Public Notices

Multivariable analysis of DNA ploidy, p53, and HER2/neu as prognostic factors in endometrial cancer

Multivariable analysis of DNA ploidy, p53, and HER2/neu as prognostic factors in endometrial cancer
of 6
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
  2380 zyxwvusr Multivariable Analysis of DNA Ploidy p53 and HER-2/neu as Prognostic Factors in Endometrial Cancer z ndrea S Lukes, B.A.,* Matthew F Kohler, M.D.,* Carl F Pieper, Dr.P.H.,t Billie J Kerns, B.S.,S Rex Bentley, M.D.,S Gustavo C. Rodriguez, M.D.,* John T. Soper, M.D.,* Daniel L. Clarke-Pearson, M.D.,* Robert C. Bast, zy r., M.D.,§Il and Andrew Berchuck, M.D.* Background. Several molecular-genetic alterations in endometrial cancers, including aneuploidy and aber- rant expression of p53 and HER-Z/neu, have been associ- ated with poor prognosis. To determine the importance of molecular-genetic factors relative to more traditional surgical-pathologic prognostic factors, a multivariable analysis was performed. Methods. Immunohistochemical staining for p53, HER-Z/neu, estrogen receptor, progesterone receptor, and epidermal growth factor receptor was performed on frozen sections from zyxwvut 00 primary endometrial cancers. DNA ploidy was determined using computerized image analysis of Feulgen-stained touch preparations. In addi- tion, information regarding surgical-pathologic features of the cancers was obtained. Univariable analysis was performed followed by multivariable analysis using Cox s proportional hazards model to identify variables predictive of poor prognosis. Results With univariable analysis, race, histologic type, stage, grade, myometrial invasion, estrogen recep- tor, progesterone receptor, ploidy, p53 and HER-Z/neu were predictive of the presence of persistent or recurrent disease. In the multivariable analysis, only International Federation of Gynecology and Obstetrics stage zyxwv P zyxwvu   zyxwv .005 , grade P 0.005 , myometrial invasion P 0.024), and ploidy P = 0.028) were significant. Conclusions. Among molecular-genetic prognostic factors, DNA ploidy was the most strongly predictive of persistent or recurrent disease. Cancer 1994; 73:2380-5. Key words: endometrial cancer, DNA ploidy, p53 tumor From the *Departments of Obstetrics and GynecologyjDivision of Gynecologic Oncology, tcommunity and Family Health/Division of Biometry, $Pathology, §Medicine, and IIImmunology, Duke Com- prehensive Cancer Center, Durham, North Carolina. A.L.S. is a Duke University medical student and the John R. McCain Fellow of the South Atlantic Association of Obstetricians and Gynecologists. Address for reprints: Andrew Berchuck, M.D., Duke University Medical Center, P.O. Box 3079, Durham, NC 27710. Accepted for publication December 30,1994. suppressor gene, HER-O/neu oncogene, prognostic fac- tors. Several surgical-pathologic features of endometrial cancers have been shown to correlate with prognosis. These include histologic grade, histologic type, depth of myometrial invasion, cervical extension, and the pres- ence of metastatic disease.'-3 This information tradi- tionally has been used to determine whether hysterec- tomy alone is likely to be curative. Patients with persis- tent disease after initial surgery or those believed to be at high risk of recurrence often receive additional post- operative therapy, most often radiation. Recently, our group and others have described mo- lecular-genetic changes that accompany malignant transformation in the endometrium. First, like normal endometrium, endometrial cancers frequently maintain expression of receptors for estrogen and progesterone. Loss of steroid receptor expression in endometrial can- cers is associated with an unfavorable progn~sis.~-~ ec- ond, although most endometrial cancers have a normal diploid zyxw NA content, those cancers that are aneuploid represent an aggressive subset of lesions that often present at an advanced stage and have a high likelihood of Third, aberrant expression of certain oncogenes and tumor-suppressor genes also has been correlated with poor prognosis. Overexpression of the HER-2/neu proto-oncogene product has been found to occur in approximately 10 of endometrial cancers and is associated with the presence of intraperitoneal meta- static di~ease.'~,'~ n contrast, we found that expression of the epidermal growth factor receptor, a related growth factor receptor tyrosine kinase, did not correlate with advanced stage or poor o~tcome.'~ verex- pression of mutant forms of the p53 tumor-suppressor gene product occurs in 20 of endometrial cancers and has been associated with poor histologic grade and the presence of advanced stage disease.  Multivariable Analysis in Endometrial Cancer/Lukes et zyxwv l. zyxwv 381 The purpose of the current study was to examine the significance of molecular-genetic prognostic factors relative to traditional surgical-pathologic prognostic factors in endometrial cancer. zyxwvu Materials and Methods Pa tien s Frozen endometrial cancer specimens were available from 100 patients who underwent exploratory laparot- omy, pelvic peritoneal cytology, total abdominal hys- terectomy, and bilateral salpingo-oophorectomy at Duke University Medical Center between 1985 and 1991. None of these patients received preoperative ra- diation therapy. Among the 100 patients, 67 underwent pelvic and periaortic lymph node sampling. Lymph node sampling was not performed in patients with well- or moderately differentiated cancers that appeared at surgery to be only superficially invasive or in patients with obvious metastatic disease in other areas. In addi- tion, lymph node sampling was not performed on some patients who were morbidly obese or in poor general health. Additional treatment after surgery (radiation, chemotherapy) was prescribed at the discretion of the individual physicians. Complete clinical information was obtained on all patients. All histologic material from these cases was reviewed to determine the histologic grade, histologic type, depth of myometrial invasion, and presence of cervical involvement or metastatic disease. Stage of dis- ease was determined using the surgical staging system of the International Federation of Gynecology and Ob- stetrics system. Stage was considered as a binary vari- able in our statistical analyses: Stages zyxwvu   and I1 versus Stages 111and IV. Myometrial invasion was analyzed using four categories: no invasion, invasion into the in- ner one third, invasion into the middle one third, and invasion into the outer one third of the myometrium. Estrogen and Progesterone Receptor Analysis Standard biochemical techniques were used for estro- gen and progesterone receptor analysis as described previ~usly.~~” result of greater than 7.5 fmol/mg tis- sue was considered positive. PIoidy Analysis DNA ploidy was determined using the CAS 200 image analysis system (Cell Analysis Systems, Elmhurst, IL to quantitate Feulgen staining in touch preparations of the cancers, as described previously.’* A sample of normal cells (DNA index zyxwvuts   1.0) was included as a control with each group of cancers tested. On the basis of prior stud- ies in the literature, cancers with a DNA index of greater than or equal to 1.25 were considered aneuploid. The significance of a threshold value of 1.15 also was exam- ined. Immunohistochemical Analysis of HER-Zlneu, p53, and Epidermal Growth Factor Receptor Immunohistochemistry was performed using frozen sections of the cancers, as described previously.’3~’5~’6 The primary monoclonal antibodies used included TAl (HER-2/neu), 1801 (p53), and 528 (epidermal growth factor receptor). In addition, positive and negative con- trol antibodies were used to ensure tissue viability and exclude the possibility of nonspecific staining. Staining for HER-2/neu and epidermal growth factor receptor was graded as light (1+), moderate (2+), or heavy 3+). As described previously, overexpression was defined as heavy (3+) staining.13,16 taining for p53 was graded as positive if nuclear staining was seen in most cancer cells. Staining for p53 is not seen in normal cells, whereas mutations in the p53 gene in some cancers lead to overexpression of p53 that is detectable immunohis- tochemically.16 Statistical Analysis Univariable analyses were performed using a chi- square test of association or Fisher’s exact test to test the association of categorical variables with the presence of persistent or recurrent disease, whereas a z   test was used for continuous variables. To analyze the simultaneous effect of all the variables and control for varied follow- up, Cox’s proportional hazards model with Breslow’s approximate likelihood method to handle ties was per- formed.” There were 29 patients with either recurrent (n = 18) or persistent (n = 11) disease. Time to event was defined as either date of recurrence for patients with a disease-free interval after primary surgery or date of initial surgery for patients who were never ren- dered free of disease. The assumptions of Cox’s propor- tional hazards model were assessed, including interac- tions and proportionality of hazards over time. Kaplan- Meier curves were created for those variables found to be significant in the multivariable analysis.” For all analyses, two-sided tests of significance were per- formed with an alpha of 0.05. All analyses were per- formed using SA software (SAS Institute Inc., Cagy, NC . Results The International Federation of Gynecology and Ob- stetrics stage distribution of the 100 patients in the cur- rent study is shown in Table l. There were 87 endome-  2382 zyxwvutsrqpo ANCER May zyxwvusr   zyxwvusrqp 994 zyxwvutsrq olume 73 No. 9 Table zyxwvu . Relationship Between FIGO Stage and Persistent/Recurrent Disease No. of No. with persistent/ FIGO stage patients recurrent disease zy  ) IA: No invasion 9 0 IB: Inner 1/2 uterine invasion 39 5 (13) IC: Outer 1/2 uterine invasion 16 3 (16) IIA: Endocervical gland involvement 0 0 IIB: Cervical strornal invasion 2 0 IIIA: Positive washings, adnexal metastases 11 2 (18) IIIB: Vaginal metastases 0 0 IIIC: Lymph node metastases 10 9 (90) IVA: Bladder or rectal involvement 0 0 IVB: Distant metastases 13 10 (77) Total 100 29 (29) FIGO: International Federation of Gvnecoloev and Obstetrics. trioid cancers, 6 adenosquamous cancers, 6 papillary se- rous cancers, and 1 clear cell cancer. Demographic information and the distribution of prognostic variables for all patients is shown in Table 2 along with results from the univariable analysis. Although race was found to be significantly associated with a poor outcome in univariable analysis, it was not significant after control- ling for other prognostic variables. In our study popula- tion, black women tended to have later-stage disease (74% with Stage 111 or IV and moderately or poorly differentiated cancers (89%). For all 100 patients, the median follow-up time from date of diagnosis until death or last contact was 3.4 years (range, 2 months to 8.4 years). Of the 29 patients found to have persistent or recurrent disease, 23 (79 ) patients died of endometrial cancer, including 9 of 11 patients with persistent disease and 14 of 18 patients with recurrent disease. The median survival time of these 23 patients was 13.4 months (range, 3-36 months). Among the other six patients with persistent or recurrent disease, one patient died of causes unre- lated to endometrial cancer, four patients were alive with evidence of disease at last follow-up, and one pa- tient was alive with no evidence of disease. Among the 71 patients in whom persistent or recurrent disease did not develop, 66 were alive at last follow-up and 5 died of other causes. The median follow-up time of these 66 patients was 4.3 years (range, 2.1-8.4 years). The results of the univariable analysis are shown in Table 1. Without considering the variation in follow-up, all variables were significantly associated with persis- tent or recurrent disease except for age and epidermal growth factor receptor status. Significant predictive value for DNA ploidy was seen when aneuploidy was defined as a DNA index greater than or equal to 1.25, but not when a threshold above 1.15 was used. In the multivariable analysis, controlling for the two demographic covariates (age and race), four vari- ables were significantly predictive of the presence of persistent or recurrent disease (stage, grade, myometrial invasion, and DNA ploidy) (Table 3). These four vari- ables were significant regardless of variable selection methods, that is, backward, forward, or stepwise selec- tion. When interpreting the risk ratios in Table 3, the assumption is made that all other variables remain con- stant, whereas the factor under consideration is varied. For example, given all other variables remaining the same, a woman with an aneuploid cancer had a 2.5-fold increased hazards risk of persistent or recurrent disease compared with a women with a diploid cancer. Given the final model in Table 3, no other prognos- tic variable was significant, including HER-2/neu or p53. However, if ploidy was excluded from this final model and then the two genes were added separately, HER-2/neu was found to be significant zy P = 0.028 whereas p53 remained nonsignificant P = 0.156). This can be explained in part by the strong association be- tween ploidy and HER-2/neu. The odds of a woman having an aneuploid cancer was much greater if her cancer overexpressed HER-2/neu (9 of 12 [75 ]) han if it did not overexpress HER-2/neu (13 of 88 [15%]) P < 0.001). In contrast, less than half of women with cancers that overexpressed p53 also had aneuploid tu- mors (10 of 21 [48%]). Kaplan-Meier curves for the four independent variables (stage, grade, myometrial invasion, and DNA ploidy) are shown in Figure 1. The relationship between ploidy and the three surgical-pathologic independent variables is demonstrated in Table 4. Aneuploidy was significantly associated with stage and grade, but not with myometrial invasion. Discussion The current study is among the first to examine the sig- nificance of surgical-pathologic prognostic factors rela-  Multivariable Analysis in Endometrial CancerlLukes zyxwvu t al. 2383 Table zyxwvutsr . Univariable Analysis of Risk Factors for Persistent or Recurrent Endometrial Cancer Recurrent or persistent disease Risk factor No Yes P value zyxwvu ean age at diagnosis (yr) Race White Black FIGO stage I or I1 111or IV Well differentiated Moderately differentiated Poorly differentiated Histologic findings Endome rioid Other None Inner 1/3 Middle 1/3 Outer 1/3 DNA ploidy Diploid Aneuploid Negative Positive Negative Positive Negative Positive Negative Positive p53 expression Negative Grade Myometrial invasion zyxwvutsr ECF eceptor overexpression Progesterone receptor Estrogen receptor HER-2/neu overexpression 66.2 65.8 NS 17 12 8 21 2 15 12 21 8 0 8 7 14 15 14 I8 11 16 13 21 8 20 9 16 zyxwvuts < 0.001 < 0.0001 0.001 < 0.01 0.014 < 0.001 NS .034 0.035 < 0.001 < 0.001 Positive 13 NS: not significant; FIGO: International Federation of Gynecology and Obstet- rics; EGF pidermal growth factor. 64 7 58 13 22 41 8 66 5 10 33 11 17 63 8 45 26 23 48 35 36 68 3 63 R tive to molecular-genetic prognostic factors in endome- trial cancer. A bias toward more aggressive cancers was present in the current study because all patients had a large enough tumor to permit collection of a sample for research while preserving adequate tissue for standard pathologic examination. As a result, small Stage IA le- sions are underrepresented relative to their expected frequency. Small superficially invasive endometrial cancers rarely metastasize or recur, however. Although the relationship of prognostic factors to Table 3. Multivariable Analysis of Risk Factors for Persistent/Recurrent Endometrial Cancer Controlling for Age at Diagnosis and Race Variable Wald chi-sauare P value Risk ratio FIG0 stage 7.87 0.005 3.9 Grade 7.88 0.005 2.7 Myometrial invasion 5.08 0.024 1.7 DNA ploidv 4.83 0.028 2.5 FIGO: International Federation of Gynecology and Obstetrics long-term survival is the best measure of their use, the relatively short follow-up of some of the patients in the current study precluded this. Instead, we used disease- free survival as a measure of poor outcome, because very few patients with persistent or recurrent disease after primary surgery will be cured. Using this end point, 29 of patients were classified as having persis- tent or recurrent disease, which allowed for a meaning- ful statistical analysis. Although a separate analysis of prognostic factors in cancers that were confined to the uterus at diagnosis would be useful, this was not feasi- ble because of the small number of recurrences in this group (8 of 66 [12 ]). Univariable analysis confirmed the significance of known surgical-pathologic factors, including stage, his- tologic grade, histologic type, depth of myometrial in- vasion, and the presence of metastatic In ad- dition, race was a significant prognostic factor in this set of patients, however, this is likely a confounding vari- able because these patients tended to have advanced stage and moderately to poorly differentiated cancers. Univariable analysis of the molecular-genetic variables also demonstrated the prognostic significance of ploidy, p53, HER- 2/neu, estrogen receptor, and progesterone receptor. Table 4. Relationship of FIGO Stage, Grade, and Myometrial Invasion With DNA Ploidy No. of Aneuploid uatients no. ( ) P Stage 8(12) 0.001 and I1 66 111or IV 34 14 (41) Histologic grade Well differentiated 24 1(4) Poorly differentiated 20 7 (35) None 10 1(10) Middle 1 3 18 6 (33) Moderately differentiated 56 14 (25) 0.035 Myornetrial invasion '(12) 0.094 nner 1/3 41 Outer 1 3 31 10 zy 32 ~ FIGO: International Federation of Gvnecolonv and Obstetncs.  CANCER zyxwvutsrq ay 1,2994, Volume 73, No 9 zyxwv 384 1 zyxwvutsrqponm o zyxwvutsr 0.8 2 0.6 2 0.4 0.2 0.0 zyxwvuts c z Stages 1 11 Stages IllllV 1 o 0.8 zyxw   0 6 z n 2 0.4 0.2 0 0 zyxw C Well Moderate Poor t I I I ~~ 0 10 20 30 40 50 60 0 10 20 30 40 50 60 Months Months C 0 6 2 0.4 0.2 1.0 Diploid .8 C 2 0.4 Aneuploid o.2 t I I zyxwvut   0.0 0.0 I 0 10 20 30 40 50 60 0 10 20 30 40 50 60 Months Months zyx igure 1 Kaplan-Meier analysis of factors found to be independent variables predictive of the presence of persistent or recurrent endometrial cancer. (Top left) International Federation of Gynecology and Obstetrics stage, (top right) histologic grade, (bottom left) myometrial invasion (0 = no invasion, 1 = inner third, 2 = middle third, 3 = outer third), (bottom right) DNA ploidy. To determine the significance of molecular-genetic prognostic factors relative to more conventional surgi- cal- pathologic prognostic factors, we performed a multivariable analysis controlling for varied follow-up times. Results showed that stage, grade, DNA ploidy, and myometrial invasion were significant variables pre- dictive of poor outcome. We were not surprised that stage was a significant variable, because the presence of advanced-stage cancer is requisite for one to have persistent disease after initial surgery. In addition, sev- eral previously published surgical-pathologic studies have demonstrated that histologic grade and myome- trial invasion are strong independent prognostic fac- tor~.',~ ith regard to the molecular-genetic prognostic factors, only DNA ploidy was significant in the multi- variable analysis. HER-2/neu is the next variable that would be added in the stepwise analysis, however. In fact, when ploidy was excluded from the final model, HER-2/neu attained sigrulicance. The strong correlation between ploidy and HER- 2/neu can explain this in part. In prior studies, ploidy of endometrial cancers has been determined using flow cytometric analysis of par- affin-embedded tissue In the current study, ploidy was determined using computed image cytometry of touch preparations of frozen endometrial cancer samples.'8 Although the two techniques are rel- atively comparable,'' advantages of image cytometry include the fact that intact cells rather than dissociated nuclei are used and that the operator can visually ex- clude normal cells from the analysis. Similar to prior studies that used flow cytometry, we found that 22% of endometrial cancers had an aneuploid DNA content, including 12% of Stage I and I cases. In view of the poor outcome associated with aneuploidy, Stage 1/11 aneuploid cancers might represent a subset in which adjuvant therapy should be considered. In addition to ploidy determination, both image and flow cytometry can be used to determine the prolif- eration index of cancers. In some,11812 ut not studies, high proliferation index also has been found to be a poor prognostic factor in endometrial cancers. Al- though we did not examine proliferation index in the current study, future studies of prognostic variables in endometrial cancer probably should include this vari- able. It remains unclear, however, whether use of these tests to identify endometrial cancers with a high risk of
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!