Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Springer 2005 Breast Cancer Research and Treatment (2005) 93: 35–40 DOI 10.1007/s10549-005-3381-1 Report Systemic effects of surgery: quantitative analysis of circulating basic fibroblast growth factor (bFGF), vascular endothelial growth factor (VEGF) and transforming growth factor beta (TGF-b) in patients with breast cancer who underwent limited or extended surgery G. Curigliano1,2, J.Y. Petit3, F. Bertolini4, M. Colleoni1, G. Peruzzotti2, F. de Braud2, S. Gandini5, A. Giraldo3, S. Martella3, L. Orlando1, E. Munzone1, E. Pietri1, A. Luini6, and A. Goldhirsch1 1 Department of Medicine, Division of Medical Oncology; 2Clinical Pharmacology and New Drugs Development Unit; Division of Plastic Surgery; 4Laboratory of Haemato-Oncology; 5Division of Epidemiology; 6Division of Senology, European Institute of Oncology, Milan, Italy 3 Key words: angiogenesis, breast cancer, bFGF, surgery, TGF-b, VEGF Summary Background. To assess if feature, extent and duration of surgery could influence levels of systemic proangiogenic cytokines vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF) and transforming growth factor beta (TGF-b). Patients and methods. We collected blood samples from 82 consecutive breast cancer patients who underwent various types of surgery, classified according to the magnitude of tissue injury in: minimal (quadrantectomy), moderate (mastectomy without reconstruction), and heavy [mastectomy followed by reconstruction with transversus recto-abdominal muscle cutaneous flap (TRAM)]. Samples were collected one day before surgery (D)1), at the end of surgical tumor removal (D0), and on 1st (D+1), 2nd (D+2) and 5th (D+5) day after surgery. Serum VEGF, bFGF and TGF-b levels were measured by the enzyme immunoassay method. Results. On average a continuous decrease was observed for all growth factors from the day before operation to the 5th day after operation. On day (D+5) an increase was observed for patients who underwent extended respect to moderate surgery. These differences were found statistically significant for bFGF and VEGF (p = 0.05 and p = 0.025 respectively). A statistically different trend for type of operation was observed also for TGF-b at 24–48 h: a minor reduction, compared to time of operation, was observed for minimal surgery, an intermediate reduction for moderate surgery and a higher decrease for extended surgery. Conclusions. Angiogenic cytokines perioperative levels could be increased on 5th day (D+5) by extent of surgery and should induce perioperative stimulation of residual cancer cells. A better understanding of the time interval during which the sequelae of events in wound healing occur may be the basis for defining new therapeutic strategies that can interfere with tumor outgrowth sparing wound healing processes. Introduction Surgery is still the main curative therapeutic modality for breast cancer. Angiogenesis plays a key role in both wound healing and the ability of a cancer to survive and grow. Investigations into the angiogenic response may help in guiding surgical approaches and in timing antiangiogenic therapy [1]. Normal wound repair generates an angiogenic response to deliver nutrients and inflammatory cells to injured tissue. The angiogenic response enables the removal of debris and is central to the development of a granulation tissue framework for wound closure [2]. The mediators of wound angiogenesis include soluble factors such as VEGF, tumor necrosis factor, TGF-b, bFGF, and platelet-derived growth factor (PDGF) identified in several wound models [3]. Angiogenic agonists (e.g. VEGF) and antagonists (e.g. thrombospondin-1) have been described at various times during repair [4–6], suggesting that the net angiogenic stimulus may be a balance of factors changing to favor either vessel growth or regression [7]. Previous studies have shown that surgical wound fluid collected within a few hours of operation is potently angiogenic. bFGF levels have been shown to peak immediately after surgery and then fall by the second post-operative day [8]. This immediate release has been suggested to function to 36 G Curigliano et al. initiate wound angiogenesis. In later wounds, VEGF is the predominant angiogenic mediator [4]. An up-regulation of VEGF production in wound repair has been demonstrated in keratinocytes in skin wounds in rat, guinea pig, and mouse models [9]. TGF family is involved in several steps of wound healing: monocyte chemoattraction, formation of granulation tissue and fibroblast stimulation, neovascularization, wound contraction and extracellular-matrix reorganization. The response of the body to a cancer is not a unique mechanism but has many parallels with inflammation and wound healing. It has been suggested that the inflammatory cells and cytokines found in tumors are more likely to contribute to tumor growth, progression, and immunosuppression than they are to mount an effective host antitumor response [10]. If genetic damage is the ‘match that lights the fire’ of cancer, some types of inflammation may provide the ‘fuel that feeds the flames’. Moreover cancer susceptibility and severity may be associated with functional polymorphisms of inflammatory cytokine genes, and deletion or inhibition of inflammatory cytokines inhibits development of experimental cancer. Work on the production of proangiogenic cytokines in early human wound fluid has been done using drain fluid from patients undergoing cancer surgery [4, 8]. These studies are based on the premise that wound fluid would be generally representative of the growth environment of the wound. However, none have studied if feature, extent and duration of surgery could affect systemic perioperative levels of angiogenic cytokines in patients with breast cancer. A better understanding of the time interval during which the sequelae of events in wound healing occur may be the basis for defining new therapeutic strategies that can interfere with tumor outgrowth sparing wound healing processes. After surgical resection of a tumor, the microenvironment of the wound site differs from that of normal tissue in several ways. Hypoxia, fibroblast activation, and various growth factors released after wounding make the wounded site different from nonwounded tissue. Patients undergoing major oncological resections might develop cytokine production disregulation and subsequent post-surgical immunosuppression, especially when the operation is of long duration. The aim of this study was to determine the effect of wounding on tumor growth. To better understand the mechanism of wound angiogenesis and its significance in tumor biology and surgical intervention, we specifically evaluated the temporal profile of serum VEGF, bFGF and TGF-b in breast cancer patients who underwent minimal, moderate or heavy surgery. European Institute of Oncology, Milan, between March 2001 and September 2001. All patients signed a written informed consent form approved by the Institutional Review Board of our Institute. Only those patients without any pre-operative treatment were included in the current study. Patients underwent minimal (quadrantectomy: Quad), moderate (mastectomy without reconstruction: Mast), and heavy surgery [mastectomy followed by reconstruction with transversus rectoabdominal muscle cutaneous flap: (TRAM)]. Pathological assessment included evaluation of the primary tumor size, histological type, lymph node status, including sentinel node biopsy procedure. Tumor grading and perivascular invasion have been assessed. Estrogen (ER) and progesterone (PgR) receptor status, Ki67 labeling index, determined with the MIB1 monoclonal antibody and HER2/neu overexpression was evaluated by Dako test (Dako, Glostrup, Denmark). Samples were obtained from peripheral vein one day before surgical resection (D)1), at the end of surgical tumor removal (D0), and on 1st (D+1), 2nd (D+2) and 5th (D+5) day after surgery. VEGF, bFGF and TGF-b levels were measured by the enzyme immunoassay method in serum samples. Approximately 10 ml of peripheral venous blood was drawn into a serum separator tube and centrifuged at 3000 revolutions per minute. Serum samples were aliquoted and preserved at )80 C until further use. Serum VEGF, bFGF, and TGF-b levels were measured using a commercially available competitive Enzyme Immunoassay (EIA) kit (Accucyte; Cytimmune Sciences, Inc., College Park, MD). Briefly, appropriate dilutions of standards (provided with the kit) and a 4-fold dilution of samples were added to the designated wells of a 96-well plates that were precoated with goat antirabbit antibody. Subsequently, competitive ligands and antibodies against the respective angiogenic factors were added serially and incubated for 3 h at room temperature. After five washings, the streptavidin–alkaline phosphatase complex was added into each well and incubated for 30 min at room temperature. Washing steps were repeated and color reagents (alcohol dehydrogenase and diaphorase) were added. Absorbance of the 0 dose was monitored at 492 nm (Molecular Devices Corporation, Sunnyvale, CA) and when the optical density reached between 1.5 and 2.0, the final reading was taken. Each sample was analyzed in duplicate and the average was considered as the final value. Data plotting and curve fitting (logit-log plot) were accomplished using a computer-assisted program (SOFTmax PRO, version 2.4, Molecular Corporation Devices, Sunnyvale, CA) on a MacIntosh computer. All samples and standards were assayed in duplicate. Patients and methods Statistical analysis Blood samples were collected prospectively from 84 consecutive pre- and post-menopausal patients who had presented with primary or recurrenced operable (T1–T4) node negative/positive (N0–N2), breast cancer to the Twenty-five observations presented measures for the growth factors below the minimum possible value therefore they have been replaced by the minimum 37 Surgery and angiogenesis in breast cancer values found in the datasets. All variables were graphically checked for normal distribution. Linear models are applied to verify statistical relevance of clinicopathological feaures on growth factors before the operation. Such analysis was restricted to patients with invasive lesions. Percentage changes from operation were evaluated at three time points: before operation, at 24–48 h after operation and at 5 days. Observations at 24 and 48 h were considered taking the latest observation because at 48 h after operation many missing values were found. For the same reason, at five days after operation growth factors, evaluated for quadrantectomy, were taken out from the analysis becase too many missing values were found and a problem of selection bias might arise. Comparisons between types of surgery were performed using linear models and adjusting for values at intervention (ANCOVA). Time effect was evaluated analyzing significance of changes from operation with t-test. Two sided p values £ 0.05 were considered significant. All calculations were performed using SAS software (SAS for Windows version 8.02. SAS Institute Inc. Cary NC USA, 2000). Table 2. Clinicopathologic features and median value of pre-operative serum angiogenic growth factors bFGF VEGF TGF-b (pg/ml) (pg/ml) (pg/ml) Median p Median 0.336 84.50 82.60 p Median p Age (years) <65 (n = 49) ‡65 (n = 7) 10.00 6.25 18.56 0.690 25.03 0.820 Tumor size (cm) <5 (n = 40) 7.38 81.34 18.31 ‡5 (n = 2) 9.91 0.775 185.50 0.361 20.99 0.735 Lymph node status No (n = 18) 4.50 Yes (n = 24) 9.55 Vascular invasion No (n = 27) Yes (n = 21) 0.617 85.60 22.54 68.00 0.390 15.84 8.14 82.60 20.95 11.51 0.305 104.10 0.898 18.11 0.828 0.256 ER & PgR status Negative (n = 15) 8.26 Positive (n = 38) 9.80 76.10 18.08 0.591 83.30 0.311 22.52 87.47 19.75 0.254 67.71 0.474 19.03 0.079 c-erb-B2 Results Forty-three (52%) patients underwent minimal (quadrantectomy), 18 (22%) moderate (mastectomy without reconstruction), and 21 (26%) heavy surgery [mastectomy followed by reconstruction with transversus rectoabdominal muscle cutaneous flap (TRAM)]. Median age of the patients was 51 years. The pre-operative median values (n = 82) of serum VEGF, bFGF and TGF-b levels were 84.50 pg/ml (range 14.97–573.66 pg/ml), 10.21 pg/ml (range 0.44–74.70 pg/ml), and 21.45 pg/ml (range 6.34–135.94 pg/ml), respectively for each type of surgery. Mean values at baseline (D)1), surgical removal Table 1. Mean value of angiogenic cytokines respect to timing of operation Pre- Post- 24 h 48 h 120 h Absent (n = 33) 9.87 Present (n = 19) 8.14 0.359 (D0), and on (D+1), (D+2) and (D+5) after surgery are reported in Table 1. Clinicopathologic features and median value of preoperative serum angiogenic growth factors are reported in Table 2. No relationship has been found between preoperative levels of angiogenic factors and features of disease. Plots of median values for VEGF, bFGF and TGF-b at each time point, related to different types of surgery, are presented in Figures 1, 2 and 3. Confidence intervals are not presented in the plots because they overlap. In Table 3 are reported median values of percentage changes of angiogenic factors in relation to extent of surgery for all time points from operation. Significance of extent of surgery and relevance of changes from operation are evaluated including p-values in the table. operative operative (pg/ml) (pg/ml) (pg/ml) (pg/ml) (pg/ml) TRAM 136.2 152.8 130.7 Mastectomy 139.9 190 208 VEGF VEGF Quadrantectomy 111.5 bFGF 102.1 94.3 81.6 111 78.1 158.0 85 250 TRAM Quadr. Mastect. 200 87.7 150 TRAM 15.7 14.9 14.1 8.0 15.8 Mastectomy 21.9 17.9 21.6 4.2 11.4 Quadrantectomy 111.5 102.1 94.3 78.1 87.7 50 0 100 TGF-b TRAM 24.3 18.7 14.8 9.4 13.3 Mastectomy 34.3 21.4 21.8 14.4 13.3 Quadrantectomy 25.4 17.2 18.6 15.1 6.9 Baseline Surgery 24 h 48 h 5 days Figure 1. Temporal trends of serum VEGF in breast cancer patients according to type of surgery. 38 G Curigliano et al. bFGF 30 TRAM Quadr. 25 Mastect. 20 15 10 5 0 Baseline Surgery 24 h 48 h 5 days Figure 2. Temporal trends of serum bFGF in breast cancer patients according to type of surgery. TGFbeta 40 35 TRAM Quadr. 30 Mastect. 25 20 15 10 5 0 Baseline Surgery 24 h 48 h 5 days Figure 3. Temporal trends of serum TGF-b in breast cancer patients according to type of surgery.. When comparing D)1 with D0, no significant differences among various types of surgery were observed for any angiogenic marker (data not reported). For TGF-b, a significant effect of time was detected; levels at D)1 were on average 48% higher than levels at D0 (Figure 3). Table 3. Median of percentage changes from operation for bFGF, VEGF and TGF-b according to extent and timing from surgery n bFGF VEGF TGF-b D)1 n 24–48 h n 5 days Quad 43 4 39 19 4 NA Mast 18 54 18 )29 13 )40 TRAM 21 6 19 )58 12 80 Overall 82 8 76 )12 29 )6 p for time 0.338 0.526 p for surgery Quad 43 0.170 9 0.329 )6 Mast 18 )20 18 )24 13 TRAM 20 4 19 )42 12 117 Overall 81 2 75 )23 30 )33 38 p for time 0.428 0.028 p for surgery 0.87 0.468 0.892 4 0.050 NA )47 0.167 0.025 Quad 42 27 41 )18 5 NA Mast TRAM 18 21 71 57 18 20 )38 )44 12 12 )38 20 Overall 81 48 79 )29 29 )27 p for time p for surgery <0.0001 0.001 0.062 0.321 0.005 0.174 At 24–48 h after surgery (D+1–D+2), a percentage decrease from D0 was observed for all growth factors. The percentage change was statistically significant for TGF-b and VEGF (on average a 29 and 23% reduction with p = 0.001 and p = 0.028, respectively). Looking at trends by type of surgery, it can be observed that the percentage reduction for any angiogenic factor was higher for mastectomy and TRAM and lower for quadrantectomy. In patients who underwent quadrantectomy bFGF showed, on average, even an increase (19%) at D+1– D+2, a but this change was not statistically different respect to the reduction found for mastectomy and TRAM. Any type of surgery was associated with significant (p = 0.005) reduction of TGF-b at D+1–D+2. Reduction of perioperative levels of TGF-b is higher for TRAM and mastectomy than for quadrantectomy (p = 0.005). On D+5 after surgery, we observed an increase of all angiogenic factors in patient who underwent TRAM, and a reduction for patients who underwent mastectomy. The increase in patients who underwent TRAM was lower for TGF-b; in fact the difference by type of operation was statistically significant only for bFGF and VEGF (at 5 and 2.5%, respectively). Discussion Tumor growth is angiogenesis dependent. Perioperative levels of endogenous stimulators (bFGF, VEGF, cathepsin, copper, interleukin 1, 6 and 8), inhibitors (plasminogen activator inhibitor-1, tissue inhibitor of metalloprotease, zinc, interleukin 10 and 12) and modulators of angiogenesis (TGF-b, tumor necrosis factor alpha) may indicate the switch to the angiogenic phenotype of neoplasia that depends on a net balance between positive and negative angiogenic factors released by the tumor. In particular, there is an alteration in the circulating levels of acute phase reactants that are believed to play an important role in the perioperative period, at the time of enhanced release of malignant cells into the circulation, with risk of metastasis induction. Endothelial growth factors have a cell mitogen effect and act as regulator of vascular permeability. Several retrospective studies reported that VEGF is significantly associated with relapse-free survival, overall survival, or both. Patients with early stage breast cancer who have tumors with elevated levels of VEGF, TGF-b or bFGF have a higher likelihood of recurrence than patients with low-angiogenic tumors, even if treated with conventional adjuvant therapy [1]. Preoperative levels of VEGF, bFGF and TGF-b described in our study are similar to previously reported [11–13]. Other studies reported a correlation between clinical pathological features of disease and preoperative levels of angiogenic factors [14]. In our study, no relationship has been observed between age, stage, biological features and levels of preoperative angiogenic factors. Median values Surgery and angiogenesis in breast cancer of VEGF, bFGF and TGF-b usually have a drop out at 24–48 h after surgery. Reduction of TGF-b levels form pre-operative to post-operative time was statistically significant. Kong et al. [13] showed that plasma TGF-b levels were elevated pre-operatively in 81% of the patients. The mean plasma TGF-b level in breast cancer patients normalized after surgery (19.3 ± 3.2 versus 5.5 ± 1.0 ng/ml, p < 0.001) in the majority of subjects; levels were persistent in case of presence of lymph node metastases or overt residual tumor. No data are reported on bFGF in correlation to timing or extent of surgery in patients with breast cancer. An overall percentage change with 23% reduction after surgery has been described for VEGF. In a previous report [15], a significant change in serum VEGF levels compared with pre-operative values has been described with time with an initial drop over the first 3 days; thereafter levels recovered. In this study has been also performed analysis of the local wound response, showing that VEGF levels in the wound environment are much higher than the serum equivalent from as early as the first postoperative day. They then peak at day 2 and remain at a higher level thereafter for several days. This observation fits well with wound vascular mechanisms in animal models. Acute wound response could act as a ‘molecular trap’ for angiogenic factors and reduction of serum levels of VEGF, TGF-b and bFGF in our patients could be a result of this ‘trapping’. Another explanation for reduction of angiogenic factors, especially for VEGF, should be considered platelet count drop after surgical injury. Since platelets are the main source of serum VEGF, it can be reasonable to suppose that their trapping in wound healing could explain VEGF levels drop after surgery. The surgical wound itself is a unique extravascular compartment with increased vascular permeability and a high surface area:volume ratio. If reabsorption occurs freely from the surgical wound site changes in local VEGF concentrations should be reflected in the circulation, i.e. serum levels. Subsequent increase of VEGF (specifically in patients who underwent TRAM surgery) should be related to massive local wound VEGF production. TRAM surgery creates a wound with a larger surface area than wide local excision. This effect may mark an interaction between residual tumor-derived local inhibitors resulting in an initially depressed normal stromal angiogenic response that recovers over time. This would be in keeping with the evidence that tumor cells secrete factors that provide negative ‘feedback’ regulation and serve to suppress vascular growth, restraining the growth of secondary tumors or metastases [25–27]. Surgical clearance of cancer involves regional extirpation, and residual tissues may still be under the influence of tumor-derived inhibitors delaying the normal angiogenic wound response. The mechanism underlying these observations requires additional investigation and may relate to the half-life of angiogenic stimulators compared with inhibitors, to a local effect on the stroma when the angiogenic drive from the tumor is removed or because 39 of impaired influx of blood and platelet release at the time of injury. This muted response in cancer patients may represent an opportunity to complete surgical treatment while minimizing stimulation of metastatic disease, a biological argument in favor of immediate reconstruction after, for example, breast cancer surgery. Experimental evidence suggests that a growth factorrich environment enables the survival of cancer cells left in an area of cancer extirpation or in the circulation [25–27]. However, as wounds age the surgical site becomes less favorable to tumor implantation, and when healing is complete injected tumor cells do not localize to the surgical site [26]. Thus, local recurrence found in conjunction with widespread metastatic disease is likely to have been established by perioperative seeding rather than as a late phenomenon. Furthermore, a growth factor-stimulated microenvironment affects growth of established residual tumor foci in vivo and cell lines in vitro [27]. Antiangiogenic therapy is currently undergoing clinical trials, and in the future perioperative systemic therapy or local therapy may include use of such therapy. However, our findings indicate that very high local concentrations may need to be antagonized. Quantification of this in vivo biological response should facilitate the design of wound healing experiments to more closely represent the ‘human’ response to surgical stress. Drainage systems may offer an opportunity to manipulate the early wound environment and reduce local cancer recurrence rates in the future. A better understanding of the time interval during which the sequelae of events in wound healing occur may be the basis for defining new therapeutic strategies that can interfere with tumor outgrowth sparing wound healing processes. References 1. Reid SE, Kaufman MW, Murthy S, Scanlon EF: Perioperative stimulation of residual cancer cells promotes local and distant recurrence of breast cancer. J Am Coll Surg 185: 290–306, 1997 2. Battegay EJ: Angiogenesis: mechanistic insights, neovascular diseases, and therapeutic prospects. J Mol Med 73: 333–346, 1995 3. Orredson SU, Knighton DR, Scheuenstuhl H, Hunt TK: A quantitative in vitro study of fibroblast and endothelial cell migration in response to serum and wound fluid. J Surg Res 35: 249–258, 1983 4. Nissen NN, Polverini PJ, Koch AE, Volin MV, Gamelli RL, DiPietro LA: Vascular endothelial growth factor mediates angiogenic activity during the proliferative phase of wound healing. Am J Pathol 152: 1445–1452, 1998 5. Reed MJ, Puolakkainen P, Lane TF, Dickerson D, Bornstein P, Sage EH: Differential expression of SPARC and thrombospondin1 in wound repair: immunolocalization and in situ hybridization. J Histochem Cytochem 41: 1467–1477, 1993 6. Di Pietro LA, Nissen NN, Gamelli RL, Koch AE, Pyle JM, Polverini PJ: Thrombospondin 1 synthesis and function in wound repair. Am J Pathol 148: 1851–1860, 1996 7. Martin P: Wound healing–aiming for perfect skin regeneration. Science (Wash. DC) 276: 75–81, 1997 8. Nissen NN, Polverini PJ, Gamelli RL, DiPietro LA: Basic fibroblast growth factor mediates angiogenic activity in early surgical wounds. Surgery (St. Louis) 119: 457–465, 1996 40 G Curigliano et al. 9. Brown LF, Yeo KT, Berse B, Yeo TK, Senger DR, Dvorak HF, van de WL: Expression of vascular permeability factor (vascular endothelial growth factor) by epidermal keratinocytes during wound healing. J Exp Med 176: 1375–1379, 1992 10. Balkwill F, Mantovani A: Inflammation and cancer: back to Virchow? The Lancet, 2001 11. Heer K, Kumar H, Read JR, Fox JN, Monson JRT, Kerin MJ: Serum vascular endothelial growth factor in breast cancer. Its relation with cancer type and estrogen receptor status. Clin Cancer Res 7: 3491–3494, 2001 12. Sliutz G, Tempfer C, Obermair A, Dadak C, Kainz C: Serum evaluation of basic FGF in breast cancer patients. Anticancer Res 15(6B). 2675–2677, 1995 13. Kong FM, Anscher MS, Murase T, Abbott BD, Iglehart JD, Jirtle RL: Elevated plasma transforming growth factor-beta 1 levels in breast cancer patients decrease after surgical removal of the tumor. Ann Surg 222(2): 155–162, 1995 14. Sancak B, Coskun U, Gunel N, Onuk E, Cihan A, Karamercan A, Yildirim Y, Ozkan S: No association between serum levels ofinsulin-like growth factor-I, vascular endothelial growth factor, prolactin and clinicopathological characteristics of breast carcinoma after surgery. Intern Med J 34(6): 310–315, 2004 15. Hormbrey E, Han C, Roberts A, McGrouther DA, Harris AL: The relationship of human wound vascular endothelial growth factor (VEGF) after breast cancer surgery to circulating VEGF and angiogenesis. Clin Cancer Res 9: 4332–4339, 2003 16. O’Reilly MS, Boehm T, Shing Y, Fukai N, Vasios G, Lane WS, Flynn E, Birkhead JR, Olsen BR, Folkman J: Endostatin: an endogenous inhibitor of angiogenesis and tumor growth. Cell 88: 277–285, 1997 17. O’Reilly MS, Holmgren L, Shing Y, Chen C, Rosenthal RA, Cao Y, Moses M, Lane WS, Sage EH, Folkman J: Angiostatin: a circulating endothelial cell inhibitor that suppresses angiogenesis and tumor growth. Cold Spring Harb Symp Quant Biol 59: 471–482, 1994 18. O’Reilly MS, Holmgren L, Shing Y, Chen C, Rosenthal RA, Moses M, Lane WS, Cao Y, Sage EH, Folkman J: Angiostatin: a 19. 20. 21. 22. 23. 24. 25. 26. 27. novel angiogenesis inhibitor that mediates the suppression of metastases by a Lewis lung carcinoma. Cell 79: 315–328, 1994 Skipper D, Jeffrey MJ, Cooper AJ, Taylor I, Alexander P: Preferential growth of bloodborne cancer cells in colonic anastomoses. Br J Cancer 57: 564–568, 1988 Whalen GF, Ingber DE: Inhibition of tumor-cell attachment to extracellular matrix as a method for preventing tumor recurrence in a surgical wound. Ann Surg 210: 758–764, 1989 Davies DE, Farmer S, White J, Senior PV, Warnes SL, Alexander P: Contribution of host-derived growth factors to in vivo growth of a transplantable murine mammary carcinoma. Br J Cancer 70: 263–269, 1994 Murthy SM, Goldschmidt RA, Rao LN, Ammirati M, Buchmann T, Scanlon EF: The influence of surgical trauma on experimental metastasis. Cancer (Phila.) 64: 2035–2044, 1989 Fisher B, Fisher ER: Experimental studies of factors influencing hepatic metastases. III. Effect of surgical trauma with special reference to liver injury. Ann Surg 150: 731–744, 1959 Gunduz N, Fisher B, Saffer EA: Effect of surgical removal on the growth and kinetics of residual tumor. Cancer Res 39: 3861–3865, 1979 Simpson-Herren L, Sanford AH, Holmquist JP: Effects of surgery on the cell kinetics of residual tumor. Cancer Treat Rep 60: 1749– 1760, 1976 Fisher B, Gunduz N, Coyle J, Rudock C, Saffer E: Presence of a growth-stimulating factor in serum following primary tumor removal in mice. Cancer Res 49: 1996–2001, 1989 Blitstein-Willinger E: The role of growth factors in wound healing. Skin Pharmacol 4: 175–182, 1991 Address for offprints and correspondence: G. Curigliano, Division of Medical Oncology, Clinical Pharmacology and New Drugs Development Unit, European Institute of Oncology, Via Ripamonti 435, 20141 Milano, Italy; Tel.: 39.02.57489482; Fax: 39.02.57489.581; E-mail: [email protected]