España

Background: Ten years ago we reported our first results of DNA quantification by image cytometry in breast cancer. The aim of the present report is to re-evaluate survival of the patients diagnosed at that time and else to compare with a new prospective study following the same methodology.
Methods: for the retrospective study 62 patients out of 135 were selected, all of whom suffered radical mastectomy due to breast cancer, while for the prospective assay we selected 145 patients with known survival and cytometry performed at the same time the cytological diagnosis was achieved. Image cytometry was carried out following the consensus report on DNA quantification standarization.
Results: The results of the quantification are classified in four types of histograms, regarding its grade of malignancy. Also survival analysis is made applying the Kaplan-Meiers´Curve. Patients that displayed high grade of malignancy showed an important decrease in survival, while cases with low grade tumors had higher rates of survival free of disease.
Conclusions: our results prove that DNA image cytometry applied to FNA cytology is a very useful, minimally invasive, and reliable tool to determine the real biological behaviour of breast neoplasm. Thus we strongly encourage surgeons and gynaecologists to use this technique and its inclusion in the breast cancer protocols.

In the year 1992, during the XIXth International Congress of the International Academy of Pathology & 10th World Congress of Academic and Environmental Pathology, held in Madrid, (Spain) we presented an oral communication called “Breast cancer prognosis by static cytometry”(1) in which we introduced our results with such technique over a group of 62 selected patients. This work was later published in an specialized journal.(2) The obtained results were so satisfactory that encouraged us to keep on using this technique, even wondering whether we could influence the patients survival since we could know the malignancy grade before surgery. Since then, the published references are so numerous that seems not adequate to cite each one, anyway w found over 250 related uniquely to the breast. From the exhaustive review of these works it is observed a remarkable evolution since the first works from Auer,(3-6) and other authors(7-20) who aimed to demonstrate the efficacy of the method, until now, that such efficacy has been corroborated and papers point out more specific aspects, describe new techniques or compare immnunohistochemistry and molecular pathology with ploidy (21-29). Through these years we gained a great experience that we wish to transmit in a comparative study between the retrospective assay of 1992 and the prospective one from 1992 until 2003. The final aim is to obtain enough biological parameters to achieve a personalized treatment for each patient.

For the retrospective study we selected 62 patients out of 135 with consequent diagnosis of malignancy by fine needle aspiration cytology (FNAC). All of them suffered radical mastectomy without any kind of adjuvant therapy. The prospective study includes a selection of 145 patients with known survival in whom DNA quantification by image cytometry was performed at the time of the primary diagnosis. The results of these studies were included in the treatment protocol of each patient, trying to modificate the type of surgery and the adjuvant therapy applied. Obviously treatments were individual thus there is not uniformity in this sense.
The possibility of using paraffin-embedded tis-sue from the same patients was evaluated, but our experience and several authors’ reports forced us to choose the study of cytological slides exclusively. In this sense, although such tissue can be processed, the processing methods leads to greater error factors, since tissue undergoes aggressive phys-ical and chemical changes that disturb cell struc-ture. Further, those samples are sections of tis-sue determining that the evaluation will not be made on whole cells and will entail differences from one cell to another. This fact pertains to the DNA distri-bution from cell to cell and to cellular morpho-metry: one is not able to evaluate classic morpho-metric parameters.
Our method was, first, to search for cases with cancer diagnosis by FNAC. Then, each was re-viewed to select those with smear characteristics that allowed quantification. In many cases a smear was sufficient for diagnosis, but the cells lacked the quantity and quality for quantification. Cell over -crowding was a reason for rejecting a smear in many cases. For assessment, cells have to be iso-lated and never overlapping. This rule applies not only to cells of the same lineage but also to other elements that could disturb the cells’ optic density and produce inaccurate results.
In our laboratory we always perform Papanicolaou and May-Grünwald-Giemsa staining over the smears obtained from FNAC. The first ones are fixed in 96% alcohol, while the former are air dried. This difference in fixation (wet-dry) causes a great difference in cell size: the cells are much larger on air-dried slides. For that reason the two stains should not be used together. Once the cases were selected, we chose only those smears fixed in alcohol and stained by Papanicolaou technique.
When the smears are reviewed and adequate ones selected, the microscope cover glasses were re-moved with xylol immersion. Then the smears were decolored in an alcoholic solution of 1% ClH. Final-ly they were restained with progressive hema-toxylin for 30 seconds .
135 cases were considered adequate for quantification, finally 49 out of them were selected for this study since we had information on their progress over the years and had a common clinical characteristic, all of them suffered radical mastectomy. We have considered radical mastectomy a reference factor regarding survival.
To perform quantification we used TEXCAN software (O. Ferrer, La Laguna, Tenerife, Spain), it allows the study of smears stained with progressive hematoxylin. Nuclei were visualized either through high resolution and a 60x wet objective, nucleus by nucleus, or through a 40x objective, which shows a greater number of cells in each view. We also used a black-and-white, charge-coupled device video camera for comparing shades of grey and a person-al computer.
A minimum of 100 study and 25 control cells were assessed. The controls were stroma cells or lymphocytes and in all cases were internal controls, thus undergoing the same physical and chemical processes as did the study cells.
From each case studied we obtained a histogram of the study cells and of the control ones and compared ploidy in both. The histogram yielded the ploidy value, coefficient of variation and mean optical den-sity related to the modal peak of the G0-G1 and G2M phases. From those data we selected three cel-lular phases for study: G0-G1, S and G2M.
The software produces a prognosis index that gives values derived from quantification, such as the 5c exceeding rate and entropy value (E), or dispersion of the histogram.
The software offers other indexes and grades of malignancy, such as the morphometric prognostic index, malignancy index of Bocking (32) and malig-nancy grade (33). They are based on histological parame-ters, such as tumor size, mitotic activity index and existence of affected lymph nodes. If our samples are compared to in vivo FNAC tumor characteristics, the size will change between palpation and the macroscopic tumor portion excised. The number of affected lymph nodes will differ. We did not use MAI. We did not consider the above factors in assessing survival. Our evaluations were derived from the type of histogram.

The software creates different histograms which based upon cellular ploidy are classified in four types previously described (2), considering types I and II as low malignancy grade, and types III and IV as high grade ones. In this new evaluation, that comprises both studies, we have considered type I those with less than 5% of cells under 5c, while type III was considered when the percentage was equal or above 5%. We are aware this classification is quite subjective, nevertheless, we have not found a more objective parameter, neither in our own research nor in other author’s reports, which allows to divide both populations. In this way, we notice there exists a borderline population which is really complicated to include in one group or the other, with the clinical consequences involved. In order to fight against this subjectivity we have reviewed our data, searching for the type I cases none of which displayed a clinical behaviour correspondent to type III.
Retrospective study showed the following distribution of the histogram types (figure 1):
FIGURE 1
As it is shown, low grade tumors represent a 37, 7% and 62, 2% corresponds to high grade ones. The agreement between these grades with its survival was exceptionally good; in the low grade group there wasn’t any death caused by their disease, nor metastases or relapse during the time of study. Looking at the Kaplan Meier Survival Curve (figure 2) for the high grade neoplasms it is clear that survival has decreased up to 88% in the first year, and there is a very intense descent during the following, establishing in 39% in the fifth year, while from the sixth year survival remains still in 31%.
FIGURE 2
In the next figure 3 are shown results obtained in the prospective study:
FIGURE 3
There is a 42, 1% of low grade cases and 57, 9% for the high grade group. Regarding survival, in figure 4 it is seen that during the first two years there isn’t any death in the low grade group, with up to 80% of survival in the fifth year and a total survival of 70% after 12 years. For the high grade population, deaths begin the first year, at the fifth one there is a survival of 68% and just 46% of patients survived after 12 years.

A detailed study of the collected data shows that, regarding malignancy grades, there is a very important increase for the low grade tumors. We believe the 20 years interval between both studies has supposed a very important advance in the diagnosis, detecting tumors in earlier stages, thus, increasing the number of low grade neoplasms. Regarding survival, at the present time we observed, curiously, deaths in the prospective study, while this fact was not present in the retrospective assay (figure 4).
On the other hand, the curve for the high grade tumors displays a very important increase in the survival respect to the one observed in the retrospective study with a difference of up to 41% at 5 years, and 15% at 12 years. Log Rank test shows a high statistical significance for p<0,010 (figure 5).
It becomes evident that along these years the social, economical, scientifically… factors have improved, thus increasing the survival. One cannot separate in an objective way which one has affected in a greater manner, nevertheless it is a fact that despite these improvements, patients with low grade tumors show high mortality rate in the prospective study; thus we believe these results clearly indicate that DNA quantification is an extremely important tool in the management of these cases, helping in the establishment of an adequate therapeutic protocol, more agree with the biological behaviour of the tumor. This is more evident when dealing with tumors that display high grade of malignancy, which showed great benefit by using this technique.
In conclusion, our results prove that DNA image cytometry applied to FNA cytology is a very useful, minimally invasive, and reliable tool to determine the real biological behaviour of breast neoplasm. Thus we strongly encourage surgeons and gynaecologists to use this technique and its inclusion in the breast cancer protocols.

1- Azua J, Romeo P, Martinez-Tello D, Serrano M. Breast cancer prognosis by static cytometry. Patología 1992; 25 (3): 17
2.- Azua J, Romeo P, Serrano M, Martinez-Tello D, Azua J. Prognostic value from DNA quantification by static cytometry in breast cancer. Anal Quant Cytol Histol 1997; 19(1): 80-86
3.-Auer G, Caspersson T, Wallgren A. DNA content and survival in mammary carcinoma. Anal Quant Cytol Histol 1980; 2(3): 161-166
4.-Auer G, Askensten U, Erhardt K, Fallenius A, Zetterberg A. Comparison between slide and flow cytophotometric DNA measurements in breast tumors. Anal Quant Cytol and Histol 1987; 9(2): 138-147
5.- Auer G, Fallenius A, Erhardt K, Sundelin B. Progression off mammary adenocarcinomas as reflected by nuclear DNA content. Cytometry 1984; 5: 420-425
6.- Auer G, Eriksson E, Azavedo E, Caspersson T, Wallgren A. Prognostic significance of nuclear DNA content in mammary adenocarcinomas in humans. Cancer-Res 1984 Jan; 44(1): 394-6
7.- Aubele M, Auer G, Falkmer U, Voss A, Rodenacker K, Jutting U, et al. Identification of a low-risk group of stage I breast cancer patients by cytometrically assessed DNA and nuclear texture parameters. J Pathol 1995 Dec; 177(4): 377-84
8.- Aubele M, Auer G, Falkmer U, Voss A, Rodenacker K, Rutquist L E, et al. Improved prognostication in small (pT1) breast cancers by image cytometry. Breast Cancer Res Treat 1995; 36(1): 83-91
9.- Azua J, Romeo P, Morales M, Castiella T, Azua J: DNA quantification as a prognostic factor in colorectal adenocarcinoma. Anal Quant Cytol Histol 1999 Dec; 21(6): 468-72
10.- Azua J, Romeo P, Morales M, Castiella T, Azua J: DNA quantification as a prognostic factor in gastric adenocarcinoma. Anal Quant Cytol Histol 1998 Jun; 20(3): 221-4
11.- Borgiani L, Cogorno P, Toso F, Gallo L, Buccaran G, Rovida S, et al. Comparative DNA analysis of breast cancer by flow cytometry and image analysis. Pathologica 1994 Aug; 86(4): 356-9
12.- Thunnissen F B, Ellis I O, Jutting U, Burger G. Interlaboratory comparison of DNA image analysis. Management group report of internal quality assurance programme in the CAAC breast cancer project. Zentralbl Pathol 1994 Feb; 139(6): 443-7
13.- Borgiani L, Cogorno P, Buccaran G, Gallo L, Rovida S, Canepa M. Human breast cancer: prognostic significance of DNA ploidy compared with c-erbB2 expression, Cathepsin D and silver-binding nucleolar organizer regions (AgNORs). Pathologica 1994 Aug; 86(4): 350-5
14.- Stal O, Sullivan S, Sun X F, Wingren S, Nordenskjold B. Simultaneous analysis of c-erbB-2 expression and DNA content in breast cancer using flow cytometry. Cytometry 1994 Jun 1; 16(2): 160-8
15.- Schulte E K, Fink D K. Hematoxylin staining in quantitative DNA cytometry: an image analysis study. Anal Cell Pathol 1995 Dec; 9(4): 257-68
16.- Ottesen G L, Christensen I J, Larsen J K, Hansen B, Andersen A J. Flow cytometric DNA analysis of breast cancers with predominance of carcinoma in situ: a comparison of the premalignant and malignant components. Clin Cancer Res 1995 Aug; 1(8): 881-8
17.- Baldetorp B, Bendahl P-O, Ferno M, Alanen K, Delle U, Falkmer U, et al. Reproducibility in DNA flow cytometric analysis of breast cancer: comparison of 12 laboratories' results for 67 sample homogenates. Cytometry 1995 Jun 15; 22(2): 115-27
18.- Susnik B, Poulin N, Phillips D, LeRiche J, Palcic B. Comparison of DNA measurement performed by flow and image cytometry of embedded breast tissue sections. Anal Quant Cytol Histol 1995 Jun; 17(3): 163-71
19.- Schulte E K, Fink D K. Hematoxylin staining in quantitative DNA cytometry: an image analysis study. Anal Cell Pathol 1995 Dec; 9(4): 257-68
20.- Williams R A, Baak J P, Meijer G A, Charlton I G. Testing the value of a measurement protocol for DNA image cytometry and comparing different cytologic preparations. Anal Quant Cytol Histol 1996 Dec; 18(6): 429-37
21.-Azua Romeo J, de Azua Blanco J, Romeo P. Image cytometry in nonproliferative fibrocystic breast changes. Ploidy evaluation and epidemiologic study. Anal Quant Cytol Histol 2001 Apr; 23(2): 129-34
22.-Martinez Arribas F, Nunez M J, Piqueras V, Lucas A R, Sanchez J, Tejerina A, et al. Flow cytometry vs Ki67 labelling index in breast cancer: a prospective evaluation of 181 cases. Anticancer Res 2002 Jan-Feb; 22(1A): 295-8
23.-Chiusa L, Margaria E, Pich A. Nuclear morphometry in male breast carcinoma: association with cell proliferative activity, oncogene expression, DNA content and prognosis. Int J Cancer 2000 Nov 20; 89(6): 494-9
24.- Fattorossi A, Battaglia A, Maggiano N, Malinconico P, Andreocci L, Mancuso S. Laser scanning cytometry can complement the flow cytometric DNA analysis in paraffin- embedded cancer samples: a paradigmatic case. Cytometry 2000 Apr 15; 42(2): 123-5
25.-Gorisse M C, Venteo L, Pluot M. A method for simultaneous quantification of monoclonal antibody Ki-67 and DNA content by flow cytometry. Application to breast carcinomas. Anal Quant Cytol Histol 1999 Feb; 21(1): 8-16
26.-Zhang Y, LeRiche J C, Jackson S M, Garner D, Palcic B: An automated image cytometry system for monitoring DNA ploidy and other cell features of radiotherapy and chemotherapy patients. Radiat Med 1999 Jan Feb; 17(1): 47-57
27.-Bloyet D, Herlin P, Masson E, Plancoulaine B, Duigou F, Angot F, et al. Image analysis software for automatic DNA ploidy assessment of archival solid tumours. Cytometry 1999 Dec 1; 37(4): 267-74
28.-Stomper P C, DeBloom J R, Budnick R M, Stewart CC. Flow cytometric DNA analyses of benign breast lesions detected by screening mammography. Clin Cancer Res 1998 Jun; 4(6): 1543-7
29.-Masood S, Bui M M, Lu L.Comparison of proliferation activity in breast carcinoma by flow cytometry analysis of S-phase and quantitative analysis of MIB-1. Ann Clin Lab Sci 1998 Nov-Dec; 28(6): 315-23
31.-Gill G W, Frost J K, Miller K A: A new formula for a half-oxidized hematoxylin solution that neither overstains nor requires differentiation. Acta Cytol 1974; 18(4): 300-311
32.- Böcking A, Adler C, Commom H, Hilgarth M, Granzen B: DNA-Cytophotometric diagnosis and grading of malignancy. Anal Quant Cytol Histol 1984; 6(1):1-8
33.- Böcking A, Chatelain R, Biesterfeld S et al: DNA grading of malignancy in breast cancer. Prognostic validity, reproducibility and comparison with other classifications. Anal Quant Cytol Histol 1989; 11(2):73-80