In a campaign mounted against theuse of prophylactic cranial irradiation(PCI) in small-cell lung cancer(SCLC), the battle cry of the anti-PCI crowd 10 to 15 years ago was“Fry now, pay later.” The problemwas that some patients survived andseemed to suffer from the treatment.In those days, the high frequency ofbrain metastasis was underestimated.It was commonly thought that withoutPCI, 20% of patients failed at sitesin the brain, and with PCI, only 5%failed.
In a campaign mounted against theuse of prophylactic cranial irradiation(PCI) in small-cell lung cancer(SCLC), the battle cry of the anti-PCI crowd 10 to 15 years ago was"Fry now, pay later." The problemwas that some patients survived andseemed to suffer from the treatment.In those days, the high frequency ofbrain metastasis was underestimated.It was commonly thought that withoutPCI, 20% of patients failed at sitesin the brain, and with PCI, only 5%failed.In the late 1980s, very few patientswere treated with upfront cisplatin/etoposideand radiotherapy in randomizedtrials of PCI. Ultimately, a meta-analysisprovided evidence that PCI increasessurvival by about 5%. However,that overview included patients wholargely had been treated with cyclophosphamide(Cytoxan, Neosar)-basedor doxorubicin-based regimens.Late Effects on the Brain
Has the improved survival calmedthe critics and doubters? Unfortunately,not yet. As survival has improvedwith better staging, better therapy includingshorter regimens of cisplatin/etoposide alone, and the emergenceof intensive early concurrent therapyas the optimal treatment for limitedSCLC, there is general agreement thatPCI not only has a place, but is amandated therapy that will add at least5% to survival.With better therapy, however, therisk of failure looms larger because ofcompeting risks for first failure-ie,if local disease and systemic controlare optimized, the sanctuary becomesa larger and more imposing cause offirst failure. This has been bolsteredby evidence from the French trial reported by Arriagada. In that trial,brain failure accounted for more than60% of first failures. PCI consisted of3 Gy delivered in eight fractions over2 weeks, for a total of 24 Gy. Despitethe larger dose per fraction, few neurocognitivedeficits were observedduring the first 2 years.Regardless of the survival advantage,the risk of late effects on thebrain remains as the muster point forthe militia against the use of PCI. Solidevidence of long-term safety is noteasy to establish. We do not have solidevidence demonstrating the optimaltotal dose or dose per fraction.The lowest brain relapse rate occurredwith 36 Gy delivered in 18 fractionsof 2 Gy.PCI in Non-Small-CellLung Cancer
Dr. Gore opens the issue of takingthis successful tactic with its checkeredpast from SCLC to NSCLC. Shenicely reviews the randomized trialsthat point the way to the use of PCI,and these data provide a solid foundationon which to base a new clinicaltrial. Originally, I was in the campthat opposed the empiric use of PCIfor NSCLC, because the drugs of the1980s were not as active in NSCLCas they were in SCLC. Moreover, theyrarely produced a complete responseby themselves, despite more optimisticreports with the current group ofdrugs. We were having substantial difficultygetting more than 5% of patientsto survive for 5 years.What has changed? With betterstaging, fewer patients present withdistant disease at diagnosis. Positronemissiontomography may further refinethe selection of patients suitablefor curative therapy. As the successof bimodal and trimodal therapyimproves control of locoregional diseaseand diminishes distant metastaticdisease, the sanctuary site withinthe blood-brain barrier looms as the more likely site of first failure. Thetiming of brain failure may be influencedby histology. If 16 weeks posttherapyis used as a cutoff point,non-squamous cell cancers are morecommon in those who experience earlybrain relapse.However, there is less dependenceon histology for those who fail after16 weeks, and the old saw-that largecell and adenocarcinomas are morelikely to relapse in the brain-startsto fall apart. Patients harboring N2 orN3 nodal metastases demonstrate anability to metastasize distantly and tothe brain regardless of histology.Whether an intervention performedearlier than 16 weeks will benefit thosewith subclinical metastasis perhaps toosmall to find even with magnetic resonanceimaging is arguable. Studiesthat include both patients with a propensityfor early metastasis and thosewith the potential for later brain metastasismight demonstrate differentresults.Cost-Effectiveness of PCI
Quality of life and cost are of interestin interventions for cancer. Aretrospective Canadian study of PCIin SCLC suggests cost-effectivenessand an improved quality of life withoutsymptoms, despite an insignificanteffect on survival. Theseinvestigators documented six lateevents in 66 treated patients. Theyreported complications and cost oftherapy, but did not mention any costsbeyond those associated with treatmentfor brain failure. Nevertheless,PCI was shown to reduce costs andimprove quality of life in this analysis.This type of study will also helpin assessing the role of PCI in NSCLC.Appropriate Dosing
Dr. Gore demonstrates that there isa clear and present danger of brainrelapse in the successfully treated patient with stage IIIA or IIIB disease.In SCLC, the question for many iswhat is the proper dose and timing ofPCI? Others worry chiefly about neurocognitivedeficits. The RadiationTherapy Oncology Group (RTOG) hasjoined an international effort to provethat a higher radiation dose (36 Gy intwo fraction schemes) is superior to alower one (25 Gy). In NSCLC, a newRTOG trial uses a dose of 30 Gy deliveredover 3 weeks vs observation.Both are excellent studies, but somewonder why the more resistant NSCLCis being treated with a dose lower thanthat used for SCLC. Each will providedata to either support the benefit of PCIor better define the frequency of toxicityassociated with its use.Conclusions
Clearly, more patients are "fried"by brain metastasis than we originallythought. What works in SCLC mayalso work in NSCLC, but we are betterprepared to ask the questions today thanwe were when studies were conductedin less well-selected patients.Some patients may remain fearfulof potential negative effects of cranialirradiation, and will refuse to entera study. Others will grasp the value ofthe treatment and be concernedenough to demand treatment with PCIoff study. We all need to support thestudies that will answer these questionswith facts rather than adhere toour old positions based on theory. Thetruth derived from data will set usfree, unless our minds are alreadymade up.
The author(s) have no significant financial interest or other relationship with the manufacturers of any products or providers of any service mentioned in this article.
Auperin A, Arriagada R, Pignon JP, et al:Prophylactic cranial irradiation for patients withsmall-cell lung cancer in complete remission.N Engl J Med 341:476-484, 1999.
Arriagada R, LeChevalier T, Borie F,et al: Prophylactic cranial irradiation forpatients with small-cell lung cancer incomplete remission J Natl Cancer Inst 87:183-190, 1995.
Gregor A, Cull A, Stephens RJ, et al:Prophylactic cranial irradiation is indicated followingcomplete response to induction therapyin small cell lung cancer: Results of a multicenterrandomized trial. Eur J Cancer 33:1752-1758, 1997.
Mac Manus MP, Hicks RJ, Ball DL, et al:F-18 flurodeoxyglucose positron emission tomographystaging in radical radiotherapy candidateswith non-small cell lung carcinoma.Cancer 92:886-895, 2001.
Gaspar LE: A phase III comparison ofprophylactic cranial irradiation vs observationin patients with locally advanced nonâsmallcelllung cancer. Presented at the 39th annualmeeting of the American Society of ClinicalOncology, Chicago, May 2003.
Tai, THP, Yu E, Dickof P, et al: Prophylacticcranial irradiation revisited: Cost-effectivenessand quality of life in small-cell lungcancer. Int J Radiat Oncol Biol Phys 52:68-74,2002.