The prognostic variability of ovarian tumor grading by different pathologists

Identification of biological tissues by rapid evaporative ionization mass spectrometry

The newly developed rapid evaporative ionization mass spectrometry (REIMS) provides the possibility of in vivo, in situ mass spectrometric tissue analysis. The experimental setup for REIMS is characterized in detail for the first time, and the description and testing of an equipment capable of in vivo analysis is presented. The spectra obtained by various standard surgical equipments were compared and found highly specific to the histological type of the tissues. The tissue analysis is based on their different phospholipid distribution; the identification algorithm uses a combination of principal component analysis (PCA) and linear discriminant analysis (LDA). The characterized method was proven to be sensitive for any perturbation such as age or diet in rats, but it was still perfectly suitable for tissue identification. Tissue identification accuracy higher than 97% was achieved with the PCA/LDA algorithm using a spectral database collected from various tissue species. In vivo, ex vivo, and post mortem REIMS studies were performed, and the method was found to be applicable for histological tissue analysis during surgical interventions, endoscopy, or after surgery in pathology.

Traditional Chinese Medicine Herbal Treatment May Have a Relevant Impact on the Prognosis of Patients With Stage IV Adenocarcinoma of the Lung Treated With Platinum-Based Chemotherapy or Combined Targeted Therapy and Chemotherapy

Background: Targeted therapy (TT), chemotherapy, and traditional Chinese medicine herbal treatment (TCM) can improve the prognosis of advanced pulmonary adenocarcinoma patients. Their independent prognostic value is unknown. Objective: To study whether TCM improves survival in stage IV pulmonary adenocarcinoma patients with platinum-based chemotherapy (PBT), or combined PBT and second-line TT. Methods: Retrospective analysis of 133 fully ambulant clinical outpatients treated with PBT alone or PBT with/without second-line TT, with/without TCM. Univariate (Kaplan—Meier) and multivariable (Cox model) survival analysis were performed, using disease-specific mortality as an endpoint. Results: Gender ( P = .002), TT ( P < .0001), and TCM ( P < .0001) had univariate prognostic value but not age, radiotherapy, or TCM syndrome differentiation ( P > .10). TCM herbal treatment ( P < .0001) and TT ( P = .03) had multivariable independent prognostic value. TCM-treated patients (n = 103, PBT+TT+TCM+ = 62; PBT+TT−TCM+ =41) had 88% 1-year overall survival rate with median survival time (MST) of 27 months, contrasting 27% 1-year overall survival and MST of 5.0 months for non-TCM-treated (n = 30) patients. Patients with chemotherapy/TT/TCM (PBT+TT+TCM+, n = 62), TCM without TT (PBT+TT−TCM+, n = 41), or chemotherapy only (PBT+TT−TCM−, n = 30), had 1-year survival rates of 94%, 78%, and 27% respectively; for these 3 groups, respectively, MST was not reached (MST of 30.9 months), 22.6, and 5.0 months ( P < .0001). Conclusions: TCM herbal treatment may improve survival of stage IV pulmonary adenocarcinoma patients treated with chemotherapy without or with second-line TT. This warrants formal phase 1 and 2 trials and ultimately properly designed prospective clinical validation trials with adequate methodology developed for data collection.

Clinical value of morphometric and DNA flow cytometric variables as independent predictors of survival in epithelial ovarian carcinoma: a 5-year follow-up study

The aim of this follow-up study is to validate the clinical significance of quantitative morphometric and DNA flow cytometric variables as independent prognostic factors of overall survival and progression-free survival in epithelial ovarian carcinoma. Tumor samples were collected from 135 patients with epithelial ovarian carcinoma at 3 hospitals in the Netherlands. Evaluated clinico-pathologic variables were age, histologic subtype, differentiation grade, clinical stage [International Federation of Gynecology and Obstetrics (FIGO)], presence of ascites, serum CA-125, and the completeness of debulking surgery. Morphometry and DNA flow cytometric techniques were assessed on each tumor sample to determine the mitotic activity index (MAI), volume percentage epithelium, mean nuclear area (MNA), standard deviation of MNA (SD MNA), nuclear perimeter (NP), and DNA ploidy. Univariate analysis showed that differentiation grade, FIGO stage, presence of ascites, preoperative CA-125 levels, DNA ploidy, and MAI, NP, and MNA were of significant prognostic value. After multivariate analysis (using forward Cox proportional hazard analysis), only differentiation grade and FIGO stage remained significant. From this study, we can conclude that morphometry and DNA flow cytometry are not independent prognosticators and therefore have no clinical value in predicting prognosis in ovarian carcinoma.

Effect of interleukin-8 gene silencing with liposome-encapsulated small interfering RNA on ovarian cancer cell growth

BACKGROUND

Interleukin-8 (IL-8) is a proangiogenic cytokine that is overexpressed in many human cancers. We investigated the clinical and biologic significance of IL-8 in ovarian carcinoma using human samples and orthotopic mouse models.

METHODS

Tumor expression of IL-8 was assessed by immunohistochemistry among ovarian cancer patients (n = 102) with available clinical and survival data. We examined the effect of IL-8 gene silencing with small interfering RNAs incorporated into neutral liposomes (siRNA-DOPCs), alone and in combination with docetaxel, on in vivo tumor growth, angiogenesis (microvessel density), and tumor cell proliferation in mice (n = 10 per treatment group) bearing orthotopic taxane-sensitive (HeyA8 and SKOV3ip1) and taxane-resistant (SKOV3ip2.TR) ovarian tumors. All statistical tests were two-sided.

RESULTS

Of the 102 cancer specimens, 43 (42%) had high IL-8 expression and 59 (58%) had low or no IL-8 expression; high IL-8 expression was associated with advanced tumor stage (P = .019), high tumor grade (P = .031), and worse survival (median survival for patients with high vs low IL-8 expression: 1.62 vs 3.79 years; P < .001). Compared with empty liposomes, IL-8 siRNA-DOPC reduced the mean tumor weight by 32% (95% confidence interval [CI] = 14% to 50%; P = .03) and 52% (95% CI = 27% to 78%; P = .03) in the HeyA8 and SKOV3ip1 mouse models, respectively. In all three mouse models, treatment with IL-8 siRNA-DOPC plus the taxane docetaxel reduced tumor growth the most compared with empty liposomes (77% to 98% reduction in tumor growth; P < .01 for all). In the HeyA8 and SKOV3ip1 models, tumors from mice treated with IL-8 siRNA-DOPC alone had lower microvessel density than tumors from mice treated with empty liposomes (HeyA8: 34% lower, 95% CI = 32% to 36% lower [P = .002]; SKOV3ip1: 39% lower, 95% CI = 34% to 44% lower [P = .007]). Compared with empty liposomes, IL-8 siRNA-DOPC plus docetaxel reduced tumor cell proliferation by 35% (95% CI = 25% to 44%; P < .001) and 38% (95% CI = 28% to 48%; P < .001) in the HeyA8 and SKOV3ip1 models, respectively.

CONCLUSIONS

Increased IL-8 expression is associated with poor clinical outcome in human ovarian carcinoma, and IL-8 gene silencing decreases tumor growth through antiangiogenic mechanisms.

Recurrent epithelial ovarian cancer: validation and improvement of an established prognostic index

OBJECTIVES

To validate and improve an established prognostic index in patients with recurrent ovarian cancer.

METHODS

A Canadian three-covariate prognostic index (tumour grade at diagnosis, initial performance status, and time to relapse/primary progression (TRP)) was validated in a well-defined cohort of comparable Danish patients. Potential parameters to be included in an improved prognostic index were revealed by univariate and multivariate analyses in the Danish validation group.

RESULTS

The Canadian index validated in the Danish patient population (n=189) found a statistical significant difference in survival between the prognostic groups good and intermediate (P<0.0001), whereas there was no significant difference in survival between the prognostic groups intermediate and poor (P=0.51). In order to improve the accuracy of the index, the candidate parameters, treatment free interval (TFI), CA125 level and performance status, at time of relapse/primary progression, were added, whereas the parameters, tumour grade, and initial performance status, from the Canadian index were excluded. As the correlation coefficient between TRP and TFI was very high (r=0.91), TRP was substituted with TFI in the improved prognostic model. The final model was: 0.8 (performance status)+0.33 log (CA125)-1.31 log (TFI). The improved model was a good predictor of one-year survival (AUC 0.85; logistic regression; P<0.0001). The median survival (with 95% CI) of the four prognostic groups (A-D) was 50.6 (34.0-not available), 25.0 (22.1-33.6), 11.3 (8.5-12.9), and 5.2 (3.5-6.3) months, respectively.

CONCLUSIONS

A novel prognostic model (the Copenhagen index) for patients with recurrent ovarian cancer is presented.

A new prognostic model comprising p53, EGFR, and tumor grade in early stage epithelial ovarian carcinoma and avoiding the problem of inaccurate surgical staging

Epithelial ovarian carcinoma rarely occurs because of a single event. Therefore, no single biological tumor factor will give accurate prognostic information for all ovarian cancer patients. On the other hand, a combination of two or more independent factors may yield an improved overall prognostic index. Because FIGO stage is included in most of the previously presented models, inaccurate surgical staging in patients with apparently early disease has been a problem. In a series of 226 patients with epithelial ovarian carcinomas in FIGO stages IA-IIC, a number of clinicopathological factors (age, FIGO stage, histopathologic type, and tumor grade) were studied in relation to the biological factors p53 and epidermal growth factor receptor (EGFR), important regulators of the apoptosis and mitosis. Immunohistochemical techniques were used. All patients received adjuvant radiotherapy or chemotherapy after the primary surgery. Expression of p53 was significantly associated with the tumor grade and disease-free survival (DFS). EGFR expression was also associated with DFS. In a Cox multivariate analysis, tumor grade, p53 status, and EGFR status were all independent and significant prognostic factors with regard to DFS. A prognostic model was proposed using these factors. A low-risk group, an intermediate-risk group, and a high-risk group were defined. DFS amounted to 89% in the low-risk group (grades 1-2, p53-negative, and EGFR-negative), 66% in the intermediate-risk group (grade 3, p53-negative, and EGFR-negative or grades 1-2, p53-positive or EGFR-positive) and 39% in the high-risk group (grade 3, p53-positive, and EGFR-positive).

Nuclear size, shape, and density in endometrial carcinoma: relationship to survival at over 5 years of follow-up. Does analyzing only cells occupying the G0-G1 peak add useful information?

OBJECTIVE

The authors, using image analysis, previously demonstrated nuclear size and summed optical density to be independent prognostic indicators of recurrence in patients with endometrial carcinoma. The same tumors were analyzed by studying the optical features in the G0-G1 peak to see if this changed the values found as well as their importance as prognostic features at greater than 5 years of follow-up.

METHODS

Tumors from 74 consecutive patients, surgically treated, with endometrial cancer, were evaluated. Survival, depth of invasion, lymphvascular space invasion, FIGO stage, grade, histology were analyzed. DNA index, progesterone receptor status, as well as nuclear size (NUSZ), shape (NUSH), and summed optical density (NUSD) were evaluated. NUSZ, NUSH, and NUSD were quantified using image analysis.

RESULTS

Fifteen patients died from disease during the observation period of the study. Mean follow-up was 82 months with a median of 84 months. Forty-nine patients had stage I cancers, five stage II, 17 stage III, and three stage IV. NUSZ and NUSD were all significantly different between the original (entire cell cycle) and the re-measured (G0G1 only) values (both P < 0.001). Multivariate analysis showed both the original (P = 0.0001) and G0G1-only (P = 0.046) NUSZ and the original (P = 0.0002) and G0G1-only (P = 0.018) NUSD to be independent prognosticators of survival.

CONCLUSION

Image analysis is able to quantify cellular and nuclear parameters not otherwise quantifiable. NUSD and NUSZ correlated with traditional prognostic indicators, were demonstrated independent predictors of survival at over 5 years of follow-up. Although the re-measured NUSZ and NUSD from only the G0-G1 peak were significantly different from the original NUSZ and NUSD, they were not as valuable as prognostic factors. Nuclear size and summed optical density measured from the entire cell cycle are independent prognostic indicators of survival at greater than 5 years of follow-up. Measuring nuclear morphometric features in the G0-G1 peak only does not add any new prognostic information.

Reproducibility and prognostic variability of grade and lamina propria invasion in stages Ta, T1 urothelial carcinoma of the bladder

PURPOSE

We assessed the reproducibility and prognostic variability of grade and lamina propria invasion in stages Ta, T1 urothelial carcinoma of the bladder.

MATERIALS AND METHODS

A total of 130 consecutive stages Ta, T1 urothelial carcinomas routinely diagnosed by 15 pathologists (original diagnosis) were reviewed by 3 independent experienced pathologists using 1999 WHO criteria (diagnoses 1 to 3 and reviewer consensus diagnosis). Interreviewer disagreement cases were blindly reviewed again. Each remaining disagreement case was discussed in a multihead microscope session to attempt to solve remaining disagreements. In cases of continuing disagreement the majority diagnosis on stage and grade was considered the consensus diagnosis. Stage progression at followup was the dependent variable. Stage progression-free Kaplan-Meier survival curves and hazard ratios of each stage and grade diagnosis were calculated and prognostic variability was determined.

RESULTS

There was complete interobserver agreement on stage and grade among reviewers in 80% and 59% of cases, while it was 87.7% and 75.4%, respectively, after the second review. More than 1 grade difference occurred in 1.5% of cases (0% after the second review). The consensus and original diagnoses agreed on stage and grade in 68.5% and 62.3% of cases, respectively. Assignment of individual cases to 1 category of the 1999 WHO classification per reviewer varied considerably. The incidence of cases classified as stage T1 grade 3 by the reviewers was between 12.3% and 18.9% (average 14.1%). Consensus diagnosis grade had the strongest prognostic value (HR 68.8, range 8.9 to 528.0). Of the 63 original diagnoses of stage T1 tumors the consensus diagnosis down staged 35 (55.6%) to Ta and up staged 8 (12.7%) to T2-3. Progression was more common in the 20 consensus diagnosis stage T1 cases (5 or 25%) than in the 55 original diagnosis stage T1 cases (11 or 20%). Original diagnosis stage T1 tumors that were down staged by the consensus diagnosis showed less progression than consensus diagnosis confirmed stage T1 tumors (17% versus 25%). The prognostically worst subgroup (T1 grade 3) also showed considerable prognostic variation among reviewers (28% to 76% at 5 years of followup), in that the consensus diagnosis again had the highest prognostic significance (HR 3.5, range 1.2 to 10.2). At the end of the study all pathologists expressed that they were regularly uncertain about stage and grade assessment in an individual case in a considerable percent of all cases.

CONCLUSIONS

Observer prognostic variability in staging and grading is considerable with potentially strong implications for patients. Interobserver variation did not decrease using the new 1999 WHO grading classification.

The framework of pathology: good laboratory practice by quantitative and molecular methods

Combined confocal laser scan microscopy (CLSM) and Fourier analysis (FA) by non-pathologists of dermal collagen bundle orientation recently gave results superior to subjective evaluation by experts. According to Good Laboratory Practice (GLP) criteria, combined CLSM/FA has not yet been adequately tested to replace current collagen evaluation, but this will not take long. Non-pathologists (clinicians) will then have taken over a laboratory test historically belonging to pathology. A general trend in this direction may develop, because pathologists seem not always to care enough about clinical significance, reproducibility and prognostic value, and new demands for innovative methods. Quantitative image analysis (QIA) and molecular methods are reproducible, inexpensive, and easy to perform; they often have greater value than classical evaluations and their cost-benefit ratio is good. However, their acceptance is not as widespread as one would expect and theoretical reasons which have been advanced do not provide a satisfactory explanation. A formal implementation study was therefore performed, in which an attempt was made to modernize a classical pathology laboratory. An external customer satisfaction investigation showed that 96% of the clinicians were 'very satisfied' (the highest rating possible) with the completed innovations, contrasting with low satisfaction at the beginning. Lack of primary innovative leadership among pathologists was judged to be the dominant cause preventing implementation. Pathologists should focus on carefully reacting to new clinical needs, using GLP criteria. Reproducibility and predictive accuracy should be major themes in any pathology practice.

The reliability of clinical methods in ophthalmology

The reliability of clinical methods is often taken for granted. There is considerable evidence, however, that variability of clinical measurements, interpretations, and judgements are greater than commonly thought. Inconsistency of clinical methods has profound and widespread implications in clinical practice. The inferential capacity of measurements depends on accuracy and reproducibility. Appropriate medical decisions cannot be reached without accurate and reliable data. This review will examine the conceptual relationship of reliability, accuracy, and validity. The role of observer reliability in ophthalmic test measurements and interpretations will be addressed in the context of three case reports.

Endometriosis-associated ovarian carcinoma (EAOC): an entity distinct from other ovarian carcinomas as suggested by a nested case-control study

OBJECTIVES

Endometriosis-associated ovarian carcinoma (EAOC) has recently received increasing attention due to its suggested biological behavior, distinctive from those of usual epithelial ovarian cancer. To elucidate some of the controversies on this intriguing entity, a series of patients with EAOC were compared to ovarian carcinoma cases without concomitant endometriosis.

METHODS

To control the confounding effect of age, a nested case-control study was designed, where all 58 EAOC patients (mean age 54.5 +/- 11.5 years) were nested with four perfectly age-matched non-EAOC patients (n = 232; mean age 54.7 +/- 11.7 years) selected among 425 women representing all FIGO stages of ovarian carcinomas without endometriosis. Pertinent clinical data and results of analysis of the tumors were subjected to statistical analyses using life-table, univariate (Kaplan-Meier), and multivariate (Cox) survival techniques to disclose dissimilarities in the key biological characteristics of these two groups as well as the independent prognostic predictors of disease outcome.

RESULTS

When compared in a case-control design with four perfectly age-matched non-EAOC patients nested to each EAOC case, the patients with EAOC proved to: (1) have a lower stage disease (both FIGO and TNM) (P = 0.000), (2) show a completely different distribution of histological subtypes (significant overpresentation of endometrioid and clear cell carcinomas) (P = 0.0001), (3) have predominantly lower grade lesions (P = 0.029), (4) be devoid of any primary residual tumor (P = 0.0001), and, most importantly (5) have demonstrated a significantly better overall survival (47/58 versus 126/232; OR 2.89, 95% CI 1.56-5.34, P = 0.0001). This better survival was evident (a) in all age groups and (b) for all histological subtypes, but (c) not explained by a better stage-specific survival in any FIGO stage. The two series also differed in their significant prognostic predictors in Kaplan-Meier and Cox analyses. In the EAOC group, the most significant (P = 0.0001) predictors of OS in univariate analysis were age, histological type, observation time for endometriosis, and distribution of endometriosis. In the non-EAOC group, such significant predictors were age, residual tumor, and type of therapy. In the multivariate (Cox) model, age and FIGO stage were the only two significant independent prognostic factors shared by these two series. In addition, histological type and type of therapy proved to be significant independent predictors in the non-EAOC series.

CONCLUSIONS

These data suggest that EAOC deviates from the non-EAOC in many of its key biological characteristics. The implications of these data in the diagnosis, treatment policy, and prognostication still require confirmation by further studies, however.

Computerised morphometrical analysis in endometrial hyperplasia for the prediction of cancer development. A long-term retrospective study from northern Norway

AIMS

To evaluate and compare the long term prognostic value of the WHO classification and the computerised multivariate morphometrical D score in endometrial hyperplasia. To test the reproducibility of the D score in two different centres.

METHODS

Histopathological WHO classification and computerised morphometrical analysis using the D score (< 0, high risk; > 1, low risk; 0-1, uncertain) in a population based study from northern Norway of archival dilatation and curettage material from 68 women with 10-20 years of follow up.

RESULTS

Of the 68 patients included in the study, 18 developed cancer. The sensitivity and specificity of the D score (< 0 v > 1) were 100% and 78%, respectively, which was better than the WHO classification (89% and 60%, respectively). The negative and positive predictive values for the D score were 100% and 58% and of the WHO classification 94% and 44%, respectively. This study found a slightly higher specificity for the D score than former retrospective studies, but otherwise the results were comparable. The D score results were reproducible between the two centres (R = 0.91; slope = 0.98; intercept = 0.3).

CONCLUSIONS

D score assessment is a reproducible and more accurate predictor of outcome of endometrial hyperplasia than the WHO classification assessed by an experienced gynaecological pathologist. Routine application of the D score might reduce over and undertreatment of endometrial hyperplasia.

Ovarian serous borderline tumors: a critical review of the literature with emphasis on prognostic indicators

BACKGROUND

The behavior of ovarian serous borderline tumors (SBTs) and significance of various prognostic factors are unclear and difficult to evaluate because of inconsistencies and confusion in the literature. Recent studies have suggested that the morphological features of the primary tumor (presence or absence of micropapillary features) and the peritoneal "implants" (presence or absence of invasive features) can reliably subclassify SBTs into benign and malignant types. The aim of the current review was to test two hypotheses. First, that the alleged malignant behavior of SBTs is poorly documented, and second, that the morphological features of the primary ovarian tumors and the associated peritoneal implants are sufficient to separate SBTs into benign and malignant types, thereby obviating the need for the category.

METHODS

245 studies reporting approximately 18,000 patients with borderline ovarian tumors were reviewed. After excluding series that lacked clinical follow-up or were not analyzable for other reasons, there remained 97 reports that included 4,129 patients. In addition to recurrences and survival, we evaluated the type of peritoneal implants, microinvasion, lymph node involvement, late recurrences, and progression to carcinoma, as these features have served as the underpinning of the concept of "borderline malignancy" or "low malignant potential."

RESULTS

Among 4,129 patients with SBTs reviewed, the recurrence rate after a mean follow-up of 6.7 years was 0.27% per year for stage I tumors, the disease-free survival was 98.2%, and the overall disease-specific survival rate was 99.5%. For patients with advanced-stage tumors, the recurrence rate was 2.4% per year. However, the majority (69%) of reported recurrences were not pathologically documented, and only 26 cases (8.4% of all recurrences) were documented to have recurred from an adequately sampled ovarian tumor. The most reliable prognostic indicator for advanced stage tumors was the type of peritoneal implant. After 7.4 years of follow-up, the survival of patients with noninvasive peritoneal inplants was 95.3%, as compared with 66% for invasive implants (P < .0001). Microinvasion in the primary ovarian tumor was associated with a 100% survival rate at 6.7 years, and lymph node involvement was associated with a 98% survival rate at 6.5 years. The few reported cases of stage IV disease, progression to invasive carcinoma, and very late (>20 years) recurrences were poorly documented. The survival for all stages among approximately 373 patients in 6 prospective randomized trials followed for a mean of 6.7 years was 100%.

CONCLUSION

Surgical pathological stage and subclassification of extraovarian disease into invasive and noninvasive implants are the most important prognostic indicators for SBTs. Survival for stage I tumors is virtually 100%. Survival for advanced stage tumors with noninvasive implants is 95.3%, whereas survival for tumors with invasive implants is 66%. Invasive implants behave as carcinomas and are most likely metastatic. The precise nature of so-called noninvasive implants is not clear, but they behave in a benign fashion. The presence of a micropapillary architecture in the primary ovarian tumor is a strong predictor of invasive implants. These data support the recommendation that ovarian tumors with a micropapillary architecture be designated "micropapillary serous carcinomas," and those lacking these features, "atypical proliferative serous tumors."

Toward the development of a universal grading system for ovarian epithelial carcinoma. I. Prognostic significance of histopathologic features--problems involved in the architectural grading system

OBJECTIVE

Because there is no universally accepted grading system for ovarian epithelial carcinoma, we attempted to compare the prognostic utility of the individual components used in some systems--both architectural and cytologic features, as well as mitotic activity and histologic tumor type--to determine which of these components fit best with survival.

METHODS

We studied 461 patients with invasive ovarian carcinoma who had uniform treatment, complete clinical data including staging and follow-up, and slides available for review. Each tumor was assigned a histologic subtype, architectural grade (based on whether the predominant pattern was glandular, papillary or solid), nuclear grade, mitotic count, and FIGO grade (based on the system for endometrial carcinoma). These features were compared with each other and with tumor stage and survival.

RESULTS

The architectural grade, nuclear grade, and mitotic count were independent variables both in stage I/II and stage III/IV disease. Each of them correlated with survival for most combinations of histologic type and stage. By multivariate analysis, in stage I/II cancer, nuclear grade and architectural grade were significantly correlated with survival, mitotic count showed only a trend, and FIGO grade did not correlate. In stage III/IV disease, nuclear grade, architectural grade 3, and mitotic count were significant, and FIGO grade was not.

CONCLUSION

The new architectural grading system proposed worked better than the FIGO system in this study. Furthermore, it could be applied to all histologic subtypes of carcinoma. The nuclear grade and mitotic count were also independent of each other, correlated with survival, and could be utilized for all histologic types. These data support the development of a grading system which combines these architectural, nuclear, and mitotic features and can be applied regardless of the histologic type of carcinoma, modeled after the Nottingham system for grading of breast carcinoma.

Toward the development of a universal grading system for ovarian epithelial carcinoma: testing of a proposed system in a series of 461 patients with uniform treatment and follow-up

BACKGROUND

Most published series of ovarian carcinoma find a correlation between histologic grade and survival, but the grading system used commonly is not specified, and several different systems exist, some of which use different criteria for different histologic types. However, several studies have shown marked interobserver variability in distinguishing among the histologic types of ovarian carcinoma. The authors attempted to derive a universal grading system for all invasive ovarian carcinomas (IOC), based on the Nottingham system for grading all types of mammary carcinoma.

METHODS

The authors studied 461 patients with IOC of different histologic types and clinicopathologic stages who were treated in a uniform manner between 1980 and 1994 with surgery and cisplatin-based chemotherapy. All slides were reviewed and the tumors graded as follows: Architectural pattern (predominant): Glandular = 1, Papillary = 2, and Solid = 3; Nuclear pleomorphism: Slight = 1, Moderate = 2, and Marked = 3; Mitotic activity (mitotic figures per 10 high-power fields [1 HPF = 0.345 mm2]) in most active region: 0-9 = 1, 10-24 = 2, and > or = 25 = 3; Grade 1 = total score (adding three values obtained earlier) 3-5, Grade 2 = 6 or 7, and Grade 3 = 8 or 9.

RESULTS

Tumor grade correlated with survival in both early and advanced stage disease and for all major histologic types of IOC except clear cell carcinoma (CCC). Results for CCC approached but did not reach clinical significance. By multivariate analysis, only this tumor grade and performance status were significant in Stage I/II IOC. For Stage III/IV tumors, the new tumor grade also was significant, as were performance status, residual tumor size, response to chemotherapy, and mucinous (unfavorable) or transitional cell (favorable) histologic type. International Federation of Gynecology and Obstetrics grade (based primarily on architectural features) did not correlate significantly with survival except in Stage III/IV serous and Stage I/II mucinous carcinomas.

CONCLUSIONS

The new grading system reported is simple, reproducible (among the current study authors), and useful for all histologic types and clinical stages of IOC. Further testing for reproducibility and clinical utility is recommended.

Quantitative prognostic features in FIGO I ovarian cancer patients without postoperative treatment

To identify FIGO I ovarian cancer patients at high risk, prognostic values of quantitative pathological features (volume percentage of epithelium, mitotic activity index, mean (MNA) and standard deviation of nuclear profile area, and volume-weighted mean nuclear volume (MNV) have been investigated in comparison with clinical features, histological grade, and type in 102 adequately staged patients with FIGO Ia, Ib, and Ic ovarian cancer of the common epithelial types. None of these patients received any postoperative treatment. Overall survival of patients alive and well was 78%, and 90% were alive. Of the clinical features, FIGO substage was the strongest prognosticator (Mantel-Cox = 7.2, p = 0.03, hazard ratio (HR) = 4.6). Histologic grade had significant prognostic value as well (Mantel-Cox = 9.7, p = 0.008, HR = 5.9), but histologic type did not. MNA and MNV were the strongest single prognostic factors for overall survival (Mantel-Cox = 12.3 for both; p = 0.0004 and 0. 0005). If MNA </= 55.6 micron2, none of the patients (n = 52) died; if MNA >55.6 micron2, 6-year overall survival was 69%. For MNV </=460 micron2, none of the patients (n = 53) died; if MNV >460 micron2, 6-year overall survival was 70%. A multivariate combination of MNA and FIGO (early cancer of the ovary prognostic score, ECOPS) had the strongest prognostic value (p < 0.0001 and Mantel-Cox value = 22.8, HR = 29.2). If ECOPS </= 5.4 (n = 66), 6-year overall survival was 97%; if ECOPS >5.4 (n = 36), 6-year overall survival was 54%. The results from this and earlier studies emphasize the strong prognostic value of easy to assess and highly reproducible morphometric nuclear features in ovarian tumors and offer a useful instrument for the definition of patient groups for future clinical trials.

Value of quantitative pathological variables as prognostic factors in advanced ovarian carcinoma

AIMS

To evaluate correlations among clinical, pathological, morphometric, stereological, and DNA flow cytometric variables and their prognostic value in advanced ovarian cancer.

METHODS

Tissue was collected from 180 patients with advanced ovarian cancer. All 180 had undergone debulking surgery and were being treated with cisplatin. Long term follow up was available for all patients. The mitotic activity index (MAI), volume % of epithelium (VPE), mean nuclear area (MNA), standard deviation of the nuclear area (SDNA), estimates of volume weighted mean nuclear volume (nu v), and variables obtained from minimum spanning tree (MST) analysis were assessed in the least differentiated tumour section in each case. DNA flow cytometry was also performed.

RESULTS

Quantitative pathological features differed significantly with respect to histological grade. The MAI, MNA, SDNA, and the number of points connected to three neighbours differed significantly among the different DNA ploidy groups. The VPE and number of points connected to two or three neighbours differed significantly between FIGO stages III and IV. Fifty two (29%) patients survived. FIGO stage, residual disease and SDNA had prognostic significance on both univariate and multivariate survival analysis. In patients with FIGO III stage disease and residual tumour nodes < or = 2 cm in diameter (67 patients, 29 (43%) survivors) a prognostic index was established based on SDNA and of the line length of the MST. The median survival time was not reached in a subgroup of patients with favourable prognosis (overall survival 57%). Median survival was 32 months for patients with an unfavourable index score (overall survival 28%).

CONCLUSION

Morphometric variables have important additional value in predicting prognosis in patients with advanced ovarian cancer.

An evaluation of prognostic factors in advanced ovarian cancer

A summary is presented of currently available prognostic factors in advanced ovarian cancer of the common epithelial types. The emphasis is on the most promising clinical, classical pathological, biochemical, immunohistochemical, molecular biological and quantitative pathological factors.

The role of computerized morphometric and cytometric feature analysis in endometrial hyperplasia and cancer prognosis

In our hospital, quantitative cell and tissue analysis is routinely applied in endometrial (pre)malignancies. Reasons for this are higher accuracy, reproducibility and objectivity when compared to subjective assessment of type and grade, the possibility of detecting changes and differences, and better compatibility with clinical requests (two-class instead of three- or four-class system). Furthermore, prognostication is at least a good or better than with the usual subjective methods. Clinical prospective intervention trials are currently being set up.

Prediction of lymph node metastases in patients with early squamous cell carcinoma of the cervix uteri by histopathological grading and flow cytometry

The present study comprises a retrospective investigation of 126 patients with stage IB squamous cell carcinoma of the cervix, and a similar prospective investigation of 53 patients with stage IB and 6 patients with stage IIA disease. Tumour biopsies from these patients were analysed by means of flow cytometry and a semiquantitative histological grading system. The study showed that a combination of a low tumour cell DNA index and a low score value of the grading system indicated a very low risk of regional lymph node metastases (0% lymph node metastases in patients with low scores vs. 24-46% metastases in patients with high scores, P < 0.001). In order to study the reproducibility of the histological grading 20 randomly selected cases were studied blindly by three of the participating pathologists and after discussion of the grading criteria. A kappa coefficient of 722 demonstrated a substantial agreement between the observers. These results suggest that by combining flow cytometry with semiquantitative histological grading, a subgroup of patients with early squamous cell cancer of the cervix uteri may be selected that could be sufficiently treated with simple hysterectomy instead of radical hysterectomy including lymphadenectomy, which, in many oncology centres, is the standard treatment of this patient category.

Advanced ovarian cancer. An overview of multi-variate analyses of prognostic variables with special reference to the role of cytoreductive surgery

BACKGROUND

Cytoreductive surgery is still widely believed to be a cornerstone in the management of advanced ovarian cancer. In the absence of data from prospective randomized trials, multi-variate analysis can assist in evaluating the role of optimal surgical cytoreduction relative to other possible prognostic variables.

MATERIALS AND METHODS

Published analyses of prognostic variables have been reviewed and the consistency and reproducibility of variables have been considered.

RESULTS

When reviewing the most recent literature it is obvious that there is a need to develop a consensus as to the most important patient characteristics which should be reported. Furthermore, these characteristics need to be standardized as do definitions such as toxicity, grading, and response criteria and duration. Subjective variables such as performance status, histological grading and typing, present a special problem as these lack reproducibility. The overview of multi-variate analyses for prognostic variables such as residual tumour size, tumour size prior to cytoreductive surgery, cisplatin therapy, grade, stage, histology, and performance status demonstrates that the impact of individual prognosticators depends on the indicators included in the analysis and that the importance of these variables could change as new technologies and treatments are introduced.

CONCLUSIONS

The collection of a single data base including all known variables reported for each patient with appropriate attention to the careful evaluation of subjective indicators should be undertaken, with a view to establishing a prognostic index. This would allow the relative risk for individual patients to be assessed.

Molecular and biological factors in the pathogenesis of ovarian cancer

BACKGROUND

The classic prognostic parameters are insufficient for predicting the prognosis of the individual patient. Knowledge of molecular and biological factors which are responsible for the development and progression of ovarian cancer may improve the prediction of prognosis.

MATERIALS AND METHODS

Recent data both on factors associated with the development and control of ovarian cancer cells and on DNA ploidy have been reviewed.

RESULTS

Observations suggest that steroid and peptide hormones have a role in disease etiology and progression, and that peptide growth factors and cytokines, oncogenes and tumor suppressor genes, by their impact on mitosis and cell number may influence the rate of mutations, which could confer malignant transformation. DNA ploidy is an objective independent prognostic factor. DNA aneuploidy indicates high risk, diploidy low risk. Only tumours shown to be DNA diploid by flow-cytometry and image cytometry are considered diploid. S-phase fraction is currently not reliable.

CONCLUSION

Understanding the mechanisms involved in ovarian cancer development and growth will allow opportunities for the rational design of effective anti-tumour treatment modalities. More objective and reproducible prognostic variables will improve the predictiveness of prognosis.

Reproducibility of two malignancy grading systems with reportedly prognostic value for oral cancer patients

Supplementary prognostic factors should be added to the TNM classification for oral squamous cell carcinomas in order to optimize its clinical value. We have recently published two prognostically valuable malignancy grading systems based on histopathology and immunohistology of the most invasive cells in oral squamous cell carcinomas (OSCCs). However, a major problem with classifications based on histologic features is frequent lack of interobserver agreement which limits the clinical value of subjective histologic classifications. Thirty-eight file cases of OSCCs were therefore graded by three pathologists according to criteria of the histologic malignancy grading system which includes 5 morphologic features, each graded from 1 to 4. Agreement was calculated by kappa statistics, which showed that interobserver agreement was not optimal, but significantly better than by chance alone. We also studied the reproducibility of grading of immunohistologic membrane expression of a tumor-associated marker (blood group antigen H), and found a similar level of agreement. We conclude that the clinical value of our grading systems will increase by improving reproducibility.

Prognostic factors for overall survival in patients with advanced ovarian carcinoma

Twenty-one pretreatment variables were investigated for prognostic influence on survival in 301 previously untreated patients with ovarian carcinoma, stage IIB-IV. Patients were randomized to sequential combination chemotherapy: cyclophosphamide, doxorubicin, 5-fluorouracil, followed by cisplatin and hexamethylmelamine, or to the 3-drug combination alternating with the 2-drug combination every other month. Median overall survivals were 25 and 22 months, respectively, P greater than 0.4. Based on the results from a Cox multivariate stepwise analysis a subset of independent significant prognostic factors was found to include: residual tumor size, performance status, alkaline phosphatase, number of metastases, histological differentiation grade and type. A prognostic index was calculated for each patient and three prognostic categories of patients were determined. The 3-yr survival rates for patients with low-, intermediate-, and high-risk scores were 62, 31, and 7%, respectively. Multivariate analysis thus contributes further information about the disease, and a knowledge of the distribution of such factors across different trials is important when comparing treatment outcome.

Reproducibility of histopathological evaluation in epithelial ovarian carcinoma. Clinical implications

An interobserver analysis of the reproducibility of malignancy grade and of histological type classified according to World Health Organization criteria was performed in 261 ovarian tumours, stages IIB to IV. All available slides per patient including metastatic tumours (16%) were reviewed independently by two pathologists. Overall observed agreement regarding type was 68%. The chance-corrected agreement expressed by the kappa coefficient was 0.55 (95% confidence interval 0.49-0.61). Overall observed agreement regarding grade was 41% compared with a kappa coefficient of 0.19 (95% confidence interval 0.12-0.26). Kappa coefficients, ranging from 0.18 to -1.00 for associations between type categories and from 0.20 to -0.32 for grade categories, indicated that any suggested pattern of association could be caused by chance occurrences. A more objective approach to histopathology is needed in ovarian carcinoma before its prognostic impact can be assessed.

Correlation between nuclear DNA content and steroid receptor status in ovarian cancer

The nuclear DNA content of 47 malignant epithelial ovarian tumours and five ovarian tumours of borderline malignancy was correlated with estrogen (ER), progesterone (PR) and androgen receptor (AR) status. Aneuploidy was observed in 60% of the malignant tumours. The DNA index of poorly differentiated tumours was higher than that of well differentiated tumours (P less than 0.05). Ploidy did not correlate with histological type, stage of disease and ER content. Of the diploid tumours, 74% was PR-positive, in contrast to 36% of the aneuploid tumours (P less than 0.05). In addition, 89% of the diploid tumours had high AR levels (greater than or equal to 30 fmol/mg cytosol protein), in contrast to only 54% of the aneuploid tumours (P less than 0.05). These observations strengthen our previous findings on the prognostic importance of PR and also suggest that androgens may have a role in ovarian cancer.

Steroid receptors and Ki-67 reactivity in ovarian cancer and in normal ovary: correlation with DNA flow cytometry, biochemical receptor assay, and patient survival

Steroid hormone receptors and reactivity for Ki-67 proliferation antigen were studied immunohistochemically in non-neoplastic post-menopausal human ovary and in 29 ovarian cancers. In the normal ovary, oestrogen (OR) and progesterone receptors (PR) were found in the surface epithelium and PR also in the ovarian stroma. Of the ovarian carcinomas 38 per cent (11/29) contained OR and 69 per cent (20/29) PR. Oestrogen receptor expression was confined to malignant cells, whereas PR was present occasionally also in the tumour stroma. In most cases, ORs and PRs were found only in a small population of cancer cells. The growth fractions assessed by the percentage of Ki-67-positive cells ranged from 1 to 59 per cent (mean 19.7 per cent) with a significant correlation (r = 0.74, P less than 0.0001) to S-phase values (mean 12.9 per cent, range 1.2-25.9 per cent) determined by DNA flow cytometry. High Ki-67 (greater than or equal to 15 per cent) and S-phase levels (greater than or equal to 7.5 per cent) correlated with advanced disease stage and patient survival but not with OR or PR status, suggesting that hormone-receptor pathways and proliferative activity are not related in ovarian cancer. Positive OR status, however, identified patients with a better prognosis (P = 0.02), suggesting a correlation with tumour differentiation. The independent prognostic value of oestrogen receptor status and Ki-67 remains to be determined, but the prognostic impact of Ki-67 was comparable to that of S-phase values.

Prognostic factors in borderline and invasive ovarian tumors of the common epithelial type

Present reports show that surgical factors, response to therapy and histological criteria are important for predicting the prognosis of patients with common epithelial types of ovarian tumors. Newer techniques such as morphometry, DNA cytometry, immunological and immunopathological techniques may help to define prognostic factors even more accurately. As a result, these recently developed methods may enhance the value of well-established classical predictors of the outcome in case of borderline or invasive ovarian tumour.