Nuclear morphometry in ductal breast carcinoma with correlation to cell proliferative activity and...

Journal of the Egyptian Nat. Cancer Inst., Vol. 15, No. 3, September: 169-182, 2003

Nuclear Morphometry in Ductal Breast Carcinoma withCorrelation to Cell Proliferative Activity and Prognosis

MAMDOUH M. RADWAN, M.D.; KAWTHER A. AMER, M.D.; NADIA M. MOKHTAR, M.D.*;MONA A. KANDIL, M.D.; ABDEL-MONEM EL-BARBARY, M.D.** and HAYAM A. AIAD, M.D.

The Departments of Pathology, Faculty of Medicine, Menoufiya University, Pathology, National Cancer Institute,Cairo University* and Anatomy, Faculty of Medicine, Menoufiya University**

ABSTRACT

Background and objectives: Morphometry is thequantitative description of biologic structures. This studywas designed to evaluate the efficiency of morphometricmeasurements in diagnosis and prognosis of patients withbreast carcinoma.

Methods: Histological samples from 61 patients ofinvasive duct carcinoma (IDC) of no special type (NST),12 cases of ductal carcinoma in situ (DCIS) and 14 controlbreast samples taken from fibrocystic change diseasewere retrospectively analyzed by computerized nuclearmorphometry. All IDC patients underwent modified radicalmastectomy without preoperative chemotherapy. The meanfollow up was 28±19 months (range 1-71). In each case,25-50 nuclei were measured and the mean nuclear area(MNA), mean nuclear perimeter (MNP), mean maximumnuclear diameter (MMND) and mean minimal nucleardiameter (Mmnd) were measured. The mean axis ratio(MAR), mean nuclear compactness (MNC), mean nuclearsize (MNS) and mean shape factor (MSHF), were calcu-lated mathematically. To measure the nuclear diameters,a new method was employed using the AutoCAD program.Morphometric parameters were compared with differentclinicopathologic features, patient's survival and cellproliferative activity as determined by Ki-67 immunostain-ing which was evaluated quantitatively.

Results: Most of the morphometric parameters weresignificantly higher in DCIS and IDC groups than benignone. In IDC group morphometric features related to nuclearsize (MNA, MNP, MMND, Mmnd & MNS) were signif-icantly correlated to most clinicopathologic features andcell proliferative activity assessed by Ki-67 immunostain-ing. However, the shape factor failed to achieve thiscorrelation. The univariate analysis using Kaplan Meiercurves indicated that short survival time was correlatedwith high nuclear morphometric values (MNA, MNP,MMND, Mmnd, MNS and MSHF). Moreover, the Spear-man correlation analysis showed that Mmnd has the highestconverse correlation with survival (r = -0.75, p < 0.0001).In multivariate analysis using the stepwise Cox regression,the Mmnd (p = 0.002, r = 0.25) and histological grade (p= 0.009, r = 0.20) were the independent prognostic factors.

169

Conclusions: Our results indicate that nuclear mor-phometric parameters can identify an aggressive tumorphenotype and provide additional significant prognosticinformation in predicting survival and tumors at risk ofprogression. The mean minimal nuclear diameter (6.60µm) might be applied as quantitative criteria for nucleargrading that separate patients into two groups with goodand poor prognosis. The measurement of various nuclearparameters would be helpful in future application ofautomated diagnosis and grading of breast carcinoma.

Key Words: Nuclear morphometry - Ductal breast carci-noma - Ki 67.

INTRODUCTION

Breast cancer ranks the first malignant tumorthat affects females and constitutes 33% of allfemale cancers as reported at the National Can-cer Institute (NCI), Cairo University [1].

The currently available prognostic factorsin breast carcinoma are not very accurate andadditional ones are needed [2]. For instance,with long follow-up, approximately 30% ofpatients with node-negative breast carcinomadie of recurrent disease [3]. Even the combineduse of tumor size and lymph node status instaging system leaves a large group of patientswith unclear prognosis [4]. Although, the gradingsystem advocated by Elston and Ellis [2] wasassociated with high prognostic potential andimproved reproducibility [5], only the mitoticactivity is evaluated in a quantitative manner,while the evaluation of nuclear pleomorphismand tubule formation is still dependent on thepathologist's subjective factors. To improve theclinical value of malignancy grading, it hasbeen suggested to quantify nuclear pleomor-

phism by measuring nuclear features such asarea, perimeter and diameters [6,7].

Baak et al. [6] successfully applied twothresholds for the mean nuclear area "MNA"(37 µm2 and 53 µm2). Thus if MNA exceeds53 µm2 the prognosis is considerably worsethan when the MNA is less than 37 µm2. Inanother attempt for nuclear grading they foundout that a value of standard deviation of thenucleolar area above 2.49 µm2 was almostexclusively found in non-survivors [8]. Withoutnucleolar morphometry these patients wouldhave been included in a favorable prognosticgroup. Similarly, Kronqvist et al. [7] suggestedtwo thresholds for nuclear area (32 µm2, 47µm2) and the mean shortest nuclear axis (4.8µm, 6.4 µm) to be used as a tool to separate thecases with favorable, intermediate and unfavor-able course of disease.

Van Diest et al. [9] suggested the combinationof mitotic activity index, tumor size and lymphnode status into a single score (morphometricmultivariate prognostic index-MPI). The MPIprovides a strong prognostic and survival infor-mation and may also be used to select patientsfor adjuvant therapy.

The purpose of this work was to study thediagnostic value of nuclear morphometry inbenign, pre-invasive and invasive ductal breastlesions and evaluating its prognostic importancetogether with other clinicopathologic parametersincluding the proliferative activity.

MATERIAL AND METHODS

Patients and tissue specimens:This study was carried out on 61 histological

specimens of invasive duct carcinoma (IDC) ofno special type (NST), 12 cases of ductal car-cinoma in situ (DCIS) and 14 control breastsamples taken from fibrocystic change diseasecases. All the cases were diagnosed in Depart-ment of Pathology, National Cancer Institute,Cairo University between 1996 and 2001. Sur-gical pathology reports for the age, tumor size,tumor grade and lymph node status were ob-tained. Survival data were collected from com-puterized breast cancer database from the dateof surgery to the date of death or last document-ed contact. None of the patients receivedpreoperative radiation therapy or other adjuvanttreatment. All the patients were treated with

modified radical mastectomy with axillary evac-uation.

Sectioning and staining:The samples were selected on the basis of

having paraffin blocks available for re-cuttingand histopathologic re-examination. For eachcase, two sections of 5µm thickness were pre-pared from formalin-fixed paraffin-embeddedtissue blocks. One section was subjected tohematoxylin and eosin stain to verify histolog-ical grade according to the criteria describedby Elston and Ellis [2] and to perform the mor-phometric measurement. The other section wassubjected to immunohistochemical staining forKi-67 (7B11) using the improved streptavidin-biotin amplified system.

Morphometric measurements:

Capturing and digitizing of images:The morphometric measurements were per-

formed using a Digitizing and Image AnalyzingDevice (DIAD) which is composed of a com-puter (IBM, windows 98), tri-ocular photograph-ic research microscope (Olympus, Tokyo, Japan)connected to a TV monitor through a sensitiveCCD video camera (Panasonic wv-CP220, Ja-pan) that is mounted to the microscopic thirdocular piece. All the measurements were per-formed in the area of the actively proliferatingcells at the invasive border of the most cellularpart of the tumor, rejecting areas showing ne-crosis and inflammation. From the subjectivelyselected areas, an average of 5-10 microscopicfields at magnification x 1000 were capturedfrom each case. For permanent recording, aconventional VCR was used to record cameraoutputs on videocassette tapes. Signals outputfrom the player were directly input to the videoblaster IE 1500 board installed in an IBM com-patible PC running Windows 98 operating sys-tem. Thus the analog video images were digi-tized and saved on the computer as files. The(Ulead image palse 2 Go), Image editor programversion 2.06 1992-1995 was used (Ulead systeminc., 1995). All the captured images were copiedand saved in power point file (*ppt) then trans-formed to jpg format using the same programto provide a highly compressed image that couldbe recognized by AutoCAD program.

Measuring:Calibration was performed before each mea-

suring session. Thus, scale factor = 52.4246

170 Nuclear Morphometry in Ductal Breast Carcinoma with Correlation

was used to open each microscopic image file.About 25-50 nuclei from each case were mea-sured using AutoCAD R14 program. Using theSketch command under 0.00001 increments,each nucleus was outlined by the computermouse and the area and perimeter of each nu-cleus were obtained. To measure the MND andthe mnd the method of resizable circles wasused. A circle was created to overlay the imageof each nucleus in AutoCAD screen by Draw/circle/2 points command. Then it is moved andadjusted to fit as a minimal outscript circle andsimilarly, the circle is resized and adjusted tofit as maximal inscript circle. The diameter ofthe former reads the MND and that of the latterreads mnd.

Calculations:

All measurements were tabulated in Excelsheets and the following parameters were cal-culated: Axis ratio = minD/MND. Compactness= nuclear perimeter2 / nuclear area. Shape factor= 4 x π x nuclear area / nuclear perimeter2.Nuclear size = 2 x (nuclear area / π)0.5 [6-10].

Immunohistochemical analysis:Sections were stained with the monoclonal

mouse antihuman Ki-67 clone: 7B11 (Zymed,08-0192) using the improved streptavidin-biotinamplified system with trypsin digestion andheat induced epitope retrieval technique.

Ki-67 immunostaining evaluation:

Ki-67 positivity was identified as any nuclearor nucleolar brown staining regardless of stain-ing intensity. Counting was started at the pe-riphery of the tumors and additional contiguousfields were selected randomly until at least 500tumor cells were counted [11,12].

To objectify the immunostaining interpreta-tion, a new quantitative method was developedusing an assembled system. It is composed ofan Olympus microscope with mounted videocamera to transfer images of the immunostainedsections with magnification x200 to a TV mon-itor through the video caliper. At the same timea video caliper is connected to the same monitorand adapted to generate a square frame that isapplied to the projected field of vision. Countingwas performed inside this square. Cell nucleiare counted only if they are fully containedwithin the frame. The video caliper is adjusted

to generate a 100 x 100 µm2 square. Knowingthat the tissue section thickness is 5 µm, thusthe counting was performed in a volume mea-suring 100 x 100 x 5 µm3. Ki-67 was quantita-tively measured by counting the number ofpositive cells and the number of negative cellsin that volume. In each case, the total percentageof the positive cells was calculated as follow:Number of positive cells / total number of cells x100

Statistical analysis:Data were collected, tabulated and statisti-

cally analyzed using SPSS version 9 statisticalprogram. The values entered into statisticalanalysis were the mean of the morphologic andmorphometric parameters. In all tests p valuesbelow 0.05 were regarded as significant. Studentt test was used for comparison between twogroups of quantitative variables and f test formore than 3 groups. Also Mann Whitney u testwas used for comparison between two groupswith non-parametric distribution and correlationcoefficient test (Spearman rank correlation test)for studying correlation between variables. Forsurvival analysis, univariate analysis was basedon Kaplan-Meier product-limit estimates andLog rank test was used to compare factorsaffecting survival. Stepwise multiple regressionanalysis (Cox regression) was used to study themost relevant factors affecting survival.

RESULTS

The clinicopathologic variables of our breastcancer cases are included in (Table1).

Comparison between different studied groupsaccording to the morphometric parameters(Table 2):

Most of the nuclear morphometric featuresshowed high mean values in cells of DCIS groupin comparison to the benign ductal epithelialcells group and the differences were highlysignificant in the MNA, MNP, MMND, Mmndand MNS (p < 0.0001) (Figs.1,2). However, nosignificant association has been found betweenthe two groups as regard MAR, MNC andMSHF (Table 2). The means of all the morpho-metric parameters were significantly higher inIDC group than in benign ductal epithelial groupand the differences were highly significant(Table 2) (Figs.1,3). Comparing between theDCIS group and the IDC group, the nuclearmorphometric parameters were higher in cells

Mamdouh M. Radwan, et al. 171

of DCIS group than the IDC cells group andagain the differences were highly statisticallysignificant in MNA, MNP, MMND, Mmnd,MSHF and MNS (p < 0.0001) (Figs. 2,3). How-ever, no significant association has been foundbetween both groups as regard the MAR (p >0.05).

Nuclear morphometric features and differentclinicopathologic parameters in IDC group:

Correlation between the mean nuclear area(MNA) and clinicopathologic features (Table3):

The mean nuclear area for the whole IDCseries was (54.54 µm2 ± 16.82). It was (45.87µm2 ± 25.7) in pT1 tumors, (49.74 µm2 ±14.47)in pT2 and (63.30±14.67) in pT3 tumors. Thedifferences between these groups were highlystatistically significant (p < 0.01). Similarly,the MNA was greater in grade III tumors (62.70µm2 ± 16.31) than grade I and II (50.25±15.63). Again the difference was statistically highlysignificant (p < 0.01). As regards lymph nodestatus, cases with positively involved lymphnodes showed larger MNA than those with noninvolved lymph nodes and the difference washighly significant (p < 0.001).

Correlation between the mean nuclear pe-rimeter (MNP) and clinicopathologic features(Table 3):

The MNP for the whole IDC series was(30.02 µm ± 6.19). It was in tumors with gradeI and II (28.44±4.35). However in grade III itwas (33.02±7.99) and the difference betweenboth groups was statistically significant (p <0.05). The group of IDC with lymph nodemetastasis showed higher MNP (30.96±6.28)than that with lymph nodes free of metastasis(25.75 ± 3.54). The difference between bothgroups was highly statistically significant (p <0.001).

Correlation between the mean maximal nu-clear diameter (MMND) and clinicopathologicfeatures (Table 3):

The MMND for the whole series was (10.26µm ± 1.57). The pT1 tumors showed MMNDof (8.92±2.67) and pT2 tumors showed MMNDof (9.90±1.29). However, pT3 tumors showedMMND of (11.06 ±1.33). The difference be-tween these groups was statistically highlysignificant (p < 0.01). The higher the histologicalgrade the larger the MMND as the MMND in

cases with grade I and II was (9.82±1.50).However, tumors with grade III showed MMNDof (11.10±1.34) and the difference was highlysignificant (p < 0.001). Also, the lymph nodestatus was directly associated to MMND. Tu-mors with positive lymph nodes showed MMNDof (10.53±1.50) and tumors with negative lymphnodes showed MMND (9.02±1.28) and therewas high statistical difference between bothgroups (p < 0.01).

Correlation between the mean minimal nu-clear diameter (Mmnd) and clinicopathologicfeatures (Table 3):

The Mmnd for the whole series was (6.60µm ± 1.11). It was (5.90±1.92) in pT1 tumors,(6.27±0.94) in pT2 tumors and (7.23±0.85) inpT3 tumors. The differences between thesegroups were highly statistically significant (p< 0.001). The Mmnd in tumors with grade I andII was (6.33±1.02). However in grade III it was(7.11±1.12) and the difference between bothgroups was highly statistically significant (p <0.01). The Mmnd was highly associated (p <0.001) with lymph node status as it was (6.83± 1.02) for the lymph node positive tumors andwas (5.55±0.93) for the lymph node negativetumors. As regard the lymph node stage, theMmnd was significantly associated with it (p< 0.05) with Mmnd of (5.96±1.24) for lymphnode stage 1 and (6.83±1.21) for lymph nodestage 2. However, tumors with lymph nodestage 3 showed Mmnd of (6.76±0.87).

Correlation between the mean nuclear axisratio (MAR) and clinicopathologic features(Table 4):

The MAR for the whole series was (0.64 ±0.04). No association has been found betweenMAR and any pathological feature (p > 0.05).

Correlation between the mean nuclear com-pactness (MNC) and clinicopathologic features(Table 4):

The MNC for the whole IDC series was(16.79±1.39). It was directly associated withthe lymph node status as it was (16.65±1.42)for tumors with positive lymph nodes and was17.43±1.11 for tumors with negative lymphnodes. The difference between both groups wasstatistically significant. However, no significantassociation between the MNC and other clini-copathologic features.


Correlation between the mean nuclear shapefactor (MSHF) and clinicopathologic features(Table 4):

The MSHF for the whole IDC series was(0.74±0.02). There was no significant associa-tion between it & any clinicopathologic feature.

Correlation between the mean nuclear size(MNS) & clinicopathologic features (Table 4):

The MNS for the whole IDC series was(8.14±1.29). The MNS was higher in tumorswith larger size, higher grade, positive lymphnodes status and lymph node stage. However,no significant association was found betweenit and age, mitosis/10HPF tumor stage, or NPI.

Ki-67 immunostaining:Ki-67 expression was progressively in-

creased from benign to IDC cases passingthrough the DCIS group cases (Table 5) (Fig.4). The differences between the three groupswere highly statistically significant (p < 0.0001).Moreover, within the DCIS group, the Ki-67positivity was higher in the comedo (high grade)carcinoma than in the non-comedo DCIS cases. In general, Ki-67 immunostaining could identifycases that have an aggressive morphologic ap-pearance, manifested by larger size, higherhistological grade, stage and higher lymph nodestage.

Correlation between the different morphometricparameters and Ki-67 reactivity in IDC group(Table 6):

A direct linear relationship was observedbetween Ki-67 reactivity and the followingmorphometric parameters: MNA, MNP,MMND, Mmnd, MSHF and MNS. The highestcorrelation was found with MMND (r = 0.36,p < 0.01). However, a negative relationship wasfound between the Ki-67 and (MAR and MNC).

Survival analysis:At time of analysis, 17 patients (35%) died

and 32 patients (65%) were alive. Median sur-vival for the whole series was 22 months (mean28 months, range 1-71 months). Using theSpearman rank correlation test, nuclear mor-phometric parameters (MNA, MNP, MMND,Mmnd, MAR, MSHF, MNS) were converselycorrelated with survival time. The highest cor-relation was found between the Mmnd andsurvival (r = -0.75, p < 0.0001) (Table7). How-ever, the MNC showed direct linear correlationwith survival.

In the univariate kaplan-Meier analysis, themean values were used as cut-off with approx-imate 95% confidence interval. All the studiedmorphometric parameters were significantlycorrelated with survival. Therefore, survivalwas shorter for cases with higher MNA, MNP,MMND, Mmnd, MAR, MNC, MSHF, MNS,results are shown in Table (8) and illustrated inkaplan-Meier curves (Figs. 5-8). Tumor size,grade, lymph node status, Nottingham prognos-tic index and Ki-67 expression also had prog-nostic value in the univariate analysis. However,there was no significant association betweenage, mitotic count and lymph node stage ortumor stage versus survival.

To develop a prognostic model for predictionof survival, we performed a stepwise Cox re-gression analysis. Only Mmnd and histologicalgrade retained independent prognostic signifi-cance. For Mmnd, the p value was p = 0.002 &risk ratio was r = 0.25 and for the histologicgrade, p = 0.009 & r = 0.20, (Table 9). Therefore,the minimal nuclear diameter (mnd) was themost powerful discriminator. Using the meanvalue 6.60, a clear separation of patients withprognostically favorable and unfavorable diseasewas achieved (p < 0.0001).


Table (1): The clinicopathologic data for patients withinvasive duct carcinoma.

Clinicopathologic variables(61 cases)

No (%)61 (100%)

Age (years):< 46≥ 46

Tumor size (cm):< 2 (pT1)> 2-5 (pT2)> 5 (pT3)

Histologic grade:123

Lymph node status:PositiveNegative

Lymph node stage:1 (negative)2 (1-3)3 (> 3)

Tumor stage (TNM):123

Nottingham prognostic index:1 (< 3.41)2 (3.41-5.41)3 (> 5.41)

32 (52)29 (48)

5 (8)33 (54)23 (38)

2 (3)38 (63)21 (34)

50 (82)11 (18)

11 (18)20 (33)30 (49)

1 (2)24 (39)36 (59)

1 (2)26 (42)34 (56)


Table (2): Morphometric parameters in different studied groups.

The parameterBenign group

(14 cases)Mean ± SD

MNAMNPMMNDMmndMARMNCMSHFMNS

38.34±7.7325.08±2.368.50±0.775.79±0.60.67±0.0216.24±0.400.77±0.027.04±0.60

DCIS group(12 cases)

Mean ± SDp value

IDC group(61 cases)

Mean ± SDp value

105.63±17.4540.73±3.2814.19±1.249.27±0.770.68±0.0816.2±0.440.78±0.0211.43±0.97

< 0.0001**#< 0.0001**#< 0.0001**#< 0.0001**#> 0.05 ns #> 0.05 ns #> 0.05 ns #< 0.0001**#

54.54±16.8230.02±6.1910.26±1.576.60±1.110.64±0.0416.79±1.390.74±0.028.14±1.29

< 0.0001**‡< 0.0001**‡< 0.0001**‡< 0.0001**‡< 0.01**‡< 0.01**‡< 0.01**‡< 0.0001**‡

Mean nuclear area (MNA), mean nuclear perimeter (MNP), mean maximal nuclear diameter (MMND), mean minimal nucleardiameter (Mmnd), mean axis ratio (MAR), mean nuclear compactness (MNC), mean nuclear shape factor (MSHF) and mean nuclearsize (MNS). DCIS = Ductal carcinoma in situ, IDC = Invasive duct carcinoma, Ns = Not significant, **Highly significant.

# p values are for comparisons between benign group and DCIS group. ‡ p values are for comparisons between benign group andIDC.

Table (3): Correlation between morphometric parameters (MNA, MNP, MMND and Mmnd) and the clinicopathologicfeatures in IDC.

Variables MNA ± SD


Tumor size (cm):pT1pT2pT3

Mitosis:123

Grade:1,23


Lymph node stage:123

Tumor stage:1,23

NPI:23

56.02±16.9852.9±16.79

45.87±25.749.74±14.4763.30±14.67

53.96±16.663.23±12.0153.42±18.15

50.25±15.6362.70±16.31

57.66±16.5640.34±9.16

46.01±15.3257.87±19.0656.54±14.97

34.44±21.6649.33±16.00

40.55±20.3845.35±19.17

p value p value

> 0.05 ns

< 0.01**

> 0.05 ns

< 0.01**

< 0.001**

> 0.05 ns

< 0.001**

> 0.05 ns

30.94±7.2329.00±4.7

26.26±7.8029.31±7.0531.86±3.70

29.47±4.2132.33±2.2330.02±6.19

28.44±4.3533.02±7.99

30.96±6.2825.75±3.54

27.24±4.5430.35±4.9831.18±7.30

29.6±4.5330.31±7.17

29.44±4.6630.48±7.22

> 0.05 ns

> 0.05 ns

> 0.05 ns

< 0.05*

< 0.001**

> 0.05 ns

> 0.05 ns

> 0.05 ns

10.41±1.4610.1±1.70

8.92±2.679.90±1.2911.06±1.33

10.31±1.4811.12±0.869.96±1.81

9.82±1.5011.10±1.34

10.53±1.509.02±1.28

9.51±1.5410.44±1.7410.51±1.39

10.2±1.5810.3±1.58

10.17±1.5910.33±1.57

MNA = Mean nuclear area. MNP = Mean nuclear perimeter. MMND = Mean maximal nuclear diameterMmnd = Mean minimal nuclear diameter.Ns = Not significant. * Significant. ** Highly significant.

MNP ± SD p value MMND ± SD p value Mmnd ± SD

> 0.05 ns

< 0.01**

> 0.05 ns

< 0.001**

< 0.01**

> 0.05 ns

> 0.05 ns

> 0.05 ns

6.68±1.046.51±1.20

5.90±1.926.27±0.947.23±0.85

6.54±1.047.10±0.787.25±0.27

6.33±1.027.11±1.12

6.83±1.025.55±0.93

5.96±1.246.83±1.216.76±0.87

6.63±1.136.58±1.11

6.58±1.186.62±1.07

> 0.05 ns

< 0.001**

> 0.05 ns

< 0.01**

< 0.01**

< 0.05*

> 0.05 ns

> 0.05 ns


Table (4): Correlation between morphometric parameters (MAR, MNC, MSHF and MNS) and the clinicopathologicfeatures in IDC.

Variables MAR


Tumor size (cm):pT1pT2pT3

Mitosis:123

Grade:1,23


Lymph node stage:123

Tumor stage:1,23

NPI:23

0.64±0.020.65±0.05

0.66±0.020.63±0.040.66±0.03

0.64±0.040.64±0.020.66±0.03

0.65±0.030.64±0.05

0.65±0.030.62±0.05

0.63±0.050.65±0.030.65±0.03

0.65±0.030.64±0.04

0.65±0.030.64±0.04

p value p value

> 0.05 ns

> 0.05 ns

> 0.05 ns

> 0.05 ns

> 0.05 ns

> 0.05 ns

> 0.05 ns

> 0.05 ns

16.67±1.7316.92±0.90

16.98±0.6116.80±1.8516.73±0.50

16.73±1.7917.15±0.8616.79±1.39

16.71±1.5016.93±1.00

16.65±1.4217.43±1.11

16.54±2.791680±0.5416.91±0.57

16.46±1.9217.02±0.8

16.56±1.8616.97±0.85

> 0.05 ns

> 0.05 ns

> 0.05 ns

> 0.05 ns

< 0.01**

> 0.05 ns

> 0.05 ns

> 0.05 ns

0.74±0.020.74±0.03

0.74±0.020.74±0.030.75±0.02

0.74±0.020.74±0.030.75±0.02

0.74±0.020.74±0.03

0.75±0.020.72±0.03

0.73±0.030.75±0.020.74±0.02

0.75±0.020.74±0.02

0.74±0.020.74±0.03

MAR = Mean axis ratio, MNC = Mean nuclear compactness, MSHF = Mean shape factor, MNS = Mean nuclear size, Ns = Not significant,** Highly significant.

MNC p value MSHF ± SD p value MNS ± SD

> 0.05 ns

> 0.05 ns

> 0.05 ns

> 0.05 ns

> 0.05 ns

> 0.05 ns

> 0.05 ns

> 0.05 ns

8.27±1.218.01±1.38

7.22±2.277.81±1.098.82±1.01

8.11±1.218.88±0.808.03±1.49

7.81±1.228.78±1.20

8.4±1.216.98±1.02

7.44±1.358.39±1.438.33±1.05

8.17±1.338.13±1.28

8.10±1.368.18±1.25

> 0.05 ns

< 0.01**

> 0.05 ns

< 0.01**

< 0.001**

< 0.05*

> 0.05 ns

> 0.05 ns

Table (5): Ki-67 expression in the different studied groups.

Group Number of cases

BenignDCISIDC

141261

0.69±1.8214.31±22.7919.79±42.9

5.074.215.42

< 0.0001**1< 0.0001**2< 0.0001**3

DCIS = ductal carcinoma in situ. IDC = Invasive duct carcinoma. 1 Comparison between benign and DCIS groups.2 Comparison between DCIS and IDC groups. 3 Comparison between benign and IDC groups. ** Highly significant.

Ki-67Mean ± SD t. test p value

Table (6): Spearman correlation between different nuclear morphometric parameters and Ki-67 immunostaining reactivityin IDC group.

Morphometric parameter


+0.33+0.33+0.36+0.31-0.06-0.03+0.10+0.34

< 0.01**< 0.01**< 0.01**< 0.01**> 0.05 ns> 0.05 ns> 0.05 ns< 0.01**

Mean nuclear area (MNA), mean nuclear perimeter (MNP), mean maximal nuclear diameter (MMND), mean minimal nucleardiameter (Mmnd), mean axis ratio (MAR), mean nuclear compactness (MNC), mean nuclear form factor (MFF), mean nuclear shapefactor (MSHF) and mean nuclear size (MNS). Ns = Not significant, ** Highly significant. r = Regression.

r p

Table (7): Spearman correlation between survival time (months) and nuclear morphometric parameters in IDC cases.

Variables


-0.73-0.74-0.71-0.75-0.36+0.28-0.27-0.73

< 0.001**< 0.0001**< 0.0001**< 0.0001**< 0.01**< 0.05*< 0.05*< 0.0001**

IDC = Invasive duct carcinoma, mean nuclear area (MNA), mean nuclear perimeter (MNP), mean maximal nuclear diameter (MMND),mean minimal nuclear diameter (Mmnd), mean axis ratio (MAR), mean nuclear compactness (MNC), mean nuclear shape factor (MSHF)and mean nuclear size (MNS). * Significant. ** Highly significant. r = Regression.

r p

Survival time (months)


Table (9): Variables included in the prognostic model derived by multivariate sepwise Cox regression analysis.

Variables

MmndHistologic grade

17.7812.96

0.0020.009

Mmnd = Mean minimal nuclear diameter, r = Risk ratio, OR = Odd ratio, CI = Confidence interval.

rpStatistical score OR 95% CI

0.250.20

7.163.95

2-25.621.39-11.18

Table (8): Univariate survival analysis of IDC cases classified by tumor cell nuclear morphometry (Kaplan-Meier estimates).

Variables Mean (95% CI)(range)

MNA (µm2):≤ 54.54> 54.54

MNP (µm):≤ 30.02> 30.02

MMND (µm):≤ 10.26> 10.06

Mmnd (µm):≤ 6.60> 6.60

MAR (µm):≤ 0.64> 0.64

MNC (µm2):≤ 16.79> 16.79

MSHF (µm2):≤ 0.74> 0.74

MNS (µm2):≤ 8.14> 8.14

61.29 (54.71-67.87)29.74 (16.85-42.62)

61.02 (53.90-68.14)19.72 (14.41-25.04)

62.39 (55.46-69.1)24.01 (15.55-32.48)

62.9 (56.5-69.29)14.79 (11.88-17.69)

56.46 (45.91-67.01)40.19 (28.81-51.58)

29.18 (20.49-37.88)56.39 (47.3-65.49)

60.33 (50.98-69.68)34.77 (24.53-45.01)

62.9 (56.5-68.29)14.96 (12.19-17.78)

p value

–19 (15.71-22.29)

–19 (14.70-23.3)

19 (15.7-22.29)–

–18 (13.02-22.08)

6143 (15.12-70.88)

19 (0-43.41)–

–43 (12.09-73.91)

–18 (13.08-22.92)

IDC = Invasive duct carcinoma, CI = Confidence interval, mean nuclear area (MNA), mean nuclear perimeter (MNP), mean maximalnuclear diameter (MMND), mean minimal nuclear diameter (Mmnd), mean axis ratio (MAR), mean nuclear compactness (MNC), meannuclear shape factor (MSHF) and mean nuclear size (MNS). * Significant. ** Highly significant.

Median (95% CI)(range

Survival rate % Test of sign(Log-rank)

83.3348

8050

83.3348

85.1940.91

76.1957.14

44.477.42

8445.83

84.6243.48

18.53

20.25

20.95

32.76

3.78

7.98

8.85

31.53

< 0.0001**

< 0.0001**

< 0.0001**

< 0.0001**

< 0.05*

< 0.01**

< 0.01**

< 0.0001**

Survival time (months)

177Mamdouh M. Radwan, et al.

Fig. (1): Print screen of AutoCAD drawing window show-ing a captured image for benign duct epithelialcells with outlined nuclei (white), surrounded byminimal outscript resizable circle (blue) andmaximal inscript resizable circle (yellow).

Fig. (2): A print screen of AutoCAD drawing windowshowing captured microscopic field for DCIScase with two outlined nuclei (white), surroundedby minimal out-script resizable circle (MAXD)(blue) and maximal in-script resizable circle(MIND) (yellow).

Fig. (3): Print screen of AutoCAD drawing window show-ing captured image for invasive duct carcinomacells with outlined nuclei (white), surrounded byminimal out-script resizable circle (MAXD)(blue) and maximal in-script resizable circle(MIND) (yellow).

Fig. (4): Ki-67 expression in invasive duct carcinoma casedemonstrated as brown stain in 84% of the nuclei,note the positively stained nucleoli and mitoticfigure (x400).

Fig. (5): Kaplan Meier curves for patients with meannuclear area (MNA) ≤ 54.54 and patients withMNA ≥ 54.54 (p < 0.0001, highly significant,Log rank = 18.53).

Fig. (6): Kaplan Meier curves for patients with meannuclear size (MNS) ≤ 8.14 and patients withMNS ≥ 8.14 (p < 0.0001, highly significant,Log rank = 31.53).

Cum

sur

viva

l

0 10 20 30 40 50 60 70 80Survival time in months

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Nuclear size

> 8.14

<= 8.14

Cum

sur

viva

l


1.2

1.0

0.8

0.6

0.4

0.2

0.0

MNA

> 54.54

< = 54.54

1.Area = 117.6543.223.31MND = 14.60mnd = 10.52

2.Area 93.71perimeter=40.39MND = 13.40Mnd = 8.2

1.Area = 30.94Perimeter=23.31MND = 7.52Mnd = 5.40






DISCUSSION

The increased sophistication in automationand computer technologies has improved thevisualization and quantification of human tumorcell parameters. The current study draws theattention to the issue of nuclear morphometry.To carry out the morphometric measurements,the semiautomatic assembled digitizing andimage-analyzing device (DIAD) was used. Themeasurement of one sample took approximately20-30 minutes. There is no doubt that automatedimage analyzers could perform these morpho-metric measurements in such a more rapid way.However, the unavailability and expensivenessof such techniques encouraged us to prefer thepartially automated approach. Moreover, inhistologic sections, the automatic image analyz-ers have the disadvantage of the complexity ofhistologic images. Thus, there is difficulty inthe automated detection of the object or featuresto be measured. For example, some nuclei maynot be easily delineated when their edges blendin light density with the cytoplasm [13]. There-fore, an automated imaging system would notbe the best choice with respect to efficiency oreven accuracy in such a situation.

In the present study, to ensure validity ofresults, measurements were performed in theactively proliferating cells at the invasive borderof the most cellular part of the tumor as breastcancer has a heterogeneous histology and vary-ing nuclear structure between areas [7,14].

In the current study, a new method was

applied to measure the MND (maximal nucleardiameter) and mnd (minimal nuclear diameter)by utilizing an easy option in the AutoCADprogram. The developed method implies draw-ing circles to overlay the nuclei images in Au-toCAD screen. The circles may be moved orresized to fit as maximum inscript circle or tofit as the minimum outscript circle. Using thesefreely mobile circles in this work was provedto be better than both Weibel's ruler [15] and El-Barbary concentric circles [16]. Being resizable,only one circle profile could be used to measuremany diameters. This overcame the limitedaccuracy when Weibel’s fixed ready-made cir-cles were used. In addition, its application ona digital image provides the advantage of zoom-ing that produces better image details and avoidsthe use of large number of fixed hard micro-scopic copies. Moreover, it avoids cumbersomeeyestrain still encountered with El-Barbaryconcentric circles [16].

In the current work, all the studied morpho-metric parameters showed significantly lowervalues in benign ductal epithelial cells as com-pared to both in situ and invasive duct carcinomacells, a result that coincides with other reports[17,18,19] and addresses the importance of mor-phometry as a diagnostic tool that could differ-entiate benign from malignant cases.

Nuclear grade of DCIS is one of the mostimportant pathological indicators that predictthe likelihood of recurrence after breast conser-vation [20]. The majority of our studied DCIScases (67%) were of nuclear grade III. Our


Fig. (8): Kaplan Meier curves for patients with histologicalgrade 1 & 2 and patient with grade 3 (p < 0.001“highly significant”, Log rank = 13.09).

1.2

1.0

0.8

0.6

0.4

0.2

0.0

-0.2

Cum

sur

viva

l


Histologic grade

Grade 3

Grade 2

Fig. (7): Kaplan Meier curves for patients with minimalnuclear diameter (mnd) ≤ 6.60 and patients withmnd ≥ 6.60 (p < 0.0001, highly significant, Lodrank = 32.76).

1.2

1.0

0.8

0.6

0.4

0.2

0.0

Cum

sur

viva

l


Minimal diameter

> 6.60

<= 6.60

nuclear morphometric measurement of DCIScases showed significantly higher numericalvalues than that of IDC cases. These findingsdo not sound strange in breast carcinoma wherethe nuclear features of grade III DCIS are moreaggressive than those of well-differentiated IDC[20].

The published results showed that in histo-logic sections of breast, the mean nuclear area(MNA) of nuclei from ductal hyperplasia with-out atypia was reported to be (30.62±6.67) [19].These finding are slightly lower than our MNA(38.34±7.73) for the benign ductal epithelialcells. Similarly, in the present study, the meannuclear area of IDC cases was (54.54±16.82).This coincides with previous range of data from(24.4±12.8) up to (67.8±18.35) [21-24]. Thedifferences in the observed values of nucleararea among different publications may be dueto factors related to individual variations andapplication of different morphometric methods.However, a strictly standardized and uniformmeasuring technique, with regular calibrationof the computerized morphometric equipmentwith a micrometer slide, could ensure reproduc-ible results [25].

The clinicopathologic data of our malignantcases confirmed the aggressive nature charac-terizing this type of carcinoma in Egyptianfemales. Where 49% of cases showed more thanthree lymph nodes invaded by metastasis, 59%were TNM stage III and 56% were of poorprognostic group according to Nottingham prog-nostic index (NPI). In the current study, themorphometric parameters related to size (MNA,MNP, MMND, Mmnd and MNS) were signifi-cantly correlated to features of aggressivenessin breast carcinoma such as tumor size, grade,lymph node status and cell proliferative activitydetermined by Ki-67. These finding agree withthose reported by Fergene et al. [26] in femalebreast. In contrast, the shape indices such as,MSHF and MNC showed no correlation toclinicopathologic features.

The most popular method used for Ki-67immunostaining evaluation is the visual scoring.This method is subjective and reviewing theliterature showed a wide range of median valuesreported for Ki-67 positivity, ranging from 7.5%to 20% [27]. Recently, quantitative evaluationof Ki-67 immunostaining has been introduced[12,28,29]. In the current study, we were able to

evaluate Ki-67 immuno-positivity quantitativelyusing a new assembled method by which thepositive cells were only counted if they werecompletely contained in a fixed volume. Thisstandardized steriologic (3 dimensional) methodensures accuracy and lowers the observer vari-ability. The mean Ki-67 positivity obtained inthe present study was higher than that of previ-ous reports [28]. This difference could be dueto the use of a different antibody and the appli-cation of the newly developed quantitativemethod.

In the present study, Ki-67 expression wassignificantly lower in benign ductal epithelialcells than in both DCIS and IDC cases. Also,there was more frequent Ki-67 staining in IDCthan in DCIS. These finding agree with thosereported by Shen et al. [30] and supports theconcept of transition from benign to malignantbreast lesions [20].

In the current work, the expression of Ki-67 succeeded to identify IDC cases with aggres-sive morphologic nature such as cases withhigher histologic grade, higher lymph nodestage and larger tumor size. These findingsagree with those reported in the literature [31,32].

According to the survival analysis of thepresent study, the well-known prognostic valueof lymph node status, tumor size, histologicgrading and the Nottingham prognostic indexwas confirmed.

Our results showed that tumor cell prolifer-ation rate determined by Ki-67 (clone 7B11)identified patients with shorter survival. Theseresults confirm previous reports [27,33,34]. How-ever, this prognostic impact was lost in multi-variate analysis when the morphometric vari-ables were included.

In the univariate analysis of the currentstudy, preliminary examination of the survivaldata according to the morphometric parametersindicated that patients with larger nuclear di-mensions and irregular shape had poor progno-sis. This was confirmed by comparison of Ka-plan-Meier survival curves.

Several authors have reported on the prog-nostic importance of estimates of nuclear areain breast carcinoma [6,7,35]. These findings wereconfirmed in the univariate analysis of thecurrent study. In contrast, Van Diest and Baak[36] did not find this correlation.


In the present study, the mean nuclear pe-rimeter (MNP) showed significant prognosticvalue in the univariate analysis. This result isin accordance with previous findings [37,38].Moreover, the MNP was found to be the mostpowerful prognostic parameter in a study doneon endometrial carcinoma [10].

The maximal nuclear diameter (MMND)was correlated to shorter survival in the univari-ate analysis of the current study. These resultsare in line with those reported in endometrialcarcinoma [10], colorectal carcinoma [37] andin breast carcinoma [7]. The latter group hasfound that patients with MND above (8.0 ± 1.6)had poor prognosis, a value that is slightly lowerthan ours (10.3 ± 1.6).

Similarly, the mean minimal nuclear diam-eter (Mmnd) of IDC in the present study (6.6± 1.1) was associated with shorter survival.This agrees with that reported by Kronkqvistet al. [7] where the Mmnd in their work (5.8 ±1.1) was slightly lower than ours. Furthermore,in another study, the Mmnd value was provedto differentiate between small cell lung carci-noma of the intermediate type and non-smallcell lung carcinoma [39].

The mean axis ratio (MAR) is the ratiobetween the shortest and the longest diameterof the nuclei. It has been previously reportedto have a prognostic impact in breast carcinoma[6] and in endometrial carcinoma [10]. In theunivariate analysis of the current study, MARshowed good correlation with survival, a featurethat was lost in multivariate analysis.

The mean nuclear shape factor (MSHF) isan indicator of nuclear shape irregularity. Ac-cording to Baak et al. [6] shape factor is dimen-sionless and its value is equal to 1.0 in circlesand < 1.0 in spheroid nuclei. In the presentstudy, the MSHF was 0.74 ± 0.02 and it showedno statistically significant correlation to any ofthe clinicopathologic features. However, it al-lowed good discrimination between cases withgood and poor prognosis but this was in theunivariate analysis only. Our results agree withthose reported by Giardina et al. [40]. In contrast,others found no prognostic significance of theshape indices [6,7].

In the current study, the morphometric eval-uation of mean nuclear size (MNS) has been

found to be of statistically significant value inthe univariate survival analysis. These resultsagree with other reports [6,35,41].

In the multivariate analysis of the presentstudy, the mean minimal nuclear diameter(Mmnd) was the strongest prognostic factorfollowed only by the histologic grade. Thisindicates that Mmnd and histologic grading aresuperior to traditional prognostic variables andmay add more prognostic information to breastcarcinoma patients. Similar results were reportedbefore by Krpnqvist et al. [7].

The use of numerical probabilities as thresh-olds for making clinical decision has been ad-vocated and applied successfully in previousstudies. The standard deviation of nucleolararea of breast cancer cells above 2.49 µ2 wasalmost exclusively found in non-survivors [8].Kronqvist et al. [7] has suggested two thresholdsfor the Mmnd (4.8 µm and 6.4 µm) that couldseparate patients into three subgroups withfavorable, intermediate and unfavorable prog-nosis. In the current study, it was found thatsurvival among patients with Mmnd ≤ 6.60 µmwas significantly better than among patientswith Mmnd > 6.60 µ. So, we suggest that thisvalue might be used as quantitative criteria forseparating patients into two groups with goodand poor prognosis. Moreover, the Mmnd andhistologic grade are suggested to be includedin future studies a combined morphometricmultivariate prognostic index to provide strongerprognostic information similar to that developedby Van Diest et al. [9].

Conclusion:For diagnostic purposes, nuclear morpho-

metric parameters could be used as a diagnostictool to differentiate between benign and malig-nant ductal breast lesions. As a diagnostic tool,nuclear morphometrics especially those relatedto size might identify cases with aggressivetumour nature manifested by larger size, highergrade, increased proliferative activity and poorsurvival. Furthermore, the mean minimal nucleardiameter (6.60 µm) could be applied as quanti-tative criteria for nuclear grading that separatepatients into two groups with good and poorprognosis. Further morphometric studies onfine needle aspiration cytology could revealnew morphometric thresholds that help inpreoperative differential diagnosis of breastlesions.


REFERENCES1- El-Bolkainy M.N.: Breast Cancer. Chapter 10. In:

Topographic pathology of cancer. National CancerInstitute, Cairo university. 1998, pp 81-87.

2- Elston C.W. and Ellis I.O.: Pathological prognosticfactors in breast cancer. I. The value of histologicgrades in breast cancer. Experience from a large studywith long-term follow-up. Histopathology, 1991, 19:403-410.

3- Yuan J., Hennessy C., Givan A.L., Corbett I.P., HenryJ.A., Sherbet G.V. and Lennard T.W.: Predictingoutcome for patients with node-negative breast cancer:a comparative study of the value of flow cytometryand cell image analysis for determination of DNAploidy. Br. J. Cancer, 1992, 65: 461-465.

4- Page D.L.: Prognosis and breast cancer: recognitionof lethal and favorable prognostic types. Am. J. Surg.Pathol., 1991, 15: 334-349.

5- Frierson H.F., Wolber R.A., Berean K.W., Franque-mount D.W., Gaffey M.J., Boyd J.C. and Wilbur D.C.:Inter-observer reproducibility of the Nottinghammodification of the Bloom and Richardson histologicgrading scheme for infiltrating duct carcinoma. Am.J. Clin. Pathol., 1995, 103: 195-198.

6- Baak J.P., Van Dop H., Kurver P.H.J. and HernansJ.O.: The value of morphometry to classic prognosti-cators in breast cancer. Cancer, 1985, 56: 374-382.

7- Kronqvist P., Kuopio T. and Collan Y.: Breast cancerprognostication: Morphometric thresholds for nucleargrading. Br. J. Cancer, 1998, 78: 800-805.

8- Baak J.P.: The relative prognostic significance ofnucleolar morphometry in invasive ductal breastcancer. Histopathology, 1985, 9 (4): 437-444.

9- Van Diest P.J., Matze-Cok P., van Galen C.M., vanGorp L.H., Los J., Schapers R.F.M., Somsen J.G. andAriens A.: Reproducibility of mitosis counting in2,469 breast cancer specimens: Results from themulticenter morphometric mammary carcinomaproject. Human Pathology, 1992, 23: 603-607.

10- Salvesen H., Iversen O.K. and Akslen L.A.: Prognosticimpact of morphometric nuclear grade of endometrialcarcinoma. Am. Cancer Soc., 1998, 83: 956-64.

11- Elston C.W., Ellis I.O., Goulding H. and Pinder S.E.:Role of pathology in the prognosis and managementof breast cancer. In Elston C.W. and Ellis IO (eds.)Systemic pathology 3rd edition. Vol. 13 The breast.Churchill Livingstone, 1998, pp 385-433.

12- Biesterfeld S., Kluppel D., Koch R., Schneider S.,Steinhagen G., Mihalcea M. and Schroder W.: Rapidand prognostically valid quantification of immuno-histochemical most positive tumor focus. A prospectivefollow-up study on breast cancer using antibodiesagainst MIB-1, PCNA, ER, and PR. J. Pathol., 1998,185: 25-31.

13- True L.D.: Morphometric applications in anatomicpathology. Human Pathol., 1996, 27 (2): 450-467.

14- Chiusa L., Margaria E. and Achille P.: Nuclear mor-

phometry in male breast carcinoma: association withcell proliferative activity, oncogene expression, DNAcontent and prognosis. Int J Cancer, 2000, 89: 494-499.

15- Weibel E.R.: Steriological methods. In: Practicalmethods for biological morphometry. San Diego, CAAcademic Press 1979. (Quoted from El-Barbary,1997).

16- El-Barbary A.A.: A modification of Weibel methodfor measuring microscopic structural diameters usinginscript and outscripts circles. Tanta Medical J., 1997,25: 92-111.

17- Ruiz A., Almenar S., Callaghan R.C., Calderon J. andLlombart-Bosch A.: Radial scar versus tubular carci-noma of breast. A comparative study with quantitativetechniques (morphometry, image -and flow cytometry).Pathol. Res Pract., 1995, 191: 547-554.

18- Ruiz A., Almenar S., Callaghan R.C. and Llombart-Bosch A.: Benign, preinvasive and invasive ductalbreast lesions. A comparative study with quantitativetechniques: Morphometry, image-and flow cytometry.Pathol. Res. Pract., 1999, 195: 741-746.

19- Nagashima T., Suzuki M., Yagata H., Hashimoto H.,Shishikura T., Imanaka N. and Nakajima N.: Cyto-morphometric differentiation of intraductal prolifera-tive breast lesions. Breast Cancer, 2000, 7 (1): 43-47.

20- Ellis I.Q., Elston C.W. and Poller D.N.: Ductal carci-noma in situ. In Elston CW and Ellis IO (eds.) Sys-temic pathology 3rd edition. Vol. 13 The breast.Churchill Livingstone, 1998, pp 249-281.

21- Baak J.P., Kurver P.H.J., De Snoo-Niewlaat A.J.E.,De Graaf S., Makkink B. and Boom M.: prognosticindicators in breast cancer. Morphometric methods.Histopathology, 1982, 6: 327-339.

22- Aaltomaa S., Lipponen P., Eskelinen M., Alhava E.and Syrjanen K.: Nuclear morphometry and mitoticindexes as prognostic factors in breast cancer. Eur. J.Surg. 1991, 157: 319-324.

23- Ladekarl M. and Sorensen F.B.: Quantitative histo-pathological variables in-situ and invasive ductal andlobular carcinoma of the breast. APMIS 1993, 101:895-903.

24- Kronqvist P., Kupio T. and Kujari H.: Nuclear mor-phometry in breast cancer: the influence of freezingof samples. Mod. Pathol., 1995, 8: 187-192.

25- Kronqvist P., Kuopio T., Collan Y., Horvath C. andTamm U.: The reproducibility of nuclear morphometricmeasurements in invasive breast carcinoma. Anal.Cell Pathol., 1997, 15 (1): 47-59.

26- Fregene T.A., Khanuja P.S., George J., Kellog C. andGimotty Pienta K.J.: Nuclear morphomety preditsfordisease recurrence in early stage breast cancer. Proc.AACR A 1994, 1354.

27- Domagala W., Markiewski M., Harezga B., DukowiczA. and Osbon M.: Prognostic significance of tumorcell proliferation rate as determined by MIB-1 antibodyin breast carcinoma: its relationship with vimentinand p53 protein. Clin. Cancer Res., 1996, 2: 147-154.


28- Ostrowski M.L., Chakraborty S. and Lucirica R.:Quantitative image analysis of MIB-1 immunoreac-tivity: a comparison with flow cytometric assessmentof proliferative activity in invasive carcinoma of thebreast. Anal. Quant. Cytol. Histol. 1995, 17: 15-24.

29- Masood S., Bui M.M. and Lu L.: Comparison ofproliferation activity in breast carcinoma by flowcytometry analysis of S-phase and quantitative analysisof MIB – 1. Ann. Clin. Lab. Sci. 1998, 28 (6): 315-323.

30- Shen K., Harn H., Ho L., Yu C., Chiu S. and Lee W.:The extent of proliferative and apoptotic activity inintraductal and invasive ductal breast carcinomasdetected by Ki-67 labeling and terminal deoxynucle-otidyl transferase-mediated digoxigenin-11-dUTPnick end labeling. Cancer, 1998, 82: 2373-2381.

31- Weinder N., Moore D.H. and Vartaniant R.: Correlationof Ki-67 antigen expression with mitotic figure indexand tumor grade in breast carcinomas using the novelparaffin-reactive MIB-1 antibody. Hum. Pathol., 1994,25: 337-342.

32- Keshgegian A.A. and Chaan A.: Proliferation markersin breast carcinoma. Mitotic figure count, S-phasefraction, proliferating cell nuclear antigen, Ki-67 andMIB-1. Anatom. Pathol., 1995, 104: 42-49.

33- Veronese S.M., Gambacorta M., Gottardi O., ScanziF., Ferrari M. and Lampertico P.: Proliferation indexas a prognostic marker in breast cancer. Cancer, 1993,71: 3926-3931.

34- Pienta K.J. and Coffey D.S.: Correlation of nuclearmorphometry with progression of breast cancer.Cancer, 1991 68: 2012-2016.

35- Uyterlinde A.M., Baak J.P.A., Schipper N.W., PeterseH., Matze E. and Meijer C.J.L.: Further evaluation ofthe prognostic value of morphometric and flow cyto-metric parameters in breast cancer patients with longfollow-up. Int. J. Cancer, 1990, 45: 1-7.

36- Van Diest P.J. and Baak J.P.A.: The morphometricprognostic index is the strongest prognosticator inpremenopausal lymph node-negative and lymph node-positive breast cancer patients. Human Pathol., 1991,22: 326-330.

37- Fernandez-Lopez F., Paredes-Cotore J.P., Cadarso-Suares C., Forteza-Vila J., Puente-Dominguez J.L.and Potel-Lesquereux J.: Prognostic value of nuclearmorphometry in colorectal cancer. Dis. Colon. Rectum.1999, 42: 386-392.

38- Ikeguchi M., Cai J., Oka S., Gomyou Y., Tsujitani S.,Maeta M. and Kaibara N.: Nuclear profiles of cancercells reveal the metastatic potential of gastric cancer. J. Pathol., 2000, 192 (1): 19-25.

39- Kavantas N., Lazaris A.C., Chatzigianni E. and DavarisP.S.: The nuclear morphometry by image analysis inthe histopathologic diagnosis of lung cancer. J. Exp.Clin. Cancer Res. 2000, 19 (2): 201-206.

40- Giardina C., Renzulli G., Serio G., Caniglia D.M.,Littini T., Ferri C., D'Eredita G., Ricco R. and DelfinoV.P.: Nuclear morphometry in node-negative breastcarcinoma. Anal. Quant. Cytol. Histol., 1996, 18: 374-382.

41- Ladekarl N.: Quantitative histopathology in ductalcarcinoma of the breast: prognostic value of meannuclear size and mitotic count. Cancer, 1995, 75:2114-2122.


Nuclear morphometry in ductal breast carcinoma with correlation to cell proliferative activity and...

Documents

Transcript of Nuclear morphometry in ductal breast carcinoma with correlation to cell proliferative activity and...