1984
CORRESPONDENCE
institution investigation review board of the university or center and that demand the written informed consent of the study participants remove any question of ethics. Furthermore, in our experience, it is very rare for a fully informed patient to prefer not to participate in a study that requires performance of marrow biopsies. We believe that the characterization of cells obtained from disaggregated spicules as intact bone marrow specimens is incorrect. By definition, these specimens are not intact. That the method of disaggregating spicules provides marrow cells free of contaminating peripheral-blood cells must be rigorously evaluated. Also, is there evidence that the procedure does not also remove some of the cells of interest, as well as that the disaggregation of the spicules results in the release of all cells in the spicule without damaging or destroying some of tl - cells? The study by Shackney et all that demonstrated a gradien, of proliferation that extended from the subendosteum to the center of the marrow suggests that it is important that the procedure used to obtain the spicules does not selectively remove spicules from different marrow regions. Has this been demonstrated? Furthermore, the use of aspirated spicules precludes the examination of subendosteal proliferation, an area in which cell proliferative rates appear to be the highest. 1 The only way to resolve each of these questions is to compare the data obtained with the spicule method with the data provided by the 2 intact biopsy method. Finally, even if all the above are resolved in the affirmative, the method described by Riccardi et al provides mean population data. Recent studies of acute myelogenous leukemia have demonstrated that proliferating leukemia cells occur as geographically restricted islands of proliferation (GRIPS) in the marrows of the majority of patients with French-American-British M3 and in a minority of patients with other types of acute myelogenous leukemia. 3 A similar situation is present in the marrows of many patients with refractory anemia with excess blasts in which the most actively proliferating cells are present in clusters.4 Recognition of these unique geographic patterns of cell proliferation was possible only because intact marrow biopsy specimens were studied. If the method described by Riccardi et al had been employed, these observations could not have been made because the geographic patterns would have been disrupted. Also, because the proliferative rates of cells are different in different regions of the marrow biopsy, any method that destroys the marrow architecture that provides a mean population value cannot provide an accurate picture of the pattern or rate of cell proliferation. Recent studies that used intact biopsy specimens have demonstrated not only apparent relationships between cytokine distribution in the microenvironment and leukemia cell proliferation,5 but also that cytokine administration to patients can affect the distribution of cytokines in the microenvironment and can also affect infiltration of the marrow by macrophages.6 These aspects of cellular proliferation and the effects of cytokine administration can be studied only if intact marrow biopsies are used. For the above reasons, when normal or leukemia marrow is studied, we believe that truly intact marrow biopsy specimens must be studied if an accurate and complete assessment of normal or leukemia cell proliferation is desired. Azra Raza Harvey D. Preisler University of CincinnatiMedical Center Cincinnati, OH
REFERENCES 1. Shackney SE, Ford SS, Wittig AB: Kinetic-microarchitectural correlations in the bone marrow of the mouse. Cell Tissue Kinet 8:505-516, 1975 2. Raza A, Yousuf N, Bokhari SAJ, et al: In situ cell cycle kinetics in bone marrow biopsies following sequential infusion of IUdR/BrdU in patients with hematopoietic malignancies. Leuk Res 16:299-306, 1992 3. Raza A, Yousuf N, Umerani A, et al: High expression of TGF-B, long cell cycle times and a unique clustering of S-phase cells in patients with acute promyelocytic leukemia. Blood 79:10371048, 1992 4. Raza A, Bokhari J, Yousuf N, et al: Cell cycle kinetic studies in human cancers. Arch Pathol Lab Med 115:873-879, 1991 5. Qadir K, Umerani A, Sanghvi R, et al: Significance of finding large numbers of macrophages (M) in bone marrow (BM) biopsies of patients with acute myelocytic leukemia (AML). Proc Am Assoc Cancer Res 33:76, 1992 (abstr 454)
Suramin: Is Adaptive Control Necessary? To the Editor: In the June issue of the Journal of Clinical Oncology, Myers et all state that ". . . major emphasis should be placed on the development of simplified methods for the safe and effective administration of [suramin]." We agree with this assertion. However, we believe that the evidence is not sufficiently compelling to warrant the routine use of adaptive control in the administration of suramin. Accordingly, we are conducting a phase I dose-escalation study of intermittent infusional administration of suramin based on pharmacologic principles, but without pharmacokinetic monitoring, to identify the maximal-tolerated dose and the pattern of dose-limiting toxicity. Suramin is given by intravenous infusion over 1 hour, beginning with 100% of the designated dose on day 1 and continuing in graded decrements on days 2, 8, and 9 of a 28-day cycle. Based on what is known about its pharmacokinetics, this scheme will allow a gradual approach to the steady state. Doses are increased in successive cohorts such that, over the first month, the first cohort received 900 mg/m 2, the second 1,440 mg/m2, and so on. Pharmacokinetic data are being collected, but not used in treatment decisions. Eleven patients, nine with prostate cancer and two with renal cell carcinoma, have been treated at two dose levels. Mean peak plasma levels were 127 + 67 pLg/mL and 248 t 59 Vg/mL for the first and second dose levels, respectively. Mean trough levels before cycle 2 were 14 t 4 mg/mL for the first cohort, and 32 t 6 jwg/mL for the second. The highest peak observed was 339 Cpg/mL. Four patients with prostate cancer have achieved initial partial responses and one patient developed objectively documented lower extremity weakness associated with hypophosphatemia, although no evidence of demyelination was found on electromyography or nerve conduction velocity testing. The recommendations for adaptive control of the administration of suramin arise from three concerns: that peak plasma levels of greater than 300 ig/mL predispose to the development of severe peripheral neuropathy similar to the Guillain-Barr6 syndrome,2 that low suramin levels stimulate tumor growth in vitro by a tumor growth factor-mediated mechanism, 3 and that interindividual variations in suramin blood levels would render the pharmacokinetics of the drug unpredictable. We have found relatively little interindividual variation in clearance among our patients, and no instances
Downloaded from ascopubs.org by RB DRAUGHTON LIBRARY on April 28, 2019 from 131.204.073.184 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
1985
CORRESPONDENCE of the demyelinating form of neurotoxicity at the doses tested to date. We also have seen encouraging preliminary evidence that suramin's antitumor activity is maintained with this dosing schedule. Thus, our experience to date suggests that suramin can be given safely without monitoring at doses that are sufficient to achieve antitumor effects in patients with hormone-refractory prostate cancer. K. Kobayashi E.E. Vokes K. Stefansson L. Janisch N.J. Vogelzang F. Berezin M.J. Ratain University of Chicago Chicago,IL REFERENCES 1. Myers CA, Cooper M, Stein C, et al: Suramin: A novel growth factor antagonist with activity in hormone-refractory metastatic prostate cancer. J Clin Oncol 10:881-889, 1992 2. LaRocca R, Meer J, Gilliatt RW, et al: Suramin-induced polyneuropathy. Neurology 40:954-960, 1990 3. Cardinali M, Sartor O, Robbins KC: Suramin, an experimental chemotherapeutic drug, activates the receptor for epidermal growth factor and promotes growth of certain malignant cells. J Clin Invest 89:1242-1247, 1992
Suramin for Prostate Carcinoma To the Editor: The editorial by Pinedo and Van Rijswijk' concludes that "there is no doubt that suramin represents an important new group of anticancer agents." As this conclusion is based on its activity in metastatic prostate cancer, I must disagree. Both the authors of this editorial and those of the substantive report2 in the same issue concede that coadmninistered hydrocortisone can produce symptomatic responses. But it is also true that 3 decreases in the level of acid phosphatase and the level of prostate-specific antigen 4 are observed with hydrocortisone therapy seemingly in a similar proportion of cases. Doubt remains. S.J. Harland Instituteof Urology & Nephrology London, United Kingdom REFERENCES 1. Pinedo HM, Van Rijswijk REN: Suramin awakes? J Clin Oncol 10:875-877, 1992 (editorial) 2. Myers C, Cooper M, Stein C, et al: Suramin: A novel growth factor antagonist with activity in hormone-refractory metastatic prostate cancer. J Clin Oncol 10:881-889, 1992 3. Miller GM, Hinman F: Cortisone treatment in advanced carcinoma of prostate. J Urol 72:485-496, 1954 4. Harland SJ, Duchesne GM: Suramin and prostate cancer: The role of hydrocortisone. Eur J Cancer 28:1295, 1992 In Reply: The preliminary data reported by Dr Kobayashi and colleagues that suggest that relatively modest doses of suramin retain activity in the treatment of hormone-refractory prostate
cancer are encouraging and consistent with unpublished data from other groups investigating the use of suramin in this disease. Treatment with suramin obviously would be simplified and could be expanded to larger groups of patients if its optimal use did not require therapeutic drug monitoring. However, rather than stating that the routine use of adaptive control with feedback in the administration of suramin is not warranted, we would prefer to state that the best way to use suramin in the treatment of prostate cancer is as yet unknown. While there are data to indicate that sustained plasma concentrations in excess of 350 pig/mL are associated with the development of neurologic toxicity, the lower bound of effective concentrations has not been established. Such information could be obtained from a concentration-controlled trial wherein different cohorts of patients are treated at different mean plasma concentrations of suramin, but from a practical standpoint it might be more reasonable simply to randomize patients between a fixed regimen of suramin doses that have been shown to be safe and a more intensive regimen requiring adaptive control with feedback. Just as it would be premature to draw final conclusions about the efficacy of a dosing regimen explored in a cohort of only nine patients, so we would caution against drawing conclusions about the interpatient pharmacokinetic variability of a drug from data from only 11 patients. We recently have published our experience with the population pharmacokinetics of suramin in an initial 1 cohort of 51 patients, and the inclusion of an additional 66 treated at the National Cancer Institute (NCI) in our data patients base has made no substantial alteration in either the mean values for suramin's pharmacokinetic parameters or their coefficients of variation. With this information it is possible to anticipate the variability that will result from various suramin dosing regimens. Figure 1 represents the fourth dosing level of the phase I trial being carried out at the University of Chicago, wherein patients 2 receive a dose of 2,400 mg/m during the first month of therapy. This dosing level yields a dose intensity that is approximately half
600 500 E 400 E
300 200 100 0 1
5
9
17
13
21
25
Days Fig 1. Simulation of population variability in plasma suramin concentrations resulting from the University of Chicago dosing regimen (level 4). The open squares represent the mean population suramin concentrations for this dose and schedule, and the error bars represent the 95% confidence limits of the anticipated population response. To avoid overstating suramin's interpatiient pharmacokinetic variability, the 95% confidence limits depicted here do not include the variability inherent in the drug assay (coefficient of variation = 7%).
Downloaded from ascopubs.org by RB DRAUGHTON LIBRARY on April 28, 2019 from 131.204.073.184 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
CORRESPONDENCE
institution investigation review board of the university or center and that demand the written informed consent of the study participants remove any question of ethics. Furthermore, in our experience, it is very rare for a fully informed patient to prefer not to participate in a study that requires performance of marrow biopsies. We believe that the characterization of cells obtained from disaggregated spicules as intact bone marrow specimens is incorrect. By definition, these specimens are not intact. That the method of disaggregating spicules provides marrow cells free of contaminating peripheral-blood cells must be rigorously evaluated. Also, is there evidence that the procedure does not also remove some of the cells of interest, as well as that the disaggregation of the spicules results in the release of all cells in the spicule without damaging or destroying some of tl - cells? The study by Shackney et all that demonstrated a gradien, of proliferation that extended from the subendosteum to the center of the marrow suggests that it is important that the procedure used to obtain the spicules does not selectively remove spicules from different marrow regions. Has this been demonstrated? Furthermore, the use of aspirated spicules precludes the examination of subendosteal proliferation, an area in which cell proliferative rates appear to be the highest. 1 The only way to resolve each of these questions is to compare the data obtained with the spicule method with the data provided by the 2 intact biopsy method. Finally, even if all the above are resolved in the affirmative, the method described by Riccardi et al provides mean population data. Recent studies of acute myelogenous leukemia have demonstrated that proliferating leukemia cells occur as geographically restricted islands of proliferation (GRIPS) in the marrows of the majority of patients with French-American-British M3 and in a minority of patients with other types of acute myelogenous leukemia. 3 A similar situation is present in the marrows of many patients with refractory anemia with excess blasts in which the most actively proliferating cells are present in clusters.4 Recognition of these unique geographic patterns of cell proliferation was possible only because intact marrow biopsy specimens were studied. If the method described by Riccardi et al had been employed, these observations could not have been made because the geographic patterns would have been disrupted. Also, because the proliferative rates of cells are different in different regions of the marrow biopsy, any method that destroys the marrow architecture that provides a mean population value cannot provide an accurate picture of the pattern or rate of cell proliferation. Recent studies that used intact biopsy specimens have demonstrated not only apparent relationships between cytokine distribution in the microenvironment and leukemia cell proliferation,5 but also that cytokine administration to patients can affect the distribution of cytokines in the microenvironment and can also affect infiltration of the marrow by macrophages.6 These aspects of cellular proliferation and the effects of cytokine administration can be studied only if intact marrow biopsies are used. For the above reasons, when normal or leukemia marrow is studied, we believe that truly intact marrow biopsy specimens must be studied if an accurate and complete assessment of normal or leukemia cell proliferation is desired. Azra Raza Harvey D. Preisler University of CincinnatiMedical Center Cincinnati, OH
REFERENCES 1. Shackney SE, Ford SS, Wittig AB: Kinetic-microarchitectural correlations in the bone marrow of the mouse. Cell Tissue Kinet 8:505-516, 1975 2. Raza A, Yousuf N, Bokhari SAJ, et al: In situ cell cycle kinetics in bone marrow biopsies following sequential infusion of IUdR/BrdU in patients with hematopoietic malignancies. Leuk Res 16:299-306, 1992 3. Raza A, Yousuf N, Umerani A, et al: High expression of TGF-B, long cell cycle times and a unique clustering of S-phase cells in patients with acute promyelocytic leukemia. Blood 79:10371048, 1992 4. Raza A, Bokhari J, Yousuf N, et al: Cell cycle kinetic studies in human cancers. Arch Pathol Lab Med 115:873-879, 1991 5. Qadir K, Umerani A, Sanghvi R, et al: Significance of finding large numbers of macrophages (M) in bone marrow (BM) biopsies of patients with acute myelocytic leukemia (AML). Proc Am Assoc Cancer Res 33:76, 1992 (abstr 454)
Suramin: Is Adaptive Control Necessary? To the Editor: In the June issue of the Journal of Clinical Oncology, Myers et all state that ". . . major emphasis should be placed on the development of simplified methods for the safe and effective administration of [suramin]." We agree with this assertion. However, we believe that the evidence is not sufficiently compelling to warrant the routine use of adaptive control in the administration of suramin. Accordingly, we are conducting a phase I dose-escalation study of intermittent infusional administration of suramin based on pharmacologic principles, but without pharmacokinetic monitoring, to identify the maximal-tolerated dose and the pattern of dose-limiting toxicity. Suramin is given by intravenous infusion over 1 hour, beginning with 100% of the designated dose on day 1 and continuing in graded decrements on days 2, 8, and 9 of a 28-day cycle. Based on what is known about its pharmacokinetics, this scheme will allow a gradual approach to the steady state. Doses are increased in successive cohorts such that, over the first month, the first cohort received 900 mg/m 2, the second 1,440 mg/m2, and so on. Pharmacokinetic data are being collected, but not used in treatment decisions. Eleven patients, nine with prostate cancer and two with renal cell carcinoma, have been treated at two dose levels. Mean peak plasma levels were 127 + 67 pLg/mL and 248 t 59 Vg/mL for the first and second dose levels, respectively. Mean trough levels before cycle 2 were 14 t 4 mg/mL for the first cohort, and 32 t 6 jwg/mL for the second. The highest peak observed was 339 Cpg/mL. Four patients with prostate cancer have achieved initial partial responses and one patient developed objectively documented lower extremity weakness associated with hypophosphatemia, although no evidence of demyelination was found on electromyography or nerve conduction velocity testing. The recommendations for adaptive control of the administration of suramin arise from three concerns: that peak plasma levels of greater than 300 ig/mL predispose to the development of severe peripheral neuropathy similar to the Guillain-Barr6 syndrome,2 that low suramin levels stimulate tumor growth in vitro by a tumor growth factor-mediated mechanism, 3 and that interindividual variations in suramin blood levels would render the pharmacokinetics of the drug unpredictable. We have found relatively little interindividual variation in clearance among our patients, and no instances
Downloaded from ascopubs.org by RB DRAUGHTON LIBRARY on April 28, 2019 from 131.204.073.184 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
1985
CORRESPONDENCE of the demyelinating form of neurotoxicity at the doses tested to date. We also have seen encouraging preliminary evidence that suramin's antitumor activity is maintained with this dosing schedule. Thus, our experience to date suggests that suramin can be given safely without monitoring at doses that are sufficient to achieve antitumor effects in patients with hormone-refractory prostate cancer. K. Kobayashi E.E. Vokes K. Stefansson L. Janisch N.J. Vogelzang F. Berezin M.J. Ratain University of Chicago Chicago,IL REFERENCES 1. Myers CA, Cooper M, Stein C, et al: Suramin: A novel growth factor antagonist with activity in hormone-refractory metastatic prostate cancer. J Clin Oncol 10:881-889, 1992 2. LaRocca R, Meer J, Gilliatt RW, et al: Suramin-induced polyneuropathy. Neurology 40:954-960, 1990 3. Cardinali M, Sartor O, Robbins KC: Suramin, an experimental chemotherapeutic drug, activates the receptor for epidermal growth factor and promotes growth of certain malignant cells. J Clin Invest 89:1242-1247, 1992
Suramin for Prostate Carcinoma To the Editor: The editorial by Pinedo and Van Rijswijk' concludes that "there is no doubt that suramin represents an important new group of anticancer agents." As this conclusion is based on its activity in metastatic prostate cancer, I must disagree. Both the authors of this editorial and those of the substantive report2 in the same issue concede that coadmninistered hydrocortisone can produce symptomatic responses. But it is also true that 3 decreases in the level of acid phosphatase and the level of prostate-specific antigen 4 are observed with hydrocortisone therapy seemingly in a similar proportion of cases. Doubt remains. S.J. Harland Instituteof Urology & Nephrology London, United Kingdom REFERENCES 1. Pinedo HM, Van Rijswijk REN: Suramin awakes? J Clin Oncol 10:875-877, 1992 (editorial) 2. Myers C, Cooper M, Stein C, et al: Suramin: A novel growth factor antagonist with activity in hormone-refractory metastatic prostate cancer. J Clin Oncol 10:881-889, 1992 3. Miller GM, Hinman F: Cortisone treatment in advanced carcinoma of prostate. J Urol 72:485-496, 1954 4. Harland SJ, Duchesne GM: Suramin and prostate cancer: The role of hydrocortisone. Eur J Cancer 28:1295, 1992 In Reply: The preliminary data reported by Dr Kobayashi and colleagues that suggest that relatively modest doses of suramin retain activity in the treatment of hormone-refractory prostate
cancer are encouraging and consistent with unpublished data from other groups investigating the use of suramin in this disease. Treatment with suramin obviously would be simplified and could be expanded to larger groups of patients if its optimal use did not require therapeutic drug monitoring. However, rather than stating that the routine use of adaptive control with feedback in the administration of suramin is not warranted, we would prefer to state that the best way to use suramin in the treatment of prostate cancer is as yet unknown. While there are data to indicate that sustained plasma concentrations in excess of 350 pig/mL are associated with the development of neurologic toxicity, the lower bound of effective concentrations has not been established. Such information could be obtained from a concentration-controlled trial wherein different cohorts of patients are treated at different mean plasma concentrations of suramin, but from a practical standpoint it might be more reasonable simply to randomize patients between a fixed regimen of suramin doses that have been shown to be safe and a more intensive regimen requiring adaptive control with feedback. Just as it would be premature to draw final conclusions about the efficacy of a dosing regimen explored in a cohort of only nine patients, so we would caution against drawing conclusions about the interpatient pharmacokinetic variability of a drug from data from only 11 patients. We recently have published our experience with the population pharmacokinetics of suramin in an initial 1 cohort of 51 patients, and the inclusion of an additional 66 treated at the National Cancer Institute (NCI) in our data patients base has made no substantial alteration in either the mean values for suramin's pharmacokinetic parameters or their coefficients of variation. With this information it is possible to anticipate the variability that will result from various suramin dosing regimens. Figure 1 represents the fourth dosing level of the phase I trial being carried out at the University of Chicago, wherein patients 2 receive a dose of 2,400 mg/m during the first month of therapy. This dosing level yields a dose intensity that is approximately half
600 500 E 400 E
300 200 100 0 1
5
9
17
13
21
25
Days Fig 1. Simulation of population variability in plasma suramin concentrations resulting from the University of Chicago dosing regimen (level 4). The open squares represent the mean population suramin concentrations for this dose and schedule, and the error bars represent the 95% confidence limits of the anticipated population response. To avoid overstating suramin's interpatiient pharmacokinetic variability, the 95% confidence limits depicted here do not include the variability inherent in the drug assay (coefficient of variation = 7%).
Downloaded from ascopubs.org by RB DRAUGHTON LIBRARY on April 28, 2019 from 131.204.073.184 Copyright © 2019 American Society of Clinical Oncology. All rights reserved.
