Clinical Factors and Biomarkers Which Affect a New Universal Grading System for Ovarian Epithelial Carcinoma

  • Published on

  • View

  • Download

Embed Size (px)


<ul><li><p>J. Obstet. Gynaecol. Xes. Vol. 27, No. 6: 313-318 2001 </p><p>Clinical Factors and Biomarkers Which Affect a New Universal Grading System for Ovarian Epithelial Carcinoma </p><p>Shin-ichi Ishioka, Satoru Sagae, Masaki Sugimura, Yoshihiro Nishioka, Kanji Kobayashi, and Ryuichi Kudo Department of Obstetrics and Gynecology, Sapporo Medical University School of Medicine, Sapporo, Japan </p><p>Abstract </p><p>Objective: To detect clinical factors and biomarkers which affect a new grading system for ovarian epithelial cancer that was proposed by Shimizu et al. and to analyze the impact of those factors on malignant behaviors of the tumor. </p><p>Methods: Review and scoring of specimens of paraffin embedded tissues of epithelial ovarian cancer were carried out according to the new universal grading system in 110 women who underwent initial treatments including the primary operation in our university hospital between January 1990 and June 1999. As biomarkers, expression of Bcl-2, Bax, Bcl-X, and the accumulation of P53 protein were also studied immunohistochemically. </p><p>This grading system functioned as a prognostic indicator. Five-year survivals of the patients was 77.6, 36.3, and 17.4%, for Grade 1, Grade 2, and Grade 3, respectively. The grade was also correlated with clinical stages, histologic subtypes, operative completeness, para-aortic lymph nodes metastasis, and the expression of Bax protein, with univariate analy- sis. Multivariate analysis revealed histological subtypes and para-aortic lymph nodes me- tastasis to be important factors which affected the grading system. </p><p>Conclusion: These results demonstrate that the new universal grading system is useful, and that this grading system might reflect potential of metastasis or dissemination of the ova- rian cancer. </p><p>Key words: histologic grade, ovarian cancer, P53, Bax </p><p>Results: </p><p>Introduction </p><p>Histologic grade is known to be an important prognostic factor in ovarian cancer.') Until now, various grading systems such as the International Federation of Gynecology and Obstetrics (FIGO) grading system:) the World Health Organization (WHO) grading system?) and the Gynecologic Oncology Group (GOG) grading system4) have </p><p>been used in different institutes. However, those grading systems have difficulties in terms of re- producibility and subjectivity. Recently, Shimizu et al . reported a new universal grading system for epithelial ovarian cancer which was based on the architectural grade, nuclear grade, and mitotic ~ o u n t . ~ , ~ ) They concluded that their grading sys- tem provided useful prognostic information for all histologic types of ovarian cancer, and that it </p><p>Received: Feb. 14,2000 Accepted: Sep. 14,2001 Reprint request to: Dr. Satoru Sagae, Department of Obstetrics and Gynecology, Sapporo Medical University School of Medicine, Minami 1- jo, Nishi 16-chome, Chuou-ku, Sappopro, Hokkaido 0604061, Japan </p><p>313 </p></li><li><p>S. Ishioka et al. </p><p>might reflect the biological behavior of tumor cells.6) </p><p>In the present study, we reviewed all the paraffin embedded tissues of epithelial ovarian cancer obtained from 110 women who under- went initial treatments including primary sur- gery in our hospital between January 1990 and June 1999, and scored the specimens according to the universal grading system. Then we assessed the usefulness of this grading system by studying relationships between this grading system and clinical factors. Furthermore, expression of vari- ous biomarkers such as P53, and bcl-2 family genes products has recently been reported to be related to the prognosis of patients with ovarian cancer.) Therefore, we also focused on the rela- tionships between this grading system and the expression of such biomarkers in ovarian cancer in this study. </p><p>Materials and Methods </p><p>Patients Patients with available histopathologic slides </p><p>who were initially diagnosed at Sapporo Medical University as having malignant epithelial ova- rian carcinoma between January 1990 and June 1999, were included in the study when they re- ceived initial treatments, including primary sur- gery, in our university hospital. Those who re- ceived primary surgery in other hospitals were excluded. Patients with borderline malignant tu- mors, non-epithelial tumors, metastatic ovarian tumors, and tumors of unknown origin, were also excluded. </p><p>Histologic analysis utilized hematoxylin and eosin-stained sections cut from 2 to 24 blocks per case. </p><p>All the slides were screened by three indepen- dent observers (S.1, S.S, K.K) without knowledge of the clinical course of the patients. When a con- sensus was not reached, S.S. decided the final grade. </p><p>Diagnosis of Histologic Type and the New Grading System </p><p>The classification of histologic subtypes was performed according to the World Health Orga- nization classification of the surface epithelial- stromal category tumor. The grade for an indi- vidual tumor was determined according to the method by Shimizu et al. Briefly, architectual grading was scored as follows; score 1 for glan- dular growth pattern, score 2 for papillary growth pattern, and score 3 for solid growth pat- </p><p>tern. Nuclear pleomorphism was scored as fol- lows; score 1 for relatively uniform vesicular nu- clei (variation in diameter &lt; 2 : l), a low nuclear: cytoplasmic ratio, no chromatin clumping or prominent nucleoli, score 2 for intermediate variation in nuclear size (between 2 : 1 to 4 : 1) and shape, nucleoli recognizable but small, some chromatin clumping, no bizarre cells, and score 3 for marked variation in nuclear size (&gt; 4 : 1) and shape, a high nuclear: cytoplasmic ratio, promi- nent chromatin clumping, thick nuclear mem- branes, large eosinophilic nucleoli, and any bi- zarre cells which may exist. Mitotic count was scored as follows; score 1 for 0-9 mitotic cells/lO HPF (1OX wide field eyepiece, 40X objective), score 2 for 10-24/10 HPF, score 3 for 25 I110 HPF. </p><p>The final grading was made as follows; total score 3-5 for Grade 1 (well differentiated), total score 6-7 for Grade 2 (moderately differentiated), and 8-9 for Grade 3 (poorly differentiated)?) For the determination of the grade, at least three slides suitable for histologic typing and grading of the tumor were selected. In particular, we tried to find slides that included the periphery of the tumor at which active growth was most likely. </p><p>lmmunohistochemical Staining Anti-P53 antibody, anti-Bcl-2 antibody, anti- </p><p>Bax antibody, and anti-Bcl-X antibody were pur- chased from DAKO. All samples were fixed in formaldehyde and embedded in paraffin. Sec- tions 6-pm-thick were routinely passed through xylene and a graded alchol series and stained for the antibodies by Envision method (DAKO). Sec- tions were counterstained with hematoxylin for microscopic examination. Whole cancer tissues on each slide were examined, and carcinomas in which 30% or more of the component cells were stained for the antibodies were evaluated as posi- tive (+), whereas tissues containing less than 30% stained cells were counted as negative (-). </p><p>Statistical Analyses Survival curves were calculated by the Kaplan- </p><p>Meier method. The significance of differences be- tween curves was evaluated with the generalized Wilcoxon test. Multivariate analysis was per- formed by multiple regression analysis with the Microsoft Excel. Other statistical analyses were carried out by x test or Students t-test. All re- sults were deemed significant at p &lt; 0.05. </p><p>324 </p></li><li><p>Ovarian Cancer and a New Universal Grade </p><p>Results </p><p>Characteristics of Patients The relationship of FIGO surgical stage and </p><p>histologic subtypes of the patients is shown in Table 1. Majority of patients with serous carci- noma had advanced stage disease at the time of diagnosis, whereas almost half of those with clear cell, endometrioid, or mucinous carcinoma were diagnosed at earlier stages. As to operative treat- ment modality, total hysterectomy, bilateral sal- pingo-oophorectomy, omentectomy, and pelvic and para-aortic lymphadenectomy were per- formed as a standard treatment modality for the patients. However, for those who had extremely advanced disease, for those who needed preser- vation of ovarian function, and for those who had severe complication, modified operative proce- dures were performed. As to chemotherapy, 73 patients received platinum-based chemotherapy, 24 patients received paclitaxel-based chemo- therapy, 9 received CPT-1 1-based chemotherapy, and 7 received no chemotherapy for severe com- plication or rejection of further treatments. </p><p>Survival of the Patients by FlGO Stage The 3-year and 5-year survivals of the 110 pa- </p><p>tients by FIGO stage are described in Table 2. Pa- </p><p>tients in advanced stages had poorer prognosis than those in earlier stages. Poorer prognosis of patients with Stage 11, might have been derived from fewer number of patients with Stage 11. FIGO stage was thought to be a significant prog- nostic factor. </p><p>Prognostic Significance of the New Grading System As shown in Table 3, this universal grading </p><p>system worked as a significant prognostic indica- tor in patients with ovarian cancer who were en- rolled in this study. Five-year survivals of pa- tients were 77.6, 36.3, and 17.4%, for Grade l, Grade 2, and Grade 3, respectively. Patients with Grade 1 showed a significantly better prognosis than those with higher grades. Among patients with Stage 1/11 and Stage III/IV, those with Grade 1 also showed better prognoses than those with higher grades, though the differences were not statistically significant. </p><p>Table 3 shows the 5-year survivals of patients in the major histologic types in each grade. For serous and mucinous types of cancer, this grad- ing system worked as a significant predictor of survival. For endometrioid and clear cell cancers, patients with Grade 1 showed a better prognosis than those with higher grades, though the differ- ences were not significant. </p><p>Table 1. noma </p><p>Relationship between histological subtypes and FIGO clinical stages in patients with ovarian carci- </p><p>~~ __ Histological subtypes I </p><p>FIGO stage </p><p>I1 111 IV Total (YO) </p><p>Serous 11 1 35 7 54 (49.1) Mucinous 8 1 5 1 15 (13.6) Endometrioid 7 2 5 0 14 (12.7) Clear cell 11 2 9 2 24 (21.8) Others* 1 1 0 1 3 (2.8) Total 38 (34.5%) 7 (6.4%) 54 (49.1%) 11 (10.0%) 110 (100) </p><p>* Others include mixed epithelial ovarian cancer and undifferentiated ovarian cancer. </p><p>Table 2. Three- and 5-year survival of the patients by the clinical stages </p><p>Clinical stage ( n = ) 3-year survival (YO) 5-year survival (%) </p><p>Iab (15) 100 100 Iab: Ic, N.S </p><p>11 (7) 43.0 43.0 IIIab (14) 83.9 47.9 IIIab: IIIc, p &lt; 0.05 IIIc (40) 32.0 25.6 IIIc: IV. p &lt; 0.05 </p><p>Ic (23) 85.6 70.6 Ic: IIIc, p &lt; 0.01 </p><p>IV (11) 19.1 0 </p><p>Five-year survivals were determined by the Kaplan-Meier method. p values are determined by the Generalised Wilcoxon test. N.S means statistically not significant. </p><p>315 </p></li><li><p>S. Ishioka et al. </p><p>Table 3. Five-year survival of the patients bv the new universal grading svstem </p><p>p-value Histology G1 G2 G3 ( n = 1 ( n = ) ( n = 1 ( n = ) </p><p>All 77.6% 36.1% 17.4% G1 : G2, p &lt; 0.01 (41) (53) (26) G2 : G3, p &lt; 0.01 </p><p>Histology Serous 100% 45.4% 19.2% GI : G2, p c 0.01 G2 : G3, p &lt; 0.05 G1 : G2, p &lt; 0.05 </p><p>(24) - </p><p>(54) (8) (22) Mucinous 60.0% 20.0% </p><p>(15) (10) (5) (0) Endometrioid 100% 66.7% 50.0% N.S </p><p>(2) - N.S </p><p>(14) (7) (5) </p><p>(24) (14) (10) (0) </p><p>(45) (27) (14) (0) </p><p>(65) (14) (25) (26) </p><p>Clear cell 52.9% 46.7Yo </p><p>Stage I, I1 84.0% 63.8% - N.S </p><p>111, IV 50.8% 22.3% 19.0% N.S </p><p>Five-year survivals were determined by the Kaplan-Meier method. p values are determined by the Generalised Wilcoxon test. N.S means statistically not significant. </p><p>lmmunohistochemical Staining As shown in Table 4, expression of Bax protein </p><p>was significantly correlated with the grade. Among 41 G1 carcinomas, 30 (73.1%) were posi- tive for Bax protein, whereas among 26 G3 carci- nomas, only 8 (30.8%) were positive for Bax pro- tein. As to the accumulation of P53 and the grade, the expression of Bcl-2 and Bcl-X protein and the grade, statistically significant relations were not obtained. However, a tendency was found for there to be a connection between negative P53 and G1, and negative Bcl-2 and G1, though it was not significant. </p><p>Univariate Analysis of Factors Which May Affect the New Grading System </p><p>As clinical factors which may affect the grade, the FIG0 clinical stage, histological subtype, size of the tumor, age of the patient, operative com- pleteness, and existence of para-aortic lymph node (PAN) metastasis, were selected. </p><p>Tumor size was divided into groups as follows; less than 10 cm in diameter and 10 cm or larger in diameter. In 3 cases, we were unable to determine the size of the tumor from the operation record. Patients were divided into groups by age as fol- lows: those less than 60 years old and those 60 or more years old. Operative completeness was di- vided into groups as follows; the size of residual tumor was less than 0.5 cm (optimal operation) and size of the residual tumor was 0.5 cm or larger (suboptimal operation). As to PAN me- tastasis, 56 of 110 cases received lymphadenec- </p><p>Table 4. Grade and the expressions of biomarkers </p><p>Grade 1 Grade 2 Grade 3 p value </p><p>P53 Positive Negative </p><p>Positive Negative </p><p>Positive Negative </p><p>Positive Negative </p><p>Bcl-2 </p><p>Bax </p><p>Bcl-X </p><p>9 12 32 31 </p><p>9 11 32 32 </p><p>30 21 11 21 </p><p>15 16 26 27 </p><p>11 N.S 15 </p><p>10 NS') 16 </p><p>8 p =0.001 18 </p><p>15 N.S 15 </p><p>a) N.S means statistically not significant. </p><p>tomy or sampling of PAN. Those who only re- ceived palpation of PAN, and those for whom exploration of PAN was not performed were ex- cluded. </p><p>As shown in Table 5, patients with advanced stages, those with the serous subtype, those with residual tumors 0.5 cm or larger, and those who had positive PAN metastasis, had significantly higher grades. </p><p>Multivariate Analysis of Factors Which May Affect the New Grading System </p><p>The universal grading system provided useful prognostic information even for our epithelial ovarian cancer cases. Univariate analysis re- vealed that the clinical stage, histological sub- </p><p>316 </p></li><li><p>Ovarian Cancer and a New Universal Grade </p><p>Table 5. Univariate analysis of the grade by clinical factors </p><p>Index Grade 1 Grade 2 Grade 3 p value Stage </p><p>I I1 Ill IV </p><p>Histology Serous Mucinous Endometrioid Clear cell Others </p><p>Tumor size &lt; 10 cm 2 10 cm </p><p>&lt; 60 years old 2 60 years old </p><p>Residual tumor &lt; 0.5 cm 2 0.5 cm </p><p>PAN meta Negative Positive </p><p>Age </p><p>23 15 0 4 3 0 </p><p>13 21 20 1 4 6 &lt; 0.001 </p><p>8 22 24 10 5 0 7 5 2 </p><p>14 10 0 2 1 0 &lt; 0.001 </p><p>15 16 12 24 26 14 N.S </p><p>27 30 13 14 13 13 N.S </p><p>32 25 6 9 18 20 &lt; 0.001 </p><p>1 6 12 5 3 1 p=0.004 </p><p>N.S means, not significant. </p><p>Table 6. Multivariate analysis of the grade by some factors </p><p>Index Regression Order of coefficient correlation") </p><p>Clinical stage 0.059082 5 Histology - 0.33487 1* Operative completeness 0.098251 4 PAN metastasis 0.32338 2, Expression of Bax 0.221427 3 </p><p>a): * indicates statistically significant correlation by multivariate analysis. </p><p>type, operative completeness, PAN metastasis, and the expression of Bax protein significantly af- fected the universal grading system. M...</p></li></ul>