Prognostic significance of DNA ploidy pattern in osteosarcomas in association with chemotherapy

Cancer Letters 137 (1999) 27±33

Prognostic signi®cance of DNA ploidy pattern in osteosarcomas in association with chemotherapy Katsuyuki Kusuzaki a,*, Hideyuki Takeshita a, Hiroaki Murata a, Masazumi Hirata a, Shin Hashiguchi a, Tsukasa Ashihara b, Yasusuke Hirasawa a a

Department of Orthopaedic Surgery, Kyoto Prefectural University of Medicine, Kamigyo-ku, Kawaramachi, Hirokoji, Kyoto 602-8566, Japan b Department of Pathology, Kyoto Prefectural University of Medicine, Kyoto 602, Japan Received 6 August 1998; received in revised form 26 October 1998; accepted 29 October 1998

Abstract In this study, we analysed the DNA ploidy of osteosarcomas at biopsy and attempted to clarify the relationship between DNA ploidy pattern and prognosis. Thirty patients with non-metastatic osteosarcoma of an extremity were studied. All underwent intensive chemotherapy with doxorubicin, cisplatin and methotrexate, in addition to wide tumor resection. DNA ploidy was detected by DNA cyto¯uorometry, using isolated and smeared cells of biopsied tumor tissue. Twelve tumors showed a diploid ploidy pattern and 18 showed a non-diploid pattern such as aneuploidy (15 tumors) and euploid±polyploidy (3 tumors). The event-free survival rate at 9 years was 63.5% in non-diploid osteosarcoma patients and 13.3% in diploid osteosarcoma patients. There was a statistically signi®cant difference between the two groups (P ˆ 0:0278). These results lead us to conclude that a non-diploid osteosarcoma may be more sensitive to chemotherapy than a diploid tumor. q 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Osteosarcoma; DNA ploidy; Chemotherapy; Prognosis; Cyto¯uorometry

1. Introduction Although intensive chemotherapy remarkably improved the prognosis of osteosarcoma patients [1± 3], the chemosensitivity of a tumor and patient's prognosis are dif®cult to predict before treatment. Many methods of chemosensitivity tests have been reported [4±6], in addition to many indicators for predicting the prognosis of osteosarcoma patients. However, most of the chemosensitivity tests are not suitable for osteosarcoma [7,8], because they are unreliable. Histologic response after preoperative chemotherapy is the most reliable indicator of prognosis [1±3,9,10]. However, a * Corresponding author. Tel.: 1 81-75-2515549; fax: 1 81-752515841.

better indicator is needed before preoperative chemotherapy. In this study, therefore, we attempted to clarify the relationship between the DNA ploidy pattern of biopsied tumor using cyto¯uorometry [11±13] and prognosis in osteosarcoma patients after intensive chemotherapy and wide tumor resection. 2. Materials and methods Thirty primary osteosarcomas were used for the study. The patients included 18 males and 12 females, with an average age of 17 (9±66) years. None of the patients had metastatic lesions at diagnosis (Surgical Stage IIB [14]). Twenty-three tumors arose from the

0304-3835/99/$ - see front matter q 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0304-383 5(98)00336-X

28

K. Kusuzaki et al. / Cancer Letters 137 (1999) 27±33

femur, four from the tibia, two from the humerus and one from the ®bula. All patients underwent a biopsy followed by preoperative chemotherapy, wide tumor resection and postoperative chemotherapy. Chemotherapy, which consisted mainly of doxorubicin (DOX) and/or cisplatin (CDDP) and/or methotrexate (MTX), was administered in four to six courses (see Table 1). The average follow-up period after the ®rst course of chemotherapy was 64 (12 (DOD case) to 184) months. 2.1. DNA cyto¯uorometry [11,12] Tumor cells were immediately isolated from biop-

sied fresh tumor tissue by mincing with mini-scissors after treatment with 0.1% collagenase (CLS II, Worthington Co., USA) at 378C for 15 min in PBS. Cells were ®ltered through a 100 mm metal mesh to eliminate intercellular matrix debris and smeared on glass slides using a centrifugal automatic smear machine (Autosmear, Sakura Seiki Co., Japan), followed by drying and ®xation with 70% ethanol. After treatment with RNase (type II, Worthington Co., USA) at 378C for 30 min in PBS, cells were stained with propidium iodide (PI; 0.0025% in citrate buffer, Sigma Co., USA, 48C, 15 min), which is a ¯uorescent dye that quantitatively intercalates between nuclear DNA strands. The nuclear DNA content of each cell was measured by an epi-illumina-

Table 1 The results of ploidy pattern, %HDC of all osteosarcomas studied, patient prognosis and chemotherapeutic agents that patients received Case

Ploidy pattern

%HDC

Chemotherapy

Prognosis

1 2 3 4 5 6 7 8 9 10 11

D D D D D D D D D D D

30.0 10.2 15.0 32.2 12.0 29.0 11.5 16.5 32.0 14.8 17.8

CDF DOD DOC DOD NED DOD DOD NED CDF DOD CDF

12

D

14.5

13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

A(3C±4C) A(3C±4C) A(3C±4C) A(3C±4C) A(3C±4C) E A(3C±4C) E A(3C±4C) A(3C ˆ 4C) A(3C±4C) E A(3C±4C) A(3C±4C) A(3C±4C) A(3C±4C) A(3C±4C) A(3C±4C)

67.2 41.0 65.0 60.0 90.5 31.0 81.9 45.0 85.0 83.5 80.0 48.0 53.0 66.0 72.0 90.0 49.5 83.0

DOX: 675 mg, CDDP: 900 mg DOX: 522 mg, CDDP: 350 mg, CBDCA: 1200 mg DOX: 405 mg, CDDP: 600 mg, MTX: 72 g DOX: 855 mg, CDDP: 1030 mg, MTX: 13.6 g DOX: 600 mg, CDDP: 2055 mg, MTX: 30 g DOX: 600 mg, CDDP: 750 mg, MTX: 14.4 g DOX: 720 mg, CDDP: 940 mg DOX: 300 mg, CDDP: 600 mg DOX: 315 mg, CDDP: 500 mg, MTX: 20 g DOX: 750 mg, CDDP: 450 mg DOX: 472.5 mg, CDDP: 570 mg, MTX: 72 g, CPA: 3600 mg, BLM: 180 mg, IFO: 40.5 g, THP: 202.5 mg, VP16: 2250 mg, CBDCA: 1350 mg DOX: 547.5 mg, CDDP: 570 mg, MTX: 90 g, IFO: 49.8, CPA: 3600 mg, CBCDA: 900 mg, THP: 135 mg, VP16: 2100 mg DOX: 585 mg, CDDP: 520 mg, MTX: 240 g, IFO: 34 g DOX: 270 mg, CDDP: 300 mg DOX: 405 mg, CBDCA: 3900 mg, THP: 405 mg DOX: 330 mg, CDDP: 300 mg, MTX: 58 g DOX: 390 mg, CDDP: 450 mg, IFO: 27 g DOX: 690 mg, CDDP: 850 mg DOX: 495 mg, CDDP: 300 mg, MTX: 69.5 g DOX: 400 mg, CDDP: 600 mg DOX: 645 mg, MTX: 40 g, CDDP: 780 mg DOX: 750 mg, CDDP: 420 mg, CBDCA1: 400 mg DOX: 450 mg, MTX: 57.5 g, CDDP: 960 mg DOX: 630 mg, CDDP: 720 mg DOX: 450 mg, CDDP: 300 mg DOX: 280 mg, MTX: 66.6 g DOX: 380 mg, MTX: 31 g DOX: 480 mg, MTX: 85.3 g DOX: 480 mg, MTX: 85.3 g DOX: 600 mg, CDDP: 850 mg

CDF CDF DOD NED CDF DOD CDF DOD CDF CDF DOD CDF CDF CDF DOD CDF CDF DOD CDF

K. Kusuzaki et al. / Cancer Letters 137 (1999) 27±33

29

sity of PI under observation of cell morphology. The cell smears on the glass slide were excited by green light (510±560 nm) and each cell emitted red ¯uorescence from PI binding to DNA. After detection of a single tumor cell in a small ®eld under high magni®cation (objective lens £ 40), the photomultiplier was used to measure ¯uorescence intensity through an interference ®lter (620 nm). In each sample, 200 tumor cells were measured and their data were automatically input to a personal computer (9800 VM2, NEC Co., Japan) to plot a DNA content histogram from which the ploidy pattern and frequency of hyperdiploid cells (%HDC) were detected. For standardization of the ¯uorescence intensity of diploid cells, human lymphocytes or contaminated granulocytes and ®brocytes in the tumor slide were initially measured. Since all of the examiners (K.K., H.T., H.M., M.H., S.H.) were orthopedic surgeons with backgrounds in pathology and cytology, individual tumor cells were identi®ed and measured by visual exclusion of contaminating normal cells, fragmented nucleus and bony or cartilaginous matrix debris emitting red ¯uorescence. 2.2. Survival rate of patients The event-free survival rate of diploid (D) and nondiploid (non-D) tumor patients was calculated by the Kaplan±Meier method and statistically analysed by the Cox±Mantel test. 3. Results

Fig. 1. The representative DNA content histograms of diploid (a: case 7), euploid±polyploid (b: case 24) and aneuploid (c: case 25) osteosarcomas. The frequency of diploid cells, expressed as %HDC, was 11.5% in case 7, 48% in case 24 and 53% in case 25. The details are given in Section 3.

tion type cyto¯uorometer (¯uorescent microscope and photomultiplier; Nikon SPM RFI-D or OPTIPHOTO with P1, Nikon, Co., Japan) as the ¯uorescence inten-

The DNA ploidy pattern obtained from all of the osteosarcomas studied was classi®ed as diploid, euploid±polyploid or aneuploid. The DNA content histogram of the diploid type showed that most of the tumor cells had a diploid (2c) nuclear DNA content and that a few were hyperdiploid (2c±4c), but no cells were hypertetraploid (4c,). Hyperdiploid cells contained both DNA synthetic (S phase) cells and G2 phase cells in the cell cycle. Fig. 1a demonstrates a representative DNA content histogram of a diploid osteosarcoma (case 7). In this case, the frequency of diploid cells expressed as %HDC is 11.5%. Euploid±polyploid osteosarcoma consisted of many diploid, tetraploid (4c) and octa-

30

K. Kusuzaki et al. / Cancer Letters 137 (1999) 27±33

Fig. 2. Event-free survival rate of 12 diploid and 18 non-diploid osteosarcoma patients. The event-free survival rate of non-diploid osteosarcoma patients was 63.5%, whereas that of diploid osteosarcoma patients was 13.3%. There was a statistically signi®cant difference between both groups (P ˆ 0:0278).

ploid (8c) cells accompanied by a few cells having an intermediate DNA value between 2c and 4c or 4c and 8c, which were in S and G2 phase of diploidy or tetraploidy. Fig. 1b shows the DNA content histogram representing a euploid±polyploid pattern obtained from the osteosarcoma of case 24. In this case, there were many octaploid cells, but tetraploid cells were rare. The %HDC of this osteosarcoma was 48%. Aneuploid osteosarcomas were composed of many aneuploid cells and diploid cells with their S and G2 phase cells. In this analysis, all of the aneuploid osteosarcoma cells had a nuclear DNA content between 3c and 4c. Fig. 1c demonstrates the DNA content histogram of an aneuploid osteosarcoma obtained from case 25. There were many aneuploid cells with S and G2 phase cells, although a diploid tumor cell population was present. The %HDC of this tumor was 53%. Table 1 summarizes the ploidy patterns and %HDC of all osteosarcomas studied, and also indicates patient prognosis, in addition to chemotherapeutic agents administered pre and postoperatively. Twelve osteosarcomas in 30 patients showed a diploid pattern, 3 were euploid±polyploid and 15 were aneuploid. The average %HDC of 12 diploid osteosarcomas was 19.7% and that of 18 non-

diploid osteosarcomas, including 3 euploid±polyploid and 15 aneuploid tumors was 66.2%. In prognosis, only 4 of 12 diploid osteosarcoma patients (33%) were continuously disease free (CDF) after chemotherapy, whereas 11 of 18 non-diploid osteosarcoma patients (61%) were CDF. Fig. 2 shows the event-free survival rate of 12 diploid and 18 non-diploid osteosarcoma patients. At 9 years after the ®rst course of preoperative chemotherapy, the event-free survival rate of nondiploid osteosarcoma patients was 63.5%, whereas that of diploid osteosarcoma patients was 13.3%. There was a statistically signi®cant difference between both (P ˆ 0:0278). 4. Discussion There have been a number of studies of DNA ploidy analysis in benign and malignant tumors [15± 16] including bone and soft tissue tumors [18±24]. Most of the studies were performed by ¯ow cytometry which permits rapid measurement of more than 10 000 cells. Flow cytometry, however, measures not only tumor cells but also contaminating normal

K. Kusuzaki et al. / Cancer Letters 137 (1999) 27±33

cells such as granulocytes, lymphocytes or ®broblast, and fragments of destructive nucleus or the intercellular matrix which are emitting auto¯uorescence [16,17,21]. Cyto¯uorometry [11±13], which was used in this study can selectively measure only tumor cells by microscopic detection of cellular morphology, although this method is time consuming and the measurable cell number is less than 300. We believe that for DNA ploidy analysis of bone and soft tissue tumors, cyto¯uorometry is better than ¯ow cytometry, because numerous intercellular matrix components such as osteoid, cartilage and collagen are present in most of the tumors. In our experience [21], it is often dif®cult to obtain a clear DNA content histogram in analysis of bone and soft tissue tumors by ¯ow cytometry, due to a lot of debris. Moreover, Ross [15] also stated that a DNA histogram generated by pathologist-directed image analysis increases sensitivity. In this study, cyto¯uorometry was performed by orthopedic surgeons with backgrounds in pathology; therefore, we are sure that our DNA histograms are more accurate than those obtained by ¯ow cytometry. Recent active studies of DNA ploidy analysis in neoplasms [15±17], most of which utilized ¯ow cytometry, revealed that aneuploidy or polyploidy is closely associated with poor prognosis in various cancers such as non-small-cell lung cancer, colorectal cancer, breast cancer, ovarian cancer, prostatic cancer, bladder cancer and renal cancer. In most of the malignant bone and soft tissue tumors, patients with aneuploid tumors similarly have a poor outcome, regardless of chemotherapy or radiotherapy [18±24]. Kreicbergs [18] reported that osteosarcoma patients with aneuploid tumors detected by microphotometry showed poor prognosis. However, this study indicated that the event-free survival rate was signi®cantly higher in patients with non-diploid osteosarcoma than in patients with diploid osteosarcoma. Kreicbergs did not describe the methods of chemotherapy used, probably because most of his patients did not receive intensive chemotherapy, because his data were old. We presume that patients today received intensive pre and postoperative chemotherapy with wide tumor resection, and that chemotherapy may potentially explain the con¯icting results obtained in the present study. Look [25] demonstrated that the disease-free survival rate of

31

the patients (n ˆ 16) with near diploid osteosarcoma was signi®cantly (P ˆ 0:003) higher than that of the patients (n ˆ 10) with hyperdiploid osteosarcoma. Since all 26 of the osteosarcoma patients in his study received chemotherapy, his result should be the same as ours. However, the criteria used to de®ne near diploidy were totally different from the criteria of diploidy, used in the present study, because all of the diploid osteosarcomas except one had other aneuploid stem lines. Look, therefore, analysed 25 aneuploid and only one diploid osteosarcomas. Gebhardt [26] reported that the probability of remaining disease-free was signi®cantly better in the aneuploid group than in euploid±polyploid (most were diploid) group (P , 0:004, F test) in a ¯ow cytometric study of 48 osteosarcoma patients who receive chemotherapy, which corroborates our results. Additional papers have also shown that aneuploid tumors were more sensitive to radiotherapy [27,28] or chemotherapy [17,29]. Look [30] demonstrated that aneuploid neuroblastomas were more chemosensitive than diploid tumors. Although we do not have enough evidence that aneuploid osteosarcoma cells are really more sensitive to chemotherapy than diploid cells, the result of the study strongly suggested to us that this is the case. To prove that, DNA ploidy alteration after preoperative chemotherapy should be analysed. Chemosensitivity is the most important factor in the prediction of prognosis in cases of osteosarcoma; however, there are no clinically useful and practical methods of assessment available before chemotherapy despite the many reported methods [4±6]. Histologic response to chemotherapy is correlated with clinical outcome [1±3,9,10], but it is impossible to obtain results before surgery. Since pglycoprotein detection [31] or doxorubicin binding ability [32] were recently reported to be useful in the prediction of chemosensitivity and prognosis in cases of osteosarcoma, we are now studying the relationship between DNA ploidy pattern and these two factors. Finally, we concluded that the DNA ploidy pattern obtained from a biopsied tumor can predict the prognosis of osteosarcoma patients treated with chemotherapy and wide tumor resection and probably predict the chemosensitivity of osteosarcomas before preoperative chemotherapy.

32

K. Kusuzaki et al. / Cancer Letters 137 (1999) 27±33

References [1] G. Bacci, P. Picci, P. Ruggieri, M. Mercuri, M. Avella, R. Capanna, A. Brach Del Prever, A. Mancini, F. Gherlinzoni, G. Padovani, C. Leonessa, R. Biagini, A. Ferraro, A. Ferruzzi, A. Cazzola, M. Manfrini, M. Campanacci, Primary chemotherapy and delayed surgery (Neoadjuvant chemotherapy) for osteosarcoma of extremities. The Istituto Rizzoli experience in 127 patients treated preoperatively with intravenous methotrexate (high versus moderate doses) and intraarterial cisplatin, Cancer 65 (1990) 2539±2553. [2] P.A. Meyers, G. Heller, J. Healey, A. Huvos, J. Lane, R. Marcove, A. Applewhite, V. Valmis, G. Rosen, Chemotherapy for nonmetastatic osteogenic sarcoma: The Memorial Sloan-Kettering experience, J. Clin. Oncol. 10 (1992) 5±15. [3] G. Bacci, P. Picci, S. Ferrari, M. Orlandi, P. Ruggieri, R. Casadei, A. Ferraro, R. Biagini, A. Battistini, Prognostic signi®cance of serum alkaline phosphatase measurements in patients with osteosarcoma treated with adjuvant or neoadjuvant chemotherapy, Cancer 71 (1993) 1224±1230. [4] R. Taetle, J.M. Honeysett, F. Rosen, R. Shoemaker, Use of nude mouse xenografts as preclinical drug screens. Further studies on in vitro growth of xenograft tumor colony-forming cells, Cancer 58 (1986) 1969±1978. [5] S.E. Salmon, A.W. Hamburger, B. Soehnlen, B.G. Durie, D.S. Alberts, T.E. Moon, Quantitation of differential sensitivity of human-tumor stem cells to anticancer drugs, N. Engl. J. Med. 298 (1978) 1321±1327. [6] N.T. Bech-Hansen, F. Saragi, D.J.A. Sutherland, V. Ling, Rapid assay for evaluating the drug sensitivity of tumor cells, J. Natl. Cancer Inst. 59 (1977) 21±27. [7] D.D. Von Hoff, J. Casper, E. Bradley, J. Sandbach, D. Jones, R. Makuch, Association between human tumor colonyforming assay results and response of an individual patient's tumor to chemotherapy, Am. J. Med. 70 (1981) 1027±1032. [8] E. Brown, M. Markman, Tumor chemosensitivity and chemoresistance assays, Cancer 77 (1996) 1020±1025. [9] G. Delling, H. Krumme, M. Salzer-Kuntschik, Morphological changes in osteosarcoma after chemotherapy ± COSS 80, J. Cancer Res. Clin. Oncol. 106 (Suppl.) (1983) 32±37. [10] P. Picci, L. Sangiorgi, B.T. Rougraff, J.R. Neff, R. Casadei, M. Campanacci, Relationship of chemotherapy-induced necrosis and surgical margins to local recurrence in osteosarcoma, J. Clin. Oncol. 12 (1994) 2699±2705. [11] T. Ashihara, M. Kamachi, Y. Urata, K. Kusuzaki, H. Takashita, K. Kagawa, Multiparametric analysis using autostage cyto¯uorometry, Acta Histochem. Cytochem. 19 (1986) 51± 59. [12] K. Kusuzaki, T. Ashihara, Y. Hirasawa, Assessment of malignancy of bone and soft tissue tumors using DNA cyto¯uorometry, in: A. Uchida, K. Ono (Eds.), Recent Advances in Musculoskeletal Oncology, Springer-Verlag, Tokyo, 1992, pp. 221. [13] S. Hamada, R. Itoh, S. Fujita, DNA distribution pattern of the so-called severe dysplasias and small carcinomas of the colon

[14] [15] [16] [17] [18] [19]

[20] [21] [22] [23]

[24]

[25]

[26]

[27]

[28] [29]

and rectum and its possible signi®cance in the tumor progression, Cancer 61 (1988) 1555±1562. W.F. Enneking, Staging musculoskeletal tumors, in: W.F. Enneking (Ed.), Musculoskeletal Tumor Surgery, Churchill Livingstone, New York, 1983, pp. 69. J.S. Ross, DNA ploidy and cell cycle analysis in cancer diagnosis and prognosis, Oncology 10 (1996) 867±890. D.P. Magennis, Nuclear DNA in histological and cytological specimens: measurement and prognostic signi®cance, Br. J. Biomed. Sci. 54 (1997) 140±148. E. Douglas, D.E. Merkel, W.L. McGuier, Ploidy, proliferative activity and prognosis, Cancer 65 (1990) 1194±1205. A. Kreicbergs, L.A. Brostrom, G. Cewrien, S. Einhorn, Cellular DNA content in human osteosarcoma ± aspect on diagnosis and prognosis, Cancer 50 (1982) 2476±2481. H.J. Mankin, J.F. Connor, A.L. Schiller, N. Perlmutter, A. Alho, M. McGuire, Grading of bone tumor by analysis of nuclear DNA content using ¯ow cytometry, J. Bone Joint Surg. 67A (1985) 404±413. A. Kreicbergs, B. Tribukait, J. Willen, H.C.F. Bauer, DNA ¯ow analysis of soft tissue tumors, Cancer 59 (1987) 128±133. H.J. Mankin, M.C. Gebhardt, D.S. Spring®eld, G.J. Litwak, K. Kusuzaki, E. Rosenberg, Flow cytometric studies of human osteosarcoma, Clin. Orthop. 270 (1991) 169±180. A. Alho, S. Skjeldal, J. Melvik, E.O. Pettersen, T. Larsen, The clinical importance of DNA synthesis and aneuploidy in bone and soft tissue tumors, Anticancer Res. 13 (1993) 2383±2388. A.M. Dierick, M. Langlois, P. Van Oostveldt, H. Roels, The prognostic signi®cance of the DNA content in Ewing's sarcoma: a retrospective cytophotometric and ¯ow cytometric study, Histopathology 23 (1993) 333±339. L.C.D. Wijnaendts, J.C. van der Linden, van Diest, A.J.M. van Unnik, J.F.M. Delemarre, P.A. Voute, Prognostic importance of DNA ¯ow cytometric variables in rhabdomyosarcomas, J. Clin. Pathol. 46 (1993) 948±952. A.T. Look, E.C. Douglass, W.H. Meyerport, Clinical importance of near-diploid tumors stem lines in patients with osteosarcoma of an extremity, N. Engl. J. Med. 318 (1984) 1567± 1572. M.C. Gebhardt, D.S. Spring®eld, H. Takeshita, K. Kusuzaki, R. Lew, H.J. Mankin, DNA ploidy as a predictor of outcome in osteosarcoma and preliminary results of adriamycin binding and p-glycoprotein immunodetection in osteosarcoma biopsy specimens, in: J.F. Novak, J.H. McMaster (Eds.), Frontiers of Osteosarcoma Research, Hogrefe & Huber Publishers, Seattle, 1993, pp. 65. J.E.D. Dyson, C.A.F. Joslin, R.I. Rothwell, P. Quirke, G.G. Khoury, C.C. Bird, Flow cytometric evidence for the differential radiosensitivity of aneuploid and diploid cervix tumors, Radiother. Oncol. 8 (1987) 263±272. H. Wijkstrom, H. Gustafson, B. Tribukait, Deoxyribonucleic acid analysis in the evaluation of transitional cell carcinoma before cystectomy, J. Urol. 132 (1984) 894±898. D.R. Morgan, J.M.S. Williamson, P. Quicker, A.D. Clayden, M.E.F. Smith, C.J. O'Brien, D.L. Allison, J.A. Child, C.C. Bird, DNA content and prognosis non-Hodgkin's lymphoma, Br. J. Cancer 54 (1986) 643±649.

K. Kusuzaki et al. / Cancer Letters 137 (1999) 27±33 [30] A.T. Look, F.A. Hayes, R. Nitschke, N.B. McWilliams, A.A. Green, Cellular DNA content as a predictor of response to chemotherapy in infants with unresectable neuroblastoma, N. Engl. J. Med. 311 (1984) 231±235. [31] N. Baldini, K. Scotlandi, G. Barbanti-Brodano, M.C. Manara, D. Maurici, G. Bacci, F. Bertoni, P. Picci, S. Sottili, M.

33

Campanacci, M. Serra, Expression of P-glycoprotein in high-grade osteosarcomas in relation to clinical outcome, N. Engl. J. Med. 333 (1995) 1380±1385. [32] M.C. Gebhardt, K. Kusuzaki, H.J. Mankin, D.S. Spring®eld, An assay to measure adriamycin binding in osteosarcoma cells, J. Orthop. Res. 12 (1994) 621±627.