Estimation of tubular differentiation in female libyan breast cancer

© 2013 Jamela et al; licensee Herbert Publications Ltd. This is an Open Access article distributed under the terms of Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0). This permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Background Breast cancer is the most common cancer in women in the world. It is also the most cause of deaths among female cancers patients [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 traditional features such as tumor size, and lymph node status [3]. The histological grading system is also associated with high prognostic potential [4], but is still subjective, and leaves a large group of patients with unclear prognosis [5]. Quantitative histopathology can be expected to be more reproducible than the subjective methods [6]. Earlier studies in breast cancer have suggested that there are variations in nuclear size features, proliferative and apoptotic activity. It is not well known, however, how these differences are reflected in tubular differentiation. Many studies have suggested that tubular differentiation could have potential as a prognosticator in different adenocarcinomas [7-9]. Other studies, however, do not agree [10-14]. The aim of this study is to focus on the role of TFD in the evaluation of Libyan breast cancer prognosis and the relation of TFD with clinicopathological features. We found this research attractive because through such a study we could compare Libyan breast cancers with those from Finnish (European) and Nigerian (Central African) patients, after we used same method described by Kronqvist et al., [15] and Ikpatt et al., [21].

doi: 10.7243/2049-7962-2-5 diagnostic biopsies were used in this study. A detailed history, clinicopathological features (age, menopausal status, tumour size, stage and grade, and lymph node status, follow up data) were collected from patients files ( Table 1 and 2). Age of the patients ranged from 25 to 85 years with a mean age at the time of diagnosis was 46.7(SD±13.4) years. 4.6%, 34.6%, 48.5% and12.3% of patients were at stages 1, 2, 3, and 4, respectively.
Follow up time ranged from 4 to 72 months, Average follow up was 32 months.
The distribution of age, axillary lymph node status, stage, histological type, and histological grades are shown in (Table  1). Mean nuclear area, mitotic and apoptotic indices were evaluated from H & E sections as described earlier [16,17].

Histological methods
All biopsy specimens were fixed in buffered formalin, and embedded in paraffin. Sections5µ m thick were stained with hematoxylin and eosin stain. The histological typing in our study was based on the International Histological Classification of Tumours [18], and grading of tumours was done according the modified Bloom-Richardson histopathological grading system [19].
Tubular differentiation was evaluated in each sample as the fraction of fields showing tubular differentiation (FTD) [15,20]. According to this method tubular differentiation was assessed in the whole tumour area. The samples were screened at x10 magnification and the presence or absence of malignant tubular structures in each microscopic field was registered. By this method the field was registered positive if a single undoubable malignant tubular structure was identified. The final result was the fraction of fields presenting tubular differentiation. This assessment method is especially recommended because it has turned out to be the most efficient and fastest way to evaluate in quantitative terms the tubular differentiation in invasive breast cancer. In a previous study [15] on Finnish material comparing several evaluation methods for tubular differentiation [20], FTD showed out to be the most practical, accurate and reproducible way to determine tubular differentiation in invasive breast cancer. In the evaluations, special emphasis was placed on histological identification of the malignant tubule. The required features for registering a tubule was a clear lumen within a tubular or alveolar pattern created by surrounding malignant epithelial cells. Special consideration was taken to avoid

Statistical analysis
The variables of the material were grouped into logical classes and descriptive statistics calculated for both grouping and continuous variables using SPSS 19.0 for  Windows. For survival analysis, Kaplan Meier curves were plotted, and differences between the curves analyzed using the log-rank test. The FTD cut-off points were tested for association with survival using a log-rank test. Survival analysis was performed with Cox proportional hazard model. Univariate analyses were carried out on all patients and on subgroups classified by menopausal status (pre and postmenopausal) and by axillary lymph node status (node positive N+, node negative N-). To evaluate the independency association of FTD and survival among the other prognosticators, the Cox multivariate regression model was used. P-values below 0.05 were regarded as significant. Student t-tests and ANOVA were also used to test differences between the groups.

Inter observer variation in estimation of the FTD
First and second authors were estimated the fraction of fields with tubular differentiation in30 samples of Libyan female BCs. Figure 1 shows a strong correlation between two observers. Clearly, there is high reproducibility and with high significance value (P<0.0001), (Pearson's correlation 0.91, Kappa statistic=0.55).

Correlation of FTD with the clinicopathological features
The clinical characteristics of the Libyan female BC patients (n = 130) and average estimate of FTD is described in Table 1 and 2. The distributions of the values are shown in Figure 2. The mean fraction of FTD in our material was 23.4% (median 20.0%, standard deviation 21.6%).
There was statistically significant correlation between the FTD and some clinicopathological features, with the strongest association observed for tumor larger than the mean (p < 0.0001). Higher values are also seen in patients with high histological grade, and those of positive lymph nodes and advanced clinical stage with (P values 0.001, 0.007 and 0.07, respectively). The difference in the FTD between invasive ductal carcinoma and other types of breast carcinomas was statistically insignificant (P = 0.46) ( Table 1).

Correlation of FTD with the survival outcome
The FTD can identify aggressive tumors and provide significant prognostic support. Kaplan-Meier curves of FTD indicated that better survival time was correlated with high FTD tumors (Figure 3). However, in multivariate analysis (all patients, with SMI, MAI, nuclear area and apoptotic count) FTD was not significant. The cut-offs (30% and 50%) can be considered to represent the most reliable thresholds for classifying patients according to tubular differentiation with (p value = 0.002, and 0.02 respectively) . and might be applied as quantitative criterium for Libyan breast cancer to separate the patients into good (≥50), moderate (30-50), and bad prognosis group (<30) (Figure 3). FTD30% thresholds could be detected also in analyses of pre-menopausal patients ( Table 2). No statistically significant threshold could be found among the other prognostic subgroups,however, FTD30% was detected as the only almost significant cut off in axillary  lymph node-negative patients. Table 3 summarizes the relative risks (RRs) of univariate analyses describing the risk of breast cancer death associated with continuous variables in results of the whole material. FTD is a predictor of survival and the patients with high FTD have a lower risk of breast cancer death. In general the FTD as a prognosticator is less efficient than the SMI, MAI, and MNA. In N+ patients the SMI and MAI were the only significant prognosticators in addition to stage. SMI was also of border line significance among postmenopausal patient ( Table 2). Multivariate analysis with FTD, MNA, FTD, AI, MAI, and SMI in all patients showed that the SMI was the only significant prognosticator (p=0.039) ( Table 4).

Discussion
The differences between African and European breast cancers have recently emerged [16,[21][22][23][24]. This study is a part of our effort to further clarifying the biology of Libyan BC and to comparing our results with the corresponding results of North European, and Central African female BC patients. We also wanted to identify more effective prognostic factors than the traditional grading system to aid therapeutic decision making. In this study it was our intention to study the Libyan breast cancer in respect to tubular differentiation. The mean value of FTD in Finland was 30.0% (SD 28.2%). This was higher than the Libyan FTD mean value 23.4% (SD 21.6%). In Nigerian material the mean FTD value was 16.7% (SD19.3%), clearly lower than the Libyan FTD figure [15,21]. The differences in FTD between Libyan and Nigerian tumours and between Libyan and Finnish tumors were significant in the whole material (p < 0.0001, 0.01). Basically our results on FTD reflect the same differences between countries in proliferative and apoptotic indices in previous work, although, the significant differences between Libyan and Finnish populations are less obvious.
These differences might be due to the fact that screening programmes were established in European countries for early detection of cancer. The African females came to the hospital in very advanced disease. 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, 80% were with LN metastasis. In the future, after establishment of screening programs in Libya, the cut point might equal that in European population. On other hand, the FTD difference between Central African, North African and European patients further strengthens our previous suggestion of biological differences and variation in genetic marker distribution between Central and North African, and European populations [16,17,22] may be involved. 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 [22,25]. The variation in haplotype marker distribution has taken place under selective environmental stresses and may be associated with radiation, viral association and life style [25].
Several authors have reported on the prognostic value of the FTD in breast carcinoma [26,7,9]. On the other hand, there are a lot of papers concluding that tubular differentiation lacks prognostic significances and is inferior to the proliferative activity and nuclear size features [10][11][12][13][14][27][28][29]. The multivariate analysis of the present study, and study of Ikpatt et al., shows that the FTD is a weak prognostic factor in relation to the proliferative and apoptotic indices. Kronqvist et al., [15] did their work by the same methodology and found that the FTD was an independent prognosticator in Finnish breast cancer. The difference in significance was smaller than in proliferative indices.
It is obvious that FTD can be used as a general prognosticator. However, mitotic activity (SMI, standardized mitotic index) is a better general prognosticator than the FTD. But FTD can add some valuable prediction particularly in premenopausal patients, when their tumours are large. The present study suggested two significant cut points for the FTD in Libyan material (30% and 50%) that could separate patients into three subgroups with favourable, intermediate and unfavourable prognosis (Figure 3). These cut points may be more suitable for the Libyan material than the cut points that used by Kronqvist et al., 2000 on Finnish material (23% and 59%).
From among five prognostic markers (MNA, AI, MAI, SMI, and FTD) the SMI and MAI were significant as general