rate of
Fig. 3. Probability of freedom from recurrence/metastasis by tumour growth pattern.
42
G. Petter, U.-F. Haustein / J. Eur. Acad. Dermatol. Venereol. 11 (1998) 37–44
Table 4 Differentiation and mitotic index
Degree of differentiation Well Moderate Poor Total Mitosis rate ≤10/3 HPF 400× >10/3 HPF 400× Total
SCC without progression
SCC with progression
No
%
No.
%
– 55 93 19
33 56 11
3 9 5
18 53 29
167 108 55 163
17 66 34
5 11
31 69
16
5%) is also quoted by other authors [1,13–15]. However, since with metastatic tumours prognosis quod vitam increasingly deteriorates, correct risk classification is very important both pre- and postoperatively. Nowadays, rarer subtypes such as acantholytic, verrucous and Bowen carcinomas can be histomorphologically distinguished from ordinary squamous cell carcinomas [16,17]. Recent publications have also referred to a desmoplastic carcinoma type which harbours an increased recurrence and metastatic spreading potential, which can be explained by the occurrence of a high proportion of subclinical tumour branch formation surrounded by a usually pronounced stroma reaction [5,18]. In our group the probability of tumour progression was also significantly increased for desmoplastic carcinomas. Bowen carcinomas and acantholytic tumours were not associated with a higher progression risk, as has been postulated in the literature [19,20]. As details of the dignity of carcinoma subtypes have so far only been scanty and partly contradictory, further investigations are required with large groups of patients. Only a few authors have so far examined tumourthickness determination as a method of evaluation prognosis for cutaneous squamous cell carcinoma [13,21,22] and found a correlation between tumour thickness and malignancy. In our investigation also, Breslow index and invasion level proved to be two important indicators of further pathological process.
The thinnest recurring carcinoma had a Breslow index of 1 mm and a Clark level of IV and the thinnest metastatic tumour had a Breslow index of 2.4 mm and a Clark level of V. These findings coincide with those of other authors [13,21]. The histological classification of growth forms with respect to tumour progression behaviour on the squamous cell carcinoma of the skin has, as far as we are aware, not yet been studied. The findings show a significant connection between the occurrence of microscopically-detectable ulceration and tumour malignancy. The metastasis/recurrence-free survival rate dropped significantly when ulceration was present. Moreover, exophytic tumours had lower malignancy compared to the endophytic squamous cell carcinomas. The degree of differentiation was stated by many authors to be a major prognosis criterion [1,9,14]. The metastasis/recurrence-free survival rates in our group were lower for moderate and high polymorphic carcinomas than for low polymorphic tumours. Moreover, there was a positive correlation between tumour thickness and the degree of cell polymorphism. However, it should also be noted that 18% of our tumours with progression only displayed low polymorphism (well differentiated). The mitotic rate in three high-power fields on the tumour growth front proved to be an indicator of the spreading ability of the carcinomas. We found only two publications referring to the mitotic rate. The authors confirmed the prognostic significance [21] and believed mitotic activity to be unsuitable as a prognostic criterion [23]. We found that tumours with mitotic indices of .10/3 HPF 400× tended more frequently to progression than carcinomas with ≤10 mitoses/3 HPF 400×. Clinical size based on T-classification is an established prognosis method. The view contained in the literature [6,15,22,24] that with increasing tumour diameter the probability of metastatic spread and recurrence increases coincide with our findings. In our opinion, however, in addition to implementation of the TNM system, pathohistological microstaging should be introduced, as 41% of progressive carcinomas in our group belonged to stage T1 and thus were appraised only as slightly malignant. Cox regression analysis was carried out to determine the most important prognostic parameters of
G. Petter, U.-F. Haustein / J. Eur. Acad. Dermatol. Venereol. 11 (1998) 37–44
our group of SCC, which were tumour diameter, ulceration and desmoplastic carcinoma subtype. However, these findings are preliminary and may change in a larger group of patients. Therefore all data are entered in the multicentre carcinoma registry of the Association of Operative and Oncological Dermatology of the German Dermatological Society, which is designed to allow multivariate analysis by using a large case series with the goal of defining generally applicable prognosis factors, as has already been done for malignant melanoma [10,25–27].
5. Summary Carcinomas with a diameter of .2 cm are significantly more frequently linked with recurrence and metastasis. However since tumours in stage T1 are also metastatic and recurrent, the T-system within clinical staging is an unreliable factor for determining future course. Other microscopic criteria were therefore analysed, with a view to obtaining more accurate malignancy evaluation. Histologically-poor prognostic signs included a desmoplastic growth type, tumour thickness of .3.5 mm, Clark-Level ≥V, ulceration, poor differentiation, and over ten mitoses/3 HPF 400× on the tumour growth front. The grading system can be augmented by quoting the mitosis rate, which further increases the reliability of malignancy assessment. In order to create a uniformly-valid system for prognosis estimation, multivariate analysis of large case-series patient collectives are essential. This could be achieved by the multicentre carcinoma registry. References [1] Johnson TM, Rowe DE, Nelson BR, Swanson NA. Squamous cell carcinoma of the skin (excluding lip and oral mucosa). J Am Acad Dermatol 1992;26:467–484. [2] Gloster HM, Brodland DG. The epidemiology of skin cancer. Dermatol Surg 1996;22:217–226. [3] Gallagher RP, Ma B, McLean DI, Yang CP, Ho V, Carruthers JA, Warshawski LM. Trends in basal cell carcinoma, squamous cell carcinoma, and melanoma of the skin from 1973 through 1987. J Am Acad Dermatol 1990;23:413–421. [4] Marks R. An overview of skin cancers. Incidence and causation. Cancer 1995;75:607–612.
43
[5] Breuninger H, Dummer R, Garbe C, Hauschild A, Kaufmann R, Panizzon R, Scho¨fer H, Sebastian G, Soyer HP, Sterry W, Tilgen W. Zur Diskussion gestellt: Entwurf zu den Leitlinien zur Qualita¨tssicherung in der dermatologischen Onkologie. Derm, Dermatologische Onkologie. Omnimed Verlag 1996;2:8–13. [6] Haydon RC. Cutaneous squamous carcinoma and related lesions. Otolaryngol Clin North Am 1993;26:57–71. [7] Marks R. The epidemiology of non-melanoma skin cancer: who, why and what can we do about it. J Dermatol 1995;22:853–857. [8] Hermanek P, Sobin LH. TNM classification of malignant tumours. Springer-Verlag Berlin, Heidelberg, New York 1987:89–98. [9] Spiessl B, Beahrs OH, Hermanek P, Hutter RVP, Scheibe O, Sobin LH, Wagner G. TNM-Atlas. Illustrierter Leitfaden zur TNM/pTNM-Klassifikation maligner Tumoren. 3. Auflage. Springer-Verlag Berlin, Heidelberg, New York 1993:171– 174. [10] Clark WH, Elder DE, Horn MV. The biologic forms of malignant melanoma. Hum Pathol 1986;17:443–450. [11] Mihm MC, Clark WH, From L. The clinical diagnosis, classification and histogenetic concepts of the early stages of cutaneous malignant melanomas. New Engl J Med 1971;284:1078–1082. [12] Cox DR, Oakes D. Analysis of survival data. London: Chapman and Hall, 1984. [13] Friedman HI, Cooper PH, Wanebo HJ. Prognostic and therapeutic use of microstaging of cutaneous squamous cell carcinoma of the trunk and extremities. Cancer 1985;56:1099– 1105. [14] Frankel DH. Squamous cell carcinoma of the skin. Hosp Pract 1992;27:99–106. [15] Yeatman J, Marks R. Non-melanoma skin cancer. Med J Australia 1996;164:492–496. [16] Lever WF, Schaumburg-Lever G. Histopathology of the skin, seventh ed. Philadelphia: J.B. Lippincott, 1990:542–563. [17] McKee P. Pathology of the skin with clinical correlations. Philadelphia: J.B. Lippincott, 1989:14.12–14.20. [18] Haneke E. Histologische Varianten des Plattenepithelkarzinoms der Haut und ihre Dignita¨t. In: Breuninger H, Rassner G, editors. Operationsplanung und Erfolgskontrolle. Berlin, Heidelberg, New York: Springer-Verlag, 1989:79–85. [19] Bernstein SC, Lim KK, Brodland DG, Heidelberg KA. The many faces of squamous cell carcinoma. Dermatol Surg 1996;22:243–254. [20] Wick MR, Pettinato G, Nappi O. Adenoid (acantholytic) squamous carcinoma of the skin. J Cutan Pathol 1988;15: 351. [21] Breuninger H, Hawlitschek E. Das Mikrostaging des Plattenepithelkarzinoms der Haut und Lippen. Lichtmikroskopisch erfaßbare Prognosefaktoren. In: Tilgen W, Petzoldt D. Operative und konservative Dermato-Onkologie. Fortschritte der operativen und onkologischen Dermatologie. Band 10. Berlin, Heidelberg, New York: Springer-Verlag 1995:110–115. [22] Rowe DE, Carroll RJ, Day CL. Prognostic factors for local recurrence, metastasis, and survival rates in squamous cell
44
G. Petter, U.-F. Haustein / J. Eur. Acad. Dermatol. Venereol. 11 (1998) 37–44
carcinoma of the skin, ear, and lip: implications for treatment modality selection. J Am Acad Dermatol 1992;26:976– 990. [23] Frierson HF, Cooper PH. Prognostic factors in squamous cell carcinoma of the lower lip. Hum Pathol 1986;17:346–354. [24] Hacker SM, Flowers FP. Squamous cell carcinoma of the skin: will heightened awareness of risk factors slow its increase? Postgrad Med 1993;93:115–126. [25] Schmoeckel C, Bockelbrink A, Bockelbrink H, Koutsis J,
Braun-Falco O. Low- and high-risk malignant melanoma-I. Evaluation of clinical and histological prognosticators in 585 cases. Eur J Cancer Clin Oncol 1983;19:227–235. [26] McGovern VJ, Shaw H.M, Milton GW, McCarthy WH. Ulceration and prognosis in cutaneous malignant melanoma. Histopathology 1982;6:399–407. [27] Breslow A. Thickness, cross-sectional areas and depth of invasion in the prognosis of cutaneous melanoma. Ann Surg 1970;172:902–908.