DNA index of ovarian carcinomas from 56 patients: in vivo in vitro studies

Out of 130 ovarian cancer patients the DNA index of cells from ovarian carcinoma was studied in 56 cases in which cytospin preparations showed the presence of atypical cells. In 24 patients the population had a diploid DNA index (1.0) and in the others the DNA index ranged from 1.2 to 2.0 (tetraploid). No hypodiploid or hypertetraploid populations were detected. Repeated samples from the same patients did not show any significant differences and primary culture did not alter the DNA index. In contrast, cell cycle phase distribution differed greatly from sample to sample, as also the ratio between DNA diploid and DNA aneuploid populations. Primary culture was successful in 57% of the tumours, with a higher percentage of success in DNA aneuploid tumours. After primary culture the ratio between DNA aneuploid cells and DNA diploid cells increased. In relation to the histological gradings of malignancy, DNA aneuploid cells clustered in the highest grade of malignancy. The mean S-phase for tumours with a DNA index of 1.0 was 3.5 and 14.1% for those with DNA index greater than 1. Ovarian carcinomas show a large difference in DNA index between patients even after primary culture.     

Prognostic importance of cellular DNA content in head-and-neck squamous-cell cancers. A comparison of retrospective and prospective series

Flow cytometric DNA-ploidy measurements were performed on formalin-fixed tumour specimens from 172 patients with squamous-cell cancers (SCCs) of the head and neck region. One hundred and two samples were chosen retrospectively and a further 70 consecutive patients were analysed prospectively in order to assess the prognostic significance of DNA ploidy and DNA index (DI). There were no statistically significant differences between retrospective and prospective groups in regard to age, sex, TNM stage, ploidy or DI. Sixty-seven percent of patients were aneuploid (65% retrospective; 71% prospective). The proportion of aneuploid tumours was significantly higher among poorly differentiated tumours. Survival analysis using Cox multivariate regression modelling revealed that DNA aneuploidy and increasing DI were significant independent prognostic factors for both relapse-free and overall survival. The relapse and death rates among aneuploid subjects were approximately 3 times as high as those for diploid subjects. Patients with a DI greater than 2.11 (hypertetraploidy) experienced a 6.6-fold higher death rate than diploid subjects. These results provide strong support for the incorporation of DNA ploidy profiles into the clinical management of patients with head and neck cancer.     

Carcinoma of the anal canal and flow cytometric DNA analysis

Using flow cytometric DNA analysis of paraffin embedded tissue, DNA histograms were successfully obtained from the anal cancers of 117 patients. DNA diploid patterns were given by 82 cancers (70%) and DNA non-diploid patterns by 35 cancers (30%): 15 DNA aneuploid, 20 DNA tetraploid. Well differentiated squamous cell cancers were mainly DNA diploid, while a larger proportion of poorly differentiated and small cell cancers were DNA non-diploid. The large majority of stage A cancers were DNA diploid. A greater proportion of tumours that had invaded through the anal sphincter or had lymph node metastases or distant spread were DNA non-diploid. Prognosis was slightly poorer for patients with DNA non-diploid cancers when compared to patients with DNA diploid tumours (P = 0.08) and significantly poorer for individuals with DNA aneuploid anal cancers (P = 0.037). However, in a multivariate analysis model, the DNA ploidy pattern of an anal cancer was not of independent prognostic significance alongside tumour histology and tumour stage.     

DNA aneuploidy and low S-phase fraction as favourable prognostic signs in metastatic melanoma

The prognostic value of cellular DNA content in melanoma metastases was investigated by flow cytometric analysis of fresh or paraffin-embedded tumour blocks from 95 consecutive patients referred to the Helsinki University Central Hospital Melanoma Team. Thirty-three per cent of the tumours were DNA diploid and 67% DNA aneuploid. S-phase fractions were lower in DNA diploid than in DNA aneuploid tumours (10.7% and 17.6%). Tumour ploidy and S-phase fraction were shown by multivariate Cox model analysis to be independent prognostic variables and major determinants of survival after first recurrence. Surprisingly, patients with DNA aneuploid tumours and with tumours with low SPF survived significantly longer than those with DNA diploid or high SPF tumours. This exceptional finding of favourable prognosis for DNA aneuploid tumours was more prominent among patients receiving intensive systemic therapy and among patients with stage IV disease, probably indicating a tendency for DNA aneuploid tumours to have higher sensitivity to systemic therapy.     

The influence of tumour cell DNA content on survival in colorectal cancer: a detailed analysis

We have investigated the influence of tumour cell DNA content (ploidy) on survival of 416 patients undergoing excisional surgery for colorectal cancer. Two hundred and eleven (51%) tumours had an abnormal DNA content (aneuploid or tetraploid). There was no correlation between ploidy status, sex, age and pathological stage, histological grade, tumour site, local tumour extension or assessment of curability. Patients with tumours with an abnormal DNA content had a poorer survival 68/211 (32%) than patients with near normal (diploid) DNA content 88/205 (43%) (test statistic 5.0, P = 0.02). The patient subgroups in which DNA content exerted an influence on survival were: stage B tumours (P = 0.0058), moderately differentiated tumours (P = 0.004), rectal tumours (P = 0.02), and mobile tumours (P = 0.02). Multivariant analysis showed that pathological stage, local tumour extension and DNA ploidy were all independent prognostic indicators whereas histological grade, tumour site and assessment of 'curability' were not. The influence of pathological stage, however, was much greater than that of local tumor extension or DNA ploidy. Tumour cell DNA content together with pathological stage and local tumour extension may be used in a prognostic index and may be important in planning adjuvant therapy.     

Hypodiploidy, Ki-67 growth fraction and prognosis of surgically resected lung cancers.

One hundred and thirty-seven lung cancer patients (123 non-small-cell lung cancers (NSCLC), 10 small-cell lung cancers (SCLC) and four carcinoid tumours) who underwent surgery in an attempt at complete resection were prospectively entered in a study whose aim was to determine the prognostic significance of a hypodiploidy or a multiploidy pattern of tumour cell DNA content and a high immunohistochemical reactivity of Ki-67, a nuclear antigen related to the cell cycle. Indirect immunoperoxidase reactivity of Ki-67 on frozen tumour tissue sections was evaluated both visually, using a classical semiquantitative scale, and by means of a computer-assisted image processor. Cell DNA content analysis was done using static computer-assisted cytometry on tumour cytological prints stained by the pararosaline Feulgen-Schiff technique. The ploidy was characterised for each tumour by DNA index (DI), percentage of hypodiploid cells and type of DNA content histogram (near diploid, hyperdiploid, hypodiploid and multiploid). Ki-67 immunostaining was negative in 64 tumours (48%) and positive in 69 (52%). DNA histogram classification disclosed 57 (42%) near diploid tumours. Among the 80 (58%) aneuploid tumours, 16 were hypodiploid, 44 hyperdiploid and 20 multiploid. The prevalence of both a positive Ki-67 immunostaining and an aneuploid DNA histogram differed according to histology as SCLC demonstrated a higher frequency of both features when compared with NSCLC and carcinoid tumours. On the other hand, Ki-67 immunostaining and ploidy did not significantly differ according to degree of differentiation, nodal status and Mountain''s stage grouping. The percentage of cells in the hypodiploid modal DNA was significantly higher for tumours which demonstrated a high Ki-67 immunostaining, suggesting a link between growth fraction and DNA content abnormalities. In univariate analysis, survival did not differ significantly according to either the Ki-67 immunohistochemical reactivity or the DNA index. Patients with a hypodiploid tumour had a shorter survival than patients with other DNA histogram patterns but, owing to the low frequency of hypodiploidy, this difference did not reach statistical significance. In Cox''s proportional hazard model, an SCLC histology, an advanced tumour status, a positive nodal status and a hypodiploid tumour (hazard ratio: 2.070; 95% confidence interval 1.041-4.116) were significant determinants of survival. We conclude that hypodiploidy in lung cancer is a distinct DNA content abnormality as it contributes significantly to prognosis. Neither visually assessed nor computer-generated Ki-67 immunostaining measurements significantly determine prognosis.     

Molecular genetic analysis of flow-sorted ovarian tumour cells: improved detection of loss of heterozygosity.

Detection of loss of heterozygosity (LOH) is usually performed on homogenised tumour specimens. In this type of analysis samples with a low percentage of tumour cells have to be excluded and possible intra-tumour heterogeneity is obscured. In this study we report the application of polymerase chain reaction (PCR)-driven LOH detection with in total 22 microsatellite markers for chromosome 1q, 3p, 3q, 4p, 6p, 6q, 11p, 11q, 17p, 17q, 18p, 18q, Xp and Xq on flow-sorted cells from fresh and paraffin-embedded ovarian tumour tissue. Titration experiments showed that LOH can be detected with as few as 100 cell equivalents of DNA. Clear examples of LOH could be detected in the sorted aneuploid fractions from one unilateral and two bilateral ovarian tumours from three patients. In two samples the sorted fraction was less than 10% of the total sample. The bilateral tumours from the same patient showed loss of identical alleles for one marker (case OV64) and two markers (case OV69), indicative of their monoclonal origin. Multiparameter flow cytometry using two different ovarian tumour markers (MOv18 and BMA180), an anti-cytokeratin monoclonal antibody (MAb) (M9), an anti-vimentin MAb (V9) and a MAb against the panepithelial antigen 17-1A on the fresh ascites cells of the fourth ovarian cancer patient was used to investigate possible intra-tumour heterogeneity. We showed the presence of at least three phenotypically different populations, of which the diploid, keratin-positive, vimentin-negative population showed a similar LOH pattern as the aneuploid population (DNA index = 1.7), indicative of its neoplastic origin. The same LOH pattern was shown in an omentum metastasis from this patient also having the same aneuploid DNA index of 1.7. The sharing of the same LOH pattern by the diploid and aneuploid tumour cell populations suggests that the observed allele loss events occurred before the development of aneuploidy. PCR on flow-sorted cells is thus an important tool to study clonal diversity in tumours.     

Predicting outcome for patients with node negative breast cancer: a comparative study of the value of flow cytometry and cell image analysis for determination of DNA ploidy.

This study was aimed at determining whether tumour DNA content measured by cell image analysis could provide additional prognostic information when compared to that provided by flow cytometry. Sections cut from paraffin blocks of tumours from 101 patients with node negative breast cancer were analysed by both methods and the results related to other prognostic variables and to patient relapse and overall survival. DNA ploidy measured by flow cytometry classified 46 tumours as diploid and 55 as aneuploid, whereas by cell image analysis 30 were diploid and 71 aneuploid (P less than 0.002). There were 20 tumours with discrepancies between the two methods; 18 of these were tumours with only one peak in flow analysis, but determined to be aneuploid with image analysis. DNA content as measured by both methods was significant for predicting relapse and survival by log-rank test, as were tumour histological grade, c-erbB-2 expression and tumour size. Multivariate analysis showed DNA ploidy measured by flow cytometry to be the only variable of independent significance (P less than 0.02) for both relapse and overall survival. Compared with cell image analysis, flow cytometry demonstrated a significantly higher proportion of diploid tumours, which may be related to differences in the internal standards applied to each method. We suggest that cell image analysis techniques can provide more sensitive information on the DNA content of tumour cells by direct measurement of nuclear DNA density of both normal lymphocytes and tumour cells in the same section. However, although image analysis appears to be more sensitive than flow cytometry in detecting DNA aneuploidy, the image technique appears to lack the specificity of flow cytometry in correlation with clinical outcome.     

Correlation of DNA ploidy, histopathology, stage and clinical outcome in gastric carcinoma.

The determination of nuclear DNA ploidy from paraffin-embedded specimens was performed by flow cytophotometry on 277 surgically resected primary gastric carcinomas to assess the relationship of various pathological findings and DNA content with survival. The preparation of samples was performed by a modification of Hedley's technique and the staining method of Vindelov. Eighty-nine (32%) carcinomas were DNA diploid, 69 (25%) were DNA tetraploid, and 119 (43%) were DNA aneuploid. DNA non-diploid patterns were significantly associated with macroscopic ulcerative appearance, location of the tumour in the proximal stomach, histological grade, and advanced stage of tumour. Patients with DNA non-diploid cancers, and specifically DNA aneuploid cancers, exhibited significantly poorer survival than patients with DNA diploid tumours. These data support the prognostic value of tumour DNA content in patients with resected gastric carcinoma.     

Flow cytofluorometric evidence for the differential radioresponsiveness of aneuploid and diploid cervix tumours

We are presently involved in a prospective study of the relationship between DNA content profiles, and their changes during treatment, determined by flow cytofluorometry, and patient prognosis and response to therapy for cancer of the uterine cervix. To date, 348 patients have been included in the study over a 54-month period. Data on these patients have shown that DNA aneuploid tumours are significantly more radioresponsive than diploid cervix tumours. Analysis of the data on 213 patients with a minimum follow-up time of 15 months has, however, failed to show an overall more favourable prognosis conferred by tumour DNA aneuploidy. Analysis of the relationship between clinical stage and disease state and tumour DNA ploidy, however, suggests that aneuploid tumours metastasize to distant sites at an earlier stage of the disease than diploid tumours and local recurrence rates for diploid tumours, in late stage disease, are double those for aneuploid tumours. Improved staining procedures, and instrument modification, has also shown that cervix tumour heterogeneity is of considerably greater frequency than at first appeared to be the case (approximately 75% of DNA aneuploid tumours show heterogeneity.     

Tumour growth rate and DNA flow cytometry parameters as prognostic factors in metastatic melanoma.

The prognostic value of flow cytometric parameters and tumour growth rate of melanoma metastases under the mouse renal capsule was investigated for tumours from 117 consecutive patients referred to the Helsinki University Central Hospital Melanoma Team. DNA flow cytometry (FCM) was interpretable for the tumours of 114 patients, and growth rate analysis for 82 patients, both results being available from 79 patients. Thirty-six percent of the tumours were DNA diploid and 64% DNA aneuploid. Tumour ploidy and S-phase fraction were shown by multivariate Cox model analysis to be independent prognostic variables and major determinants of survival after first recurrence. Patients with DNA diploid or aneuploid tumours survived a median 16 and 27 months, respectively. A high growth rate of tumour sample in vivo under the mouse renal capsule tended to be a sign of poor prognosis, although not reaching statistical significance. Combining the results of FCM, tumour growth rate and TNM stage, we propose a highly efficient prognostic scoring method. Patients with a score above 0.75 had a median survival of 11 months compared to 30 months among patients scoring under 0.75 (P less than 0.0001). This score was the most significant (P less than 0.0001) prognostic factor in the Cox model when TNM stage, age, ploidy, SPF, and tumour growth rate were analysed as covariates.     

Prognostic value of ploidy of primary tumour and nodal secondaries in colorectal cancers

Tumour ploidy is of prognostic value in colorectal cancer, DNA aneuploid tumours having a worse outlook. Nearly all studies have concentrated on the DNA content of the primary tumour. We have examined the ploidy of the primary tumour and its lymph node metastases in 71 cases of Dukes' stage C disease, to see whether this provides greater prognostic information than the primary alone. Analysis was performed using formalin-fixed, paraffin-embedded tumour sections. Ploidy of primary and metastases was different in 20 cases (28%), aneuploid nodes being seen with diploid primaries and vice versa. Ploidy of both the primary (χ² = 4.86, P = 0.03) and secondary (χ² = 4.86, P = 0.03) tumours predicted survival in univariate analysis. Combining the ploidy of primary and nodes, three prognostic groups could be defined — diploid primaries with diploid metastases (hazard relative to both aneuploid, 0.36) had significantly better survival than cases where the ploidy of the primary and nodes were mixed (relative hazard 0.47–0.56), which did better than cases with aneuploid primary and nodes. This study demonstrates that ploidy variation between primary and secondary tumours is common, and better prognostic information may be gained by studying both.     

Selection of tumour cell subpopulations occurs during cultivation of human tumours in soft agar. A DNA flow cytometric study

To examine whether selection of tumour cell subpopulations occurs during cultivation in soft agar, we compared in 23 human tumours of different histological types the DNA content of cells from colonies formed in soft agar (method of Courtenay and Mills, 1978) with that of the original tumour cells. The ploidy as well as the fraction of cells in S phase were determined from DNA histograms after staining of the nuclei with a propidium-iodide procedure and flow cytometric recordings. In 8 of 17 aneuploid tumours analysed, specific aneuploid subpopulations disappeared during cultivation or new aneuploid populations, not demonstrable in the original cell suspensions, appeared in the colonies. In 9 cases identical aneuploid populations were found in the colonies and the tumours. In one of 6 diploid tumours examined, aneuploid cell populations not revealed in the original cell suspension, were found in addition to diploid cells, whereas 5 tumours gave rise to colonies containing a purely diploid population. The results show that in a variety of human malignant tumours cultivation in soft agar may select specific aneuploid tumour cell populations.     

An immunohistochemical assessment of cellular proliferation markers in head and neck squamous cell cancers

Prognostic information is essential for the evaluation, judgement and optimal treatment of patients with squamous cell cancers (SCCs) of the upper aerodigestive tract. Using immunohistochemical and flow cytometric techniques, we have studied the significance of cellular expression of the Ki-67 antigen, epidermal growth factor receptor (EGFR), the transferrin receptor (TFR) and DNA ploidy status in a prospective analysis of patients with SCCs of the head and neck region. All 42 fresh tumour samples (five well differentiated; 28 moderately differentiated; nine poorly differentiated) expressed both EGFR and TFR to varying degrees. Receptor expression was most marked on the peripheral invading margin of cancer cell islands although staining was also demonstrated in a random fashion within cellular islands and consistently along the basal cell layer of overlying stratified squamous epithelium. The percentage of cancer cells that reacted with the Ki-67 monoclonal antibody was assessed as low (less than 10%) in 15 samples (35.8%), intermediate (10-30%) in 19 samples (45.2%) and high (greater than 30%) in eight samples (19.0%). Eleven of 15 samples (73%) with a low percentage reactivity were DNA diploid, whereas seven of eight samples (87.5%) with a high percentage reactivity were DNA aneuploid. Poorly differentiated SCCs were significantly more often aneuploid than were either moderately or well differentiated tumours. Our results suggest that EGFR and TFR are widely distributed on SCCs, especially on proliferating cells at the invading tumour margin. In addition, there is a close spatial correlation between cells expressing EGFR, TFR and those expressing the Ki-67 antigen. Tumours in which the staining intensity for both EGFR and TFR was intense invariably expressed the Ki-67 antigen in a high proportion of cells. Further patient follow-up will be important in determining whether intense EGFR and TFR staining, combined with a high percentage reactivity with Ki-67 antibody and DNA aneuploidy, will ultimately define a subset of head and neck cancer patients with a poor clinical outcome.     

DNA analysis by flow cytometry, response to endocrine treatment and prognosis in advanced carcinoma of the breast

The relationship between DNA content of mammary cancer and subsequent response to endocrine therapy was studied in 136 patients with advanced disease. All were treated with tamoxifen or ovarian ablation as first-line systemic therapy after relapse and were evaluable for response according to UICC criteria. DNA characterisation by flow cytometry was used on formalin fixed paraffin-embedded samples of tumour. Tumours were grouped according to DNA index into diploid (n = 52, 38%), 'tetraploid' (n = 46, 34%) and 'other DNA-aneuploid' (n = 38, 28%). The highest proportion of oestrogen receptor positive tumours (ER + ve) was found in the 'tetraploid' tumours (38/46, 85%, Chi-square = 8.53, P less than 0.02), and response rates, (SD + PR + CR), were 26/52 (50%), 34/46 (74%), and 15/38 (39%) respectively (Chi-square = 10.88, P less than 0.005). Patients with diploid or 'tetraploid' tumours survived longer and stayed in remission longer than those with 'other DNA-aneuploid' tumours. We suggest that 'tetraploid' or 'near tetraploid' human mammary tumours may comprise a distinct group of endocrine responsive tumours within the overall group of aneuploid tumours. The conventional interpretation of DNA histograms, grouping into diploid and aneuploid, may be masking important features of some tumour groups.     

Cell DNA content--correlation with clonogenicity in the human tumour cloning system (HTCS)

Thirty-six ovarian and renal-cell tumours were analyzed by flow cytometry (FCM) for DNA content and in parallel were assayed for colony formation in a human tumour cloning system (HTCS). While 15/19 (79%) tumours with an abnormal (aneuploid) DNA stemline formed colonies in the HTCS, only 2/17 (12%) of diploid tumours formed colonies. All samples contained tumour cells as assessed by routine cytological examination. The capacity to form colonies in the HTCS was not correlated in these tumours with grade or stage of disease or tumour type. The level of aneuploidy expressed as the FCM DNA index did not correlate with the cloning efficiency in HTCS. These findings suggest that tumour growth in the HTCS reflects a biologically important potential, related in at least some tumours to an abnormal DNA stemline.     

Proliferative behaviour of an oestrogen sensitive rat mammary tumour: evidence for a paracrine interaction between tumour and stroma.

An oestrogen-sensitive rat mammary tumour (OES HR1) has been grown in normal female rats and in female and male rats supplemented with oestrone. In some rats, after the tumour was established, both exogenous and endogenous sources of oestrogen were removed--a treatment which inhibited further growth of the tumour. The proliferative characteristics of the tumours were measured by injecting the rats with deoxybromouridine (BrdU) 4 h before removing the tumour. Extracted nuclei were reacted with anti-BrdU and the labelling index and DNA content measured by flow cytometry. A correlation between the number of (diploid) host cells present and the number of (aneuploid) tumour cells in S-phase of the cell cycle was observed. This result suggests that there are paracrine interactions between tumour and host cells. We also observed that, on oestrogen ablation, the labelling index was significantly reduced while the percentage of cells in S-phase changed far less. The demonstration that there are cells in S-phase which are not proliferating highlights a possible problem with the measurement of proliferation in human tumours from a DNA histogram.     

Heterogeneity in renal cell carcinoma and its impact no prognosis--a flow cytometric study.

In the process of tumour progression genetic instability is the basis for the evolution of tumour cell clones with various genotypic and phenotypic characteristics causing heterogeneity. Renal cell carcinoma has a long prediagnostic growth period, which increases the probability of clonal evolution. We have studied 200 consecutive renal cell carcinomas, addressing the interrelationship between intratumour heterogeneity and clinicopathological factors. DNA ploidy patterns were analysed in multiple samples from each tumour using flow cytometry and compared with clinical stage, tumour invasion, metastatic rate and survival. Eighty-five of 192 evaluable tumours (44%) were homogeneous concerning DNA ploidy (62% diploid, 38% aneuploid). Among 107 heterogeneous tumours a majority (79%) contained aneuploid as well as diploid cell clones. Homogeneously diploid tumours had a lower incidence of local tumour spread compared with tumours with aneuploid cell clones (P < or = 0.001), but the frequency of distant metastasis at time of diagnosis was similar. The presence of aneuploidy in at least one sample from a tumour was a significant adverse prognostic factor (P < 0.001), whereas the degree of heterogeneity had no influence on survival. The frequent heterogeneity demonstrated indicates that multiple samples must be investigated to evaluate properly the malignant character of renal cell carcinoma.     

The prognostic significance of DNA flow cytometry in breast cancer: results from 881 patients treated in a single centre.

In this single-centre study of 881 patients, S-phase fraction (SPF) was shown to be a significant prognostic marker in terms of overall survival (OS), relapse-free survival (RFS) and survival after relapse (SAR). Further, SPF had independent prognostic significance when considering a range of other clinicopathological variables, namely tumour grade and stage, nodal status, patient age, tumour size, menstrual status and treatment details. For OS and RFS, SPF was the second strongest predictor of the clinical course of the disease after nodal status, and for SAR it was the strongest prognostic marker. SPF correlated positively with histological grade but was the stronger predictor of survival. The distribution of SPF values was markedly different for the two ploidy classes of tumour, with DNA aneuploid tumours having a significantly higher average SPF. However, SPF retained its independent prognostic ability when DNA diploid and aneuploid tumours were analysed separately, DNA ploidy itself also proved to be an independent prognostic marker but the survival difference between the two ploidy classes was much less than that seen for different levels of SPF. Tumours with several DNA aneuploid populations (multiploid tumours) tended to have a worse prognosis than other aneuploid tumours but this trend did not reach statistical significance. In this and other studies from this centre, SPF has proved to be a robust predictor of clinical outcome in carcinoma of the breast.     

Lack of prognostic significance of DNA ploidy and S phase fraction in breast cancer.

DNA Ploidy and S phase fraction (SPF) were measured in Stage I and II breast cancers from patients with at least 8 years of follow-up, to assess the prognostic significance of these data. Disaggregated sections of formalin-fixed, paraffin-embedded tumour were analysed by flow cytometry. SPF was calculated using a rectangular model of S phase, after subtraction of background debris using an exponential model. 64% of tumours were DNA aneuploid. The median SPF was 4.5% for DNA diploid, and 10.9% for DNA aneuploid tumours. There were small reductions in survival at 10 years for DNA aneuploid tumours, and for tumours with above median SPF, but these were not statistically significant. The relative hazard for DNA aneuploid tumours was 1.20 (95% CI 0.81-1.76), and for high SPF was 1.31 (95% CI 0.87-1.98). Neither factor was statistically correlated with survival in multivariate analysis. Technical and theoretical factors limit the accuracy and reproducibility of flow cytometric data, and may explain the lack of prognostic information given.