Chapter 59 Cytophotometric DNA-Content Analysis in Malignancies of the Female Genital Tract Gregory J. Naus and Michael P. Meyer Table Of Contents | |
Gregory J. Naus, MD Michael P. Meyer, MS Hyg |
INTRODUCTION CYTOPHOTOMETRIC METHODS DNA HISTOGRAM CHARACTERISTICS CARCINOMA OF THE OVARY CARCINOMA OF THE ENDOMETRIUM CARCINOMA OF THE CERVIX CONCLUSIONS REFERENCES |
INTRODUCTION |
Abnormalities in DNA-content (aneuploidy) and cell cycle distribution are features of some malignant neoplasms that have been associated with aggressive biologic behavior. These abnormalities can be identified with cytophotometric methods, which measure the amount of DNA in large numbers of individual cells. After giving a brief discussion of cytophotometric methods, we will examine the relationship of DNA content and cell cycle abnormalities in malignant neoplasms of the female reproductive system with respect to prognosis, stage, and histopathologic appearance. |
CYTOPHOTOMETRIC METHODS |
Cytophotometric DNA quantitation is based on the use of DNA-specific dyes, which bind to DNA in a stoichiometric way such that the amount of dye present is directly proportional to the amount of DNA (RNA may have to be removed to ensure specificity with some dyes). These dyes have photometric characteristics, either fluorescent or absorptive, that allow them to be quantitated. By measuring the amount of dye present in the cell nucleus, the amount of DNA can be determined irrespective of whether it is mitotically active or in interphase. Flow cytometry is one of the two methods commonly used for dye quantitation. Tissues—fresh, fixed, or deparaffinized—are prepared for flow cytometric analysis by disaggregation of cells, usually using enzymatic or mechanical methods. The resulting single-cell suspension is then incubated with an appropriate fluorescent DNA dye. After being incubated, cells are driven single file in a narrow fluid stream past a light beam (usually a laser) with a predetermined wavelength and intensity. The intensity of the resulting dye fluorescence is measured and recorded as each cell passes the light beam. This method allows the DNA content of tens of thousands of cells to be measured quickly and accurately. At the same time, the characteristics of how the cell deflects or scatters the laser light beam itself are measured and recorded. These light scatter characteristics are related to the size and internal granularity of the cell and can be used to identify specific cell populations for analysis. In addition, simultaneous measurements of other fluorescent probes, such as monoclonal antibodies directed against nuclear, cytoplasmic, or membrane antigens, also can be obtained and correlated with cell cycle measurements. The second method available for DNA dye quantitation is light microscope-based static cytophotometry, in which fixed histologic sections or smear preparations are incubated with an absorptive DNA dye that is then quantitated after visual identification of appropriate cells. Although relatively labor intensive, advances in instrumentation with computer-assisted image processing have made static cytophotometry a reliable and accurate method for DNA quantitation that is especially adaptable to small tissue samples. |
DNA HISTOGRAM CHARACTERISTICS |
Results of cytophotometric cellular DNA-content measurements are represented as frequency histograms in which the horizontal and vertical axes represent the amount of DNA and frequency, respectively. Cytophotometric DNA frequency histograms for most nongonadal, non-neoplastic tissues are characterized by the presence of two peaks (peaks 1 and 2) with scattered cells in between. Peak 2 usually is smaller than peak 1 and has twice the amount of DNA (Fig. 1A and 1B). These characteristics are a reflection of the underlying cell cycle, in which most cells are not actively synthesizing DNA for replication (G1/O phase); such cells are represented in peak 1. A second group of cells has completed synthesis of DNA in preparation for cell division (G2/M phase); such cells are represented in peak 2. Peak 2 is smaller than peak 1 because there are generally fewer cells in G2/M than in G1/O. The position of peak 2 is a reflection of the tetraploid DNA content of the G2/M cells. In addition to these two groups of cells, some cells have begun synthesizing DNA for replication, but have yet to complete the process (synthesis [S] phase). Such cells are represented by the scattered cells located between peaks 1 and 2. Figure 1A and B represents examples of such DNA histograms, and they differ only in that Figure 1B shows higher proliferative activity than Figure 1A, which is reflected as an increased proportion of S and G2/M cells. DNA-content frequency histograms from malignant neoplasms may be indistinguishable from those generated from non-neoplastic tissues (differences may exist that are beyond the resolution of cytophotometric techniques). Cytophotometric DNA analysis, therefore, cannot be used to exclude the presence of malignancy. Malignancies with DNA-content histograms similar to that shown in Figure 1A are categorized as euploid (or not demonstrably aneuploid), whereas those with an increased proportion of S and G2/M cells are categorized as euploid with a high proliferative fraction (see Fig. 1B). Figure 1C is an example of the third type of DNA-content histogram that can be found in malignant neoplasms. Such histograms are characterized by the presence of a population of cells with an abnormal DNA content. This abnormality is reflected in the DNA histogram as an additional peak. Tumors with this type of DNA histogram are categorized as aneuploid. (Although cytogenetic terms such as aneuploid, euploid, and tetraploid are not entirely appropriate when applied to cytophotometric DNA analysis, their use is established in the literature and is used in the remainder of the chapter.) Aneuploid histograms can be subclassified based on the position of the aneuploid cell population. This position is expressed relative to that of the euploid G1/O peak. It is called the DNA index and is derived by dividing a peak's position by that of the euploid G1/O peak. For example, the DNA index of the aneuploid peak represented in Figure 1C is 1.5 (75 divided by 50). Most aneuploid cell populations are located within the normal S region and therefore have DNA indices between 1 and 2. Some aneuploid cell populations are located in the normal G2/M region and have a tetraploid DNA index of 2; some are located below the normal G1/O region or above the normal G2/M region and have DNA indices of less than 1 and greater than 2, respectively. In addition, some tumors contain more than one aneuploid cell population and are classified as multiploid aneuploid. Finally, it should be noted that for many tumor types the distribution of DNA indices is not random, but bimodal, with aneuploid peaks clustering in the near-diploid and subtriploid to tetraploid regions. In the remainder of this chapter, we examine the prognostic value of classifying malignancies of the female genital tract based on DNA-histogram type. |
CARCINOMA OF THE OVARY |
A clinically significant relationship between cytophotometrically determined DNA-content abnormalities and aggressive biologic behavior has been established in ovarian carcinoma.1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24 Kallioniemi and associates2 examined the DNA content of 157 ovarian malignancies (89 stages I and II; 68 stages III and IV) and showed a significant negative association of both abnormal DNA content and cell cycle distribution with survival (clinical follow-up for censored observations 2 to 17 years). Tumors were categorized as low-risk, intermediate-risk, or high-risk based on a combination of both of these DNA histogram characteristics. Low-risk tumors (20%) were nonaneuploid with a low S-phase fraction, whereas high-risk tumors (32%) were characterized by either multiploidy or aneuploidy combined with a hypertetraploid DNA index or a high S-phase fraction. The remaining 48% of intermediate-risk tumors had DNA-content histograms that did not meet either of these sets of criteria. The risk of death for patients with high-risk and intermediate-risk DNA-histogram type tumors was 9.6 and 5.6 times higher, respectively, than that for patients with low-risk DNA-histogram type tumors. This association between survival and DNA histogram type was retained even after the effects of other significant covariants (e.g., stage, residual tumor, age, treatment, histopathologic type) were removed. In addition, although the incidence of aneuploidy was higher in high-stage tumors, the prognostic value of cytophotometric DNA analysis was retained in both low-stage (I and II) and high-stage (III and IV) tumors. Khoo and colleagues15 reported similar results for 125 ovarian adenocarcinomas in which multivariate analysis identified only ploidy and postsurgical residual disease as independent prognostic variables. The prognostic value of DNA-content analysis also has been confirmed when only high-stage ovarian carcinomas are examined.4,5,6,7,8,9,10,15,16,17,18,21,22,23 Friedlander and co-workers10 examined the prognostic value of cytophotometric DNA-content analysis in 91 stage III and IV ovarian epithelial malignancies with a median clinical follow-up of 20 months (range, 9 to 36 months). Aneuploid cell populations were resolved in 69% of tumors. These aneuploid tumors had a median survival of approximately 55 weeks, whereas the 29 tumors without resolvable aneuploid cell populations had not yet reached their median survival at 100 weeks. Multivariate analysis of DNA content, stage, chemotherapy group (sequential or combined chlorambucil and cis-platinum), histopathologic grade and type, amount of residual tumor, and performance status showed that DNA content was the major determinant of survival for these advanced-stage tumors, with stage and performance status also remaining significant prognostic factors. Blumenfeld and associates4 reported similar results among 84 patients with stages II to IV ovarian epithelial malignancies with an average clinical follow-up of 33 months (range, 16 to 88 months). They divided tumors into two risk groups based on cytophotometric DNA analysis. High-risk tumors consisted of those with nontetraploid aneuploid cell populations (52%). The remaining 40 diploid and tetraploid tumors were categorized as low-risk. The median survival for patients with high-risk and low-risk tumors was 19 and 48 months, respectively. Univariate analysis demonstrated that DNA content, stage, age, histopathologic grade (grade 1 vs grades 2 and 3 combined), residual disease after surgery, and single-agent versus triple-agent chemotherapy were also significant prognostic factors. Multivariate analysis of these factors, however, showed only DNA content, stage, and age to be independent prognostic factors, DNA content being the most important determinant of survival for stage III tumors. Kaern and associates21 also demonstrated DNA-content abnormality to be an independent prognostic variable in 169 stage III and IV ovarian adenocarcinomas with a minimum follow-up of 2 years. In a study of 184 stage III and IV ovarian carcinomas, however, Pfisterer and colleagues22 reported that although aneuploidy was associated with decreased survival, they did not find it to be an independent prognostic variable using multivariate analysis. Although the incidence of aneuploidy in stage I and II ovarian carcinomas appears to be significantly lower than the 50% to 80% reported in advanced-stage disease,25 its prognostic value is retained.1,2,4,7,8,10,12,14,20,25 Kallioniemi and associates2 found that 42 of 89 (47%) stage I and II ovarian carcinomas were aneuploid and that these aneuploid low-stage tumors had a significantly poorer prognosis than the nonaneuploid tumors. This was most apparent for multiploid tumors and for those aneuploid tumors with DNA indices greater than 2. In addition, further prognostic information was obtained when nonaneuploid tumors were divided on the basis of S-phase fraction. Similarly, Vergote and co-workers14 demonstrated that nondiploidy could be resolved with flow cytometry in archival tumor tissue in 49% of 279 stage I ovarian carcinomas and that it was a powerful independent prognostic factor of disease-free survival in these patients. In addition to the demonstrated association between aneuploidy and prognosis in ovarian adenocarcinoma detailed above, an association between aneuploidy and high histopathologic grade has been reported by most,1,11,14,15,21,22,25,26,27,28 but not all,4,29 investigators, including those who based the histopathologic assessment on quantitative morphology.5,6,7,24 A significant association of cytophotometric DNA content and cell cycle abnormalities with histopathologic grade was reported in ovarian carcinoma by Kallioniemi and associates.25 This association was stronger when DNA-histogram type was compared with histopathologic grading based on nuclear characteristics (e.g., nuclear-to-cytoplasmic ratio, nuclear chromatin pattern, and nucleolar abnormalities) than when histopathologic grading was based on architectural features (e.g., percentage of microscopically solid tumor). This relationship between nuclear grade and aneuploidy was retained for both low-stage and high-stage tumors. In addition, a significant relationship between tumor histopathologic type (World Health Organization histopathologic types: serous, mucinous, clear cell, endometrioid, and undifferentiated) and the incidence of aneuploidy was reported by these investigators. This was due primarily to a lower than average incidence of aneuploidy among mucinous tumors (39%) and a higher than average incidence of aneuploidy in undifferentiated tumors (75%) as compared with serous, clear cell, and endometrioid tumors (61%, 54%, and 61%, respectively). This association between histopathologic type and DNA-histogram abnormalities was not, however, independent of nuclear grade and was eliminated when groups were adjusted for histopathologic nuclear grade. This relationship between histopathologic type and DNA-histogram abnormalities has been confirmed by most4,10,14,22,23,26 but not all15 investigators. Although the prognostic values of histopathologic and cytophotometric DNA analysis are interrelated in ovarian carcinoma, cytophotometric DNA analysis is frequently the stronger prognostic variable.1,2,4,10,15,21 Although no significant relationship between cytophotometric DNA analysis and histopathologic grading was found for the 84 advanced-stage ovarian carcinomas reported by Blumenfeld and associates,4 univariate analysis showed a significant prognostic value for both of these factors. However, only DNA analysis retained its prognostic value when analyzed simultaneously with stage, age, residual disease after surgery, chemotherapy, tumor type, tumor size, and surgical treatment. Similarly, Kallioniemi and associates2 reported that although DNA analysis, architectural histopathologic grading, nuclear histopathologic grading, and mitotic index were all significant prognostic variables in ovarian carcinoma, only cytophotometric DNA analysis, nuclear histopathologic grading, and mitotic index retained prognostic value after they adjusted for covariance due to stage, residual tumor, histopathologic type, age, and chemotherapeutic treatment. In addition, when they analyzed all factors simultaneously, only cytophotometric DNA analysis and stage remained independent prognostic variables. Finally, the incidence and prognostic value of cytophotometric ploidy determination has been reported to be associated with residual disease.15,17,18,22 Multivariate analysis of 125 ovarian carcinomas reported by Khoo and associates15 showed not only that the amount of residual disease and the ploidy status were independent prognostic variables, but that the prognostic value of ploidy analysis was particularly significant in patients with minimal residual disease. Although the prognostic value of cytophotometric DNA-content analysis is well established in ovarian carcinomas, its clinical value in epithelial tumors of low malignant potential (LMP) remains undemonstrated.30,31,32,33,34,35,36,37 Conclusions from limited reports include the following:
In addition to the lower incidence of aneuploidy in ovarian LMP tumors, DNA indices in these tumors have been reported to be generally lower than those in ovarian carcinomas, clustering in the near-diploid region.1,34 Finally, relatively little is known about the possible clinical utility of cytophotometric DNA-content analysis in nonepithelial ovarian neoplasms. Although investigations of its prognostic value in granulosa cell tumors have been reported,38,39,40,41 overall numbers are small and conclusions contradictory. |
CARCINOMA OF THE ENDOMETRIUM |
A clinically significant relationship between cytophotometric DNA histogram abnormalities and aggressive biologic behavior also has been demonstrated in endometrial carcinoma.42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62 In addition, although the incidence of high-stage tumors is relatively low, the relationship between the incidence of cytophotometric DNA abnormalities and the extent of disease at diagnosis found in ovarian carcinoma also is evident in endometrial carcinoma.44,48,63 Lindahl and co-workers63 showed that although 80 of 189 (42%) endometrial carcinomas had aneuploid cell populations, such populations were resolvable in 72 of 178 FIGO clinical stage I and II tumors (40%), but could be identified in 8 of 11 FIGO clinical stage III and IV tumors (73%). Similarly, Britton and colleagues48 reported a rate of aneuploidy of 17.4% and 48.6% for stage I and II and stage III and IV tumors, respectively. The prognostic value of cytophotometric DNA analysis was demonstrated by Lindahl and co-workers45 in 110 stage I and II endometrial carcinomas with at least 2 years of clinical follow-up. Of these tumors, 40% were demonstrably aneuploid. Patients with aneuploid tumors had a 2-year relapse rate of 28%, compared with a 6% 2-year relapse rate for patients with tumors without resolvable aneuploid cell populations. When 166 endometrial carcinomas were considered irrespective of the length of clinical follow-up, the projected percentage of patients free of disease at 40 months was approximately 90% and 55% for euploid and aneuploid tumors, respectively. Univariate analysis showed that in addition to DNA-content analysis, both histopathologic grade and depth of myometrial invasion were significantly associated with relapse rate and with the presence of aneuploidy. Multivariate analysis of cytophotometric DNA analysis, histopathologic grade, depth of myometrial invasion, and estrogen receptor status, however, showed that only cytophotometric DNA analysis retained its prognostic value, with a relative risk of relapse of more than 16 for aneuploid tumors as compared with nonaneuploid tumors. This was 9.6 times higher than the relative risk of relapse based on grouping by histopathologic grade (well and moderate versus poor histopathologic differentiation). Similar results with respect to ploidy and prognosis were reported by Britton and co-workers.48 They studied 256 endometrial carcinomas in which only 10% of diploid tumors relapsed compared with a 39% relapse rate for nondiploid tumors. Multivariate analysis of stage, grade, depth of myometrial invasion, histologic type, peritoneal cytology, and cytophotometric ploidy analysis showed that only histologic type and ploidy were significant independent prognostic variables. In contrast, Wagenius and associates50 and Stendahl and colleagues62 reported that although univariate analysis confirmed the prognostic value of ploidy analysis, cytophotometric estimation of S-phase fraction was a more significant prognostic variable and that S-phase fraction but not ploidy retained independent prognostic value in multivariate analysis. These results are supported by others,46,57 and although there appears to be a relationship between aneuploidy and proliferative activity, with aneuploid tumors tending to have higher proliferative activity than nonaneuploid tumors, the prognostic value of cell cycle analysis may be at least partially independent of DNA content in endometrial carcinoma. Although the prognostic value of cytophotometric DNA analysis appears to be greater than that of histopathologic grading in endometrial carcinoma, there is a significant association of both the incidence and level of aneuploidy (DNA index) with histopathologic assessment of differentiation.47,49,50,54,55,56,57,58,60,63 Lindahl and co-workers63 reported that nontetraploid aneuploidy was resolvable in 29% of 130 well and moderately differentiated endometrial carcinomas, but could be identified in 62% of 92 poorly differentiated tumors. In addition to this difference in the incidence of aneuploidy per se, the distribution of the DNA indices of aneuploid tumors also was associated with histopathologic grade. Specifically, peridiploid DNA indices (less than 1.3) were characteristic of well and moderately differentiated tumors, whereas poorly differentiated tumors tended to have DNA indices greater than 1.3. In addition, similar to ovarian carcinoma,2,3,4,5,6,7,8,9,10,11,12,13,24,25 the incidence of multiploidy increased with decreasing histopathologic differentiation. This relationship between biologic behavior and DNA index has been confirmed by some,42,43,44,47,56 but not all,60 investigators. Iverson44 identified aneuploidy in 3 of 30 grade 1, 4 of 12 grade 2, and 7 of 9 grade 3 endometrial carcinomas. The DNA indices for histopathologically well differentiated aneuploid tumors were all less than 1.3, whereas DNA indices greater than 1.3 were found in all moderately and poorly differentiated tumors. Atkin42 reported that among 107 patients with endometrial carcinoma surviving for 5 years and 79 patients with endometrial carcinoma not surviving for 5 years, the incidence of aneuploid tumors with DNA indices greater than approximately 1.4 was 9% for well, 16% for moderately, and 37% for poorly differentiated tumors. Overall, 18% of these endometrial carcinomas had aneuploid cell populations with DNA indices greater than 1.4. These accounted for 35% of all 5-year fatalities, resulting in a 5-year survival rate of only 18% for patients with these tumors. In contrast, the remaining 152 tumors with DNA indices less than 1.4 had an overall 5-year survival rate of 66%. Atkin also demonstrated that the prognostic value of tumor characterization based on DNA index was retained at all levels of histopathologic differentiation and also that histopathologic differentiation retained some prognostic value in both ploidy groups. The prognostic value of the level of aneuploidy also has been shown to be a significant factor in ovarian carcinoma. Klemi and associates1 categorized DNA-content risk groups based on the presence of aneuploid cell populations with DNA indices greater than 1.3. A significant relationship of both prognosis and histopathologic grading with the presence of these aneuploid cell populations was demonstrated. In addition, classification of tumors based on DNA index was a more important prognostic factor than aneuploidy per se. In summary, in both ovarian and endometrial carcinomas the characteristics reflected in cytophotometric DNA analysis and histopathologic grade are related, but not identical, and each contains some degree of unique prognostic information. Proliferative activity also has been related to histopathologic grade in endometrial carcinoma.46,64,65,66,67 Specifically, increased proliferative activity has been associated with decreasing histopathologic differentiation. An association between cytophotometric DNA abnormalities and steroid receptor content has also been suggested.45,55,64,65,68 Although the exact nature of this association is uncertain, the prognostic value of cytophotometric DNA analysis appears to be independent of steroid receptor concentration.45 Finally, an association between an increased incidence of aneuploidy and deep myometrial invasion has been suggested by a majority of investigators.45,47,49,50,52,56,58,59,68 Lindahl and co-workers52 reported that DNA index and depth of myometrial invasion are significant and independent prognostic variables in stage I and II endometrial carcinoma and can be used in combination to define high-risk and low-risk populations with a relapse rate of 38% and 4%, respectively. The clinical value of cytophotometric DNA-content analysis in uterine sarcomas is uncertain.69,70,71,72,73 Malmstrom and associates69 analyzed 37 uterine sarcomas (including 18 leiomyosarcomas and 14 malignant mixed mu¨llerian tumors) and demonstrated a significantly higher incidence of aneuploidy and a higher S-phase fraction in high-stage and poorly differentiated tumors. In addition, multivariate analysis showed that both ploidy and S-phase fraction were significant and independent prognostic variables. Because of the small number of total cases reported, however, the prognostic value of cytophotometric analysis in uterine sarcomas is best considered preliminary. |
CARCINOMA OF THE CERVIX |
Many investigators have examined the prognostic value of cytophotometric DNA content analysis in cervical carcinoma.42,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99 The cytophotometric DNA characteristics associated with aggressive biologic behavior in adenocarcinoma of the cervix appear similar to those identified in adenocarcinomas of the ovary and endometrium. Leminen and colleagues82 reported an aneuploidy rate of 31% in 125 cases of cervical adenocarcinoma and adenosquamous carcinoma, with a notably high incidence of triploidy identified among these tumors. Both ploidy and S-phase fraction were shown to be independent prognostic variables in predicting 5-year survival. In addition, a significant relationship between ploidy and tumor size, histopathologic grade, and clinical stage was demonstrated. Although many reports have investigated the value of cytophotometric DNA-content analysis in cervical squamous cell carcinomas, results are less consistent than those reported in adenocarcinomas of the female genital tract. In addition, there appears to be little association between cytophotometric DNA abnormalities and histopathologic grade or stage in cervical squamous cell carcinoma. When DNA content is analyzed with flow cytometry, ploidy is identified as a significant prognostic factor in cervical squamous carcinoma by some investigators,80,81,85,88,90,94 but not the majority.83,84,86,87,89,91,92,93,95,96,97,98,99 Results with respect to the prognostic value of cytophotometrically determined increased proliferative activity, however, are more consistent with most,77,84,85,86,89,93,96,97 but not all,90,91,98,99 reports showing a significant association between biologic behavior and proliferative activity. Tribukait77 reported that although aneuploidy did not correlate with survival in 120 cases of cervical carcinoma, when these tumors were categorized on the basis of the proportion of cells in S-phase (less than 11%, 11% to 20%, and greater than 20%), there was a significant correlation between these categories and survival. Tribukait suggested that the prognostic value of cytophotometric DNA analysis ultimately may be a reflection of differences in proliferation rate and that, because the rate of proliferation and aneuploidy may be correlated in some carcinomas, both factors may have prognostic value in these tumors. In cervical carcinomas, however, in which nonaneuploid tumors have S-phase fractions that may be similar to those found in aneuploid tumors, only measurements of cell cycle demonstrate prognostic value. These findings were supported by the same group of investigators in 307 cervical squamous cell carcinomas in which S-phase fraction, but not aneuploidy, was demonstrated to be a significant and independent prognostic factor.84 Although a majority of reports suggest that the prognostic value of cytometric DNA-content analysis is reflected in the measurement of proliferative activity rather than aneuploidy, it has been suggested that the DNA index may be a significant prognostic factor in cervical squamous cell carcinoma.80,81,90,94 Similar to many carcinomas, the distribution of aneuploidy in squamous cell carcinoma of the cervix is bimodal.77,79,80,81,100 (A similar distribution of DNA indices also has been shown in high-grade squamous intraepithelial lesions of the cervix,77,101,102 suggesting that DNA-content abnormalities may occur early in the evolution of cervical squamous neoplasia.) It has been reported in other tumors (e.g., ovarian, endometrial) that although aneuploid tumors in general may be more aggressive than diploid tumors, aneuploid tumors with high DNA indices (with the exception of tetraploidy) show particularly aggressive biologic behavior. Similar results were reported by Jakobsen,81 who studied 171 FIGO stage IB to III cervical carcinomas in which aneuploid cell populations with a DNA index greater than 1.5 were resolved in 80 tumors. Recurrence rates for tumors with and without such high DNA index aneuploid cell populations were 55% and 20%, respectively. Lai and co-workers90 reported similar results in 411 stage IB and II cervical carcinomas with primary surgical treatment, in which 5-year recurrence-free survival rates were not significantly different when tumors were stratified based on aneuploidy per se. Only when tumors were stratified on the basis of a DNA index greater than 1.3 was an independent significant prognostic value demonstrated. In addition, compared with surgically treated cervical tumors, the predictive value of the DNA index may be different in tumors treated with primary radiation therapy.42,79,84,88,103,104,105,106 Strang and associates,84 reporting on 307 patients receiving primary radiation therapy, showed that although aneuploidy per se was not a significant prognostic variable, multivariate analysis demonstrated a poorer survival for near-diploid tumors. Similar results also have been reported by other investigators using Feulgen microspectrophotometry42,79 and flow cytometric DNA-content analysis.88,103,104,105,106 These results suggest an association between aneuploidy and increased radiosensitivity, a finding that is both clinically relevant and also potentially confounding in determining the prognostic value of aneuploidy. Finally, differences in tumor stage may be an additional confounding factor in comparing tumors treated by surgery versus radiation. In summary, although there appears to be prognostic value in cytophotometric DNA-content analysis for cervical squamous cell carcinoma, the relative value of ploidy and cell cycle measurements remains unclear and may be confounded by differences in methods of tissue procurement, DNA-histogram analysis methods, definitions of aneuploidy and cell cycle parameters, stage, and treatment modalities. In our laboratory, by using simple DNA-histogram analysis methods and a conservative definition of aneuploidy, we have found that a combination of both DNA index and elevated S-phase fraction predicts aggressive biologic behavior in stage IB to IIA cervical squamous cell carcinoma.85 |
CONCLUSIONS |
Cytophotometric DNA analysis is a clinically important method of tumor characterization. As detailed above, there is a clinically significant association between DNA-content abnormalities and proliferative activity and aggressive behavior in adenocarcinomas of the ovary, endometrium, and cervix. Although such abnormalities are more frequent in advanced-stage disease, this prognostic value is retained in both low-stage and high-stage tumors. In addition, although an association exists between tumor characterization based on cytophotometric DNA analysis and histopathologic assessment in these carcinomas, the prognostic value of cytophotometric DNA-content analysis is generally greater and, in some cases, complementary to standard histopathologic characterization. Although there also appears to be a prognostic value of cytophotometric DNA-content analysis in squamous cell carcinoma of the cervix, the relative significance of abnormalities of DNA content and increased proliferative activity is uncertain. Although this relative uncertainty may be the result of true differences among different organs or tumors of various histologic types, it also may be the result of differences in tissue source, differences in methods of histogram analysis, and differences in treatment modalities (e.g., surgery versus radiation therapy). The identification of these potentially confounding variables underscores the need for simple, reproducible, and standardized laboratory methods based on minimal underlying assumptions, as well as the importance of a rigorous definition of clinical variables, including treatment modality.107,108,109 The prognostic value of cytophotometric DNA-content analysis in less common carcinomas of the female genital tract, as well as sarcomas and germ cell neoplasms, remains to be determined. |