This excellent, concise review of the status of lung cancer chemoprevention complements a benchmark review of the whole field of chemoprevention by Drs. Benner, Lippman, Hong and colleagues that appeared recently in the Journal of Clinical Oncology  Careful scrutiny of both these articles yields a number of issues for consideration.
The field of chemoprevention has advanced considerably in a relatively short period. This progress has been greatly facilitated by the significant, sustained contributions of the group from M.D. Anderson to the chemoprevention of upper aerodigestive cancers. The sequence of chemoprevention trials conducted at M.D. Anderson is worth analyzing, because the development of better approaches to preventing aerodigestive cancers is the most successful model to date for this type of research.
Efforts to improve aerodigestive cancer chemoprevention began with the use of retinoids in the oral cancers. This cancer site provides easy-to-follow mucosal changes that allowed for definitive evaluation of retinoid benefit.
Studies of this site also afforded an opportunity to identify a more optimal dose. With all chemoprevention approaches, tolerance of side effects by study participants will not be as great as is accepted by patients being treated for advanced cancer. As the M.D. Anderson group lowered the empiric dose to optimize the activity/toxicity relationship, they also included both pill count and pharmacologic monitoring so that they could obtain a definitive determination of patient compliance. Optimizing dose and formally measuring compliance were refinements in clinical trial design that represented a pragmatic effort to increase the scientific credibility of this emerging field.
One Continuous Organ Site
Invoking the work of Slaughter and colleagues, who first used the term "field cancerization" to describe the extensive mucosal injury caused by chronic cigarette smoking, the M.D. Anderson group has appropriately focused attention on these multifocal epithelial changes . If one views the head and neck and proximal lungs as a single continuous organ site, the common biology of these adjacent organs does provide strategic opportunities to combine early detection and intervention approaches. The most obvious example of this is the shared tendency of both head and neck cancer and lung cancer to manifest as multiple metachronous primary cancers.
The lifetime risk of developing a second cancer among the M.D. Anderson study population is 30%, which defines an extraordinarily high-risk cohort. The randomized trial of isotretinoin (Accutane) in treated head and neck cancers was made possible by the risk profile of that cohort . As pointed out by Benner et al, the greatest reduction of new cancers in that trial was seen in primary cancers of the aerodigestive sites. Remarkably, with only 100 previously treated patients randomized, a significant, persistent survival benefit and a reduction in new cancers were observed.
At the time of the trial's inception, this was perceived as an unusual type of study. As a result, Dr. Hong never acquired an outside funding source to support what proved to be a seminal study. To see that magnitude of difference in a lower-risk population, such as the smokers in this study who had not developed cancer (smokers who have already developed cancer are at greater risk of developing a second aerodigestive tract cancer than smokers who have never had cancer are of developing a first cancer), a study population ten- to a hundredfold larger would have been required. Such a large number of participants would have made the study impossible from a cost perspective.
Intermediate End Point Markers Sought
The next logical step for this group was to move further back in the natural history of the disease, to determine whether a "preneoplastic"' change, ie, bronchial metaplasia, could be reversed in heavy smokers with the use of retinoids. Given the overall 87% mortality associated with a new diagnosis of lung cancer at that time, it seemed supportable to use more invasive diagnostic measures to evaluate high-risk individuals. The serial bronchoscopy trial was the M.D. Anderson group's attempt in that direction.
The protocol called for sampling of the entire respiratory epithelium in all participating smokers who had previously been found to have bronchoscopically documented metaplasia. Sampling was done by routinely biopsying six predetermined sites. This permitted a number of promising assays to be tested for their utility in reflecting the status of the bronchial epithelium relative to its involvement with cancer.
The study thus represented an effort to identify an intermediate (or surrogate) end point marker that could be used to simplify the conduct of retinoid chemoprevention trials. Trials based on survival outcome require a long time frame. Trials based on changes in a validated intermediate end point marker could result in faster and more economical progress in this new area.
Unfortunately, the metaplasia study, published recently by Lee and colleagues from M.D. Anderson, raised more questions than it answered . No significant reduction in metaplasia was observed in the retinoid arm compared to the control arm. The rigorous trial design permitted the authors to find that the greatest reduction in metaplasia occurred in those individuals on either the experimental or control arm who quit smoking cigarettes during the study. It is reassuring to again confirm the benefit of smoking cessation. However, the more provocative finding was either that the use of retinoids is not beneficial in this population or that bronchial metaplasia is not a valid indicator of progressive pulmonary carcinogenesis.
In an earlier NCI-sponsored chest x-ray/sputum early detection trial, 80% of high-risk individuals with abnormal sputum cytology were found to revert back to normal cytology on follow-up evaluation without any intervention. This suggests that the morphologic changes in the respiratory epithelium are dynamic, and that markers reflecting features more tightly linked to carcinogenesis will need to be identified for valid use as intermediate end points.
As suggested by Benner et al, investigators involved with the Euroscan effort may also be well positioned to identify useful intermediate end point markers by evaluating the clinical material derived from their important ongoing studies. As the authors mention in their review, validation of such tools may diminish the likelihood of another clinical trial having an outcome like that of the Finnish smoker trial. This major vitamin intervention study resulted in an unexpected excess of lung cancer deaths in the group of participants who received beta carotene. Since this study did not perform biological monitoring of the participants, such as with relevant intermediate end point markers, it is impossible to unravel the mechanistic basis of this unanticipated result.
Aerosol Delivery of Retinoids?
We recently reported that free retinoid binds rapidly and tightly to serum albumin. Although this interaction had been previously described, the effect of the retinoid-albumin interaction on bioavailability at the tissue level had not been previously characterized . Free retinoid is 10 to 100 times more potent in mediating antiproliferative effects in vitro than it is when bound to albumin. Although we showed that the binding interaction can be modulated by the preadministration of a fatty acid source, the real lesson may be that we need to greatly simplify the delivery of retinoids for upper aerodigestive chemoprevention by giving the drug in aerosol form. This approach uses the same route that seems to be so effective in the original pathogenesis of the cancer. Using the prevention trial constructs pioneered by the M.D. Anderson group, we should be able to rapidly define the benefit of this maneuver.
The other future studies outlined by the authors will also build on the broad foundation described above. The promise of chemoprevention research is much more evident due to the accomplishments of the M.D. Anderson investigators over the last decade. The critical contributions of many other researchers outside of M.D. Anderson
to the development of chemoprevention approaches, including Drs. Sporn, Meyskens, Alberts, and Pastorino, are graciously acknowledged by Benner et al. Fortunately, a legacy of adaptive and productive collaborations in this area has allowed for rapid translation of beneficial developments to the management of the preclinical phase of these lethal cancers. Support of the ongoing prevention trials described by Benner et al will help define the magnitude of real benefit we can expect from new chemoprevention approaches. In the current, turbulent world of clinical oncology, success stories such as this one merit close consideration.
1. Lippman SM, Benner SE, Hong WK: Cancer chemoprevention. J Clin Oncol 12:851-873, 1994.
2. Slaughter DP, Southwick HW, Smejkal W: Field cancerization of oral stratified squamous epithelium: Clinical implications of multicentric origin. Cancer 6:963-968, 1953.
3. Hong WK, Lippman SM, Itri LM, et al: Prevention of second primary tumors with isotretinoin in squamous carcinoma of the head and neck. N Engl J Med 323:795-801, 1990.
4. Lee JS, Lippman SM, Benner SE, et al: Randomized placebo-controlled trial of isotretinoin in chemoprevention of bronchial squamous metaplasia. J Clin Oncol 12:5, 937-945, 1994.
5. Jakowlew SB, Avis I, Mathias A, et al: Presaturation of serum albumin with triglycerides enhances 13-cis retinoic acid growth inhibition of small cell lung cancer cells. Proc Amer Soc Clin Oncol, in press, 1995.