M sive primary malignant uterine tumors. Fortunately, they represent

Transcription

1 2543 Prognostic Parameters for Survival of Patients with Malignant Mesenchymal Tumors of the Uterus Marin Nola, M.D., Ph.o. Damir Babic, M.o. Jadranka Ilic, M.D. Matko MaruSic, M.D., Ph.D. Branka Uiarevic, Ph.0. Mladen PetroveEki, M.D., Ph.o.* Ante Sabioncello, ~ h.0.~ Draien KovaE, M.D., Ph.D.4 Stanko Jukic, M.D., Ph.D. Department of Gynecological and Perinatal Pathology, Zagreb University School of Medicine, Zagreb, Croatia. Department of Clinical Laboratory Diagnosis, Zagreb University Hospital Center and School of Medicine, Zagreb, Croatia. Institute of Immunology, Zagreb, Croatia Department of Pathology, Rijeka University School of Medicine, Rijeka, Croatia. The authors thank Dr. Ana MaruSii: for her critical review of the article and Ms. iaklina Cavar for her expert technical assistance. Address for reprints: Marin Nola, M.D., Ph.D., Department of Gynecological and Perinatal Pathology, Zagreb University School of Medicine, Petrova 13, Zagreb, Croatia. Received June 11, 1996; revision received August 26, 1996; accepted August 26, BACKGROUND. Malignant mesenchymal uterine neoplasms are the most aggressive type of primary uterine tumors, with most patients dying within a few years of diagnosis. Thus, it would be very important to define prognostic factors for predicting the malignancy potential of at least some of their subtypes. METHODS. Flow cytometric cell cycle analysis (proliferative activity, DNA ploidy, and DNA index) was performed on archival paraffin embedded blocks from 80 patients with malignant mesenchymal uterine neoplasms (endometrial stromal sarcomas, malignant smooth muscle tumors, and malignant Mullerian mixed tumors). The Cox proportional hazards regression model was used to assess relative effects of the following factors on patient survival: clinical stage, mode of therapy, DNA+proliferative activity, DNA index, histologic type, cellularity, degree of atypia, mitotic activity, and depth of myometrial invasion. RESULTS. There were 9 low grade stromal sarcomas, 17 high grade stromal sarcomas, 8 smooth muscle neoplasms with uncertain malignant potential, 23 leiomyosarcomas, and 16 homologous and 7 heterologous malignant Mullerian mixed tumors. In univariate analysis for stromal sarcomas, statistical significance was found for DNA ploidy+proliferative activity (P i 0.001), histologic type (P = 0.005), and DNA index (P < 0.001). In multivariate analysis, DNA index appeared to be the only significant parameter influencing patient survival (P = 0.005). In univariate analysis for malignant smooth muscle neoplasms, statistical significance was detected for mitotic activity (P = 0.049) and International Federation of Gynecology and Obstetrics classification (P = 0.021), but in multivariate analysis, clinical stage appeared to be the only significant parameter influencing patient survival (P = 0.032). In univariate analysis for malignant Mullerian mixed tumors, statistical significance was found for the depth of myometrial invasion (P = 0.039), DNA index (P = 0.037), and clinical stage (P = 0.013), but in multivariate analysis, only the depth of myometrial invasion (P = 0.036) and clinical stage (P = 0.025) were of statistical significance. CONCLUSIONS. The most powerful prognostic indicator for stromal sarcomas was the DNA index, for malignant smooth muscle neoplasms it was the clinical stage, and for malignant Mullerian mixed tumors it was the depth of myometrial invasion. Cancer 1996; American Cancer Society. KEYWORDS: DNA index, DNA ploidy, endometrial stromal sarcoma, flow cytometry, malignant Miillerian mixed tumors, malignant smooth muscle neoplasm, proliferative activity, uterine sarcoma. alignant mesenchymal tumors of the uterus are the most aggres- M sive primary malignant uterine tumors. Fortunately, they represent only 3-5% of all malignant uterine neoplasms. The most important diagnostic method is curettage of the uterus, but very often the diagnosis is made histologically after surgery for uterine myoma. There have been reports on various factors indicating biologic behav American Cancer Society

2 2544 CANCER December 15, 1996 / Volume 78 / Number 12 ior of the tumors in patients with mesenchymal uter- The degree of atypia was scored as minimal, modine neoplasm: stage of disease,'-4 degree of at~pia,~.~ erate, and marked. The assignment of the degree of cell~larity,~,~ depth of myometrial inva~ion,~,~ mitotic atypia was based on an assessment of nuclear pleolymphatic and vascular space morphism, nuclear size, chromatin density, and the presence of heterologous elements in the t~mor,~'~ size, prominence, and configuration of nucleoli.5 presence of coagulative tumor cell necrosis,5 and concentration of estrogen and progesterone receptors.'~" DNA ploidy, proliferative activity, and DNA index of the tumor cells have been rarely investigated.*-l' The prognostic significance of these tests in patients with malignant mesenchymal tumors of the uterus was the subject of the current study. PATIENTS AND METHODS The study included 99 patients with endometrial stroma1 sarcomas, malignant smooth muscle neoplasms, and malignant Mullerian mixed tumors (MMMT) treated in the study department from 1978 to Only the patients who had not been treated prior to surgery were analyzed. Two patients with insufficient tissue for flow cytometry and 17 patients with insufficient data were excluded from further analysis. Analysis of clinical stage, mode of therapy, DNA content (DNA ploidy, proliferative activity, and DNA index), and histologic review of tumor histologic type, cellularity, degree of atypia, mitotic activity, and depth of myometrial invasion (for MMMT) were performed for 80 patients. When there were discrepancies in the histologic assessment between the two study pathologists, a third pathologist was consulted. Clinical stages of the disease were defined according to the 1988 International Federation of Gynecology and Obstetrics (FIGO) classification; Stages I and I1 were considered favorable and Stages I11 and IV unfavorable." Treatment was comprised of primary total abdominal hysterectomy and bilateral salpingooophorectomy with adjuvant radiation or chemotherapy. For the purpose of this study, the patients were divided into two groups: one that received surgical therapy only and one that received surgical and adjuvant therapy. Endometrial stromal sarcomas were classified into low and high grade according to their resemblance to stromal cells of proliferative phase endometrium.lj According to the mitotic number and degree of atypia, malignant smooth muscle neoplasms were divided into leiomyosarcomas and smooth muscle neoplasms with uncertain malignant p~tential.'~ MMMT were divided into homologous and heterologous tumors, depending on existing mesenchymal elements3 Cellularity was evaluated by comparing the tumor tissue with the adjacent nonneoplastic myometrium. When the cellularity was increased, it was differentiated into minimal, moderate, and marked.5 Smooth muscle cells in tumors with minimal atypia showed minimal variation in nuclear size and shape. Nucleoli were small. Cells in tumors with moderate atypia showed a greater degree of variation in nuclear size and shape when compared with tumors with minimal atypia. Nucleoli were occasionally prominent. Tumors with marked atypia were obviously pleomorphic and comprised of numerous cells with enlarged hizarre nuclei with dense coarse chromatin and prominent nucleoli.5 The number of mitotic figures per 10 high-power fields was determined in the most active areas of the tumor. For the endometrial stromal sarcomas and MMMT, only two groups were formed according to mitotic figures. Tumors showing fewer than 10 mitoses per 10 high-power fields were included in the first group, and those with 10 or more mitoses per 10 highpower fields in the second group. Malignant smooth muscle tumors showing 3 to 4 mitoses per 10 highpower fields were included in the first group, those with 5 to 9 mitoses per 10 high-power fields in the second group, and those with 10 or more mitoses per 10 high-power fields in the third gro~p.'~ Preparation of Specimens for Flow Cytometric Analysis Paraffin embedded specimens from uterine sarcomas removed during surgical therapy were processed by the technique described by Hedley et al.i5 In addition to 3 40-pm thick sections required for flow cytometry, 2 4-pm thick sections were cut, one from the front and one from the back of the tissue block. They were stained with hematoxylin and eosin, and the pathohistologic parameters were evaluated. The tissue was deparaffinized with two changes of xylene and rehydrated in a sequence of decreasing ethanol concentrations followed by a final wash in distilled water. Tissue was then suspended in 2 ml of 0.5% pepsin (Sigma, St. Louis, MO) in 0.9% sodium chloride (ph 1.5) and incubated in a 37 "C water bath for 2 hours. Cells were filtered through a 42-pm mesh and centrifuged 5 minutes at 800 x G. One hundred microliters of citrate buffer and 900 pl trypsin solution were added to the pellet and incubated for 10 minutes at 37 "C. The extracted nuclei were stained for DNA content by a modification of the method of Vindelov et al.'" Ribonuclease S (Sigma), in the final concentration of 1 mg/ml, was added to approximately 1-2 x 10' nuclei and incubated in a 37 "C water bath for 30 minutes. The nuclei were centrifuged for 5 minutes at 800 x G,

3 Prognostic Parameters in Uterine Sarcoma/Nola et al resuspended in propidium iodide (Sigma), and incubated for 30 minutes at room temperature. In the preparation of paraffin embedded blocks for flow cytometric cell cycle analysis (proliferative activity, DNA ploidy, and DNA index), care was taken to separate tumor tissue from normal tissue in each specimen." Flow Cytometry Cellular DNA content and proliferative activity were analyzed on a FACScan flow cytometer (Becton Dickinson, San Jose, CA) using excitation wavelength at 488 nanometers and 15-milliwatt argon ion laser. For each DNA analysis, 10,000 nuclei were counted. DNA histograms were analyzed with a Consort 30 program (Becton Dickinson). The coefficient of variation was always less than lo%.'' A tumor was classified as DNA diploid if only one G,, /GI peak was observed in the expected region of the histogram. If more than one G,/G, peaks were seen, the tumor was considered DNA aneuploid (i.e., DNA index was not used for the definition of DNA aneuploidy). DNA tetraploid tumors were included in the DNA aneuploid group. Those tumors were separated from DNA diploid tumors because the percentage of cells in the G,/M fraction was higher than 10% (upper limit for the G2 /M fraction).i8 Proliferative activity was defined as the sum of the cell percentages in the S- and G2 /M-phases. The authors always analyzed paraffin embedded specimens from the most aggressive tumor parts. Control tissues were comprised of normal uterine tissue. Proliferative activity was calculated for DNA diploid cell populations only, because DNA aneuploid and DNA diploid tumors with proliferative activity higher than 20% were shown to have a worse prognosis according to the statistical analysis. Therefore, DNA ploidy and proliferative activity represented a single parameter in the current study. DNA index was defined as the ratio of the mode (or mean) of the relative DNA content of the G,, /GI cells of the sample divided by the mode (or mean) of the relative DNA measurement of the diploid G,,/G, reference cells.'" Statistical Analysis Survival time was measured in months from the date of diagnosis to the date of death or last follow-up. Only the deaths caused by the tumor were considered failed observation; all other cases were considered censored. Actuarial survival probability curves were constructed using the Kaplan-Meier method and compared by the log rank nonparametric test. Categoric (nominal) parameters were analyzed as dummy variables (0, l) in the case of a binary split, or were coded as 1, 2, etc., when appropriate. The Cox proportional hazards re ".", MONTHS FIGURE 1. Actuarial survival for endometrial stromal sarcomas (A), malignant smooth muscle tumors (B), and malignant Mullerian tumors mixed (GI. gression model (forward stepwise variable selection) was used to assess the relative effect of clinical stage, mode of therapy, histologic type, cellularity, degree of atypia, mitotic activity, depth of myometrial invasion, DNA ploidy, proliferative activity, and DNA index on survival prognosis. Data analysis was performed on a PC computer using the Number Cruncher Statistical System (NCSS, Kaysville, UT). The DNA index and proliferative activity (in DNA diploid tumors) cutoff values were determined as the values of the largest chi-square test by the proportional hazards regression method. In all the tests performed, only a P value < 0.05 was considered significant. RESULTS The authors analyzed 9 low and 17 high grade stromal sarcomas (total, 26), 8 smooth muscle neoplasms with uncertain malignant potential and 23 leiomyosarcomas (total, 31), and 16 homologous and 7 heterologous MMMT (total, 23). Five-year survival after therapy in patients with endometrial stromal sarcoma was 39.8%, 31.7% in patients with malignant smooth muscle tumors, and 41.4% in patients with MMMT, with no significant differences among the groups (Fig. 1). Univariate analysis for stromal sarcomas (Table 1) revealed a significant association between patient survival and histologic type (P = 0.005), DNA index (P < 0.001) (Fig. 2), and DNA ploid+proliferative activity (P < 0.001) (Fig. 3) of the tumor. Multivariate analysis disclosed that the DNA index was the only statistically significant parameter (P = 0.005), with a 1.25 cutoff value (Table 1) (Fig. 2). In patients with malignant smooth muscle neoplasms (Table 2), both univariate (P = 0.021) and mul- B

4 2546 CANCER December 15, 1996 / Volume 78 / Number 12 TABLE 1 Clinicopathologic Characteristics and Cell-Cycle Parameters in 26 Patients with Endometrial Stromal Sarcomas Overall survival Cell cycle analysis Prognostic parameters Statistics (P) No. of DNA ploidy DNA-index No. MST Group Subgroup patients (%) + PAa (x _f SD) failed (rnos) Univariate Multivariate Clinical stage I t I1 17 (65.4) ? t lv 9 (34.6) Histologic type LGSS 9 (34.6) ? HGSS 17 (65.4) 3/ Cellularity Minimal 5 (19.2) Moderate 4 (15.4) ? Marked 17 ( I Degree of atypia Minim a I 11 (42.3) Moderate 5 (19.2) ? Marked 10 (38.5) t Mitoses per 10 high-power fields < (38.5) (61.5) ? Therapy Surgical 6 (23.1) t Surgical t adjuvant 20 ( ? DNA ploidy t PA Favorable 11 (42.3) lli < 0, Unfavorable 15 (57.7) ? DNA index 26 (100) < PA proliferative activity, SD: standard deviation; MST median sunival time; LGSS: low grade stromal sarcoma: HGSS: high grade stromal sarcoma. "No. of patients with fa\'arable/unfavorable finding. Unfavorable findings were DNA aneuplaidy or DNA diploidy with greater than 20% of cells in the S- and GJM cell cycle phases. tivariate (P = 0.032) analysis detected a statistically significant association between tumor clinical stage and survival. High mitotic activity was also a marginally adverse prognostic parameter for those patients (P = , whereas cell cycle parameters had no prognostic value (Figs. 2 and 3). In patients with MMMT (Table 3), univariate analysis revealed an association between survival and depth of myometrial invasion (P = , clinical stage (P = , anddnaindex (P= 0.037). Multivariate analysis confirmed that the depth of myometrial invasion (P = 0.036) and clinical stage (P = 0.025) were the dominant prognostic factors. Flow cytometric cell cycle parameters were presented as DNA index and DNA ploidy combined with cell proliferative activity (Tables 1-3). Higher DNA index was a significantly unfavorable prognostic factor in patients with endometrial stromal sarcomas and MMMT (Fig. 2); in this respect, it appeared dominant over other factors in the former patient group (Table 1). Tumor cell aneuploidy (including tetraploidy) or high cell proliferative activity, (greater than 20% of cells in the S- and G,/M-phases) were adverse prognostic factors in patients with endometrial stromal sarcoma only (Fig. 3), but even in these patients their importance was surpassed by that of the DNA index (Table 1). DISCUSSION Various approaches have been shown to have little effect on the treatment of mesenchymal uterine neoplasms, and most patients die within a few years of diagnosi~.~,~~~~' Thus, it would be important to find prognostic factors for predicting the malignancy potential of these tumors. In principle, it would then be possible to select patients at low and high risk and plan the type and volume of therapy for each individual patient. This would be important not only for enhancement of the 5-year survival but also for improvement of the quality of life and exploration of possibilities for experimental treatment, such as autologous bone marrow transplantation for ovarian carcinoma.22 The current study revealed the importance of combining flow cytometric analysis with pathohistologic and clinical parameters to achieve accurate prognosis for uterine sarcomas. In the past, mitotic activity was the only parameter that distinguished low and high grade endometrial stromal sarcomas.'* Recently, the term "low grade endometrial stromal sarcoma" was used for neoplasms comprised of cells that resembled stromal cells of the

5 Prognostic Parameters in Uterine Sarcoma/Nola et al E d 0.8 f$ d a c d s 0.8 f$ 0.6 a a MONTHS MONTHS FIGURE 2. Actuarial survival according to DNA index for a group of patients with (top) endometrial stromal sarcomas, (middle) malignant smooth muscle tumors, and (bottom) malignant Mullerian mixed tumors (A: DNA index ; 6: DNA index > 1.25). proliferative phase endometrium, whereas high grade endometrial stromal sarcoma was comprised of cells that were more atypical but still resembled endometrial stromal cells." Flow cytometric data are not an absolute criteria for the division of stromal sarcoma into low and high grades, nor do they always correlate with clinical behavior, but according to the current study results, flow cytometric analysis (especially DNA index) appears to add an important prognostic variable for predicting survival in patients with endometrial stromal sarcoma. This finding corresponds to some other reports,'-'o which did not as- FIGURE 3. Actuarial survival according to the DNA ploidy and proliferative activity for the patients with (top) endometrial stromal sarcomas; (middle) malignant smooth muscle tumors, and (bottom) malignant Mullerian mixed tumors (A: DNA aneuploidy, DNA diploidy with proliferative activity 2 20%; B: DNA diploidy with proliferative activity < 20%). sess the values of DNA index as in the current study. The DNA index and proliferative activity of tumor cells definitively measure different biologic parameters; DNA index defines cell DNA content, whereas the proliferative activity shows the proportion of proliferating cells. Most of the authors emphasized the validity of the value of DNA aneuploidy or proliferative activity in the prognosis of various types of malignant According to the current study results, the increase in the DNA index represented a better parameter for prognosis than DNA

6 2548 CANCER December 15, 1996 / Volume 78 I Number 12 TABLE 2 Clinicopathologic Characteristics and Cell Cycle Parameters in 31 Patients with Malignant Smooth Muscle Tumors Overall survival Cell cycle analysis Prognostic parameters Statistics (0 No. of DNA ploidy DNA-index No. MST Group Subgroup patients (%) t PAa (X i SD) failed (rnos) Univariate Multivariate Clinical stage I t I1 27 (87.1) i , t IV 4 ( i Histologic type UMP 8 (25.8) i LMS 23 (74.2) i Cellularity Minimal 5 (16.1) ? Moderate I1 (35.5) i Marked 15 (48.4) ? Degree of atlpia Minim a I 13 (41.9) f Moderate 11 (35.5) Oil1 1.62? Marked 7 (22.5) i Mitoses per 10 high-power fields (16.1) i (22.6) i (61.3) i Therapy Surgical 7 (22.6) t Surgical t adjuvant 24 (77.4) 9/ ? DNA ploidg t PA Favorable 12 (38.7) i Unfavorable 19 (61.3) ? DNA index 31 (100) 12/ ? PA proliferative activity: SD: standard deviation; MST: median survival time: LIMP: Smooth muscle neoplasm with uncertain malignant potenrial; LWS: leiomyosarcoma. "No. of patients u'ith falrorableiunfa~,orable finding. Unfawrable findings were DNA aneuploidy or DNA diploidy with greater than 20% of cells in the S- and CJhl-cell cycle phases. ploidy or proliferative activity in endometrial stromal sarcomas. In the current study sample (Table 11, tumor cell cycle analysis revealed that 1 of 9 patients (1 1%) with low grade stromal sarcoma was at higher risk than shown by classic prognostic parameters; this could have prompted the change in treatment toward a more aggressive regimen. The significance of the prognostic value of cell cycle parameters should be interpreted with due caution, as is the case for any potentially significant prognostic factor. The prognostic importance of any clinical or laboratory parameter should be assessed in relation to other tumor (and patient) characteristics, not only statistically (i.e., in a multivariate analysis [Tables 1-31), but also with respect to their biologic relations. For example, if all the patients with high DNA indexes had advanced stage tumors, DNA index would be of no prognostic value. However, in this respect, when the patients with a single type of mesenchymal tumor were subdivided according to the specificities of their clinical stages, modes of therapy, etc., the final small number of patients in the respective subgroups made the analysis of cell cycle parameters impossible. To overcome this, the authors listed the data on all analyzed tumor characteristics together (Tables 1-3) and determined their significance by comparing their statistic significances with respect to patient survival. Accordingly, because the data listed in Table 1 (endometrial stromal sarcomas) showed that clinical stages did not differentiate patients with respect to survival, it may be concluded that the clinical stages of endometrial stromal sarcomas did not affect (high) prognostic significance of the cell cycle parameters, especially the DNA index. In contrast, in the case of malignant smooth muscle tumors (Table 2), only the clinical stage of the tumors was a significant prognostic factor, whereas cell cycle parameters were not. Obviously, in the case of smooth muscle uterine tumors DNA flow cytometric analysis is of no use. Only in the case of MMMT (Table 3) were both tumor clinical stage and DNA index significantly related to patient survival. Inasmuch as the numbers of patients in the respective subgroups were too small to be tested directly, and the clinical stage of the tumor appeared significant in both univariate and multivariate analysis (DNA index in the univariate only), it must be assumed that the stage-related patient selection might have affected the prognostic significance of the DNA index. It should again be suspected that the analysis of tumor cell cycle parameters does not add to the

7 Prognostic Parameters in Uterine Sarcoma/Nola et al TABLE 3 Clinicopathologic Characteristics and Cell Cycle Parameters in 23 Patients with Malignant Miillerian Mixed Tumors Overall survival Cell cycle analysis Prognostic parameters Statistics (8 No. of DNA ploidy DNA index No. MST Group Subgroup patients (%) t PA" (x I SD) failed (mod Univariate Multivariate Clinical stage I t I1 I6 (69.6) 6/ i I11 t IV 7 (30.4) 0/ f Histologic type Homologous 16 (69.6) Heterologous 7 (30.4) i Degree of atgpia Minimal 4 (17.4) I Moderate 7 (30.4) f Marked 12 (52.2) I Mitoses per 10 highpower fields < ) ? (74) 5/ f Depth of myumetrial invasion < (69.6) 5/ i > (30.4) 1/ Therapy Surgical 2 (8.7) f Surgical t adjuvant 21 (91.3) ? DNA ploidy t PA Favorable 6 (26.1) i Unfavorable 17 (73.9) i DNA index 23 (100) 6/ PA: proliferatin, activitl SD: standard deviation; MST: median survival time. ' No, uf parienr\ with favorableiunfavnrable finding. Unfavorable findings were DNA aneuplaidy or DNA diploidy with greater than 20% of cells in the S- and G)M cell cycle phases. improvement of disease prognosis. In patients with MMMT, a similar line of reasoning can be inferred for the highly significant role of the depth of myometrial invasion (Table 3). Indeed, higher clinical stage of the tumor has been associated with deeper myometrial invasion4 By the same token, because the mode of therapy did not influence patient survival in any of the studied tumors (Tables 1-31, it probably could not affect the data obtained with the cell cycle parameters. REFERENCES Larson B, Silfversward C, Nilsson B, Pettersson 1:. Prognostic factors in uterine leiomyosarcoma. A clinical and histopathological study of 143 cases. The Radiumhemmet series Actn O ~CO~ 1990;29: Chang LK, Crabtree GS, Lim-Tan SK, Kempson RL, Hendrickson MR. Primary uterine endometrial stromal neoplasms. A clinicopathologic study of 117 cases. Am J Surg PlZthl 1990; 14: Zaloudek C, Norris HJ. Mesenchymal tumors of the uterus. In: Kurman KJ, editor. Blaustein's pathology of the female genital tract. 4th edition. New York: Springer-Verlag, 1994: Ali S, Wells M. Mixed Mullerian tumors of the uterine corpus: a review. Irzr J Gynecol Cancer 1993;3:1-11. Bell SW, Kempson RL, Hendrickson MR. Problematic uterine smooth muscle neoplasms. A clinicopathologic study of 213 cases. Am J Surg Puthol 1994; 18: Major Fl, Blessing JA, Silverberg SG, Morrow CP, Creasman WT, Currie JL, et al. Prognostic factors in early-stage uterine sarcoma. A Gynecologic Oncology Group study. Cancer 1993; 71: Vetek N, Nola M, MaruSiC M, IliC I, Babic D, Petrovecki M, et al. Prognostic value of steroid hormone receptors concentration in patients with endometrial carcinoma. Actu Obstet Gynecol Scand 1994; 73: Dunton JC, Kelsten ML, Brooks SE, Viglione MJ, Carlson JA, Mikuta JJ. Low-grade stromal sarcoma: DNA flow cytometric analysis and estrogen progesterone receptor data. Gynecol Oncol 1990; El-Naggar AK, Abdul-Karim FW, Silva EG, McLemore D, Garnsey L. Uterine stromal neoplasms: a clinicopathologic and DNA flow cytometric correlation. Hum Patholl991; 22: August CZ, Bauer KD, Lurain J, Murad T. Neoplasms of endometrial stroma: histopathologic and flow cytometric analysis with clinical correlation. Hum Puthol 1989; 20: Tsushima K, Stanhope CR, Gaffey TA, Lieber MM. Uterine leiomyosarcomas and benign smooth muscle tumors: usefulness of nuclear DNA patterns studied by flow cytometry. Muyo Clin Proc 1988;63: DiSaia PJ, Creasman WT. Clinical gynecologic oncology. 4th edition. St. Louis: Mosby Year Book, Evans HI.. Endometrial stromal sarcoma and poorly differentiated endometrial sarcoma. Cancer 1982;50: Kempson RL, Hendrickson MR. Pure mesenchymal neoplasms of the uterine corpus. In: Fox H, editor. Haines and Taylor obstetrical and gynecological pathology. 3rd edition. New York Churchill Livingstone, 1987: Hedlep DW, Friedlander ML, Taylor IW, Rugg CA, Musgrove EA. Method for analysis of cellular DNA content of paraffinembedded pathological material using flow cytometry. / Histoclzem Cytochem 1983; 31:

8 2550 CANCER December 15,1996 I Volume 78 I Number Vindelov LL, Christensen IJ, Nissen NI. A detergent-trypsin method for the preparation of nuclei for flow cytometric DNA analysis. Cytornerry 1983:3: Coon IS, Landay AL, Weinstein RS. Advances in flow cytometry for diagnostic pathology. Lab Invesr 1987; 57: Hiddemann W, Schumann J, Andreeff M, Barlogie B, Herman CJ, Leif RC, et al. Convention on nomenclature for DNA cytometry. Cytornetry 1984;5: Wersto RP, Libit RL, Koss LG. Flow cytometric DNA analysis of human solid tumors: a review of the interpretation of DNA histograms. Hum Pathol 1991:22: Nordal RN, Kjorstad KE, Stenwig AE, Trope CG. Leioniyosarcoma (LMS) and endometrial stromal sarcoma (ESS) of the uterus. A survey of patients treated in the Norwegian Radium Hospital Int J Gynecol Cancer 1993:3: Larson B, Silfversward C, Nilsson B, Pettersson F. Endometrial stromal sarcoma of the uterus. A clinical and histopathological study. The radiumhemmet series Ezir J Obstet Gynecol Reprod Biol 1990;35: Tepler I, Cannistra SA, Frei E 111, Gonin R, Anderson KC, Demetri G, et al. Use of peripheral-blood progenitor cells abrogates the myelotoxicity of repetitive outpatient high dose carboplatin and cyclophosphamide chemotherapy. J Clin Oncol 1993;11: Merkel DE, McGuire WL. Ploidy, proliferative activity and prognosis. Cancer 1990: 65: Vetek N, Nola M, MaruSiC M, Babic D, Uiarevid B, Sabioncello A, et al. Tumor cell-cycle in patients with stage I endometrial carcinoma. Gynecol Oncol 1994;53: Van Dam PA, Watson JV, Lowe DG, Shepherd JH. Flow cytometric DNA analysis in gynecological oncology. IntJ Gynecol Cancer 1992: 2: Peters WA 111, Howard DR, Andersen WA, Figge DC. Deoxyribonucleic acid analysis by flow cytometry of uterine leiomyosarcomas and smooth muscle tumors of uncertain malignant potential. Am J Obstet Gynecol 1992: 166: Nola M, BabiC D, IliC-Forko 1, UiareviC B, PetroveEki M, Jukii: S. Tumor cell cycle parameters in patients with stage I endometrial carcinoma. Croatian Med J 1995; 36: