The Clinical Value of Flow Cytometric DNA Content Analysis in Patients with Soft Tissue Sarcomas

Purpose. The purpose of this study was to evaluate: (1) the correlation between grade and ploidy or S-phase fraction (SPF), (2) the prognostic value of DNA flow cytometric study in soft tissue sarcomas. Patients /Methods. In all, 47 tissue samples from soft tissue sarcoma patients, surgically treated in the same center, were included. Flow cytometric analyses were performed according to a modified version of the original method of Hedley et al. Results. DNA ploidy status could be determined in 44 samples out of 47 (success rate 94%). Of these 44, S-phase fraction could be calculated in 34 samples (77%). In the study group as a whole, aneuploidy was significantly correlated with high grade. Survival analyses were carried out in 21 patients with soft tissue sarcoma, all surgically treated in the same center, without chemotherapy or radiotherapy. In univariate analyses, DNA ploidy was found to be a significant factor for overall survival (OAS) and metastasis-free survival MFS. Mean OAS for aneuploid tumors and diploid tumors were 35 and 65 months (p=0.034), and mean MFS 23 and 61 months, respectively (p=0.005) . Discussion.There is a relation between histological grade and ploidy in soft tissue sarcomas. It appears that low-grade tumors are generally diploid, whereas high-grade tumors tend to be aneuploid. In a subgroup of patients treated only with surgery, DNA ploidy was found to be an important factor for predicting OAS and MFS.


Introduction
Sarc om as are a hetero gen eo us g rou p of tum ors causing diagnostic and therapeutic difficulties. 1,2 T he m a in stay o f th erap y is su rger y. A d ju va n t chemotherapy has been shown to be effective in pediatric rhabdomyosarcom a, osteosarcom a and Ew ing's sarcom a but rem ains controversial in the other adult sarcom as. 3± 5 It is im portant to identify the subset of patients w ith a poor progno sis because they are candidates for experim ental adjuvant chemotherapy regim ens in ongoing studies. U nfortunately this selection is not easy and ham pered by the heterogeneity of m o r p h o lo gy a n d clin ical b ehav io r. Tum o r siz e, h isto lo g ica l g rad e, lo ca lization , tu m o r n ecrosis, vascular invasion, clinical and pathologic response to neoadjuvant chemotherapy, age and sex, have been all proposed to be prognostic factors for sarcom as. 6± 8 Am ong them , histological grade and tum or size are universally accepted as im portant factors and m ost adjuvant trials use a com bination of these two factors for patient selection. U nfortunately histological grade of sarcom as has low reproducibility. 9± 11 Flow cytom etric analysis of ploidy and S-p hase fraction (SP F ) has been show n to b e usefu l in prognostication in various tumors. 12 Such studies have also been done for sarcom as and som e have found it to be an independent prognostic factor. In som e studies it was also shown that¯ow cytom etric analysis could help in grading these tum ors. 13± 18 In this study we aim ed to evaluate: (1) correlation between grade and ploidy or S-p hase fraction, and (2) the prognostic value of D N A¯ow cytom etric study in adult soft tissue sarcom as.

Patients
Patients with histologically proven soft tissue sarcom as between 1989 and 1998 in the Ankara U niversity Fa cu lty o f M ed icin e, Ib n i-S in a H o sp ita l, w ere included in the study. T he paraffin-em bedded tissue sam ples of patients were re-evaluated by the sam e pathologist, for diagnostic veri® cation and grading. Patients w ith diagnostic and grading disparance were excluded. Forty seven tissue sam ples from patients with soft tissu e sarcom a were selected for study. Patient characteristics, treatm ent and follow up were noted. Patient characteristics are shown in Table 1.

Method
Flow cytom etric analyses were perform ed according to principles of a modi® ed version of the original method of Hedley et al. 19 First, histological specim ens with H emotoxylin± eosin were evaluated. From tissues composed of at least 20% m alignant cells, 5± 6 sections of 50 m m were cut. Areas with tum or necrosis were avoided. The sections were deparaffin ized with xylene and rehydrated by changing concentrations of ethanol: 100% , 100% , 95%, 70% and 50%, respectively. After overnight resting in distilled water, the tissue was washed twice w ith phosph ate-buffe red saline (PBS). T he pieces were digested w ith 1 m g/m l protease (Sigm a Type XX IV ) at 37Ê C for 60 m in, shaking by hand every 5± 10 m in. Follow ing the digestion, 2 m l of cold PBS was added to the solution. T he solution was ® ltered through a 37-m m nylon m esh. After two washings w ith cold PBS, cell suspensions having 53 10 5 to 13 10 6 nuclei per m l were prepared with trypsin buffer and centrifuged, then the nuclei were incubated with RN Ase solution for 10 m in. Finally the nuclear suspensions were stained w ith propidium iodide and kept in the dark for at least 10 m in at 4Ê C. T he sam ples were then ® ltered thorough a 37-m m nylon m esh before cytom etric analysis. D N A analyses were carried out with a FACSort ow cytom etr y (Becton & D ickin son FAC S or t). Excitation of propidium iodide occurred at 488 nm . At least 20 000 nu clei from each specimen were analyzed. D N A histogram s having only one G0/G1 peak with a coefficient of variation (CV ) less than 5% were de® ned as diploid. D N A peaks with C V of m ore than 10% were not evaluated. The histogram s with CV between 5 and 10% were accepted as wide diploid and collected in the sam e group with diploid ones. T he sam ples were classi® ed as aneuploid if there were at least two distinct G o/G1 peaks, the latter having at least 10% of the total count. Cell cycle analyses were perform ed by M O D FIT software of Becton & D ickinson. D N A histogram s having C V of m ore than 8% were not evaluated for SPF m easurem ents. 20 Statistical analysis C or re latio n betw een the D N A p loid y an d S P F m easurem ents and other param eters were exam ined using the Pearson test and survivals were calculated by the K aplan± M eier m ethod. U nivariate overall su r vival (OA S), disease-fr ee su r viva l (D F S ) and m eta sta sis-free su r v ival an aly sis ( M F S ) w ere perform ed using the log-rank test. All calculations were perform ed using the SPSS 7.0 for W indows statistical package.

Results
D N A ploidy status could be determined in 44 sam ples out of 47 (success rate 96%). O f these 44, S-phase fraction could be calculated in 34 sam ples (77% ). O ut of 44 soft tissue sarcom a sam ples, 21 were aneuploid (two were grade 1, ® ve were grade 2 and 14 were grade 3), 23 were diploid (nine were grade 1, nine were grade 2 and ® ve were grade 3). Ploidy status distributions according to grade of all other sam ples are shown in Table 2.
Correlation between the results of D N A analyses and other param eters are shown in Table 3. Aneuploidy was signi® cantly correlated w ith high grade. Tetraploid (t) and m ultiploid (m) samples were given separately under the aneuploid group. SPFS-phase fraction.
Tum ors with high grade or aneuploid were prone to be m etastatic in presentation.
Since sarcom a patients evaluated in this study were histologically heterogeneous and treated with different chemotherapy and radiotherapy protocols, survival analyses were perform ed only in a subg roup of 21 patients w ith soft tissue sarcom a, all surgically treated without chem otherapy or radiotherapy. In univariate analyses, D N A ploidy was found to be signi® cant factor for OAS and M FS. M ean OAS for aneuploid tum ors and diploid tum ors were 35 and 65 m onths, respectively (P=0.034). M ean M F S for aneuploid tum ors and diploid tum ors were 23 and 61 m onths, respectively (P=0.005; Table 4, F igs 1± 2). SPF was a signi® cant factor for D F S. M ean D F S for tumors w ith S PF £ 10% and S PF >10% w ere 53 an d 20 m o n th s, resp ective ly (p = 0.03 1) . G ra d e w a s a signi® cant prognostic factor only for D FS. M ean D FS were 52, 19 and 13 m onths for grade 1, 2 and 3 tum ors, respectively (p=0.007) ( Table 4).

Discussion
This study show s a correlation of grade and D N A ploidy in sarcom as. Aneuploidy is found to be related both to we ® n d that high g rade and m etastatic behavior in sarcom as. In keeping w ith the literature, D N A an alysis m ay pred ict the aggressiveness of sarco m a s. 14 E sp ecially in so ft tissu e sarco m as, although grade is accepted as the m ost im portant prognostic factor, criteria used to determine the grade of a sarcom a are subjective, not standard and poorly reproducible. 9± 11 In this respect D N A analyses with ow cytometry may be helpful for assessing the clinical behavior of sarcomas objectively. In this way these tum ors can be separated into subsets to de® ne those patients w ho should or should not be candidates for experim ental adjuvant chem otherapy.
To perfor m a su r vival an alysis, w e selected a hom ogenous group of soft tissue sarcom a com posed of 21 patients with localized disease at presentation   28 Con¯icting results m ight be due to the huge variety of histological subtypes of sarcom as or the technical difficulties in¯ow cytom etric D N A analysis of sarcom as. In our study the success rate of 83% in¯ow cytom etric D N A analysis was consistent with others in the literature. 25 Tum or necrosis is the m ost im portant factor causing technical difficulties in¯ow c yto m etr ic D N A an alysis o f sarco m a s. Although all sam ples were re-valu ated histologically before study, it is still possible that som e sam ples m ight be contam inated with necrotic debris and reactive cells.
In co nclu sion, there is a relation ship between histological grade and ploidy in sarcom as. It appears that low-g rade tum ors are generally diploid, whereas high-g rade tum ors tend to be aneuploid. At least in the sarcom a group treated with only surgery, we showed that D N A ploidy is an im portant factor predicting OAS and M FS. W ith im provem ent of the technique and standardization with further studies, ow cytom etric D N A analysis m ay provide an objective an d rep ro d u cib le p ro gn o stic var ia b le fo r com parison of se ries fro m differe nt centers an d possibly for the selection of high-risk patients who m ay be candidates for adjuvant therapy.