Cancer Chemoprevention: Progress and Promise
Scott Michael Lippman, J. J. Lee, Anita Lyn Sabichi · JNCI Journal of the National Cancer Institute · 1998
In the relatively brief history of cancer chemoprevention, hundreds of chemoprevention studies have been reported. The vast majority were uncontrolled studies that were designed and reported in a nonrigorous fashion, which perhaps resulted largely from the zealousness of early chemoprevention researchers to nurture growth of an exciting novel approach through its infancy. Nevertheless, limitations characterizing many of these trials have tended to cloud chemoprevention's record of progress and promise. This tendency overlooks the solid cohort of definitive randomized trials that have successfully addressed many important clinical questions. What distinguishes chemotherapy from chemoprevention? The simplest answer is that the former involves agents that treat cancer, and the latter involves agents that prevent cancer by either preventing or treating premalignant lesions. This distinction, however, is blurred by several recent results. Some agents, such as retinoids and tamoxifen, can both prevent and treat cancer. Several newer agents, including angiogenesis inhibitors and ras inhibitors, also show activity and promise in both preinvasive and invasive carcinogenesis. The distinction between premalignant and malignant lesions also is becoming blurred by new findings in molecular biology. For example, clonal expansion now is seen in preinvasive stages as well as in invasive stages of carcinogenesis. Furthermore, it is not always clear whether an agent is treating subclinical (e.g., microscopic) cancer or is treating premalignant lesions and thus preventing progression to cancer. Agents for which this issue has been raised are retinoids in the prevention of second primary head and neck cancers and tamoxifen in the prevention of breast cancer in the Breast Cancer Prevention Trial (BCPT). Every trial that we discuss in this commentary involves specific natural or synthetic chemical agents used to reverse, suppress, or prevent carcinogenic progression to invasive cancer. None of the studies involves a nutritional intervention with foods ingested in a normal diet. This points out the important distinction between the field of cancer chemoprevention and the field of diet and cancer (1) . The natural agent β-carotene is an example of a compound that can be the subject of both fields. Ingestion of fruits and vegetables high in β-carotene can be a dietary intervention; however, when taken as an extracted and purified natural compound, β-carotene qualifies as a chemoprevention agent. There have been more than 60 randomized chemoprevention trials reported in the English language literature ( 1 – 3 ); nine of these trials, we believe, qualify as definitive trials. As presented in Table 1 and Fig. 1 , these nine trials involved 15 primary interventions or end points, which were subjected to rigorous hypothesis testing ( 4 – 14 ). What is a “definitive” chemoprevention trial? The criteria used for selecting definitive trials discussed in this commentary include the following elements of rigorous study design: 1) primary end point of cancer incidence; 2) twosided hypothesis testing; 3) randomization with placebo control versus interventions; and 4) large scale (n ⩾ 1000), with the definitive sample size and duration based on anticipated event rates in the intervention arm (treatment effect) and placebo arm. Every definitive trial included in this commentary had a duration greater than 3.5 years and was supported by the National Cancer Institute. These design elements contribute to the ability of a trial to achieve either a statistically significant definitive result (positive/protective or negative/harmful) or a definitive null result (negative/neutral) that is truly informative (either of which is necessary for making public policy recommendations) (15) . Two of the 15 definitive reported to from statistically significant are from was statistically significant between and placebo Two are from statistically significant the arm was statistically than the placebo arm with to cancer and ( Table 1 and Fig. 1 ). Several when the record of definitive chemoprevention in Fig. 1 . by these studies in primary (e.g., agents, and of the including both of the statistically significant involved interventions with and are the definitive trials in Fig. 1 . The statistically significant were in the recent trials, both reported in the The recent to the of definitive studies was the with statistically significant that tamoxifen the of breast cancer . and for 15 from definitive cancer chemoprevention trials. used to are in more in Table 1 the to an with the in Table 1 and The is a trial that has a chemoprevention agent to the of the of As agent has been by the and for in and and National Cancer for tamoxifen are tamoxifen for the of breast is perhaps a of high of breast cancer. of this including the duration of tamoxifen and the of for tamoxifen are discussed in definitive chemoprevention not or of also an of of cancer prevention and carcinogenesis. not with the trials also with the trials. definitive trials and and definitive trials by from This commentary on not in the of of chemoprevention, several of which have agents and trials in ( 1 – 3 ). discuss the 15 definitive of chemoprevention trial end points that have been to and the in of progress and that these have to the field of also discuss novel of chemoprevention, including in the study of on molecular and for trials. The and studies criteria for definitive cancer chemoprevention trials on the of and for selecting trials to include in this commentary were based on for ( ). for were the and and from recent in either or of the clinical chemoprevention trials included had a randomized design and were reported in in English from to Table 1 and Fig. 1 the 15 definitive points, including the statistically significant points are presented in of or 1 in the and the and 15 the with are in of The statistically significant trials were to study cancer prevention in the breast and The by the National Breast and is a trial of with tamoxifen versus with placebo in high for breast cancer. The more than as reported ( 14 a statistically significant in breast cancer This the and to that and in the study be for more than 1 than The hypothesis of the the tamoxifen prevent breast supported by clinical and from trials that with tamoxifen has statistically the of breast cancers by ( ). criteria for the of the included 60 or more years and years with a or greater of breast cancer such as breast and history of breast was based on the reported by ( ). For in the the for breast cancer was greater than of had greater than The result of the was that a in of invasive breast cancer tamoxifen There was a in and statistically significant on a of 4 of invasive breast cancer in the placebo with of breast cancer in the tamoxifen Breast cancer was and were in to of breast and in the with placebo and tamoxifen, of tamoxifen that from the was a in the following of and In the placebo had these in the tamoxifen had such its in tamoxifen in the statistically significant in the of Every from the either or has been with from the on which the trial was have been in the tamoxifen is statistically significant between the study of the In to the of tamoxifen also were tamoxifen were years had of cancer and in the placebo were 15 invasive cancers and of and of in in the tamoxifen were invasive cancers of and and of and of The reported were and The of whether to tamoxifen is and several by findings of the on the for In of the of years had of tamoxifen had the from tamoxifen years or had an of cancer and In of tamoxifen statistically the of invasive and breast cancers in were years in the of years have the more for years or as of cancer and versus in the of breast cancer and that also be are and of tamoxifen, including and in the the of and Cancer and criteria for tamoxifen and to this Several tamoxifen in prevention of breast What is the and duration of tamoxifen that of breast cancer of The National Breast and study of that and years of tamoxifen are in to and prevention of breast ( ). Furthermore, the of tamoxifen for years the Nevertheless, on the of and on years to be can findings of the be to (e.g., of or or to the of 60 years the former breast cancer in that in the that tamoxifen is in high history was a of in the the latter trials ( discussed are tamoxifen in ( , ). tamoxifen in the treat cancer than prevent progression of premalignant This is an both of which can of of the and the of tamoxifen on with the on in the In of the these and preinvasive breast in , in was by in the tamoxifen arm. tamoxifen in the of with history of a statistically significant greater than in breast cancers ( 14 ). In it that both the and prevention are in tamoxifen the of primary cancer in that it the clinical of in relatively The preinvasive findings an issue the of these lesions in , and in which have to be ( 14 ). have been that the in in and in have been as with the invasive This points out the of the in the in to molecular These and have been by recent of trials on tamoxifen and breast cancer prevention ( , ). trial has a statistically significant between tamoxifen and placebo in preventing breast cancer. is to early of these trials, however, with the of the In with the both trials are by 1) sample and 2) and with relatively of breast cancer and 3) of in and of the study There also were important between or the study and the such as in the trial and in and history of breast cancer. the the National Cancer has its in the early and in the definitive answer to the of whether tamoxifen breast cancer in the study Nevertheless, the trials the of tamoxifen in and whether tamoxifen prevention large tamoxifen trial Breast Cancer of of this of the trials, and of the these many as the public for tamoxifen and the clinical of tamoxifen to be for the trial to the tamoxifen are well a of tamoxifen, has been for the the of and or and to be in by the National Breast and the important of in and in criteria of are to of the The or arm of and tamoxifen for to tamoxifen, is a it have a ( ). recent of a ( of an trial that be in the of new breast cancers and with This study was designed to in preventing in years and breast cancer than of the a of the has had breast cancers and in cancers with the placebo arm. these result from a the breast cancer be in a definitive such as the This with a control arm of of growth in definitive chemoprevention which has on placebo The Cancer trial had a can prevent of a new or a new of the The trial involved ( ). not and the for was statistically significant ( The clinical of this trial is by the of for the relatively of of the The raised important the between and of the the definitive trial be the of 1) The be and statistically from based on to 2) end points be For example, be in and be and with (e.g., and early in of and clinical and important of be such as the in in the The issue of in is the in cancer not be with in the is in the (e.g., of the with in preventing the issue of biology. This also is a issue in cancer control and early The statistically significant trials, the and Trial ( 4 , and the Cancer Prevention ( , and a of the trials, the ( involved the compound such a in definitive chemoprevention 1) The was for a of with the of clinical in of cancers and when chemoprevention was as a new approach for cancer 2) and ( that β-carotene be an agent. 3) This compound was relatively and to be it was to be of clinical ( 1 – 3 ). of β-carotene the the in on as the for These are and to that a diet high in fruits and vegetables can and that β-carotene is of this diet ( 1 – 3 ). also now that β-carotene has not cancer in trials more than ( Table 1 and Fig. 1 and can of cancer in as in the statistically significant trials more than ( 4 – ). The of cancer chemoprevention trials not in the of β-carotene ( ). Nevertheless, the β-carotene trials to progress in the field of chemoprevention, not by cancer with this compound, which is the in and also by of the and for testing natural as chemoprevention The findings of the trial and the and Trial ( 4 – that of β-carotene in the former trial and β-carotene in the latter cancer in are by and the of this in this is it be to the reported of β-carotene high and which in the of ( ). cancer were with the of were not in the of than a in the trial or in the of former in the and There is from these and studies the ( discussed that β-carotene cancer in or former The trial and the and Trial that β-carotene were with cancer rates in the β-carotene This is with that a high of foods and high β-carotene are with cancer ( 3 , , ). not cancer a high β-carotene in the can be a of a diet high in fruits and studies in have that β-carotene was for cancer such as were ( ). The of β-carotene chemoprevention study in the is that definitive clinical trials are necessary to These trials be based on and clinical as by the for the in to the The trial in Table 1 and Fig. 1 a of trials, including or in cancer in cancer and β-carotene and in prevention of of these from trials that have the field of chemoprevention by definitive clinical questions. of these trials ( , the trial in the ( the trial ( , , and the ( , – of and of which are discussed The trial involved in and interventions in a design ( ). This trial the hypothesis that agent interventions can the of from cancers on that in are with and As in Table 1 and Fig. 1 , statistically significant on cancer was by important however, involved the of in a β-carotene and which was with a statistically ( a statistically in the and a statistically cancer cancer in second trial included arm of of 14 and the and that cancer in the versus a placebo arm ( ). This trial was designed to prevent cancer in with There was statistically significant on the primary or The trial a statistically in and cancer this study from the study were for and that have with the it to and of the Furthermore, it is to either with to such as the with in the The study of in was in its primary end point of preventing and of the in of cancer and in of the ( ). This however, had and statistically significant of in preventing and The cancer an of in the cancers in the versus cancers in the placebo The cancer an of with cancers in the versus cancers in the placebo The cancer an of with cancers in the versus cancers in the placebo this trial a statistically in and statistically significant in cancer In both the and trials, end points of the were In both trials, primary cancer end points were these end points were with to rates to cancer ( or to ( ). these qualify the of these trials for of As with the these are and testing of the involved to and ability to In it is that these trials, the these both involved in that have a of The which included more than is both in of cancer and statistically in of cancer in ( , ). the however, this trial a of in preventing cancer ( ). had statistically significant in cancer and of and in with in the not ( ). cancer was statistically in The a of more than the that β-carotene on for can cancer and that can prevent The β-carotene intervention was including statistically significant in of cancer or or in of This included were former and a of ( ). The intervention on however, was early intervention of of statistically significant of in ( ). The of both randomized and of this trial is that is not with of cancer a of years ( , ). This is with reported by several that cancer ( 1 , ). This from the to the and relatively is from a trial be subject to than were findings of the This that be used in the of or for cancer prevention in this in the trials that we have with and is in as the to the trials with the the of results. the are subject to an or These be in of trial results. In the definitive primary which are in the are or be ( findings can be for hypothesis randomized trials more than have been not are by sample to a of trials with and and end points premalignant lesions ( 1 – 3 ). There are trials, however, not have many of these limitations and statistically significant and that are the important of these trials is the relatively large Prevention This study involved and a statistically significant in of in ( ). The hypothesis in this trial was that of can or the of The primary was to of new a between the study and the second years The primary included and placebo both The of was There were had study in the had an of for dietary and not the and was to of the studies that we have to involved interventions in either prevention of second primary or The interventions were as 1) in head and neck cancer (n ( 2) in cancer (n ( 3) the in cancer (n ( and 4) to prevent second primary of the in ( ). The trial in this not in the of second primary involved in (n The intervention was more in than in the reported of in a ( ). There also are several randomized trials not also to be in for and for chemoprevention to of these trials more than of trials in this are to is more to of than This trials. Two trials were in the of and in and of the (n ( and of in of the (n ( ). trial and in (n ( ). There were trials in (n ( (n ( and β-carotene (n ( ). trials ( – β-carotene in prevention (n and trials trial ( involved β-carotene (n and the ( , involved (n of both There are clinical or of promise in the exciting field of chemoprevention 1) 2) premalignant 3) trials, including the Cancer Prevention Trial and trials second primary 4) prevention control of that to for are to novel for of premalignant to the and clinical of and molecular study of and agent This involves of novel agents or agent on molecular for clinical molecular and for definitive chemoprevention trials. from definitive trials, for new in chemoprevention The statistically significant and with the in cancer ( and the of the and of the promise for cancer prevention ( , , have to to both agents in a definitive cancer prevention trial ( ). Two in and the retinoids in premalignant these the following 1) sample size and on 2) 3) and 4) activity in and trials. trial of the in the a high in ( and has been by and randomized studies of ( 1 , , ). trial of in the premalignant a high ( that was by from randomized trials ( 1 – 3 , ). this has a of including on and the of (e.g., and of and of and ( – ). Several studies also promise for cancer There are trials designed to trial in prevention of second primary of the head and neck and of the The trial in of and of the intervention with be reported the the have and ( ). The head and neck second primary trial testing by the end of ( 1 , 3 ). In an is testing and in a design in more than with early cancer of the head and neck or early cancer ( 1 , 3 ). important for this involves molecular for between second primary and ( 60 ). of chemoprevention of second primary include the and breast second primary study is testing and of be reported of that a have been in the of ( ). The Cancer Prevention Trial is to the in several both are of a large are by the National Cancer and are based on an The Cancer Prevention an study by the and randomization of more than by the for trial is the ( ). This trial was based on the that and that a the which is the in the that to which has a for the studies also for cancer prevention ( ). was by the and in for the of The primary of the Cancer Prevention Trial is to whether the of cancer a of is by criteria included than years with a normal and a of in than or to design of the Cancer Prevention Trial include the to in of of on this and the of a to the of a to the primary end point of cancer The trial is by the and of the Cancer Prevention are and result in a design issue of this trial involves the clinical of the primary end point and its on the of sample As with the several Cancer Prevention Trial end points be of including cancer and of These and design of the Cancer Prevention Trial have been discussed in ( ). There are of that of chemoprevention however, the of cancer chemoprevention, the of specific agents to prevent cancer such important and novel 1) agents to prevent or control with cancer 2) agents to prevent or control and 3) to treat premalignant The of novel chemoprevention is the prevention of or the control of with specific or that are with of of such or and and or and and and and and cancer ( and ( and ); and and cancer. of for in that this which in in ( ). The and the of cancer from and from through to and through Two study for cancer prevention are of (e.g., to prevent cancer and control of (e.g., to prevent cancer ( ). This of study of agents and can such as from the of clinical including chemoprevention of cancer. The second of novel chemoprevention involves agents to prevent or control in have been with of the and the of ( , ). to control to these include and of the This approach to of molecular and chemoprevention in treating a of cancer The of novel chemoprevention involves of agents, including of agents and to treat for cancer The approach is for it to the with ( ). is a novel approach of a agent (e.g., or by with which in premalignant of a well that is by ( ). is 1) agents are taken in and 2) of the to the has been in a for this approach the of is well with and by As by study of the to be are to the the that the and to control the of which on the of and agents and are to to control of ( ). chemoprevention trials include and that are based on in to the primary end including of agents on from cancers than the primary end to and This is the of the field of which for making public based on trial results. trials the for and are the which primary and on cancer and and the which an of and Furthermore, of the cancer agents are either by the and or definitive study in in in in and inhibitors in The of promise included is molecular study of and chemoprevention agent In to molecular of and discussed an important of this involves the of more in and in (e.g., and ( for selecting the from the agents ( 1 that have for As discussed β-carotene was to large clinical trials a of its had been these trials and by β-carotene in and testing prevent clinical testing of agents and for testing of the new also agents have been in the molecular of and to the approach to agent are designed to is in clinical trials as agents more or molecular are For example, the inhibitors (e.g., are than with both with and ( – ). however, is that also as more agents or that contribute to for example, inhibitors, which are in clinical trials, than the or out to be involved with preventing cancer. is an example of an new that through a that is of or and to be to cancer prevention Table several agents on the of molecular such as the inhibitors, inhibitors, and inhibitors, including retinoids and inhibitors, and and agents used in ( – ). Agents from of the in Table are in for including cancer prevention and or not to cancer. of of these is a for cancer chemoprevention, for example, or molecular that can in preinvasive are are as by the of agent inhibitors for the more than 15 and hundreds of ( , in to the ability of (e.g., and to the molecular (e.g., growth and ( Table ). study of the the of This involves ( and to or and elements in in to of new and are many molecular for involved with the This study is by the of cancer prevention and of such as The of molecular in breast cancer prevention on with on the breast and and on and promise for the in testing of agents with agents that with agents that This to that are more in and (e.g., to of premalignant lesions to ( , 3 , , ). of of and promise for cancer chemoprevention ( – ). of involves testing for such as and in to and breast cancer reported study that cancer in the and ( , ). not including of study cancer also activity in The has activity in the of ( , which is with As discussed in high of breast cancer (e.g., or are by the result of breast cancer in the of ( 14 included study in this field include of cancer in such as have as molecular for such as of with molecular by and definitive trials. have been well ( 1 – 3 , – ); this commentary a new to of include the of the end point to cancer and to the intervention and the to which the end point the between the intervention and cancer The of end points in chemoprevention is for than for cancer. In the former end points of high and are as end points end points be in cancer, to that the of invasive is a that to be ( ). In cancer chemoprevention, are a end point that is as a or end and are to as The of molecular studies of this field by in preinvasive that are to cancer be a or molecular have to be to be with the of a definitive trial that has a end This is to the have to be in the between an and cancer The and of study are that studies to are the studies the were designed to ( ). There be a to this as by the of and its to agents such as inhibitors and ( , , ). The of as an in chemoprevention trials premalignant in that can be relatively for for such as the and can be by the in of these Two be for study of and 1) premalignant as for example, the and 2) such as the In the be or premalignant for In the of the ability to or which chemoprevention study is more in more such as the than in the cancer a more ( ). is as of is to be a for the of head and neck cancer or cancer. None of the (e.g., of of or of are in this definitive study to be with the of cancer. be to of as a for definitive chemoprevention trials and sample of of in the head and neck or the result significant of be definitive for this and for with a (e.g., of in ). result statistically significant of the for for this agent and to a more agent. This is a of chemoprevention study in the and the head and neck of and in and in on in this however, are to be such as that not in a the of on