Znt. J. Cancer: 58,781-786 (1994) Publication of the International Union Against Cancer Publication de I Union lnternationale Contre le Cancer

0 1994 Wiley-Liss, Inc.

c-myc ONCOGENE AMPLIFICATION AND CYTOMETRIC DNA PLOIDY PATTERN AS PROGNOSTIC FACTORS IN MUSCULOSKELETAL NEOPLASMS Carlos BARRIOS', Javier S. CASTRESANA', Ursula G. FALKMER2s4, Ingvar ROsENDAHL3 and Andris KREICBERGS'

'Department of Orthopedics, 2Department of Oncology and Pathology, and 30ncologic Center, Karolinska Institute and Hospital, Stockholm, Sweden.

The relationship between c-myc oncogene amplification in neoplastic cells as determined by means of Southern-blot analysis, and their nuclear DNA content as assessed by com- bined flow and image cytometry, was investigated in fresh tumor specimens from 33 patients with musculoskeletal neo- plasms. Amplification, without rearrangement of the c-myc proto-oncogene, was detected in 4 out of 7 bone sarcomas and in 6 out of 26 soft-tissue sarcomas, but in none of 3 benign giant-cell bone tumors. Among the 10 cases with c-myc amplification, 2 were found to be cytometrically DNA diploid, 2 DNA tetraploid, and 6 DNA aneuploid. Conversely, there were I0 tumors displaying extremely aneuploid DNA patterns with- out c-myc oncogene amplification. Thus, there was no relation- ship between c-myc amplification and DNA ploidy; neither did the percentage of S-phase cells, as determined by means of image cytometry. correlate significantly with the occurrence of c-myc amplification. A surprising sex-bias was observed; all 6 cases of c-myc-amplified soft-tissue sarcomas occurred in fe- males, whereas none of the l l males with such sarcomas showed this amplification. When the clinical follow-up data of the patients were scrutinized, it was found that the DNA ploidy pattern of the neoplastic cell nuclei, in combination with the S-phase values, as well as the occurrence of c-myc amplification, yielded prognostic information, being statistically significant 2 years after the diagnosis. o 1994 Wiley-Liss, Inc.

The c-myc oncogene is thought to play a crucial cell- biological role, not only in neoplastic transformation, but also in cell proliferation under normal conditions. The gene has been reported to encode a nuclear DNA-binding protein that regulates the transition of cells from a quiescent state into the cell cycle (Cole, 1986). These facts also apply to conditions in the human being.

Amplification with enhanced expression of the c-myc gene has been detected in certain human malignant tumors (Bro- deur et al., 1984; Bishop, 1987). Increased numbers of c-myc transcripts have been reported to be associated with highly aggressive malignant neoplasms (Schwab et al., 1984; Field and Spardidos, 1990). In musculoskeletal sarcomas, however, data on oncogene involvement are still limited. Recent works suggest that c-myc alterations may be implicated in these tumors (Ikeda et al., 1989; Ladanyi et al., 1993). To date, the clinical significance of such changes remains unknown.

Occasional cytometric DNA studies of musculoskeletal tumors indicate that the DNA ploidy pattern of the neoplastic cell nuclei offers valid prognostic information (Kreicbergs et al., 1987; Bauer et al., 1988). Whether the combination of DNA ploidy data and those of c-myc amplification offers a powerful prognostic tool, supplementary to the fundamental clinical and histopathological data in musculoskeletal sarcomas, is still not known. The relationship between the DNA ploidy pattern and N-myc amplification has earlier been investigated in neuroblas- tomas (Oppedal et al., 1989; Look et al., 1991). The combina- tion of both these factors seems to be of high prognostic value, and it has been claimed that even well-established clinical features, including age and stage, may-in comparison-lose their prognostic significance.

In one of our laboratories, we have recently studied the sporadic occurrence of c-myc amplification in musculoskeletal

tumors (Barrios et al., 1993, 1994). Against the background given above, we thought it worth-while to try to analyze the c-myc amplification data, as well as the cytometric DNA ploidy pattern in these tumors, and to assess their prognostic value for the individual patient.

MATERIAL AND METHODS Clinical and histopathological data

Fresh specimens from 36 patients with musculoskeletal tumors treated by surgery at the Karolinska Hospital, 1986- 1988, were analyzed for c-myc gene amplification and their cytometric nuclear DNA content. No patient had received preoperative radiation or chemotherapy. Clinical and histo- pathological data are summarized in Table I. The classification used is the conventional one applied by the pathologists at our hospital (Enzinger and Weiss, 1988). In every patient, the ultimate histopathological diagnosis was made at clinico- pathological conferences where the growth pattern and cytodi- agnostic details were evaluated by expert pathologists against the background of clinical and X-ray diagnostic data. The clinical follow-up information for each patient was obtained in the registry of the Department of Orthopedics and the Regional Cancer Registry of the Stockholm County Council. The closing date for follow-up was November 30, 1993. The mean follow-up was 4 years and 6 months. The patients' survival time was calculated from the date of diagnosis to that of death if due to neoplastic disease. Patients who died without a reported recurrence were excluded, as well as patients dying from other major diseases.

DNA extraction, blotting and hybridization Immediately after surgical removal, small tumor samples for

oncogene analysis were frozen in liquid nitrogen and stored at -70°C until extraction. High-molecular-weight DNA was iso- lated by the phenol-chloroform method as described in a standard handbook (Sambrook et al., 1989). The DNA was digested with the restriction endonucleases EcoRI and Hin- dIII, electrophoresed in 1% agarose gels, and transferred to nylon membrane (Zeta-Probe, Bio-Rad, Richmond, CA) by the conventional Southern method. Filters were hybridized with a DNA-specific probe (a 1.2-kb Pst I fragment excised from a pBR322 clone, Ryc 7.4) covering the third exon of the human c-myc oncogene, kindly provided by Dr. C.M. Croce, Philadelphia, PA (Watt et al., 1983). Variations in DNA amount per lane were adjusted by standardizing the c-myc gene signal to a c-myb probe signal (Barrios et al., 1993~). A hybridization signal was considered to be amplified if it showed at least 3 times the intensity produced by an equal amount of DNA from cultured human fibroblasts (Hollstein et al., 1988). The intensity of the bands was assessed by scanning densitom- etry (Mod. 1650, Bio-Rad).

4T0 whom correspondence and reprint requests should be sent, at the Department of Oncology and Pathology, Karolinska Hospital, S-171 76 Stockholm, Sweden. Fax: 46-8-331696.

Received: February 8,1994 and in revised form May 13,1994.

782 BARRIOS ETAL.

Cytometric DNA assessments Flow cytometiy (FCM). The preparative procedures, includ-

ing staining and internal standardization, were previously described in detail (Vindelov et al., 1983). Briefly, cell suspen- sions were obtained from the freshly excised tumor specimens by the fine-needle aspiration technique. In most cases, 2 samples were prepared. The FCM assessments were per- formed with a cytofluorometer (Ahrens, Hamburg, Germany) equipped with a 100-W Hg lamp. Excitation and emission filters were 530-560 nm and > 580 nm, respectively. For each histogram, at least 20,000 counts were recorded.

Image cytometiy (ICM). Simultaneously with the preparation of the FCM samples, several imprints were made from the tumor specimens. These were air-dried, fixed in 10% neutral- buffered formalin for at least 15 hr, and Feulgen-stained (Falkmer et al., 1990). The ICM assessments were made using an image analysis system (Ahrens), equipped with a Nikon microscope, plan objective 40/0.95, and a video-CCD camera (Panasonic 2000). For each histogram, at least 200 structurally well-preserved and cytodiagnostically identified neoplastic cells were assessed at the wavelength of 546 nm. In each case, 20-30 nuclei from lymphocytes or granulocytes were used as an internal “2c” standard.

Evaluation and interpretation of the DNA histograms FCM. The DNA index (DI) was calculated by means of the

internal reference cells (Vindelov et al., 1983). DNA histo- grams, showing one cell population between 0.95 and 1.05, were defined as DNA “diploid,” whereas DNA histograms with >20% of the cells in the tetraploid region, viz. 1.90-2.10, were interpreted as DNA “tetraploid”. DNA histograms, displaying one or more well-defined peaks of at least 5% of all counts collected, outside these “diploid” and “tetraploid” regions, were interpreted as DNA “aneuploid”. Calculations of the DI of the peaks, their coefficient of variation (CV), and a computer-assisted estimation of the cells in the “S-phase- fraction (SPF),” were made. Histograms with a background noise of > 20% of all counts (cell suspensions with too few cells and/or excessive cellular debris) and/or with a CV of the main peak > 8% were not evaluated at all.

ZCM. The modal values (MV) of the tumor-cell population were calculated in relation to the DNA distribution pattern of the internal standard cells (Falkmer et aZ., 1990). DNA histograms showing an MV between 1 . 8 ~ and 2 . 2 ~ were defined as DNA “diploid”. If > 20% of the cells measured showed an MV in the tetraploid range, viz. 3.7c-4.3c, the histograms were interpreted as being DNA “tetraploid”. DNA histograms with one or more well-defined peaks of at least 5% of all counts registered outside these regions were interpreted as being DNA “aneuploid”. For calculations of the MV of the peaks obtained, their coefficient of variation (CV) and the percent- ages of cells in the “S-phase and G2-regions,” the “scatter- fraction” (SF), a computerized evaluation program was used.

Statistical methods The relationship between c-myc amplification and the other

variables, such as sex, age, histopathological grade, DNA ploidy and SF, was assessed by the chi-squared test for trends in proportions (Armitage and Berry, 1987). The probability of survival in musculoskeletal sarcomas of the patients was calculated by the conventional Kaplan-Meier method. Patients who were alive at the end of the follow-up period, and patients who died due to other causes than musculoskeletal sarcomas, were excluded (n = 17). The log-rank test was used to assess the differences in survival between patients with c-myc- amplified tumors and those with c-myc non-amplified ones (Peto et aL, 1977). The same test was used to calculate the differences in survival between the patients with tumors of DNA euploid type and low SF values, and those with tumors of DNA euploid or aneuploid and high SF values.

RESULTS

The clinico-histopathological findings of all tumors investi- gated for DNA ploidy and c-myc amplification are shown in Table I. Amplification without rearrangement of the c-myc proto-oncogene was detected in 6 out of 26 soft-tissue sarco- mas and in 4 out of 7 bone sarcomas (Table 11). Both the primary osteosarcomas investigated showed c-myc amplification, whereas 2 out of 5 chondrosarcomas, 3 out of 8 malignant fibrous histiocytomas, none of the 5 neurofibrosarco- mas, 1 of 4 liposarcomas, 1 of 2 fibrosarcomas, and the only synovial sarcoma were c-myc amplified. Assessment by scan- ning densitometry revealed 4- to 16-fold amplification intensi- ties. As for the 3 benign bone tumors analyzed, hybridization with the c-myc probe disclosed no increase in number of gene copies. The results described were reproducible when DNA from the tumor samples was cleaved by Hind111 and EcoRI enzymes.

The cytometric DNAvariables found by both FCM and ICM are shown in Table 11. As regards the FCM DNA results, there were-despite several repeated attempts at preparation-1 1 histograms out of 36 which were not evaluable, according to the criteria mentioned above. In contrast, the ICM data obtained on several imprints from the same tumor were found to be conclusive in all 36 neoplasms investigated. In 5 out of the 25 tumors, assessed by both methods, there was a disagree- ment between the DNA ploidy indicated by FCM and that shown by ICM. In 3 of these 5 cases, the FCM DNA histograms were assessed as diploid, whereas those given by ICM were assessed as tetraploid. Thus, the disagreement was only of low degree; all 3 tumors were actually “euploid”. In tumor 8, the FCM DNA histogram displayed only one diploid peak, whereas the ICM DNA histogram showed a predomi- nant, aneuploid peak, viz. one with a considerably higher c-value, located at 5 . 9 ~ . In tumor 17, the last of these 5, the FCM DNA histogram contained one aneuploid peak, whereas the corresponding ICM histogram showed 2 aneuploid peaks, one of them with a high c-value, viz. 5 . 2 ~ . Again, this disagree- ment was only of low grade; finally, both the FCM and ICM were of aneuploid type.

The results obtained by means of the ICM technique showed that 8 out of the 33 malignant tumors harbored a DNA diploid stemline. Also, the cell nuclei of 3 histopathologically benign giant-cell bone tumors displayed a diploid DNA distribution pattern. In the 10 malignant neoplasms with c-myc amplification, DNA ploidy of cell nuclei was as follows: 2 diploid, 2 tetraploid and 6 aneuploid. In the 23 malignant tumors without c-myc amplification, DNA ploidy of cell nuclei was as follows: 6 diploid, 5 tetraploid and 12 aneuploid. Thus, no appreciable differences were observed between these 2 main groups.

When the SPF in the FCM DNA histograms was deter- mined, only 14 out of 36 histograms were of sufficient quality to allow a satisfactory evaluation to be made: this failure was due to the uncertainty caused by the presence of high amounts of background noise (see above). When the percentages of SF cells were calculated in the DNA histograms obtained by means of ICM, this fraction appeared to be well related to the DNA ploidy pattern. Thus, the parenchymal cell nuclei of the DNA aneuploid neoplasms seemed to have higher SF values than those of the DNA diploid tumors, as shown in Table 11. The DNA tetraploid tumors showed both high SF (> 20%) and low SF (<20%) values. The SF values in the DNA histograms of the benign giant-cell bone neoplasms were all found to be <20% of the total number of cells measured. Consequently, high SF values indicated a high proliferative potential.

As could be expected, patients with tumors in which the cytometric DNA distribution pattern of the neoplastic nuclei was euploid (i.e. diploid or tetraploid) with a low (<20%) SF showed a significantly better 2-year survival than patients with

DNA PLOIDY AND c-myc ONCOGENE IN MUSCULOSKELETAL TUMORS 783

TABLE 1 - CLINICO-HISTOPATHOLOGICAL FEATURES OF THE MUSCULOSKELETAL TUMORS INVESTIGATED FOR c-myc ONCOGENE AMPLIFICATION

Soft-tissue sarcomas Course of neoolastic disease

Site Case Age Typeof number Sex (yrs) lesion

Histopathological classification

1 F 2 F 3 M 4 F 5 F 6 F 7 M 8 F 9 M

10 F 11 M 12 F 13 F 14 F 15 M 16 M 17 M 18 M 19 F 20 F 21 F 22 F 23 F 24 M 25 F 26 M

32 47 62 80 60 61 76 79 20 29 30 73 77 39 70 66 86 59 63 65 47 20 75 37 74 36

P. P. P. L. R. L. R. L. R. P. P. P. P. L. R. L. R. P. P. P. P. L. R. P. P. P. P. L. R. P. P. L. R. P.

Gluteal Groin Thigh Upper arm Calf Thigh Thigh Thigh Shoulder Thigh Thigh Popliteal Shoulder Forearm Thigh Shoulder Forearm Thigh Thigh Pelvis Gluteal Calf Thigh Intercostal Sacrum Thigh

Malignant fibr. histiocytoma Malignant fibr. histiocytoma Malignant fibr. histiocytoma Malignant fibr. histiocytoma Malignant fibr. histiocytoma Malignant fibr. histiocytoma Malignant fibr. histiocytoma Malignant fibr. histiocvtoma Neurofibrosarcoma Neurofibrosarcoma Neurofibrosarcoma Neurofibrosarcoma Neurofibrosarcoma Liposarcoma Liposarcoma Liposarcoma Liposarcoma Fibrosarcoma Fibrosarcoma Synovial sarcoma Angiosarcoma Lymphangiosarcoma Rhabdomyosarcoma Mesenchymal chondrosarc. Chordoma Hemangiopericytoma

I11 I11 I11 I11 IV IV IV IV I11 I11 I11 I11 I11 I I1 I11 IV I1 IV I11 I11 I11

I11 I11 I11

-

A A D A D D D D A D D D A A A A A A D D D D A A A D

- 74 69 + 19 81 37 + 42 + 20 0.5 +

72 16 + 54 +

-

-

-

-

25 + 77 70

- - - -

76 65

228 61

- -

1 + 5 +

16 + 25 + 62 79 +

115 + 38 +

-

Malignant bone tumors

number sex (yrs) lesion Site Case Age Typeof Course of neoplastic disease

.. Grade Histopathological

classification

27 M 77 P. Prox.femur Chondrosarcoma 28 M 52 L. R. Abdom.wall Chondrosarcoma 29 M 60 P. Femur Chondrosarcoma 30 M 70 P. Pelvis Chondrosarcoma 31 M 76 P. Prox.femur Chondrosarcoma 32 M 43 P. Prox.femur Osteosarcoma 33 F 17 P. Pubis Osteosarcoma

- 0.5 -

I D I1 A 63 I1 A 68 I1 A 60 I1 D 17 + I11 D 12 + IV A 84 +

- -

Benign bone tumors

Grade Histopat hological classification Site Case Age Typeof

number Sex (yrs) lesion

34 F 29 P. Iliac bone Giant-cell tumor - F 41 P. Ulna Giant-cell tumor -

- 35 36 F 67 L. R. Dist. femur Giant-cell tumor

1P, primary; L.R., local recurrence.

an euploid DNA pattern and a high (> 20%) SF; it was also significantly better than that of patients with an aneuploid DNA content the neoplastic nuclei (Fig. 1). However, the log-rank test showed no statistical significance after 5 years of follow-up, probably mainly due to the low number of patients. The group of patients suffering from neoplasms without c-myc amplification had a significantly better 2-year survival than the patients whose tumors showed a clear-cut c-myc amplification (Fig. 2). Here again, the log-rank test showed no statistical significance after 5 years of follow-up.

In contrast to these clinical relationships, a statistical analy- sis of the existence of some kind of correlation between the occurrence of c-myc amplification, high SF-values and histo- pathological grade showed that no such correlations were present. An unexpected observation was that a sex difference seemed to exist as regards c-myc amplification in the neoplastic cells; the 6 soft-tissue sarcomas in which this cell-biological feature was present all occurred in female patients (Tables I,

11). In contrast, no such amplification was found in any of the soft-tissue sarcomas in the 11 male patients.

DISCUSSION

The finding that as many as one third of all FCM DNA histograms were methodologically deficient is rather unusual. In fact, the incidence is about 10% higher than that usually observed in our laboratory in DNA cytometric analyses of other series of solid tumors, such as carcinomas of the breast (Falkmer et al., 1990; Falkmer, 1991). In carcinomas of this kind, the carrying through of FCM DNA assessments is often limited by one factor only, viz. the amount of tumor material available. In the present series of musculoskeletal sarcomas, the amount of tumor tissue for FCM analysis was essentially unlimited. Instead, the major obstacle was the appearance of technically unsatisfactory DNA histograms. Thus, even after several repeated preparations of cell suspensions from the

784 BARRIOS ETAL.

TABLE I1 - CYTOMETRIC DNA RESULTS AND LEVEL OF c-myc ONCOGENE AMPLIFICATION O F THE MUSCULOSKELETAL TUMORS INVESTIGATED

Soft-tissue sarcomas Oncogene

amplification Flow cytometry Image cytometry -

Case DNA number Ploidy index

1 -

1 Dl NA2 2

3 NA -

6 7 8 9

10 11 12 13 14 15

A NA D A

NA T A A

NA A

1.5

1 0.9

1.9 0.8 2.9

1.2

-

-

-

16 D 1 17 A 1.5 18 D 1 19 NA - 20 NA 21 A 1.8

1.3 22 A 23 NA

-

- 24 D 1 25 D 1 26 D 1

27 D 1 NA - 28

29 NA 30 D 1

Malignant bone tumors

-

31 D 1 .o 32 NA - 33 A 1.3

34 D 1 Benign bone tumors

35 36

D 1 D 1

“S-phase” “GlM’ cells

(%x)

21 - - 24 NA NA

3 6

NA NA NA

NA 6 3 9

-

-

-

- -

NA NA

9 NA

7

6

-

- -

NA 7 -

9 6 6

Ploidy

T T A A T A T A A A T A A D A D A D A A A A A T D D

D D A T D A A

D D D

Modal value

(c-units)

3.8 4.0 3.0 3.6 3.9 3.0 4.0 5.9 1.8 3.1 3.8 1.6 5.8 2.1 2.5 2.0

3.2,5.2 2.0 2.5 3.0 3.6 2.6 5.5 4.0 2.0 2.0

2.1 2.0 2.6 3.9 2.0 4.3 2.6

2.0 2.0 2.0

Level “S-phase”

(9 “Gi M” cells c-nryc cv (”/.) I,?&\

7 8 5 > 20 7 > 20 6 > 20 8 > 20 7 > 20 6 > 20 7 > 20 5 9 6 > 20 9 > 20 8 > 20 7 > 20 8 4

10 > 20 5 2

10 > 20 5 10 9 > 20 6 > 20 8 15 7 > 20 6 > 20

- - + 8 + 8 -

- -

3 2 5 2

- 4 1

> 20 8 6 8 8 > 20 4 2 + 8 4 13 + 4 6 > 20 + 16 7 > 20 + 12

6 15 4 12 6 10

- - -

- - -

ID, diploid; T, tetraploid; A, aneuploid.JNA, not assessable.

tumor parenchyma, the background values found in the DNA histograms were still too high and the CVs too broad; only 40% of all the FCM DNA histograms could be evaluated for their SPF. One explanation may be found in the fact that these tumors consist of neoplastic cells with extreme cellular and nuclear polymorphism, which are often large and which have bizarre-shaped nuclei. Non-spherical giant nuclei can cause turbulence in the hydrodynamically focused flow stream in the FCM procedure, resulting in broad peaks and high CV values. Admixture of necrotic parenchymal cells with viable tumor parenchyma cannot be avoided during the preparation of cell suspensions for FCM, causing high background noise.

In contrast, the ICM procedure, performed on imprints of the tumor parenchyma, allows selective measurement of un- damaged nuclei, avoiding any kind of background. Also, the assessment of large-sized or bizarre-shaped nuclei can easily be made. Therefore, in this group of tumors, the efficacy of the ICM technique is higher than that of the FCM procedure.

The 20% incidence of discrepancies between the data obtained on the nuclear DNA distribution pattern of neoplas- tic cells by both the FCM and the ICM techniques is in line with previous studies (Falkmer et aL, 1990). One explanation is that, when applying ICM assessments, it is more often possible to identify tetraploid or aneuploid peaks with extremely high c-values. This depends on the more selective measurement of

microscopically identified tumor cells, leaving out all kinds of admixture of non-tumor cells. This permits the detection of even small numbers of cells with extremely high amounts of aneuploid nuclear DNA contents.

The results of the present study show that the neoplastic cells of musculoskeletal sarcomas with c-myc oncogene can display widely different DNA ploidy patterns. This statement also applies to sarcomas without c-myc amplification. The failure to detect any statistical relationship between c-myc amplification and DNA ploidy pattern can be explained by differences in the detecting level of the methods applied. Whereas tumor characterization by means of molecular- biological techniques permits detection of subtle genetic alter- ations, DNA ploidy assessments, either by FCM or by ICM, provide only a rather crude estimate of the presence of some kind of DNA change. However, the 16-fold increase in c-myc gene copies, as observed in the tumor cells of one of the malignant fibrous histiocytomas and one of the osteosarcomas, was found to be correlated with a high level of DNA values as assessed by means of ICM, viz. 3 . 9 ~ and 4.3c, respectively. Although the number of gene copies can be expected to be increased in “aneuploid” tumors, c-myc amplification was only found in one-third of these tumors. Considering the high sensitivity of the hybridization technique applied, the absence of c-myc amplification in the remaining aneuploid tumors is

DNA PLOIDY AND c-myc ONCOGENE IN MUSCULOSKELETAL TUMORS

% 785

J 20

Years otter diagnosis

0 I I I

2 4 6 8 10 Numberof atients

a t risk Group1 10 9 8 6 2 Group2 23 13 9 8 1

FIGURE 1 - The 33 patients with musculoskeletal sarcomas were classified into 2 groups according to the DNA ploidy pattern found in their tumor-cell nuclei. Group 1 = DNA euploid and low SF values ( <20%); group 2 = DNA euploid or aneuploid and high SF values (> 20%). After 2 years of follow-up, the probability of survival for group 1 is twice as good as that for group 2. After 5 years of follow-up, the difference is no longer statistically signifi- cant, probably due to the small number of patients.

probably a reliable finding. “DNA aneuploidy” should merely be considered as a reflection of major genetic instability, and no relationship to the presence or the activity of specific oncogenes has, up to now, been established.

As mentioned in the opening paragraphs, an activation of c-myc has been proposed to reflect proliferative activity (Cole, 1986). In the present study, 7 out of 10 amplified tumors also showed a high SF, in all cases > 20% as determined by ICM. When only the amplified soft-tissue sarcomas are considered, 5 out of 6 showed SFvalues > 20%, whereas as a whole 17 out of 26 showed SF values > 20%. Here, the statistical test for trend was not significant, probably due to the small number of cases. Tumor 14, a primary liposarcoma of histopathological grade I, seems to be an exception, showing an ICM DNA diploid pattern and an extremely low S F value (re-analyzed data), but a c-myc amplification of as much as 4-fold.

The fact that the 6 cases of c-myc-amplified soft-tissue sarcomas occurred in females, whereas none out of the 11 males with soft-tissue sarcomas showed this amplification, is an observation for which we have no explanation. Its significance should perhaps be investigated further.

The prognostic information offered by the results of cytomet- ric DNA ploidy analysis with calculations of SF values has already been well established in earlier studies of the clinical course of the individual patient with a musculoskeletal sar- coma (Kreicbergs et al., 1987; Bauer et al., 1988). Also, for the relatively small group of patients in the present study, the

b-u-

Years after diagnosis

0 0 2 4 6 8 10

13 10 2 4 4 1

2ol, , I , , ~

Number of patients at risk

c-myct 10 4 c-myc- 23 17

FIGURE 2 - The 33 patients with musculoskeletal sarcomas were subdivided into 1 group of patients with tumors showing c-myc amplification (c-myc+) and another with tumors showing no amplification (c-myc-). The first group showed a significantly worse 2-year survival rate than the second group. Five years after diagnosis the difference observed between the 2 groups was no longer statistically significant, again probably due to the small number of patients.

clinical observations clearly show that the main pieces of prognostic information offered by the DNA data (ploidy and SF-values), are available during the first years after diagnosis. This can be illustrated by the observation that the clinical course of patients having sarcomas with high SF-values was significantly worse during the first 2 years than that for patients having sarcomas with low SF-values, irrespective of whether they were classified as DNA diploid, tetraploid or aneuploid.

As regards the prognostic information yielded by the finding of a c-myc amplification in the neoplastic cells, the survival data also show quite clearly that patients with c-myc-amplified sarcomas had a rapidly fatal course during the first 2 years after diagnosis, compared with the group of patients whose sarcomas did not present c-myc amplification. Due to the small number of patients investigated here, it is still an open question whether DNA ploidy data, combined with c-myc amplification, can offer more significant pieces of prognostic information.

ACKNOWLEDGEMENTS

This work was supported by grants from the Swedish Medical Research Council (Projects 102 and 718), the Swedish Cancer Society (Project 910364), the Cancer Society of Stock- holm and the Gustaf V Jubilee Fund, Stockholm. J.S. Castre- sana received funds from the Ministry of Education, Research, from the Universities of the Basque Government in Spain, and also from the Margit and Folke Pehrzon Foundation in Stockholm, Sweden.

REFERENCES

ARMITAGE, P. and BERRY, G., Statistical methods in medical research (2nd ed.), p. 125, Blackwell, Oxford (1987).

Amplification of the c-myc proto-oncogene in soft tissue sarcomas. Oncology, 51,13-17 (1994). BAUER, H.C.F., KREICBERGS, A., SILFVERSWARD, C. and TRIBUKAIT, B., DNA analysis in the differential diagnosis of osteosarcoma. Cancer, 61, 1430-1436 (1988). BISHOP, J.M., The molecular genetics of cancer. Science (Wash.), 235,

c.3 CASTRESANA3 J’s’, Ru1z9 J’ and KREICBERGS, A’, Amplification of c-myc oncogene and absence of c-H-ras point muta- tions in human bone sarcoma. J. orthop. Rex, 11,556-563 (1993). BARRIOS, C., CASTRESANA, J.S., RUIZ, J. and KREICBERGS, A,, 305-311 (1987).

786 BARRIOS ETAL.

BRODEUR, G.M., SEEGER, R.C., SCHWAB, M., VARMUS, H.E. and BISHOP, J.M., Amplification of N-myc in untreated human neuroblas- toma correlated with advanced disease stage. Science (Wash.), 224, 1121-1124 (1984). COLE, M.D., The myc oncogene: its role in transformation and differentiation. Ann. Rev. Genet., 20,361-384 (1986). ENZINGER, F.M. and WEISS, S.W., (eds.), Soft tissue tumors (2nd ed.), C.V. Mosby, St. Louis (1988). FALKMER, U.G., Methodological aspects on the value of flow and image cytometric nuclear DNA assessments of the neoplastic cells of prostatic adenocarcinomas. Acta oncol., 30,201-203 (1991). FALKMER, U.G., HAGMAR, T. and AUER, G.U., Efficacy of combined image and flow cytometric DNA assessments in human breast cancer. Anal. Cell Pathol., 2,297-312 (1990). FIELD, J.K. and SPANDIDOS, D.A., The role of ras and myc oncogenes in human solid tumors and their relevance in diagnosis and prognosis. Anticancer Res., 10,l-22 (1990).

MANDARD, A., PARTENSKY, C. and MONTESANO, R., Amplification of epidermal growth factor receptor gene but no evidence of ras muta- tions in primary human esophageal cancers. Cancer Res., 48, 5119- 5123 (1988). IKEDA, S., SUMII, H., AKIYAMA, K., WATANABE, S., ITO, S., INONE, H., TACKECHI, H., TANABE, G. and ODA, T., Amplification of both c-myc and c-raf-1 oncogenes in a human osteosarcoma. Jap. J. Cancer Res., 80,6-9 (1989). KREICBERGS, A., TRIBUKAIT, B., WILLEMS, J. and BAUER, H.C.F., Flow DNA analysis of soft tissue tumors. Cancer, 59,128-133 (1987).

HOLLSTEIN, M.C., SMITS, A.M., GALIANA, C., YAMASAKI, H., BOS, J.L.,

LADANYI, M., KUM PARK, C., LEWIS, R., JHANWAR, S.C., HEALEY, J.H.

and Huvos, A.G., Sporadic amplification of the MYC gene in human osteosarcomas. Diagn. mol. Pathol., 2,163-167 (1993). LOOK, AT., HAYES, F.A., SHUSTER, J.J., DOUGLASS, E.C., CASTLE- BERRY, R.B., BOWMAN, L.C., SMITH, E.I. and BRODEUR, G.M., Clinical relevance of tumor cell ploidy and N-myc gene amplification in childhood neuroblastoma: a pediatric oncology group study. J. din. Oncol., 9,581-591 (1991). OPPEDAL, B.R., OIEN, O., JAHNSEN, T. and BRANDZAEG, P., N-myc amplification in neuroblastomas: histopathological, DNA ploidy, and clinical variables. L din. Path., 42,1148-1152 (1989). PETO, R., PIKE, M.C., ARMITAGE, P., BRESLOW, N.E., Cox, D.R., HOWARD, S.V., MANTEL, N., MCPHERSON, K., PETO, J. and SMITH, P.G., Design and analysis of randomized clinical trials requiring prolonged observation of each patient. Brit. J. Cancer, 35,7-9 (1977). SAMBROOK, J., FRITSCH, E.F. and MANIATIS, T., Molecular cloning. A laboratory manual (2nd ed.), Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY (1989). SCHWAB, M., ELLISON, J., BUSH, M., ROSENAU, W., VARMUS, H.E. and BISHOP, J.M., Enhanced expression of the human gene N-myc conse- quent to amplification of DNA may contribute to malignant progres- sion of neuroblastoma. Roc. nat. Acad. Sci. (Wash.), 81, 4940-4944 (1984). VINDELOV, L.L., CHRISTENSEN, I.J. and NILSSON, N.J., Detergent trypsin method for the preparation of nuclei for flow cytometric DNA analysis. Cytometry, 3,323-327 (1983). WATT, R., STANTON, L.W., MARAN, K.B., GALLO, R.C., CROCE, C.M. and ROVERA, G., Nucleotide sequence of cloned cDNA of human c-myc oncogene. Nature (Lond.), 303,725-728 (1983).