Abstract
Background: The relation of nuclear morphometry measurements with clinicopathological features was evaluated along with prognosis in invasive female breast carcinoma in Libyan patients. Data was compared with corresponding results on Finnish, and Nigerian female breast cancer patients. Patients and Methods: Histological samples from 131 patients of breast carcinoma were retrospectively studied by computerized nuclear morphometry. In each case, 50 nuclei were measured and the mean nuclear morphometric features were calculated and compared with different clinicopathological features, and patient's survival. Results: There was statistically significant correlation between the mean nuclear area (MNA) and most clinicopathological features, with the strongest association observed for nuclear grade (p<0.0001). There was also correlation between nuclear area and tumor stage (p<0.04), tumor size (p<0.03) and lymph node (LN) status (p<0.001). A corresponding relationship was found between other size related features and clinical factors. The univariate analysis and survival analysis indicated that short survival time was associated with high nuclear morphometric values. MNA had negative correlation with length of survival (Pearson's test r=-0.29, p=0.019). Morphometric shape features did not show significant association with clinical features or survival. Conclusion: The results indicated that nuclear size features are reliable prognostic indicators in Libyan female breast carcinomas, as they were among Finnish and Nigerian females. The nuclear morphometric parameters can identify the aggressive tumor phenotype and provide significant prognostic information in predicting survival and tumors at risk of progression. The cut-off (71.0 μm2) of MNA might be applied as quantitative criterium for Libyan nuclear grading to separate patients into good and poor prognosis groups.
Breast cancer is the most common cancer in women in the world. Breast cancer is also the condition causing most deaths among patients with female cancers (1, 2). Breast cancer is actively studied, but many aspects still remain unclear, including the special features associated with individual countries. The breast cancer prognosis can be evaluated by combining morphometric features, tumor size and lymph node status (3). The histological grading system is associated with high prognostic potential (4), but is still subjective, and leaves a large group of patients with unclear prognosis (5). True measurements, statistically assessed, can be expected to be more reproducible than the subjective methods (6).
Quantitative histopathology offers a wide range of methods for unbiased assessment, as was shown by nuclear morphometry (7, 8), which was able to distinguish between benign and malignant lesions. It is suggested that the quantitative methods in combination with other objective prognostic criteria can improve the evaluation of the patient's prognosis, and possibly predict response to therapy. One of the important prognostic and predictive markers is the nuclear morphometry, which has proven valuable in many connections (9, 10).
It was decided to focus on the role of nuclear morphometry in the evaluation of Libyan breast cancer prognosis and the relation of morphometry with clinicopathological features.
This research was found attractive because through such a study Libyan breast cancers could be compared with those from Finnish (European) and Nigerian (Central African) patients.
Patients and Methods
The study was performed on Libyan female breast cancer samples. All cases were diagnosed at the Department of Pathology, African Oncology Institute, Sabratha, and Tripoli Medical Centre, Libya and during the years 2000-2006. The patients were excluded from this study on the basis of the following exclusion criteria: histopathology was done elsewhere than in the study centres; patient history and medical files, or specimens, were not found; the follow-up was less than 3 months; paraffin blocks were not available for re- cutting. This left 131 samples for the study; 116 patients were treated with modified radical mastectomy with axillary clearance; 15 patients were unfit for surgery due to distant metastases, but diagnostic biopsies were used for morphometry.
A detailed history, clinicopathological features (age, menopausal status, tumor size, stage and grade, and lymph node status), and the follow-up and survival data were collected from patients files. The mean age at the time of diagnosis was 46.9 years, and 4.6%, 33.6%, 49.6%, and 12.2% of patients were at stages 1, 2, 3, and 4, respectively.
Treatment and follow-up. Patients were followed-up until death or the end of the observation period at the mid of July 2007. Some patients were lost from follow-up. The follow-up period ranged from 4 to 72 months, the average being 32 months. The patients were seen at 3-6 month intervals, and bone isotope scan, and chest, and abdominopelvic CAT scan were performed every 6-12 months. In most instances, the causes of death were obvious on clinical grounds alone. Autopsy was not performed in any case. Breast cancer was recorded as the underlying cause of death for 34 patients, and unrelated to death for 3 patients.
At least one section of 5 μm thickness was stained with hematoxylin and eosin stain for re-grading according to the Bloom and Richardson grading system, and for the morphometric measurements. The tumor diameters were measured after surgical removal in 3 dimensions and the largest diameter was entered in the database of the study. There were 96 invasive ductal carcinomas (73.3%), 13 invasive lobular carcinomas (9.9%), 7 mixed ductal and lobular carcinomas (5.3%), 6 medullary carcinomas (4.6%), 3 papillary carcinomas (2.3%), 5 mucinous carcinomas (3.8%), and 1 metaplastic carcinoma (0.8%).
The survival period was defined as the time from diagnosis of the tumor either to the time of death, or the latest date on which the patient was known to be alive.
Morphometry. The most representative sections were analyzed by using an interactive digitizing image overlay drawing system run by the Prodit morphometry program (Prodit 3.1, Promis Inc, Almere, and Buro medische Automatiserving, De Meern, the Netherlands). The system consists of a light microscope, a personal computer (Compaq Deskpro 386/20e; Compaq Computer Corporation, Houston, TX, USA), a video camera attached to the microscope (JVC TK-870U; JVC, Japan) and a digitizer board (PIP-512B video digitizer board; Matrox Electronic Systems, Dorval; Quebec, Canada). Analog images of the nuclear profile were outlined on the monitor screen using a computer mouse, and consequently a digital database was created of the nuclear features in computer memory. The instrument was calibrated in 2 perpendicular directions with a micrometer scale before each session of measurement. Measurement was carried out at ×2600 magnification on the monitor screen (×40 objective lens magnification, ×10 video ocular and ×1.25 internal magnification). At the end of each measurement, the computer automatically produced the basic statistics of the variables measured (11). These included size variables (e.g. mean area, perimeter), and several shape factors.
The clinical characteristics of 131 Libyan patients with breast cancer, studied in this article.
The morphometric features evaluated in this study were mean nuclear area (MNA; μm2), standard deviation of nuclear area (μm2), mean nuclear perimeter (μm), mean nuclear diameter (μm), mean nuclear short axis (μm), and mean nuclear long axis (μm).
Statistical analysis. The variables of the material were grouped into logical classes and descriptive statistics calculated for the continuous variables using SPSS 12.0 for Windows. For survival analysis, Kaplan-Meier curves were plotted, and differences between the curves analysed using the log-rank test. The nuclear morphometric thresholds were the cut-off points showing curve separation with the highest statistical significance. Univariate analysis was performed and multivariate analysis was the Cox's regression model to evaluate additional prognostic value of nuclear morphometry to other prognostic variables. P-values below 0.05 were regarded as significant. Student t-tests and ANOVA were also used to test differences between the groups. A correlation between the morphometric parameters and survival was evaluated using Pearson's correlation test at a level of significance p<0.05.
Results
Clinicopathological features. The clinicopathological characteristics of the Libyan breast cancer patients are shown in Table I. The mean age at diagnosis of breast cancer was 46.5 (13.6) years. A large fraction of the patients were premenopausal (61.8%). Large tumour size (mean 5.6 cm, SD 2.1 cm), and frequency of lymph node involvement (79.4%) characterized the material. The average follow-up time was 32.9 months, and a survival rate was 81.3% after 3 years and 58.6% after 5 years.
Means of morphometric nuclear variables in the whole Libyan material (n=131), and in subgroups defined by menopausal status, tumor size and nodal status. The feature values are presented as means with the SD shown under the mean. The p-values refer to significance of difference between the subgroups.
Means of the morphometric nuclear variables (SD) in subgroups defined by clinical stage, and histological grade and type, in the Libyan breast cancer material (n=131).
Nuclear morphometric parameters were analysed in the whole material, and in groups defined by the menopausal status, lymph node status, tumor size, clinical stage, histological grade and histological type as shown in Tables II and III.
Higher nuclear size values are seen in large tumours and tumours with higher histological grade or stage. The difference in the nuclear parameters between invasive ductal carcinoma and lobular carcinoma were statistically significant (p<0.001 for all nuclear morphometric parameters). Also features between special type of carcinoma and infiltrating lobular carcinoma were significant (p<0.001). The features among infiltrating ductal carcinomas (IDC) were slightly higher than in special types of carcinoma, but the difference was not statistically significant (Figure 1).
On the other hand, the difference in nuclear parameters between the lymph node-positive and negative tumours was highly significant. Lymph node-positive tumours had higher values. Grade 3 cases showed a highly significant difference as compared with Grade 2 cases (p<0.001 in all parameters). The MNA and the mean standard deviation of nuclear area values were higher in the premenopausal patients than in the postmenopausal group, but the difference was not statistically significant.
The nuclear morphometric parameters were higher in advanced stages (stages 3, 4) than in less advanced stages (Figure 2). However, the difference was not significant. A positive correlation between the nuclear morphometric parameters and clinicopathological features was observed. The MNA and the SD of MNA showed a strong correlation with the grade and clinical stage.
Mean±2 SD of the mean nuclear area of different histological types of breast cancer in Libyan female patients (IDC=infiltrating ductal carcinoma ILC=Infiltrative lobular carcinoma, other=other types of infiltrating carcinoma; not IDC or IL.
Survival analysis. At 5 years, 34 patients were known to have died, and 33 patients were known to be alive. Median survival for the whole series of 131 patients was 33 months (mean 33.9 months, range 4-78 months). In Pearson's correlation test, nuclear morphometric parameters were negatively correlated with survival time. Determination of decision cut-points for MNA in the Libyan material resulted in an obvious cut-point at 71 μm2. At 71 μm2 the groups with higher or lower means were prognostically most significantly different. The analysis detected only one cut-point surrounded by less significant cut-points. MNA was more significant than other morphometric features in respect to significant potential cut-points.
All the studied morphometric size parameters were significantly correlated with survival. Shape parameters were not significantly associated with survival. The analysis using Kaplan Meier curves of MNA indicated that short survival time was correlated with high mean nuclear area (Figure 3). Moreover, the Pearson's correlation analysis showed that MNA had the highest negative correlation with survival (r=0.29, p=0.019).
Mean nuclear area (mean Area)±2SD in histological sections of breast cancer of Libyan female patients in different clinical stages (stages1-4). Clearly the mean nuclear area is in correlation with the stage (Pearson's r=0.173, p=0.04).
Survival curves associated with the mean nuclear area as measured from 131 Libyan breast cancers. The cut-point at 71 μm2 was the most significant cut-point, and the corresponding survival curves are shown here. The survival curves are significantly different after 5 years (Log Rank test, p=0.044). The upper curve started with 65 patients, the lower curve with 66 patients. At 5 years the upper curve had 21 survivors; the lower curve had 12 survivors.
Discussion
Several authors have reported on the prognostic importance of estimates of nuclear area in breast carcinoma (12). These findings were confirmed in the univariate analysis of the current study.
Tables II and III give the mean and standard deviation values of all the nuclear morphometric parameters for the different clinicopathological features. There was considerable difference in nuclear area and other parameters values reported by Kronqvist et al. (12) and Ikpatt et al. (13).
The mean of Finnish's MNA value was 38.6 μm2 (SD 15.0) which was within the range of previous European data from 24.4 (SD 12.8) up to 67.8 μm2 (SD 18.35) (14-17). This was lower than the Libyan mean MNA value 74.25 μm2 (SD 23.74). The Nigerian mean MNA value was highest; 89.2 μm2 (SD 34.0). These differences might be due to the fixation technique employed. Another explanation may be that screening programmes have been established for years in European countries for early detection of cancer. The African females came to the hospital in very advanced stage and grade. For example the clinicopathological data of Libyan patients confirmed the aggressive nature characterizing this type of carcinoma. In Libya 49% of cases showed stage III, 79% were with LN metastasis. In the future, after establishment of screening programs in Libya, the cut-point difference will probably decrease. Biological factors as explanation of difference should not be excluded, either.
There were no differences in the used morphometric methods between the 3 studies. The same equipment was used and the technique was standardized and uniform, with regular calibration of the computerized morphometric equipment with a micrometer slide, which should ensure reproducible results (18).
Nuclear size features were in correlation with grade, stage, and nodal status in all 3 studies, and these findings support the other corresponding studies (10, 19). The fact that African and European breast cancers are different in many respects is one of the dominant discussion aspects today (20, 21).
These findings on Libyan breast cancer fall between the findings in Nigerian and Finnish populations (13). This is very much in line with the idea that the differences may have a genetic basis (22). The variation in the distribution of different genetic marker haplotypes makes this easily understandable. There is a clear difference between the marker haplotype distribution in western central Africa and northern Africa. A similar difference is to be found between North-Africa and Europe (23). The variation in haplotype marker distribution has taken place under selective environmental stresses. However, there is no need to believe that the environmental influences are not operating at the moment. So far the viral association with African breast cancer is far from proven, but if found, can potentially be an explanation for differences (22).
In the current work, all the studied morphometric parameters showed significantly lower values in lobular carcinoma as compared to invasive ductal carcinoma and other types of carcinoma. This result coincides with Ikpatt's results (13). In the multivariate analysis of the present study, the mean nuclear area was the strongest prognostic factor among the nuclear parameters. Kronqvist et al. (12) suggested that the mean short axis was still better. However, the difference in significance was small. Kronqvist et al. (12) has suggested two thresholds for the MNA (32 μm2 and 47 μm2) that could separate patients into three subgroups with favourable, intermediate and unfavourable prognosis.
In this study, it was found that survival among patients with MNA <71 μm2 was significantly better than among patients with MNA >71 μm2. So, it is suggested that this value might be used as quantitative criteria for separating patients into two groups with good and poor prognosis in Libyan female patients. Ikpatt et al. also found only one decision cut-point, far higher than the presented cut-point.
According to Giardina et al. (24), the nuclear shape parameters allowed good discrimination between cases with good and poor prognosis. However, the present results agree with those reported by Kronqvist et al. (12), and Baak et al. (9), which showed neither statistically significant correlation to clinicopathological features, nor prognostic significance indices of diagnostic importance (25). The studied features of Giardina et al. however, were different than the shape features studied in this paper.
For prognostic purposes, nuclear morphometric parameters can be used as a prognostic tool. Size parameters in particular show association with aggressive tumour nature, and poor survival. Furthermore, in Libyan patients the mean nuclear area (71 μm2) could be used as a criterium for separation of patients into two groups with good and poor prognosis. A positive correlation between the nuclear morphometric parameters and clinicopathological features was observed, and the MNA showed the strongest correlation with grade and clinical stage.
Acknowledgements
We would like to thank the Libyan Board of Medical Specialities for their help and advice. This study has been supported by the Libyan Ministry of Health.
Footnotes
-
↵* Both authors contributed equally to this work.
- Received November 20, 2008.
- Revision received January 27, 2009.
- Accepted February 17, 2009.
- Copyright© 2009 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
References
In this issue
Jump to section
Related Articles
Cited By...
- mTOR controls growth and internal architecture of human breast cancer spheroids
- DNA double-strand break-capturing nuclear envelope tubules drive DNA repair
- Label-free differential imaging of cellular components in mouse brain tissue by wide-band photoacoustic microscopy
- Computer-assisted Diagnosis of Breast Cancer by Cell Network Matrix Extraction and Multilayer Perceptron Analysis
- Cytoplasmic volume and limiting nucleoplasmin scale nuclear size during Xenopus laevis development
- PKC-mediated phosphorylation of nuclear lamins at a single serine residue regulates interphase nuclear size in Xenopus and mammalian cells
- Fast label-free multilayered histology-like imaging of human breast cancer by photoacoustic microscopy
- Nuclear size is sensitive to NTF2 protein levels in a manner dependent on Ran binding
- Concentration-dependent Effects of Nuclear Lamins on Nuclear Size in Xenopus and Mammalian Cells
- Estrogen Receptor, Progesterone Receptor, and Nuclear Size Features in Female Breast Cancer in Libya: Correlation with Clinical Features and Survival
- Nuclear Morphometry in Prognostication of Breast Cancer in Saudi Arabian Patients: Comparison with European and African Breast Cancer