GYNECOLOGIC
ONCOLOGY
39, 181-185 (1%))
DNA Ploidy Level as Prognostic Factor in Low Stage Carcinoma of the Uterine Cervix G. G.
KENTER,*
C. J.
CORNELISSE,~
E. J.
AARTSEN,$ W. MOOY,§ AND G. J. FLEURENt
J.
HERMANS,~~
A. P. M. HEINTZ#
Departments of *Gynaecology, TPathology, and [[Medical Statistics of the University of Leiden and the Departments of SGynaecology and SPathology of the Antoni van Leeuwenhoek huis of the Netherlands Cancer Institute Received
March 26, 1990
Plow cytometry was used to measure DNA content of tumor cells in paraffin-embedded archival material from 89 patients with squamous cell carcinoma of the cervix uteri stages IB and DA. Patients were all treated with radical hysterectomy and transperitoneal lymphadenectomy. Twenty-four percent received radiotherapy postoperatively because of tumor spread into parametria or positive lymph nodes. Ploidy grade was compared to other prognostic factors. DNA-aneuploidy was seen in 44 (49%), DNA-diploidy in 16 (18%) and 29 (33%) of the tumors were DNA-periploid. Sixty-nine (78%) patients were in PIG0 stage IB, 20 (22%) in DA. In 19 (21%) lymph node mebstases were found, vasoinvasion in 25 (28%). Overall disease-specific S-year survival was 80%. There was a significant effect of positive lymph nodes (90% vs 47%) and vasoinvasion (87% vs 64%) on the 5year survival (resp. P < 0.01 and P = 0.02). No correlation was found between stage (81% for IB vs 80% for DA) or DNA-ploidy grade (81% for aneuploidy, 82% for periploidy, 79% for diploidy) and survival (req. P = 0.9 and P = 0.8). Ploidy grade was equally divided over other prognostic factors. In skpwise Cox regression analysis DNA-ploidy grade showed no independent effect on survival. We conclude that DNA flow cytometry in this ~eadan~ mess, material was of no additional prognustic value. Q 1990 Inc.
INTRODUCTION Although the incidence of cancer of the cervix uteri has decreased, probably because of better screening programs, there is growing evidence that cervical cancer is nowadays more often seen in younger women [l]. Cervical cancer is the second cause of death in women under the age of 50. The results of surgical treatment in stages IB and IIA ‘ Present Utrecht.
address:
Department
of
Gynaecology,
University
of
are good; the 5-year survival is usually more than 85% [2,3]. Still about 15-20% of these patients have to be treated with radiotherapy after radical hysterectomy because of lymph node metastases or tumor growth into the parametria. These parameters are known to worsen the prognosis [4]. Like other prognostic factors, such as tumor size [5] and vasoinvasion [6], these can be measured only postoperatively. Complications of surgical treatment [7] and radiotherapy [8] are considerable, and those of combined treatment are even worse [9]. This indicates the need for a more precise quantitative method in the prediction of prognosis to be able to individualize treatment as much as possible. With the introduction of DNA flow cytometry on tumor tissue a new method has become available that is possibly of prognostic importance. With this method DNA content of nuclei of tumor cells can be measured in fresh as well as archival paraffin-embedded tissue samples [lo]. Reports have been published in which tumor DNA-ploidy seemed to be of prognostic value with respect to survival for several (gynecologic) malignancies [l 11. This report evaluates the prognostic value of DNA flow cytometry in a series of patients with stage I and IIA cervical cancer. MATERIALANDMETHOD Paraffin-embedded archival material from 213 patients treated between 1%7 and 1981 for a carcinoma of the uterine cervix stage I or IIA was available. All patients were treated by means of radical hysterectomy with transperitoneal lymphadenectomy in the Antoni van Leeuwenhoek Hospital of The Netherlands Cancer Institute. Preoperative staging was carried out according to FIGO. Details about the results of the surgical pro-
181
0090-8258BO $1.50 Copyright 8 l!EJOby Academic Press, Inc. All rights of reproduction in any form reserved.
182
KENTER ET AL.
FIG. 1. Examples of a diploid (A), peridiploid (B), and aneuploid (C) histogram.
cedure, the complications, and the survival have been published previously [7]. Histological material was reviewed as discussed elsewhere [12]. There were 106 cases with squamous cell carcinoma of the cervix, in which there was sufficient tumor tissue in the surgical material to perform the procedure, selected for flow cytometry. A 40-pm section was cut. Sections were deparafhnized and prepared as described by Hedley et al. TABLE 1 Relation betweenStage,Node Status, Vasoinvasion,and Ploidy Grade Diploid Stage IB IIA Nodes Negative Positive Vasoinvasion Negative Positive
Periploid
Aneuploid
N
14 2
21 8
34 10
69 20
13 3
21 8
36 8
70 19
11 5
21 8
32 12
64 25
16
29
44
89
1101.Samples were measured on a ICP 22 flow cytometer (Ortho, Westwood, MA). Tumor DNA content was classified as aneuploid if the DNA profile showed two or more distinct G 0.1 peaks. (Fig. 1). The resolution of the DNA profiles obtained from deparaffinized tissue samples frequently shows a larger range of coefficient of variation (CV) values than those from fresh tumor tissue which interferes with the minimal detectable DNA content difference [13]. A high CV may obscure the presence of a low aneuploid population (e.g., with a CV > 5% a split peak is seen only when the DNA content difference between two populations is at least 10%). According to Hedley et al. [lo] we defined tumors with a single G 0.1 peak with a CV 2 5.5% as a different (peridiploid) group, to exclude the possibility that low aneuploid tumors with DNA differences in the lo-20% range might be overlooked. DNA histograms were classified as DNA-diploid, -aneuploid, or -peridiploid without knowledge of patient survival or clinicopathologic features. Follow-up for all patients was at least 60 months. Survival was measured from the time of surgery to death due to disease. Statistical analysis was carried out using the BMDP package. Survival curves were compared with the log-rank test. A stepwise Cox regression model [141 was used to evaluate the effect of ploidy grade on survival additional to stage, vasoinvasion, and lymph node metastases. RESULTS Evaluable DNA profiles were obtained for 89 of 106 selected patients. Seventeen profiles could not be evaluated, probably because of the advanced age of the blocks (from before 1973). To see if the possibility of flow cytometry itself had any impact on survival rates we compared the results of 89 patients in whom flow cytometry was performed with the rest of the group. No differences of statistical interest were found. There were 69 (78%) stage IB patients and 20 (22%) stage IIA patients. Lymph node metastases were present in 19 (21%) patients. Postoperatively, 21 (24%) patients were treated with radiotherapy because of tumor spread into the parametria and/or positive lymph nodes. Vasoinvasion was found in 25 (28%) patients. Flow cytometry showed DNA-diploid profiles in 16 (18%), peridiploid histograms in 29 (33%), and 44 (49%) DNA-aneuploid tumors. The relation between ploidy grade and the other prognostic variables is shown in Table 1. The worst prognosis categories (positive lymph nodes and vasoinvasion) are more or less equally present in the three ploidy classes. Overall disease-specific 5-year survival was 80%. Table 2 shows the effect of several parameters on survival rates. In the case of lymph node metastases the 5-year
DNA PLOIDY IN CERVICAL TABLE Effect
183
CANCER
2
of Node Status, Vaseinvasion,Stage,and Ploidy Grade on S-YearSurvival
Nodes Negative Positive Vasoinvasion Negative Positive Stage IB IIA Ploidy grade Diploid Periploid Aneuploid Total group
N
Recurrence (4
S-year survival m
70 19
7 10
90 47
co.01
64 25
8 9
87.5 64
0.17
69 20
13 4
81 80
0.8
16 29 44
3 5 9
81 82 79
0.9
P value
80
89
survival rate dropped from 90% without to 47% with positive nodes (P < 0.01) (Fig. 2); vasoinvasion decreased the 5-year survival from 87% without to 64% with invasion (P = 0.02). The stage at time of surgery had no effect on survival: 81% in stage IB and 80% in stage IIA (P = 0.8). Ploidy grade showed no statistical differences in survival rates: 81% for patients with diploid tumors, 82% for those with periploid tumors, and 79% in cases with aneuploidy (P = 0.9) (Fig. 3). In the stepwise Cox regression analysis the effects of the following parameters on survival were studied: lymph node metastases, vasoinvasion, stage, and ploidy grade. The presence of lymph node metastases had by far the strongest effect on survival. The other parameters (vasoinvasion, stage, and ploidy grade) turned out to have no additional prognostic value. DISCUSSION With the development of flow cytometry it has become possible to rapidly measure DNA content of tens of thousands of cells within a short amount of time. Initially only fresh tumor material could be analyzed and this was of little use in the establishment of the value of DNA flow cytometry for prognosis because of the long followup periods required. After the introduction of a simple technique for the preparation of cell suspensions from paraffin-embedded tissue by Hedley et al. [15], it became possible to perform retrospective studies on archival material from a series of patients with adequate follow-up. In several tumor tissues flow cytometry was performed with different results: in carcinomas of the breast, the lung, the bladder, and the colon, high percentages of
0 ’
I
I
I
I
I
I
20
40
60
80
100
120
1 140 month
FIG. 2. Corrected survival curve with negative (N) and positive (P) nodes.
aneuploidy were found [ 161, although no independent effect on prognosis was shown [15]. In studies conceming gynecologic tumors, a relation was found between DNA-ploidy level and survival for 203 patients with endometrial carcinoma stage I [17] and for 110 patients with cancer of the ovaries [18]. Jacobsen et al. [19] found an evident relation between the percentage of aneuploidy in preoperative biopsies and the recurrence rate in patients with low-stage carcinoma of the cervix. Most of these patients were treated with radiotherapy. In a smaller subgroup of 21 women treated with radical hysterectomy, more positive lymph nodes were seen in patients with aneuploid tumors (none vs 50%) [20]. As seen in Table 1 this relation could not be reproduced in our study. They also found a relation between ploidy grade and severity of CIN: 0% aneuploidy in normal or mildly dysplastic tissue vs 78% aneuploidy in biopsies with severe dysplasia [21], according to the results of Dudzinski et al. [22]. In a follow-up study [23] of 100 patients with CIN the aneuploid lesions seemed to be persistent or even progressive in a significantly higher percentage compared with diploid lesions. Flow cytometry was used in an attempt to predict ra-
0.6
0
20
40
60
80
100
120
140 month
FIG. 3. Corrected survival curve with aneuploid (A), periploid (P), and diploid (D) tumors.
184
KENTER ET AL.
diocurability [24,25], but no correlation was found between tissue healing response and ploidy grade [26]. Comparison of DNA content with other histopathologic features in patients with cervical carcinoma treated with radiotherapy [27-291 showed very heterogenic relations. The percentage of aneuploid tumors in our study (49%) is far lower than that in Jacobsen’s report [20] concerning low-stage carcinoma of the uterine cervix (80%), but agrees better with the results of Dyson er al. [30], who found 62% in stages IB and IIA, and Strang et al. [31], who found 41%. Jacobsen was the only one who carried out flow cytometry on fresh material, which might explain the difference in results. The only significant relation in the study of Dyson et al. [30] was a higher percentage of lymph node metastases in aneuploid tumors compared with diploid (9.8% vs 1.7%). Follow-up was too short to relate to prognosis in this study. Davis et al. [32] studied 56 cases with cervical carcinoma stages I-III treated with radiotherapy and a follow-up of 2 years. As in the present study, no correlation with prognosis was found. Rutgers et al. [33] came to the same conclusions in a comparable study among 112 women, although they mentioned a relation between a combination of DNA-ploidy and histologic grade of the tumor. There is no study in the literature that actually resembles the present one: patients with a minimal follow-up of a5 years treated with radical hysterectomy and lymphadenectomy. In most of the studies flow cytometry was carried out in biopsies from patients treated with radiotherapy. In these cases the pretreatment lymph node status was not known. In our center we prefer surgical treatment for lower-stage carcinomas especially in the younger patients, because of the preservation of ovarian function. Because in this procedure a proper lymphadenectomy is carried out, good insight is given into the possible spread of the tumor. From the present study and the literature it is well known that the presence of lymph node metastases has the strongest effect on survival. The results of DNA flow cytometry in this study give no information about prognosis in terms of either independent or additional effects on survival or relation to other prognostic factors. The conclusion of this study is that in low-stage carcinoma of the uterine cervix DNA-ploidy level is of no extra prognostic value. REFERENCES 1. Hoogendoom, D. Trends in kankersterfte, Ned. Tijdschr. Geneeskd. 127, 1661-1668 (1983). 2. Hoskins, W. J., Ford, J. H., Lutz, M. H., and Averett, H. E. Radical hysterectomy and pelvic lymphadenectomy for the management of early invasive cancer of the cervix, Gynecof. Oncol. 4, 278-290 (1976). 3. Lee, Y., Wang, K., Lin, M., Liu, C., Wang, K., Lan, C.,
Chuang, J., Chen, A., and Wu, C. Radical hysterectomy with pelvic lymph node dissection for treatment of cervical cancer: A clinical review of 954 cases, Gynecol. 0~02. 32, 135-142 (1989). 4. Fuller, A. F., Elliot, N., Kosloff, C., and Lewis, J. L. Lymph node metastases from carcinoma of the cervix, stage IB and HA: Implications for the prognosis and treatment, Gynecol. Oncol. 13, 165-174 (1982). 5. Burghardt, E., Pickel, H., Haas, J., and Lahousen, M. Prognostic factors and operative treatment of stages IB to IIB cervical cancer, Amer. J. Obstet. Gynecol. 156, 986-988 (1987). 6. Boyce, J. G., Fruchter, R. G., Nicastri, A. D., Deregt, R. H., Ambiavagar, P. C., Reinis, M., Macasaet, M., and Rotman, M. Vascular invasion in stage I carcinoma of the cervix, Cancer 53, 1175-1180 (1984). 7. Kenter, G. G., Ansink, A. C., Heintz, A. P. M., Aartsen, E. J., Delemarre, J. F. M., and Hart, A. A. M. Carcinoma of the uterine cervix stage I and IIA: Results of surgical treatment: Complications, recurrence and survival, Eur. J. Surg. Oncol. 15, 55-60 (1989). 8. Einhom, N., Patek, E., and Sjoberg, B. Outcome of different treatment modalities in cervix carcinoma stage IB and IIA, Cancer 55, 949-955 (1985). 9. Barter, J. F., Soong, S. J., Shingleton, H. M., Hatch, K. D., and Orr, J. W. Complications of combined radical hysterectomy-postoperative radiation therapy in women with early stage cervical cancer, Gynecol. Oncol. 32, 292-2% (1989). 10. Hedley, D. W., Friedlander, M. L., Taylor, I. W., Rugg, C. A., and Musgrove, E. A. Method for analysis of cellular DNA content of pat&in-embedded pathological material using flow cytometry, J. Histochem. Cytochem. 31, 1333-1335 (1983). 11. Koss, L. G., Czemiak, B., Herz, F., and Wersto, R. P. Flow cytometric measurements of DNA and other cell components in human tumors: A critical appraisal, Hum. Puthol. 20, 528-548 (1989). 12. Kenter, G. G., Ansink, A., Aartsen, E., Delemarre, J., Hart, G., and Heintz, A. P. M. Low stage invasive carcinoma of the uterine cervix stage I-IIA. Morphological prognostic factors, Eur. J. Surg. Oncol. 14, 187-192 (1988). 13. Schutte, B., Reynders, M. M. J., Bosman, F. T., and Bkijham, G. Flow cytometric determination of DNA ploidy level in nuclei isolated from paraffin embedded tissue, Cytometry 6, 26-30 (1985). 14. Matthews, D. E., and Farewell, V. T. Using and understanding medical statistics, Karger, Base1 (1985). 15. Hedley, D. W., Friedlander, M. L., and Taylor, I. W. Application of DNA flow cytometry to paraffin-embedded archival material for the study of aneuploidy and its clinical significance, Cytometry 6, 327-333 (1985). 16. Barlogie, B., Raber, M. N., Schumann, J., Johnson, T. S., Drewinko, B., Swartzendruber, D. E., Gohde, W., Andreeff, M., and Freireich, E. J. Flow cytometry in clinical cancer research, Cuncer Res. 43, 3982-3997 (1983). 17. B&ton, L. C., Wilson, T. O., Gaffey, T. A., Lieber, M. M., Wieand, H. S., and Podratz, K. Flow cytometric DNA analysis of stage I endometrial carcinoma, Gynecol. Oncol. 34, 317-322 (1989). 18. Rodenburg, C. J., Comelisse, C. J., Heintz, A. P. M., Hermans, J., and Fleuren, G. J. Tumour ploidy as a major prognostic factor in advanced ovarian cancer, Cancer 59, 317-323 (1987). 19. Jacobsen, A., Bichel, P., and Vaeth, M. New prognostic factors in squamous cell carcinoma of the cervix uteri, Amer. J. Cfin. Oncol. 8, 39-43 (1985).
DNA PLOIDY IN CERVICAL 20. Jacobsen, A. Ploidy level and short time prognosis of early cervix cancer, Radiother. Oncol. 1, 271-275 (1984). 21. Jacobsen, A., Kristinsen, P. B., and Paulsen, H. K. Flow cytometric classification of biopsy specimens from cervical intraepithelial neoplasia, Cyrometry 4, 166-169 (1983).
28*
22. Dudzinski, M. R., Haskill, S. J., Fowler, W. C., Currie, J. L., and Walton, L. A. DNA content in cervical neoplasia and its relationship to prognosis, Obsret. Gynecol. 69, 373-377 (1987).
29 ’
23. Fu, Y. S., Reagan, J. W., and Richart, R. M. Definition of precursors, Gynecol. Oncol. 12, S220-S231 (1981).
30.
24. Dyson, J. E. D., Joslin, C. A. F., Rothwell, R. I., Quirke, P., Khoury, G. G., and Bird, C. C. Flow cytofluorometric evidence for the differential radioresponsiveness of aneuploid and diploid cervix turnours, Radiother. Oncol. 8, 263-272 (1987).
31.
25. Peters, L. J., Brock, W., and Johnson, T. Predicting radiocurability, Cancer 55, 2118-2122 (1985).
32.
26. Jacobsen, A. Prognostic impact of ploidy level in carcinoma of the cervix, Amer. J. Clin. Oncol. 7, 475-480 (1984).
33
27. Johnson, T. S., Adelsen, M. D., Sneige, N., Williamson, K. D., Lee, A. M., and Katz, R. Cervical carcinoma DNA content, S
CANCER
185
fraction, and malignancy grading. I. Interrelationships, Gynecol. Oncol. 26, 41-56 (1987). Adelson, M. D., Johnson, T. S., Sneige, N., Williamson, K. D., Freedman, R. S., and Peters, L. J. Cervical carcinoma DNA content, S-fraction, and malignancy grading. II. Comparison with clinical staging, Gynecol. Oncol. 26, 57-70 (1987). Strang, P., Lindgren, A., Frankendal, B., and Stendahl, U. DNA patterns and aggressive histopathologic features in 159 patients with cervical carcinoma, Znt. J. Gynecol. Pathol. 7, 56-63 (1987). Dyson, J. E. D., Joslin, C. A. F., Quirke, P., and Bird, C. C. Flow cytofluorometric analysis of serial biopsies of tumour of the uterine cervix, Eur. J. Cancer Clin. Oncol. 20, 1249-1259 (1984). Strang, P., Stendahl, U., Frankendahl, B., and Lindgren, A. Flow cytometric DNA patterns in cervical carcinoma, Acta Radiol. Oncol. 25, 249-254 (1985). Davis, J. R., Aristizabal, S., Way, D. L., Weiner, S. A., Hicks, M. J., and Hagaman, R. M. DNA ploidy, grade, and stage in prognosis of uterine cervical cancer, Gynecol. Oncol. 32, 4-7 (1989). Rutgers, D. H., and Peperzeel, H. A. Prognostic relevance of DNA-index of human cervical carcinomas, in Clinical cytometry and hisfometry (G. Burger, J. S. Ploem, and K. Goertler, Eds.), Academic Press, New York/London, pp. 365-368 (1987).
ONCOLOGY
39, 181-185 (1%))
DNA Ploidy Level as Prognostic Factor in Low Stage Carcinoma of the Uterine Cervix G. G.
KENTER,*
C. J.
CORNELISSE,~
E. J.
AARTSEN,$ W. MOOY,§ AND G. J. FLEURENt
J.
HERMANS,~~
A. P. M. HEINTZ#
Departments of *Gynaecology, TPathology, and [[Medical Statistics of the University of Leiden and the Departments of SGynaecology and SPathology of the Antoni van Leeuwenhoek huis of the Netherlands Cancer Institute Received
March 26, 1990
Plow cytometry was used to measure DNA content of tumor cells in paraffin-embedded archival material from 89 patients with squamous cell carcinoma of the cervix uteri stages IB and DA. Patients were all treated with radical hysterectomy and transperitoneal lymphadenectomy. Twenty-four percent received radiotherapy postoperatively because of tumor spread into parametria or positive lymph nodes. Ploidy grade was compared to other prognostic factors. DNA-aneuploidy was seen in 44 (49%), DNA-diploidy in 16 (18%) and 29 (33%) of the tumors were DNA-periploid. Sixty-nine (78%) patients were in PIG0 stage IB, 20 (22%) in DA. In 19 (21%) lymph node mebstases were found, vasoinvasion in 25 (28%). Overall disease-specific S-year survival was 80%. There was a significant effect of positive lymph nodes (90% vs 47%) and vasoinvasion (87% vs 64%) on the 5year survival (resp. P < 0.01 and P = 0.02). No correlation was found between stage (81% for IB vs 80% for DA) or DNA-ploidy grade (81% for aneuploidy, 82% for periploidy, 79% for diploidy) and survival (req. P = 0.9 and P = 0.8). Ploidy grade was equally divided over other prognostic factors. In skpwise Cox regression analysis DNA-ploidy grade showed no independent effect on survival. We conclude that DNA flow cytometry in this ~eadan~ mess, material was of no additional prognustic value. Q 1990 Inc.
INTRODUCTION Although the incidence of cancer of the cervix uteri has decreased, probably because of better screening programs, there is growing evidence that cervical cancer is nowadays more often seen in younger women [l]. Cervical cancer is the second cause of death in women under the age of 50. The results of surgical treatment in stages IB and IIA ‘ Present Utrecht.
address:
Department
of
Gynaecology,
University
of
are good; the 5-year survival is usually more than 85% [2,3]. Still about 15-20% of these patients have to be treated with radiotherapy after radical hysterectomy because of lymph node metastases or tumor growth into the parametria. These parameters are known to worsen the prognosis [4]. Like other prognostic factors, such as tumor size [5] and vasoinvasion [6], these can be measured only postoperatively. Complications of surgical treatment [7] and radiotherapy [8] are considerable, and those of combined treatment are even worse [9]. This indicates the need for a more precise quantitative method in the prediction of prognosis to be able to individualize treatment as much as possible. With the introduction of DNA flow cytometry on tumor tissue a new method has become available that is possibly of prognostic importance. With this method DNA content of nuclei of tumor cells can be measured in fresh as well as archival paraffin-embedded tissue samples [lo]. Reports have been published in which tumor DNA-ploidy seemed to be of prognostic value with respect to survival for several (gynecologic) malignancies [l 11. This report evaluates the prognostic value of DNA flow cytometry in a series of patients with stage I and IIA cervical cancer. MATERIALANDMETHOD Paraffin-embedded archival material from 213 patients treated between 1%7 and 1981 for a carcinoma of the uterine cervix stage I or IIA was available. All patients were treated by means of radical hysterectomy with transperitoneal lymphadenectomy in the Antoni van Leeuwenhoek Hospital of The Netherlands Cancer Institute. Preoperative staging was carried out according to FIGO. Details about the results of the surgical pro-
181
0090-8258BO $1.50 Copyright 8 l!EJOby Academic Press, Inc. All rights of reproduction in any form reserved.
182
KENTER ET AL.
FIG. 1. Examples of a diploid (A), peridiploid (B), and aneuploid (C) histogram.
cedure, the complications, and the survival have been published previously [7]. Histological material was reviewed as discussed elsewhere [12]. There were 106 cases with squamous cell carcinoma of the cervix, in which there was sufficient tumor tissue in the surgical material to perform the procedure, selected for flow cytometry. A 40-pm section was cut. Sections were deparafhnized and prepared as described by Hedley et al. TABLE 1 Relation betweenStage,Node Status, Vasoinvasion,and Ploidy Grade Diploid Stage IB IIA Nodes Negative Positive Vasoinvasion Negative Positive
Periploid
Aneuploid
N
14 2
21 8
34 10
69 20
13 3
21 8
36 8
70 19
11 5
21 8
32 12
64 25
16
29
44
89
1101.Samples were measured on a ICP 22 flow cytometer (Ortho, Westwood, MA). Tumor DNA content was classified as aneuploid if the DNA profile showed two or more distinct G 0.1 peaks. (Fig. 1). The resolution of the DNA profiles obtained from deparaffinized tissue samples frequently shows a larger range of coefficient of variation (CV) values than those from fresh tumor tissue which interferes with the minimal detectable DNA content difference [13]. A high CV may obscure the presence of a low aneuploid population (e.g., with a CV > 5% a split peak is seen only when the DNA content difference between two populations is at least 10%). According to Hedley et al. [lo] we defined tumors with a single G 0.1 peak with a CV 2 5.5% as a different (peridiploid) group, to exclude the possibility that low aneuploid tumors with DNA differences in the lo-20% range might be overlooked. DNA histograms were classified as DNA-diploid, -aneuploid, or -peridiploid without knowledge of patient survival or clinicopathologic features. Follow-up for all patients was at least 60 months. Survival was measured from the time of surgery to death due to disease. Statistical analysis was carried out using the BMDP package. Survival curves were compared with the log-rank test. A stepwise Cox regression model [141 was used to evaluate the effect of ploidy grade on survival additional to stage, vasoinvasion, and lymph node metastases. RESULTS Evaluable DNA profiles were obtained for 89 of 106 selected patients. Seventeen profiles could not be evaluated, probably because of the advanced age of the blocks (from before 1973). To see if the possibility of flow cytometry itself had any impact on survival rates we compared the results of 89 patients in whom flow cytometry was performed with the rest of the group. No differences of statistical interest were found. There were 69 (78%) stage IB patients and 20 (22%) stage IIA patients. Lymph node metastases were present in 19 (21%) patients. Postoperatively, 21 (24%) patients were treated with radiotherapy because of tumor spread into the parametria and/or positive lymph nodes. Vasoinvasion was found in 25 (28%) patients. Flow cytometry showed DNA-diploid profiles in 16 (18%), peridiploid histograms in 29 (33%), and 44 (49%) DNA-aneuploid tumors. The relation between ploidy grade and the other prognostic variables is shown in Table 1. The worst prognosis categories (positive lymph nodes and vasoinvasion) are more or less equally present in the three ploidy classes. Overall disease-specific 5-year survival was 80%. Table 2 shows the effect of several parameters on survival rates. In the case of lymph node metastases the 5-year
DNA PLOIDY IN CERVICAL TABLE Effect
183
CANCER
2
of Node Status, Vaseinvasion,Stage,and Ploidy Grade on S-YearSurvival
Nodes Negative Positive Vasoinvasion Negative Positive Stage IB IIA Ploidy grade Diploid Periploid Aneuploid Total group
N
Recurrence (4
S-year survival m
70 19
7 10
90 47
co.01
64 25
8 9
87.5 64
0.17
69 20
13 4
81 80
0.8
16 29 44
3 5 9
81 82 79
0.9
P value
80
89
survival rate dropped from 90% without to 47% with positive nodes (P < 0.01) (Fig. 2); vasoinvasion decreased the 5-year survival from 87% without to 64% with invasion (P = 0.02). The stage at time of surgery had no effect on survival: 81% in stage IB and 80% in stage IIA (P = 0.8). Ploidy grade showed no statistical differences in survival rates: 81% for patients with diploid tumors, 82% for those with periploid tumors, and 79% in cases with aneuploidy (P = 0.9) (Fig. 3). In the stepwise Cox regression analysis the effects of the following parameters on survival were studied: lymph node metastases, vasoinvasion, stage, and ploidy grade. The presence of lymph node metastases had by far the strongest effect on survival. The other parameters (vasoinvasion, stage, and ploidy grade) turned out to have no additional prognostic value. DISCUSSION With the development of flow cytometry it has become possible to rapidly measure DNA content of tens of thousands of cells within a short amount of time. Initially only fresh tumor material could be analyzed and this was of little use in the establishment of the value of DNA flow cytometry for prognosis because of the long followup periods required. After the introduction of a simple technique for the preparation of cell suspensions from paraffin-embedded tissue by Hedley et al. [15], it became possible to perform retrospective studies on archival material from a series of patients with adequate follow-up. In several tumor tissues flow cytometry was performed with different results: in carcinomas of the breast, the lung, the bladder, and the colon, high percentages of
0 ’
I
I
I
I
I
I
20
40
60
80
100
120
1 140 month
FIG. 2. Corrected survival curve with negative (N) and positive (P) nodes.
aneuploidy were found [ 161, although no independent effect on prognosis was shown [15]. In studies conceming gynecologic tumors, a relation was found between DNA-ploidy level and survival for 203 patients with endometrial carcinoma stage I [17] and for 110 patients with cancer of the ovaries [18]. Jacobsen et al. [19] found an evident relation between the percentage of aneuploidy in preoperative biopsies and the recurrence rate in patients with low-stage carcinoma of the cervix. Most of these patients were treated with radiotherapy. In a smaller subgroup of 21 women treated with radical hysterectomy, more positive lymph nodes were seen in patients with aneuploid tumors (none vs 50%) [20]. As seen in Table 1 this relation could not be reproduced in our study. They also found a relation between ploidy grade and severity of CIN: 0% aneuploidy in normal or mildly dysplastic tissue vs 78% aneuploidy in biopsies with severe dysplasia [21], according to the results of Dudzinski et al. [22]. In a follow-up study [23] of 100 patients with CIN the aneuploid lesions seemed to be persistent or even progressive in a significantly higher percentage compared with diploid lesions. Flow cytometry was used in an attempt to predict ra-
0.6
0
20
40
60
80
100
120
140 month
FIG. 3. Corrected survival curve with aneuploid (A), periploid (P), and diploid (D) tumors.
184
KENTER ET AL.
diocurability [24,25], but no correlation was found between tissue healing response and ploidy grade [26]. Comparison of DNA content with other histopathologic features in patients with cervical carcinoma treated with radiotherapy [27-291 showed very heterogenic relations. The percentage of aneuploid tumors in our study (49%) is far lower than that in Jacobsen’s report [20] concerning low-stage carcinoma of the uterine cervix (80%), but agrees better with the results of Dyson er al. [30], who found 62% in stages IB and IIA, and Strang et al. [31], who found 41%. Jacobsen was the only one who carried out flow cytometry on fresh material, which might explain the difference in results. The only significant relation in the study of Dyson et al. [30] was a higher percentage of lymph node metastases in aneuploid tumors compared with diploid (9.8% vs 1.7%). Follow-up was too short to relate to prognosis in this study. Davis et al. [32] studied 56 cases with cervical carcinoma stages I-III treated with radiotherapy and a follow-up of 2 years. As in the present study, no correlation with prognosis was found. Rutgers et al. [33] came to the same conclusions in a comparable study among 112 women, although they mentioned a relation between a combination of DNA-ploidy and histologic grade of the tumor. There is no study in the literature that actually resembles the present one: patients with a minimal follow-up of a5 years treated with radical hysterectomy and lymphadenectomy. In most of the studies flow cytometry was carried out in biopsies from patients treated with radiotherapy. In these cases the pretreatment lymph node status was not known. In our center we prefer surgical treatment for lower-stage carcinomas especially in the younger patients, because of the preservation of ovarian function. Because in this procedure a proper lymphadenectomy is carried out, good insight is given into the possible spread of the tumor. From the present study and the literature it is well known that the presence of lymph node metastases has the strongest effect on survival. The results of DNA flow cytometry in this study give no information about prognosis in terms of either independent or additional effects on survival or relation to other prognostic factors. The conclusion of this study is that in low-stage carcinoma of the uterine cervix DNA-ploidy level is of no extra prognostic value. REFERENCES 1. Hoogendoom, D. Trends in kankersterfte, Ned. Tijdschr. Geneeskd. 127, 1661-1668 (1983). 2. Hoskins, W. J., Ford, J. H., Lutz, M. H., and Averett, H. E. Radical hysterectomy and pelvic lymphadenectomy for the management of early invasive cancer of the cervix, Gynecof. Oncol. 4, 278-290 (1976). 3. Lee, Y., Wang, K., Lin, M., Liu, C., Wang, K., Lan, C.,
Chuang, J., Chen, A., and Wu, C. Radical hysterectomy with pelvic lymph node dissection for treatment of cervical cancer: A clinical review of 954 cases, Gynecol. 0~02. 32, 135-142 (1989). 4. Fuller, A. F., Elliot, N., Kosloff, C., and Lewis, J. L. Lymph node metastases from carcinoma of the cervix, stage IB and HA: Implications for the prognosis and treatment, Gynecol. Oncol. 13, 165-174 (1982). 5. Burghardt, E., Pickel, H., Haas, J., and Lahousen, M. Prognostic factors and operative treatment of stages IB to IIB cervical cancer, Amer. J. Obstet. Gynecol. 156, 986-988 (1987). 6. Boyce, J. G., Fruchter, R. G., Nicastri, A. D., Deregt, R. H., Ambiavagar, P. C., Reinis, M., Macasaet, M., and Rotman, M. Vascular invasion in stage I carcinoma of the cervix, Cancer 53, 1175-1180 (1984). 7. Kenter, G. G., Ansink, A. C., Heintz, A. P. M., Aartsen, E. J., Delemarre, J. F. M., and Hart, A. A. M. Carcinoma of the uterine cervix stage I and IIA: Results of surgical treatment: Complications, recurrence and survival, Eur. J. Surg. Oncol. 15, 55-60 (1989). 8. Einhom, N., Patek, E., and Sjoberg, B. Outcome of different treatment modalities in cervix carcinoma stage IB and IIA, Cancer 55, 949-955 (1985). 9. Barter, J. F., Soong, S. J., Shingleton, H. M., Hatch, K. D., and Orr, J. W. Complications of combined radical hysterectomy-postoperative radiation therapy in women with early stage cervical cancer, Gynecol. Oncol. 32, 292-2% (1989). 10. Hedley, D. W., Friedlander, M. L., Taylor, I. W., Rugg, C. A., and Musgrove, E. A. Method for analysis of cellular DNA content of pat&in-embedded pathological material using flow cytometry, J. Histochem. Cytochem. 31, 1333-1335 (1983). 11. Koss, L. G., Czemiak, B., Herz, F., and Wersto, R. P. Flow cytometric measurements of DNA and other cell components in human tumors: A critical appraisal, Hum. Puthol. 20, 528-548 (1989). 12. Kenter, G. G., Ansink, A., Aartsen, E., Delemarre, J., Hart, G., and Heintz, A. P. M. Low stage invasive carcinoma of the uterine cervix stage I-IIA. Morphological prognostic factors, Eur. J. Surg. Oncol. 14, 187-192 (1988). 13. Schutte, B., Reynders, M. M. J., Bosman, F. T., and Bkijham, G. Flow cytometric determination of DNA ploidy level in nuclei isolated from paraffin embedded tissue, Cytometry 6, 26-30 (1985). 14. Matthews, D. E., and Farewell, V. T. Using and understanding medical statistics, Karger, Base1 (1985). 15. Hedley, D. W., Friedlander, M. L., and Taylor, I. W. Application of DNA flow cytometry to paraffin-embedded archival material for the study of aneuploidy and its clinical significance, Cytometry 6, 327-333 (1985). 16. Barlogie, B., Raber, M. N., Schumann, J., Johnson, T. S., Drewinko, B., Swartzendruber, D. E., Gohde, W., Andreeff, M., and Freireich, E. J. Flow cytometry in clinical cancer research, Cuncer Res. 43, 3982-3997 (1983). 17. B&ton, L. C., Wilson, T. O., Gaffey, T. A., Lieber, M. M., Wieand, H. S., and Podratz, K. Flow cytometric DNA analysis of stage I endometrial carcinoma, Gynecol. Oncol. 34, 317-322 (1989). 18. Rodenburg, C. J., Comelisse, C. J., Heintz, A. P. M., Hermans, J., and Fleuren, G. J. Tumour ploidy as a major prognostic factor in advanced ovarian cancer, Cancer 59, 317-323 (1987). 19. Jacobsen, A., Bichel, P., and Vaeth, M. New prognostic factors in squamous cell carcinoma of the cervix uteri, Amer. J. Cfin. Oncol. 8, 39-43 (1985).
DNA PLOIDY IN CERVICAL 20. Jacobsen, A. Ploidy level and short time prognosis of early cervix cancer, Radiother. Oncol. 1, 271-275 (1984). 21. Jacobsen, A., Kristinsen, P. B., and Paulsen, H. K. Flow cytometric classification of biopsy specimens from cervical intraepithelial neoplasia, Cyrometry 4, 166-169 (1983).
28*
22. Dudzinski, M. R., Haskill, S. J., Fowler, W. C., Currie, J. L., and Walton, L. A. DNA content in cervical neoplasia and its relationship to prognosis, Obsret. Gynecol. 69, 373-377 (1987).
29 ’
23. Fu, Y. S., Reagan, J. W., and Richart, R. M. Definition of precursors, Gynecol. Oncol. 12, S220-S231 (1981).
30.
24. Dyson, J. E. D., Joslin, C. A. F., Rothwell, R. I., Quirke, P., Khoury, G. G., and Bird, C. C. Flow cytofluorometric evidence for the differential radioresponsiveness of aneuploid and diploid cervix turnours, Radiother. Oncol. 8, 263-272 (1987).
31.
25. Peters, L. J., Brock, W., and Johnson, T. Predicting radiocurability, Cancer 55, 2118-2122 (1985).
32.
26. Jacobsen, A. Prognostic impact of ploidy level in carcinoma of the cervix, Amer. J. Clin. Oncol. 7, 475-480 (1984).
33
27. Johnson, T. S., Adelsen, M. D., Sneige, N., Williamson, K. D., Lee, A. M., and Katz, R. Cervical carcinoma DNA content, S
CANCER
185
fraction, and malignancy grading. I. Interrelationships, Gynecol. Oncol. 26, 41-56 (1987). Adelson, M. D., Johnson, T. S., Sneige, N., Williamson, K. D., Freedman, R. S., and Peters, L. J. Cervical carcinoma DNA content, S-fraction, and malignancy grading. II. Comparison with clinical staging, Gynecol. Oncol. 26, 57-70 (1987). Strang, P., Lindgren, A., Frankendal, B., and Stendahl, U. DNA patterns and aggressive histopathologic features in 159 patients with cervical carcinoma, Znt. J. Gynecol. Pathol. 7, 56-63 (1987). Dyson, J. E. D., Joslin, C. A. F., Quirke, P., and Bird, C. C. Flow cytofluorometric analysis of serial biopsies of tumour of the uterine cervix, Eur. J. Cancer Clin. Oncol. 20, 1249-1259 (1984). Strang, P., Stendahl, U., Frankendahl, B., and Lindgren, A. Flow cytometric DNA patterns in cervical carcinoma, Acta Radiol. Oncol. 25, 249-254 (1985). Davis, J. R., Aristizabal, S., Way, D. L., Weiner, S. A., Hicks, M. J., and Hagaman, R. M. DNA ploidy, grade, and stage in prognosis of uterine cervical cancer, Gynecol. Oncol. 32, 4-7 (1989). Rutgers, D. H., and Peperzeel, H. A. Prognostic relevance of DNA-index of human cervical carcinomas, in Clinical cytometry and hisfometry (G. Burger, J. S. Ploem, and K. Goertler, Eds.), Academic Press, New York/London, pp. 365-368 (1987).
