|Home | About | Journals | Submit | Contact Us | Français|
A two-tier grading system based on nuclear grade divides ovarian serous carcinomas into low- (nuclear grade 1) and high-grade (nuclear grade 3). In most instances the separation is straightforward but at times, the morphologic distinction between them can be difficult. We studied eleven ovarian serous carcinomas with features that were “intermediate” (nuclear grade 2) between low and high grade. All the cases were high stage and had a poor clinical outcome. None of the tumors showed mutations in KRAS, BRAF and ERBB2 genes which characterize most low-grade serous carcinomas. In contrast, 10 (90.9%) of 11 cases contained non-synonymous TP53 mutations characteristic of high-grade serous carcinomas. In summary, the molecular genetic profile and behavior of serous carcinomas with grade 2 nuclei are virtually the same as those of serous carcinomas with grade 3 nuclei, supporting the use of the two-tier grading system for classifying ovarian serous carcinomas.
Recently, it has been proposed that the three-tier (well, moderately and poorly differentiated) grading system for ovarian serous carcinomas be replaced by a two-tier (low-grade and high-grade) system based on clinical, morphological and molecular studies. 6, 10, 12, 16 The two-tier system was initially proposed based on morphologic and molecular genetic features 12 and subsequently was shown in a multivariate analysis to be useful in predicting outcome. 6 In the two-tier system, high-grade serous carcinomas exhibit a papillary or solid growth pattern and the tumor cells contain markedly atypical, large pleomorphic nuclei and frequent mitotic figures, many of them abnormal. 5, 6 These nuclei are classified as grade 3. In contrast, the majority of low-grade carcinomas display a micropapillary pattern. The tumor cells contain small, uniform nuclei with minimal atypia and low mitotic activity corresponding to grade 1. Clinically, high-grade serous carcinomas are aggressive neoplasms frequently affecting women in the peri-menopausal or postmenopausal age group while low-grade serous carcinomas are relatively indolent affecting younger women. 4 Low-grade serous carcinomas are more refractory to platinum-based chemotherapy as compared to high-grade serous carcinomas 9 probably because of their low proliferative rate. They develop slowly in a stepwise fashion from non-invasive micropapillary serous carcinomas which in turn arise from atypical proliferative serous (borderline) tumors. Molecular genetic studies have demonstrated that low-grade micropapillary serous carcinomas contain either KRAS, BRAF or ERBB2 mutations in approximately two thirds of cases but only rarely harbor TP53 mutations. 3, 7, 13 In contrast, high-grade serous carcinomas are believed to develop from intraepithelial carcinoma in ovarian surface inclusion cysts or the fallopian tube. High-grade serous carcinomas frequently contain TP53 mutations (>80%) 8 but on rare occasion harbor mutations in KRAS, BRAF and ERBB2 genes. 3, 7, 13
Although high-grade and low-grade serous carcinomas are usually easily distinguished, it may be difficult for carcinomas with nuclear features that are intermediate between grade 1 and grade 3. These intermediate grade tumors have relatively uniform nuclei that are midway in size between grade 1 and grade 3 tumors and therefore qualify as grade 2. Compared to typical high-grade nuclei (Fig. 1), grade 2 nuclei are more uniform in size but are larger and more atypical than grade 1 nuclei. Moreover, the mitotic index in these intermediate grade tumors is increased compared to low-grade serous tumors but lower than high-grade serous carcinomas. In this study, we analyzed a small group of serous carcinomas displaying these intermediate features corresponding to what would be classified as moderately differentiated (nuclear grade 2) carcinomas for mutations of TP53, KRAS, BRAF and ERBB2 genes in order to determine whether these tumors displayed a molecular genetic profile that paralleled the genotype of low or high-grade serous carcinomas. In addition, we evaluated their clinicopathological features to determine if their behavior was it similar to low or high grade tumors.
A total of eleven cases of ovarian serous carcinomas with grade 2 nuclei were studied. Four cases were identified from the archival files of the Department of Pathology at the Johns Hopkins Hospital; two cases were collected from the Department of Pathology at the Emory University School of Medicine, and five cases from Washington Health Center. Tumor sampling was approximately one section/cm of the greatest tumor dimension. Hematoxylin and eosin stained sections were reviewed and the corresponding paraffin blocks retrieved. Collection and analysis of the anonymized specimens were approved by the Institutional Review Boards from the four institutions. The morphologic criteria used for nuclear grading are shown in Fig. 1. Grade 1 nuclei were relatively small and uniform in size whereas grade 3 nuclei were large, pleomorphic with coarse clumping of chromatin. Grade 2 nuclei in this study were relatively uniform in size; they were larger than grade 1 nuclei and smaller than grade 3 nuclei. Nucleoli were more prominent than in grade 1 tumors (Fig. 1). Mitotic activity was determined by counting 50 high power field (HPF; 40x) using an Olympus BH2 microscope. Ten high power fields were counted in the most active areas.
Genomic DNA was isolated from fresh tumor tissue in three cases and from paraffin sections in the remaining 8 cases. Purification of genomic DNA was performed by proteinase K digestion followed by a Qiagen DNA purification kit for fresh tissues, and the Forma Pure kit (Agencourt, MA) for sections from formalin fixed paraffin embedded samples. Genomic regions containing the exon 1 of KRAS, exon 15 of BRAF, exon 20 of ERBB2, exons 4 to 9 of TP53; and exons 1-3, 9 and 20 of PIK3CA were amplified by the polymerase chain reaction (PCR). Purification of PCR products was performed using a Qiagen PCR purification kit. Sequencing was performed at the Johns Hopkins DNA sequencing facility. The sequences of primers for PCR and nucleotide sequencing have been previously reported. 7, 8
The clinical features of the eleven serous ovarian carcinomas with grade 2 nuclei are summarized in Table 1. The ages of the patients ranged from 43 to 81 years with an average of 65.4 years. The frequency of serous carcinomas with grade 2 nuclei was estimated to be approximately 4% of all serous carcinomas based on reviewing archival specimens at the Johns Hopkins Hospital and the Emory University Hospital. Ten cases were primary ovarian serous carcinomas, and one was a primary peritoneal serous carcinoma. Those cases were originally reported as papillary serous cystadenocarcinoma without specifying either low- or high-grade. All patients presented with FIGO stage IIIC disease. Eight patients died of disease and two patients died from surgical complications one a month and the other six months after surgery. Follow up information was not available on one case. The survival time for the eight patients varied from 6 to 30 months. Histopathologic features of the tumors are presented in Table 2. Six tumors (55%) had no necrosis; five tumors (45%) had multiple small foci of tumor necrosis. Four tumors (36%) displayed a solid and slit-like pattern and six displayed a micropapillary pattern involving at least 25% of the tumor. Mitotic counts ranged between 4.4 and 18 per 10 HPFs with a medium of 10.2, and a mean of 10.7 per 10 HPFs. Representative hematoxylin-and-eosin stained sections from two tumors are shown in Fig. 2.
To determine the molecular genetic features of the serous ovarian carcinomas with grade 2 nuclei, KRAS, BRAF, ERBB2, PIK3CA and TP53 genes were sequenced because somatic mutations of KRAS, BRAF and ERBB2 are the most common molecular genetic changes in low-grade micropapillary serous carcinomas while TP53 mutations are the most frequent mutation in high-grade serous carcinomas. PIK3CA kinase gene was also sequenced in this study because the gene has been analyzed in ovarian cancer tissues in several reports. 2, 7, 17, 18 We found no KRAS, BRAF, ERBB2, mutations in any of the grade 2 tumors while 10 of the 11 tumors carried non-synonymous TP53 mutations (Table 3 and and4).4). Wild-type PIK3CA sequences were present in all the specimens examined.
Grading systems for ovarian serous carcinomas include the FIGO system which is based solely on architecture 1, the WHO grading system based on cytological and architectural features 15 and the Silverberg-Shimizu grading which takes into account architectural and nuclear features and mitotic activity. 11 More recently, a two-tier grading system that divides serous carcinomas into low and high-grade has been proposed. 6, 12 This grading system is primarily based on nuclear grade with mitotic activity used as a secondary feature. In contrast to the three tier system, the two-tier grading system has a molecular genetic underpinning. This is important because it provides biologic support rather than being based entirely on morphologic criteria.3, 10 Using the two-tier system, the majority of serous carcinomas can be classified into either low or high-grade, but there are occasional tumors with nuclear features that are intermediate between conventional low-grade and high-grade. The current study provides clinical and molecular genetic evidence that tumors with grade 2 nuclei are more closely related to high-grade rather than low-grade serous carcinomas. Specifically, the aggressive behavior of the tumors, the high proportion of TP53 mutations (90%) and the absence of KRAS, BRAF, ERBB2 mutations are similar to tumors with grade 3 nuclei as compared to tumors with grade 1 nuclei.
The results of this study have important clinical implications because the architectural features such as solid growth pattern (poorly differentiated), papillary (moderately differentiated) or micropapillary/macropapillary (well-differentiated) may overlap. The findings in the present study support those in our previous study in which we showed that serous carcinomas with micropapillary features and grade 3 nuclei were similar to conventional high-grade carcinomas based on absence of KRAS mutations. 14 Thus, nuclear grade 2/3 versus grade 1 is the major microscopic feature that distinguishes low from high-grade serous carcinomas. As there were only eleven cases in this study, more cases are needed to corroborate our conclusion.
We found that the mitotic activity ranged widely among the 11 carcinomas with grade 2 nuclei (Table 2). There were seven tumors with mitotic counts below 12 mitoses/10 HPFs, a cutoff proposed to distinguish low-grade versus high-grade serous carcinomas in a previous study. 6 Thus, mitotic activity should be used as an ancillary feature in the two-tier grading system. In fact, we believe the proposed mitotic count cut point of 12 mitotic figures/10 HPFs should be revised. In our experience low grade serous carcinomas almost never have mitotic counts greater than 5 mitotic figures/10 HPFs. A larger study will be necessary to come up with firm guidelines.
In conclusion, we analyzed a series of ovarian serous carcinomas with grade 2 nuclei to delineate their clinicopathologic and molecular genetic relationship to low- and high-grade serous carcinomas. All the tumors were high stage and clinically aggressive. Mutations of KRAS, BRAF and ERBB2 which characterize most low-grade serous carcinomas were not identified in any of them. In contrast, 10 (90%) of 11 tumors contained non-synonymous TP53 mutations characteristic of high-grade serous carcinoma. The clinical and morphologic data in this study showing close similarity between ovarian serous carcinomas with grade 2 nuclei and those with grade 3 nuclei together with our recent report showing that subclassification of high-grade serous carcinoma into moderately and poorly differentiated serous carcinomas is not relevant 16 provide strong support for the use of a two-tier grading system for ovarian serous carcinomas.
Supported in part by NIH/NCI RO1 CA116184.