costs and effectiveness of prostate cancer screening in elderly men potx

However, his conlsion is extreme 1y sentive to assumptions abou 1 he eectiveess of treating prose cancer, and 2 the rte at which un treed cancers spread 0 ohe pats ofthe body and uti- “

E veventive

health services iender

Medicare

or tHE

Costs AND EFFECTIVENESS OF PROSTATE CANCER SCREENING IN

ELDERLY MEN

c—— .-— 2M

Foreword

‘verte ast 5 years interest in strategies to promote health and prevent disease song eet people has grown substantially This wend has a ast prialy reuhed from the desire 1 moderate rising health care costs among this sg:

‘ment of he population As thas done in the cae of his background paper, the House Commitee on Ways and Means ha periodically asked the Office of

“Techuology Assessmeat wo analyz the cost and electiveness of providing selected pee

‘vetive health series elderly men unr the Medicare program The Seaute Commitee

‘on Labor and Human Resources had ear requested that OTA provide infomation onthe

‘vale of preventive services tothe American people Past wok by OTA on prevention folly people has focused on sues of he costs

sd effectiveness of peumococcl and infenz vaccine, and screening fr bes, evi «al, and colorectal cance and for glaucoma and elevated cholesterol This background pa

pe focuses on the procedures of digital rectal examination andthe more recently devel- oped, less ovaive prostate specific angen bled est—both used op detect prstate

‘The background pape summaries the evidence on the efectveness an cost of ros: tate cancersereening and eatment in ley men and explores the implications for Medi- care of offering his preventive technology aa Medicare benefit This analysis 0s the hat policy cies in deciing wheter to expend federal resource for ceening sad

‘weament aswell asi ther atendant complications before scenic research hs defini vey established the effectiveness of diferent technologies atemping 0 cure disease + tected in varying sages and circumstances,

224 '8ogerC.NendnonOfA Dtecer

hye bomney

soon 2 i

oom aoger Det

Project staff CHAE E UCK Poet Oroctor

Fomso Colmes Reseach Auton!

ADMANSTRATVE SAE louie Sey Ofte Ait,

‘covey Menta Act Scitay Mona neh Woes Rocosing peciast

‘Con Swarm PC Speci chert tron vera Pocesng Seach

Monae A Bay Mosocrusts Genes Kem oon, Mec

‘Cvtopner M Coley Mowers Gener Howe, axon, Manaraset Cig Reming Heat Outcomes Asocates, Voncouve Weshinglon

Jeph & Gaeding Urry of Mcrigan Medical Corte,

‘hn Roo Mcrigon Matis Fans hemetonelLngovy Cntr (US), Deperivent ot Comsrnty Mec, Mt, Sect Macca Canter New You New or choo sander eterno Longe Contr (US) Oeporiment ot Camera Mecca, Mt Sct Mecca Canter New Yor, Now

‘Coxe Upped remota Longe Cena (8) Depermentot ommunty Meche, Mt SneiMasec Corte Now a Naw You {seo Ramey Daparment of etna Atk Maciel Carte,

‘Stephen D.Fen Deparmant Veter Alta Mace Conte, Sacto, Wexhngton

Acknowledgments

{oe Ancervn US xscuhe Oc the resent Ofce of Marogement ens Bigot L1 —MeM Cit sdợn ConcerEoderloopyUnven 2G, 9xedemn

‘Pater ben Dope often Urey of Conctcut

.eưgdjL.Anđfoe, lí 0Ndenof elo Supe Weshington Urry

‘ove Bx Deportrant of Porro Mayo Cc Morn town Nota rate of Heth US Deparment Heathen Hanon Sandee eds Bates Oeparnen of arpa Macca chao, Unventy of Waar Moun agene Carton Anetcon ogc Asoctton Sor CologectWeccne

Neney Caton Mortons Corpor

.8etfdí Chodst Dopormarfr ulley.creolt Metcne, ery of Crcon0

.Megen Cohennelconlrgopedl Aeocele'

| Mont Coten Dion of teserc,KoserPemenente aca! Car rogram

teu Dene ntenatena esa Heh Cou fering Cachet £nhoep Bom Jen LFouciy Food ond Dug Acniteton, US Dopinert of Heath and Hamer ences _ 8a ffehentrddetnoogy Grew he na oasis, Oklon of Resch Pamonant Medial toendd GkesonPonobsgr.Mneopoke Mi

_¬ốũGÖGÕẤGÚẶÚẮẶỘẶĂệừơ¬ơaă—

‘cere Green Of Heatn ochleay Axes ny of fin Cater BC:

‘Scot? Hoo Dopertmart of clog hoa et Main, Wey Sot niventy

Den Wane, ASN benef Cor

BNWggg Henle ne

Sytem Dison ct Concer Reventon ond Contok Nose Concer se

abet Lowence Heh Science fooddeirFoundoion

4 Meno Meinis © Copotmentot Hendin tances Ofte of Creare —

oman Monte Deparme ot tnogy.Shodlof Macc oye Sat mery

‘ive Lagat rool Rc Hem Corbi tne, Pou iteg Deoorrent of fmt Macc, Medical cco, Une of Clo

Âa 1 1 anna esta Mcrigan Concer Fesrdon Pn Commereheie Cancer Carr

‘Amol toy Notorel Corea rte at: enema East Vigra Gade Seo

| Smpun Deporiment ot Hoth Folcy ae Asrnkhton Scoot Plc Hach, ety

te sendont oust uta oor Hy gene Arte Bogum

tuey heĐetre emdfcơnAeocofcn — of Rete Fenors

evn Wo George Westngton voy Medct Contes

Wet, nor erate Son ati Concer Cate

Pratte Caner rlenenton Ver Conan Ht atic eh cade dsocon eto)

‘ret, ung, Clacton pede aestatenecte agen Ovaton Sree et

gi

ai sorweca section ome oust

C89 cdlononreneieeeooekr leo AonnkrơionCòperoivetrdbợc RuechGime

Contents

eyfrdnox Prostate Concern Cer Men Technologies o Dstect rsate Concer Thọ Efoctưenemof Yeotment Sones is ond Costs of Sreerng Rerearch fo Resove Uncertaries CChaploctwo <1) PROSTATE CANCER IN OLDER MEN,

11 Sereerina Vous Diogross

12 _Retlonal for Ety Detection and reatmont 13 SpeciatiuesinScieoning Medicare Age Men

14 Conticing Guldanes on EotyDetocton

18 Bosc Boogy of Patate Concer

16 RekFoctor 18 The Prevolence for Posate Concer of Prcsole Concer

18 Protte Cancer Mantatty CChopter Time | 21 TECHNOLOGIESTO DETECT ROSUIE CANCER

23 Dogtet Rectal Examination

24 roscteSpectc Antigen

28 Combboton of OR on PSA

28 Fotows Teng {30 Sereaning he Medicare Popuiton

i i TREATING PROSTATE CANCER

33 stotociet A The Btecthenes to Detemine Concer Hoge of reoknent

1 1 Followup teetment ater Curve erapy

—

‘open A

‘Memos Used 0 Estmote Ukstnoocs of Cancet fo Portico DRE ond PSA Append

‘Stuer of estate Specie Anigen fer Prdate Concer Steering

‘endear Detection

‘Append

‘Sucerof torwectl Uracund or eote Concerning ary Dotecson ond

Summary!

restate cancer isa common and serious ma

lignney among Medicare-age men In

1995, 24,000 new cases and 40,400 deaths

are atcipated from his disease; menage 65

nd oer bear moet ofthe burden of less

Tnreceat years, the prosate cancer diagnosis rat asin

‘creased dramatically, witha slower increase in age-spe-

cif moral At eas inp, the increasing incidence

undoobedly reflects more aggressive efforts at early etction of prostate cance, pariculaly through the wse

‘of anew blood test, prostate specifi antigen (PSA) ‘his background paper examines the implications

‘of potetal Medicare benefit to cover prose cancer

screening using a combination of the PSA and digital

rectal examination (DRE), a time-honored test per

formed inthe physician's office

KEY FINDINGS

‘The Otice of Technology Assessment (OTA) con-

‘des that researc has not yet been completed to de-

{ermine whether systematic early screening fr pros- tate cancer extends lives The evidence of benefit for other preventive services alreay covered by Meare (ep, reas and ceva ence screening, in ven and

‘poeumeccocal vaccines) is substantially more devel

‘ped and stronger than for prostate cancer screening Be- ‘cause senile knowledge is limited, but the conse-

‘quences of prostate cancer and Its treatment are serlous, an informed and reasonable patient could equally well decide to have screening oF forgo it Henge, each patent, in consultation with his physician,

‘must use his own vals to weigh the potential benefits

of screning againe the risks of incontinence, impo- tence, and ober adverse outcomes that may el from treating cancers uncovered ‘Given the tate of current knowledge about pros- by screening, tate cancer itmay be reasonable for Medicare to con tider reimbursement of the screening tet Reim- Dbursement could be seen as ensuring that

‘out-of-pocket screening expenses (however small not c2 Các cnưctzcrtccteretre=ecettesrecrecdrsctnezcetvstreraxzed7).eoaneecercetzeeeotezzzcc

© cee: wo Enemas or roam Cnc Ste en Mon

Impede welinformed đbeuslan and deilonmak-

Ing between physician and patient Such a Medicare

serening benefit could be unrestricted as a similar

benefits fo cervical and breast cancer eresning How-

ve, an unesicted, permanent benefit might imply

that science actually ha cablabed túc beef of ery

đdecúon, An slemire would be to offersrening 0°

temporary basis sobject to reconsideration a evidence

torn cinial wk shout he effectiveness of seeing

nd teatnen becomes avilable Such a benefit could

also be coupled with efor bythe federal goverment

to avolve st many pints at posible ineffectiveness

‘esearch ant ensure ptiens and physicians are well

infomed shout potest benefits and rik of tating

cancers discovered by sereeins “Te tecnica! analysis inthis backgroend paper

shows that in terms ofthe expected cost per feyexr saved prone cancer scroenng cool indeed be cost

elective aoe dneseserecning service lead cov- cre by Medicare However, his conlsion is extreme

1y sentive to assumptions abou 1) he eectiveess of

treating prose cancer, and 2) the rte at which un

treed cancers spread 0 ohe pats ofthe body and uti-

“manh case dea, Relaivly small changes in these

sssumplcas make he sue prostate cancer seeing

‘benefit appear very expensive witout ay health bee:

‘i, and the mu values for these assumptions ae un

sen ko medkal šenles đc ta tbe lack of sppope-

se research noted above As ako inticated above,

‘nement of detected cancers wood ek in complica

‘ion inching death, sbsanal mts of impotence nd

inceninence and her disease

Why Might Screening Inmstivey one would expect that eal detection f= Not Be Benefcior?

{ors souk find mare prostate ances before they have

spread ote of be prostate plan, which shoal incr

lead to more rosa ance cures with aggressive teat-

‘ment Indeed, evidence shows ta paints with ances

<dacovered by screening ed to do wel Fonhermor, mostmen who have a postive PSA test folowedby sut-

ey thateveals th cancers not yet spe beyond the prostate gland suongly beeve shat aly detection and treatment have sved their ies One ofthe factors tat may acto strengthen his belie i te file ber of men who become impotent r incontinent ta

‘ul of surgery The bel tha surgery was necestary 9 vo fl less could be an imporan meat of se: cepting these woubletome symptoms lowove itl not clea ht hse outcomes ae the seo of srening and subsequent weament Coed out comes may reflect the fc hat srening advances the point of diagnosis, witout changing the destined course ofthe eancer (latte bias) o that srening may ‘efeenially find lower growing cancer ready det tine to do well (eng bis) Becante of tee biases,

‘ext diagnosis would appear to improve survival even

if weamert were worthless (rhamfuD ‘Thete problems are compounded by theft hatin mos eases, provate cancer is alow-growing diese Moat men whose lcliaed prostate cancers ae sco ered by screening might never sur any effec of tei ease, imately dying from some other canse Hence,

‘ven weatment i imately proven toe beneficial for men with very aggresive localized prostate cancers, it

‘would sie unnecessary fr most The dilemma for po- eymahers artes fom the fact hat cure đgnoodc

‘measure ae not sfciento determine prior and pe- ely which cancers are ikely to came har, Wer there orice or cots associated with treatment, tight more lealy make sense to wet all cancers fund However, inlight of hse eaten risks andthe cuentuncerin-

‘y about treatment benefit, the decision aout srening and any sobsequent treatment must curently rest with

‘hepatica in consuhaton vi his physician As our

un-derstanding of this disease and of our bly to intervene

{nit grows, selene will be able to provide more defi

tive guidance to bth clinical and policy decisions

PROSTATE CANCER IN OLDER MEN

Screening Recommendations

‘While te American Cancer Society (ACS) and the

American Urological Assocation recommend adding

5A to annua digital rectal examination for earl deEc-

ion of prostate cancer, the US Preventive Services

“Task Fexce and Canadian Task Force on th Periodic

Health Examination, citing lek of evidence of benefit

fiom controled studies, donot? Al of these groops ges tht esearch has yet to document hat on popula:

ton-wide bass, PSA testing reduces the risk of dying

‘iow prostate cancer The differences in recommenda-

ion reflect different philosophies about wheter elini-

‘al medicine an publi policy shoul encourage he use

‘of potentially beneficial, but unproven, cancer preven tion sstegiex before comoled studies defintely es

‘abi that they do more good han harm,

Prostate Cancer Biology and Risk Factors

“The prostate isa goiballsize gland dat helps

produce semen, the Mid ejaculated with spemn I is

found below the biadỏer and surrounds the orethra

‘trough which urine pases as itis voided Most early

prostate cancers seem tobe slow-growing, with doubling

1

times of wo year or mor The future course of posta cancer is pected by tumor grade (he extent to which cancerous calls are diferent from nmal cells) and sage (extent of cancer spread); patient age does not seentoinfhence the rte t which tumor spread and be come life-threatening Determining the stage of prostate cancer without surgery is unreliable 3 Once prostate can- cer spreads to bones or other organs, hormonal tat-

‘ments can only achieve temporary remisions often measured in mons +

‘Those most at risk for prostate cancer are Aicsn

‘American men and men with family history of prostate cancer Recently, rior vasectomy’ and a high-fat dit Ihave been proposed as possible ational risk factors {In addition, the probailiy of harboring an asymptomat-

je prostate cancer increases as men age: about 22 percent

of ren nthe 60s and 39 percent of men in thei 7s For those cancers geste than 05 min volume (hich are mere likely to cause fue problems) the age-spe- cite probabilities of having prostate cancer are abou 9 and 15 percent, respectively

TECHNOLOGIES TO DETECT PROSTATE CANCER

‘DREand PSA are both feasible wets for early dtee- tion of prostate cancer Tansrecal ulrasoend (TRUS) and wansrectl need biopsy (TRNB) are followup tests sed to forter investigate suspicions esos on DRE or

TT TH

TH nong 2Hữmr tr atcccenehcdeorrecreergd) 8 teen 2neecrfeceaazesverservost

© core w0 erccwors Or Promise Coen Stem oR Mo

PSA, The te false-negative rates of DRE al PSA sre

unknown, because studies have generally not dete

mined what proportion of men with nonsuspcious DRE

‘and PSA resus in fact harbor eancet

Digital Rectal Examinations ‘Among older mes, digital rectal examinations are

es ily to detect smal an peobably insigifea an-

cers than PSA, bt it is more likely to detect cancer that

have already spread beyond the prostate Avalble data

Indicate that a sspeions DRE raises the Hkelhood that

8 patient has intracapsular (and possibly curable) pros

tale cancer I 1/2-to2fod above he average rik faced

bbymenof thesame age Inareceat large study, DRE wat

suspicious in 15 perent of male volunteers overage 5,

and 21 percent of men with a suspicious DRE bad pres

tate canoer at biopsy However, these high percentages

‘were dependent upon aJow threshold for considering he

[DRE abnormal, and upon the performance of multiple

"viet on volumers witha suspicious DRE In fst,

shout tal the cancers found by TRNB i his stdy were

found elsewhere inthe prostate than the palpably susp-

—

Prostate-Specifie Antigen ‘The prostate specie antigen i protein produced

bby prone sue and meaturabe in blood Itean be ele

‘vain men both with and without prostate cancer, and

the level at which aPSA mestorement shouldbe consi

ed suspicious i controversial, On the ho mo com

‘monly usedasays, levels above nanograms pe ilili-

(er (ngimL) of blood ae often considered abnormal Available daa suggest that a PSA elevation fom 4.110

100 nanograms per milter (ng of boot raises the

‘ikethood that aman harbors an intracapsular prostate cancer one and one-hl to threefld above the average

fk fr men is age Methods to impeove te bi of

SA to dlseriminae between men wih and withou can cer re under active investigation a resent, he no consensus on an optimal method PSA does a particular

ly poor job at separating men with benign prostatic hy petplasa (BPH), x common nonfal disease of aging from men vith inmacapsular,possbly curable poste

‘Combined DRE and PSA Screening

‘What is gine by đong boh DRE and PSA rater than just DRE? Research indicates that by adding PSA, texting to DRE in onetine sening program, ad by adopting an aggresive strategy of systematic prostatic biopsies for suspicious results on eter tet, rosa cancers canbe fund in about 4.2 percent of menage 6S (4s opposed to about 2.4 percent with DRE alone, cos of performing nlite biopsies in 19 percent At age 75, cancer would be found in abot 7.2 percent of

"me (as epposed to 35 percent with DRE alone), with

27 percent of men requiring biopry, Some ofthe cancers

‘at are found in screening programs ae discovered be-

‘cause of the high percentage of men who undergo malt ple systematic biopsies, rather than because ofthe dis

<rimiaing capacity of he tests themselves

Followup Testing

"TRUS ie no accurate enough to serve asa pinary

screening est TRNB is the et uly weed to confi

whether cancer i present, and TRUS i ofen sed 10

hepdinet hen the samples se taken during biopsy

Maay experts now recomend tha patients witha spi

cious DRE or PSA undergo mohipe (fours rosa

se biosien (onal done i single session) TRNB is ncomforabe ands alow bt finite ik of beeing

1 infetion

‘THE EFFECTIVENESS OF TREATMENT

For the early detection of prose cance to anrove

uomes, game for cancers found a sereeing

eed ob efletive In ter words, owed oft

presence of cancer wll no sve any ives unless teting

‘howe cancers makes aifference Teri considerable conrovey regarding opi weatnent for cancer that

oes not appear to ave spread beyond the prose

ad Unlogiet geeerdly sgae that radical prostatec-

tomy, procedure remove the eatieprosia lan e-

suksinthe est cucomes fortesemen As aresult ates

‘ofthis procedure have zen dramatically in een! yer, in response tothe precipitous increase in dignoss of

ely prostate cance However, expectant management {also called “watchfal wating) in which the clinician

treats symptoms and compicions without atempting

‘cur, and radiation therapy are two other commonly

ed treatment sees Prostate cancer management

vende 0 be more comervative in Wester European

countries than nthe United States No ta tại shows

hich of he various weament sieges saves he moet

es fan) has yet bon completed

une) sama GÉĐ

CConoveny about weatment eetvenss exists

‘because ofa lak of well controled studies comparing

‘he main semgis for managing localized prostate can- cec To dete only completed studies are based on ob- servational ses Toth exten that any fee sais show tht pains receiving a particu treatment op- tien do beter than thse receiving another weament,

‘one canst definitively conclude thatthe observed result

‘eas due only to weatment and not doe to oer iter ence between the paint group Determining Cancer Stage

Before men begin teatmeat fora prosate cancer

<dscovered by DRE or PSA, they woud often nderz0 some saging ests 0 help determine the best weament sry Patents with cancers tat hve aleady spread

‘outside the capsule ofthe prostate glnd, and paictlsly

‘cancers that have spread to Iymph nodes inthe pelvic sea oF to bones are much less Likely tobe helped by sxgresive wemments with carve iment Computer ined tomography (CT) or magnetic resonance imaging (OARS) sans and surgieal examination of pelvic lymph

‘mds commonly employed to determine ifthe cancer bs pend ae not particularly accurate er this pps

‘Asset, even if CT of MRI sen suggests spread, Clinicians often proceed to weatent ut of fea of with- holding a potential cure Despite some substantial ms lasifcation rues, recent mathemati! models de signed to predic cancer spread suggest clinicians could ue ome sapng tests more sparing?

‘Expectant Management Expecuamt management isa suulegy of reserving

‘weament for symptoms or complications relued 10

€ Coss wo ericnvenss or Prosu Caices Semin hv Eien Mew

prot cance, without mcessarilyatempting aca It

{scommonly used ia Westem Europe, and unl ces,

for many men with cancers found incidentally during

surgery for BPH, Men ested expecta risk develop:

ing symptoms due to loel progression of ter cancer

(euch as bladder outflow obstruction) oe fom spred of

the prose cance to othe pars of the bey (which nay

‘ead to deat)? The prognosis for men with clinically lo-

called prostate cancer depends on the agaresiveness of

the cancer, parcululy its grade A recent synthesis of

un fom several dis of expectant managements:

sets a 10-year cancer-specific death ate of 13 percent

formen with well and moderately differentite poste

cancer (the most commen types found by eatly detection

‘with DRE and PSA) compared wih a 66 percent death

rte for men with port ifereniated cancers 0

Radiation Therapy

Radiation therapy for prostate cancer, mos con

monly delivered as extemal beam x-iradiaton, a

tempts to deliver maximal dose of radiation tothe

‘mor while minimizing he side effects from expxsre 10

ter neaby radiation tensive sues, Paint venlly

motive five weekday estments ovr sco seven weeks (e304035 meaiments tom, Atnough much cent i=

‘crane has focused on surgical teatment of prostate ean-

er (radial prostatecony), as a 2s 1990 radiotherapy

as the moat comin treater administered for every

sage of prostate cancer in the United States!" ‘The comparative elecitenes of motherapy ver-

sus radical pesatecomy or expectant management as

ot been well sued The medial Iterture suggests

‘worse outcomes for patents with oalize pasate can: cer ete with rtitherapy compared with these oer

‘wo srateies, but results are confounded by radioters:

y series including more older patients whose tumors hve less favorable prognostic characteristics, While arog ve raised concems about the high proper:

‘ion of patients tested with radiotherapy having subse- quently poiive biopsies for cance rising PSA levels poscaraument, selected sves suggest very good out comes in terms of re of future mela disease and cancer death bough radiation therapy is mee Wkly {fo rest in bowel injury thn i radical prostatectomy, ter sd effects ae less common han thote associated

‘with prostnectomy

Radical Prostatectomy Radical prostatectomy ents removing the emre

‘rosie wit is fascial coverings and he seminal ves les Moe aggressive erly detection elo fr prostate

‘ance in recent years have ben accompanied by precip nos rises in population-based cates of radical prostate tommy Recent modifications in surgiea chase, esl Ông Ín an “anaemic” radical promatectoms, have

‘reduced the risk of surgical complications in some cen- vers While some men wit prostate cancer treated sugi- cally have done extremely well, dhe benefit of radial prostatectomy is uelear only one controled say has

‘compared its outcomes against cher treatment srte- ies This single randomized wil, which showed no df= fereoce in morality berween radial prostatectomy and

expectant management, was tbo small 0 detect ch:

callyimporunt benefit rom surgery if realy existed

“The isks of radical prstaeciomy include operative

eat, perioperative medical completions, incon

ence, impotence, and urea stcture formation Ina

recent survey of random sample of all Medicare po

‘ents who underwent this procedre in ihe United Sates

‘between 988 and 190,31 percent of men were wearing

‘ais to Help deal with wetnes, 60 percent reporad n> full or patil erections since te surgery and 20 pereent

indicted they had been weed fora sitar, The atib-

abl! 30 day posuurical death rt was 6 perceat

Followup Treatment

Men whos intl cancer has spread wo ober pas

ofthe body, ormen who are found to have cancer that has

spread postoperatively can be tested with hormonal

(androgen deprivation! erap After inal weatment

by radical prostatectomy, lncians also often consider

joan radiation or androgen depivation therapy for

‘men considered at higher isk of harboring residual can

‘cer Cancers tat have spread to other pars of the ody tend to be sesponsive inal to hormonal eaten, but

then become unresponsive (refractory) Thee are no

at from wei contrlled studies that indicate that any

adjuvant therapies improve survival

BENEFITS, RISKS, AND COSTS

OF SCREENING

In the absence of controled studies documenting

that early detection of posi cancer does more good

owe) sama GD

than ham, this analysis wsed a quantasve decision

‘model to estimate risks, benefits, and costs of an ely

<etction program under diferent sets of assumpsions

Te examiaed te implications of an Wustrative, one time sereening program for thee cabo of 100,000 men, ages 65, 70, and 75 respectively

‘Realistically, a Medicare benefit would most likely

‘cover parole sereening, for example, « DRE and PSA

‘every year asthe ACS curently recommends, ‘oor tre years as Medicare curently does fr beast or every and cervical cancer seeening respectively, Understand {ng he te effects ofan actual Medicare benef would

so require acoumtng forthe ft that tome mea would have already received srening before their 65th birth

ay However, as this analysis demonstrates, cunt un- erstunding does not allow a definitive asessmeat of he cos-effectiveness of even a one-time benefit with is el- ively simplified se of sssumptins, ch essa more complex, but realistic periodic benefit The uncertainty conceming weatment effectiveness andthe muerte at hich smaller ances eventually spread and cause deat

‘verwtelm other assumptions i the mod

‘Modeling an itustrative Screening Benof

‘The model employs a quantitative tool known 28 Markov proces! to calculate what happens to men io each of the vee age groups examined once they ae screened for prostate cancer tiny incorporates

‘many sumptions favorable early detection and tet

"nen nehuing 1) relatively high metastatic rts hat predic higher shan actully observed lifetime probs a'aiex2gk he ttelft 0h imrrtdronvtlccưglrebolspkcrdpoco2ưng0t1mcưnfaoermlegrlasre

—

@ cos wotmcmoes oF Prose Cure Sctome Iceni

bily of prostate cancer death i the exons) Sand 2)

1 100-percent cure rte by sugery for cancers that have

‘ot spread beyond the prosat (esting in oveal cue

rates of 7,70 and 56 percent for all well mederaely,

and poorly difereniated cancers respectively) The

analysis eximats the impacts ofa on-me verening

‘rogram under these assumptions, and then examines

how relanng the favorable asumptions abou treatment

cia changes the resis,

Heath Effects of Screening

‘Using the baseline assumptions, the made! predic

4 very favorable mix of potentially corable cancers

would be discovered by early detection efor with DRE

and PSA A large mmber of prostate biopsies would be

performed asa esulf his program; much bigherpro-

ton of patients would require further invasive evah-

ation asa result of ternal esting than for other com-

monly used cancer screening srtgies, such as guaac

testing for colorectal cancer or anmography cancer The proponion of men screened who underg> fr beast

biopsy ould range from 19 percent at age 65 1027 per-

cent at age 75, Treating cases of clinically localized

rostate cancer wit radical prostatectomy out 300 oot of every 100,00 men screened iaconi= would render

‘ent, about 1.400 to 1,600 ‘screened ienptent, ad an ational 400 to 500 ou - out of every 100,000 men

ry 100000 both incontinent and impotent About

another 20 out of every 100,000 screenees would de

“dd

However atthe same tine, early deletion might

save as many as 4,353 lifeyears inthe 65-year-old c0-

hort of 100,000 mea, 2.74 lie-yar inthe 70-year-old

coor and 1.415 life-years inthe 75-year-old cobot!®

‘The benefits diminish considerably asthe assumgeion of -lavely high ates of metastasis and treatment effec- tiveness ae rela,

“The costs per year of life saved withthe favorable sumptions (compared o doing nosereening at al) was competitive with other commonly-used early detection maneuvers ranging from $14,200 per year of life saved age 65 to $51,290 per year of life saved a age 75 However, hese results are exremely sensitive othe as sumption made sbout the effectiveness of wetment and

he ue a which etracgpsuar cancers spread and case death, Reducing the estimates of ftue tiết of mets ses modesty to levels found elsewhere in the published Teraue and assuming treatment cures only half of ai Iracapsulr cancers greater than 0S mln vole sub- stanly uss the estimated costs pe year fli saved, under these assumptions, these estimates would sange from $94,458 a age 65 to $506,909 a ge 5 ‘As indicated eae, current scene evidence is nefientioknow the ire it of metastasis or wbeth treatment actualy enbances survival, and hence,

whee or not prose srening (even under the sini

Sed assumptions needed wo analyz neti program)

{ssimilarto oer eat destin programs fr Medicare

in ts com per fe year sve, or subsatally pemive Regudlets of whether screening sul the mor ex-

{quent weament eed if and regudes ofthe cast of

any such Heal benefit, i is cenain that popolaticn

based screening wood subject men othe rks of npo- tence, fcontinence, ad ter beakh problems cated

by serening and eatment

RESEARCH TO RESOLVE

UNCERTAINTIES

‘ery data rom conoid suis ae avable

to determine whether the nei of erly detection and

‘weameat of prose cancer ouweigh te risks One

case-conro study suggested ch igial rectal exams do

not rece the rik of developing late-stage prostate an-

‘ez And on al of inadequate size showed no die

‘ence nthe survival of men rete wih expecta man-

agement versus radial prostatectomy However,

‘estarchrs ar now intiating a number of wellesigned

randomized wil of adequate siz 10 ade thi st

‘Tas comparing expectant management versus ages sive weatnet with radical prostatectomy or radiation therapy for men with known cially localized protte

‘ance are underway or boat o tan in Scandinavia, the ‘United Kingdon, and the United Stats Tras comp ing inwaive screening with DRE and PSA versus 90 serening oc “usual care” ae beng ited in both Ex- rope andthe Unite States, Unfrtuatly, from he pr spective of policymakers, tbe relatively indolent mare of many posta cancers means that 101015 years may tbe required io see enough prostate cancer deaths among renin hese dist obtain degen eompaions of

‘he suatepies being sted

‘This analysis of te estimate risks, beefs, and cot of early detection of prostate cancer highlights the

‘ceri sorounding this topic Any decison in he shorter about whether Meise shuld cover (to,

‘ence encourage) prostate cancer serening must weigh ‘te remures required and he known complications hat wil esl fom seroesing and weatnent against an un- cenaia esa net

Prostate Cancer in Older Men

rosa cancer is 2 major bea problem in

the United States In 1985, 244,000 new

<xses (up 44,00 from 1994) of prostate can

‘xr and 40,400 death (up 2,00 from 194)

de wo this dscae are expected among all

‘American men (19) However, mast cases of prostate

cancer and deaths from the disease occur in dr men

Of the 32,378 US prostate cancer deaths observed in 1990, 12423 (8 percent occured in men ages 551074

snd 19,622 (61 percent) in men ages 75 and above See

table? fora comparison ofthe number of prostate no

ce deaths with ober causes of death for older men (4)

“The dong probably of dying of prostate cance for

‘men nthe Unie States is 2.5 103 percent (308,314)! Patients who are diagnosed because they report

symptoms (sch as bone pun or difficulty urinating)

‘every have cancer spread outside of the prose

‘land, and are incurable, Although these paints may initially show some improvement trough game,

‘hese responses often Jo not Ist, and fllowap treat ments have ben disappointing (131)

Given this burden of ness and the dicaty a

treating symptomatic dss, ey detection wing @

simple cial pocedae called dig rectal examina tion (DRE) anda blood test called prostate spec ant {en (PSA) measorement would seem to be a common- sense strtegy for reducing the mobiiy and morality fom prostate cancer in the United States This back- round paper examines the validity ofthis conelsion

‘This chapter gives an overview of the rationale or screening and provides background onthe nae of, prostate cancer Chapter 3 discusses technologies forthe screening and diagnosis of prosate cancer, and chapter “reviews evidence onthe effectiveness of eating the iscase Chapter 5 presets sme illustrative analyses of, the potent costs and eletvenese of onetime pros 7.— ions fr a potent Medicare screening benefit

‘SCREENING VERSUS DIAGNOSIS Before proceeding, i i seul consider what is meant by the tem screning and how it fers from diagnosis While screening is an atempt to ideotify a

‘condition inthe abseace of symptoms, diagnosis pe- formed in response 1a pants symptoms This disie-

‘ion hat important public policy implications since the

© cons wo seco Or Pose Ce Scart MEER Mi

TABLE 21: MUN@ERS OF DEATS BY LEADING CAUSE, US MEN AGES 551074 AND 75, 1990,

{eera Medicare program that provides health insurance

te almost all Americans overage 65 pays for cupaint iagnsis, but it only pay for Limited eypes of disease

sreening Cuently, prostate cancer screening isnot

‘among the services covered by Media In this bak-

round pape, the use of prose cancer detection

technologies in mass secening programs as well as by incian in their offices are considered together at ear

ly detection?

RATIONALE FOR EARLY DETECTION

‘AND TREATMENT

‘Theoretically, surgical removal ofthe entire pros-

tate (radial prostatectomy) or radiation therapy (cur

tive radiotherapy) should cue prostate cancer tht is

confined within the prostate capsle The survival pob-

bilities fr patients with early-stage prostate cancer re

early and dramatically beter than for patients with

Ine-stage disease, suchas i commonly sen inthe ab-

sence ofsrening Screening tests are currently avai-

tle that result inthe detection of disease that is more

ten localize to the prostatic apse than would be the case among men presenting with symptoms Therefore,

‘nis tempting to conclude that screening for rotate an:

zr wl result inthe curative treament of pre symptom

‘tic cancers destined to cause foure meobidity and

‘morality, reducing the burden of ines: among older men (95.295) However, thishypethesishasnosyeeen

‘ese in well-controlled scent research and, despite

it atrctvenes, might nt be comet

‘Wir night serening Flo result inducing pros- tae cancer morality and meobidity? These potential problems are both general to screening for any cance, and eltvely specifi to prostate eancer Data frm un- congolled seoening toes that report the probability of detected cancers progressing to more serious stages (nage shit data) do no ecesaily pedi longterm r- uctons in cancer monty This is because of "ead time bias,” the phenomenon ofa sreening test finding cancers cavierin ther courses without changing thei! timate outcomes, and because of length bis,” in which test may prefereatally find low-risk, slow-growing

cancers (81, 136), As deteried by Sacket and cl

‘eagues (292), 0 the bass of stage si at, early

iagnosis wil always appear to improve suvial even

lace te pain at increased fame rik of moe serious

cancer Advocates of srening belive that the stem

ing ests caren avaebe for prostate cancer cannot

every detect hve sal, anes ences (12,295)

however aggressive susepes of performing systematic

biopsies ofthe prostate following suspicions screening

tests wil ncreae thei deteton G19) The tue, unveaed, natural history of cancers d=

covered by seeening (whether they woud ukimate-

2y ease any hart the pind) i known Because

‘many poste cancers grow relatively slowly, he eve benefit of ening cancers detected by srening main:

anknowa The fact that many prostate cancers, even

those detected by screening, tave sheady spread

rough the prostate capsule, forber dt any benefit

of seening Furhermore, according 10 one theory

eww fom observations of breast cance (and unesed

{or prostate cancer, poste cancers destined 1 cause

morality may actully spread cutie the prose early

‘on ven when they appear fo be confined othe prostate

‘pon examination of sve removed in a pronatctmy (17, 240), And finaly, aggressive curative nesment of

rosie cancer cares risk self hese ths, which in-

lade postoperative heat dase, impotence, ocon=

‘SPECIAL ISSUES IN SCREENING MEDICARE-AGE MEN This repo focuses on sereening Medicar-age men, 65 and olde, Beemse prostate cancer prevalence

ad metliy creates subtly with ge, Medicare beneficiaries would appear especialy likly to benef ftom screening (assuming weatment wotks) However, {hese men also havea higher isk of dying om meal problems oter than porate cancer, and they hae ewer years of life expectancy ding whic o reap the poten

‘ial benefits of serening (ee table 22) Funhermore, See fe risks of aggressive prostate cancer weament 9 create with age, making thes nen puy higher

“le for any expected benefit of seening, Te dii- cally of curentsrening technology in distinguishing

‘exween psealycoble posts cance and the a cancerous condition BPH, whose prevalence increases

© các sotneneesornoseCocotorone neo,

wa age, also reduces the value of screening Final

‘elder men are also a higher ssk of harboring lire can:

‘ees and cancers witha poor prognosis tat have acady

spread outside the prostate (233)

‘CONFLICTING GUIDELINES

‘ON EARLY DETECTION

Atpresent, the American Cancer Society (ACS) and

the American Urological Association (AUA) recom-

‘mend DRE and PSA determinations to evaluate the pros-

tar gland for cancer staring st age 50 (age 40 for men atinreased ris) although ACS acknowledges that, “re-

<ucton in morality rom serening has ot yet been oe

‘umente” (11,237) ACS recommends annual exams Ta

adtion, the American Medical Assocation (AMA)

recommends that PSA should be covered every three

yet for men overage 80 as part of standard insurance ‘benefits package (10

ACS and AUA do nt specify a definite “stoping

age” for serening although ACS recommendation 2c-

knowledges that, “genenlly, men with aie expectancy

of atleast ten yeas after detection may benefit fom ex-

amination” These guidelines, which were adopted after

the introduction of PSA into usual urologic practice, are

‘consistent with recent published reviews that suggest

lysine reserve erly detection and aggressive teat-

seat for men with life expectancy of more than ten

yeas (50,204); in he United Sates, for men with aver

age commit, his threshold would come at bout age

73 AMA recommends coverage of PSA testing up

through age 70 10)

“The 1993 US Preventive Services Task Force up- ate (952) andthe 1991 Canadian Task Force on the Pes odie Heath Exainaton (57) found evidence insu ent 10 recomment for or against DRE, and fur evidence exclude PSA, from the peioic heal ex- amination The College of American Pathologie r2-

‘ommends that PSA not be used for serening among the

‘ener asymptomatie farcases where prose male population reserving its wse

cancer is suspected (20)

‘The National Cance Instat (NCI) se to tem

‘mend hatsnen overage 50 recive a DRE, bat nota PSA test Recently, however, NCI has decided not 0 make any recommendation concerning cancer srecning,det- ing instead tte evidence-based policy guideline de- velopment proceses used hy the U.S Preventive Ser vices Task Force and the US Agency fr Heal Care Policy and Research (AHCPR) (199) +

Reasons for Conflicting Recommendations Inthe absence of well-controlled duc; tạ ca lish the vist and Reni of screening fr prostate can cet oF een large, controled tls that document the

‘benefit of aggressive crave eure for cancer ạt

‘hasnt pres bon the prostate iis possible to te Pret the nonexperimental data ạt áo của lờ spDot any ofthese guidelines However, diferences in pet- spectves among policymakers, clinicians, and paints also contribute tothe eure contoversy about prostate

‘cancer screening For example, Adami and colleagues (@) recenly concluded tht, piven the possibly that

‘early detection of prostate cancer does move arm than

ood, even randomized wal f screening for prose

‘cancer might be unetid

rom policy perspective, some expers emphasize

neal imperative w avoid the harms of erly detec

tion effons in general, and mass sereeing in paul,

unless there is efnkive roo ofanet benefit omclini-

cal as (4,80, 167,302,322) Oers emphasize he

need todo everthing posible o lower the risk cancer

‘nt the ress of thoae stadia re avaiable (12, 13,68,

131,217,258) Sucke (291) has refered othe prtago-

nse represented in these basic iological disper as

cider advocates of te ecient method (sna), oF

advocates of eening (evangelists Te former pe

spective it incorporated ‘many grovps for determining the nt benefit of preven into set of extra used by

tive maneuvers in general and cancer serening in pati

lat, including the Canadian Task Frce om the PeisiE

‘ean Examination (6), the U.S Preventive Services

“Task Force (351), and the World Heals Organization

(368) No maner what expen groups recommend for

populations, on the level of individual patients and clin

‘ins ferences of opinion and variations in actual

practic wll exs (219, 238,240, “The rapid increas in medical cae ost in recent

years ha placed greater scrutiny onthe fecitenet of

‘medial interventions nthe past, medical interventions

‘hatseemed concepully sound wer often administered

un clinical wials proved they did not work 141 More

‘cen, the burden of proof fr some interventions as

begun sift to those who want to ute the eaten,

suggesting that these interventions be withbelé uni

sic! wins establish that they work (112) Although

recommendations my also vary depending on wheter

they consider the heath are ost associate wih aly

xX

URE 21: CROSS SECTONAL MLUSTRATON {OF NORMAL MALE FEVIC REGION

_ Rec

detection, none ofthe guidlines described above diect-

|y tok thes costs nto account BASIC BIOLOGY OF PROSTATE CANCER

‘The prostate sa gol-ball sind gland whose prims- + fneton the manufactur of semeo, he Mid ejace- aed with spem 1s found below a man’s Bader and surrounds the urethra through which rine pastes on it

‘way fom the blader (se figure 2-1) Prostatic carcno- sma (porate canes ie a eave slow-growing mlig- ney, withthe pote fr spread relied to both vol= lume of the tumor and degree of cel ifereniaton (he centntto which the cancerous cells are diferent rom the

‘oral cells fom which they arose) wich bemesves

a elated,

@ cox: 0 ences Or Frost Coes Stans MEER Min

In careful sues of atopy material, MeNeal and

colleagues have documented that tumors less han

approximately 0.5 ml are commonly found among older

ren, and are rarely associated with penetration ofthe

prose capsule (called capsular penewation) (233)

[Above 0.5 ml, penetration ofthe prostatic expslebe-

sin tobe seen and over metastases (spread ofthe can- ce) begintobe seen with umors above | ma pari

sary above 3 mL, along with more frequent capsular

enerton and invasion of he suroundig tue Ole

tiem ave lager tumors, ad larger tumors are more

‘ely tobe ess wel differenti Clinically ocalzed

cancers are eximated to ive a doling time of to

years o more (299, 325, 328), Based on epidemiologic

observations, Stamey and colleagues (328) doubt that

cancers less than OS ml a volume are key to cause

iaure morbidity and morality given his long doubling

time: however, al lage prostate cancers were undoubs

‘ly small a some point

Prostate cancers are described by tumor grade (ihe

extent of cell ferentiation) and stage (how advanced

‘he cancer has become) In sues ofthe nara history

‘of prostate cance, grade and stage are use to predict,

‘malignant behavior, The most common grading system

‘isthe Gleason score, which els asum of 21010 based

‘on the two most common ptems of el diferentistion

inthe tue sample Tumors assigned sores of 210 4are

considered “wel diferentited"; 5 to 7, “moderately

.iferenisedf; sai to 10, "pool deren”

“The two predominant staging systems for prose cancer are the Whitmore (A-D) system and te Ture [Node-Metasasis (TNM) system (245)6 Table 23 de- Scribes the wo predominant systems Although increas ing sages of prostate cancer peBenlly indicate a poorer prognosis, different stages can behave simialy (ie, Stage TIWA2 and T2'B1 (340), As will be dscosed

tế d unreliable, and many ances thought tobe lealized 0

‘he prostate are foun tobe more advanced upon srgey In ation, the grade fs tumor evatuated rom a biopsy {a procedure for eovinga smal sample of torte termine itis eancerus) may diverge fom te graded termined from an examination of the suricaly removed prosaie (7), These phenomena make it dificult 10

‘compare the prognosis of prostate ance patents staged an rated by afferent methods

cians tempts to stage paints” cancers are

RISK FACTORS FOR PROSTATE CANCER

“The cause of prostate cancer isnot known, although evidence poins to both genetics and environment as avin roles (62, 85,273, 310):

1 ge isthe mest important risk factor, withthe inc ence ofboth prostate cancer diagnosis and death in- creasing sharply with ge (able 2-9

+Family bstory i als determinant of risk Men with

‘ne immmedine relative with prostate cancer have

‘twofold inreased risk, which increases 1 roughly

Cưng2 monsrCmcsnOunMe GP

trical sage

‘Stoemcsea pos wih encod ma orn te

we “net eorcegienebbeofhagroidleere

* TreNetom Veosergemdbpoxrn ‘Shiuegecatnan! Senter mwa cone cored by ule oreo

Đ x a Tê sen rn rosa aE ones eon nc wig yroh rece 2m race

Tivefold with to aected family members (323, + Afiean American mea, who have generally been 380), A recently decribed hereditary clustering of unteresented in voluntary prosiate cancer screning

‘rove cancer in families may be responsible for programs (104), havea 1.3 01.6 old higher ik of shout 4 percet of eases in en under age SS and 10 prutecancer than do now-ASican-Amercan men peoent of prsat cance eases over (9, 60)

Research at hon stattial asoaton between

letary fat, pariclarly animal fat fom rod meat,

and prostate cance (142, 286) Although fat may not

ety cause prostate cance, it may combat indi-

realy by affecting cenain hoomone levels in men

en

Several sie ave founda weak sisal aesocin-

tion between prior vasectomy and prostate cancer

(14, 141,288 However, because the assoclation is

‘weak, beeme contradictory dta exis (14), and be-

cause there 8 no convincing biologies!exparation

for this resus, eawsalty cannot be considered proven

(053,165,

“The lack of data on vs factors that could change

(except perhaps reductions in dietary fat intake) makes

‘he potential fr preventing rosa cancer before i de

‘velops modest a this pont However, considerable ia

terest at arisen in uying prevent prostate cancer with

rugs A randomized linia al of prostate cancer pe-

eatin using Finasteride, a drug employed in eating

some cases of BPH, i jst geing underway (49)

‘THE PREVALENCE OF PROSTATE CANCER

Im ower fo analyze the potential impact of» sren- Ing program a is atemped in chap 5, tis mcesary

tw know the age-specific prevalence of litt prostate cancer inthe population Table 2-5 presents xine for prostate cancer prevalence derived from a syuhesis of amopsy sedis (24,13, 138,134, 19, 222,293, 305) Aogetber with McNcal's analysis ofthe volume of can cers found at autopsy 231) 1 presents estimates of the robebilis of menage 65 and olde aig ino one of

‘he four fllowing states fea: no cance, cancers 05

‘ok or les in volume, cancers greater han OS mL sil confined othe prostate, and cancers greater han 0 ma spread beyond the poste capsule, Appeniix A describes the methods used wo derive lable 25, These probabilities can only be considered es- timates because patems coming 1 atopy may not be representative ofthe general population, and becatse scarce data exist scribing distributions of mops ean-

‘cersby hostage, and tumor volume and extent However, auopsy studies were exchided fom tis analyse unless Pints wih cancers suspected before death were spe

‘ically excluded PROSTATE CANCER MORTALITY

‘Te discussion of treatment effectiveness in chapter

4 reviews epidemiologic data onthe natural history of nueated, clnicaly-significant prostate cancer The sge-danlirdized morality rae fr prostate cancer in- eased from about 21 0 25 per 10,000 mals inthe

‘Unted States between 1960 and 1988 (39; meanshie,

‘he inidence of poste cancer in the United States has Increased mich more dramatically at fst ein par wider ws ofthe surgical procedure anual sec tion ofthe prostate, fr symptomsof BPH (274) lcreas- ingen detection effort have sstined this tend in r=

Cương: manmCacnuOosa E>

emt year (105) These tends are flected in an

Increased wndency o diagnose cancer at ạt advanced

stapes, and improved stage-specific fveyear survival

rates 238,330,

‘These statis alo emphasize he danger of wsing

sage hit” datato make conclusions about underying

cancer morality, shift toward more localized cancers

snd beter outcomes for inva paints in recent

years as actually been accompanied in the rate of prostate cancer morality, rom aeatonal by asmall increase perspective However, since aggressive early detection effons ae a relatively new phenomenon, some yeas may be required before this strategy result in any de

‘crease in population-based rates of prostate cancer rmonaliy

E>

3

Technologies To

Detect Prostate Cancer

‘i mostcommonly used technologies fr detet-

ng and diagnosing prostate cancer are digital

rectal examination (DRE), posate-peifcan-

tigen (PSA) measurement, transrectal viưa-

sound (TRUS), and wansecal needle biopsy of,

the prostate (TRNB), For primary-care based cas-find-

ing and mass screning, TRUS and TRNB would be lo

-iscaly difficult wo include as primary screening ets

‘iven thei relative complenty and invasive nae

“Moreover, the marginal vale of TRUS above DRE and

PSA seems tobe smll(18,91,215) and he rskand is-

comfort of TRNB would seem to obviate its ác sa ti:

mary screening test Therefore, this chapter considers

the use of DRE andlor PSA as primary screening test,

and TRUS and TRNB as followup, conimatory tet

‘To analy the impact of screening, itis necessary to

‘know the “operating characterises” ofeach screening

technology In general, the opening characteristics,

hich fer tothe ability of testo finda cancers that

would cause harm and to find only those cancers, a ex-

pressed in ems ofthe sensitivity and specificity of the

tes (Box 3-1 describes these concepts) Unfortunately, the “ue operating characteristics of DRE and PSA

cannot be defined since few studies have evaluated hem in populations where the tru underlying prevalence of

clinically signfican prostate cancer is known, The fact that small volume, well liferetated cancers shouldbe

‘considered a “nondiseaie™ and tat iti lavely easy

to detect advance cancer which may offer no thrape- tic benefit farther complicates the design and analysis of

‘these sues ‘What re sully avalable are sues of he “positive redicve value” of tests, the propoton of postive 0 spicion tet ests that imately tur otto be ean

ca (Se box 3-1; in hese studies, patents with "pega- tive” tet reeults do nt receive fllowup TRNEB (even

‘though they may harbor significant prostate cancers that the screening test didnot find), Farhermore, these tuc ies use different combinations of primary screeing tests and diferent strategies of followup evaluation Finally,

‘the studies dont uniformly provide age-specific pedic- tive values, which are important to an analysis of screen ing older men,

‘To overcome these problems, this analysis presents

“ethood ratios” of disease (292) for DRE and for PSA These likelihood ratios are estimates of how many times more likely a patient witha patil test rest is tohave a given type of ance than if he patient didnot Ihave the test The probabilities of cancer with notes are the prevalence estimates found in ble 25 Appendix C

€ conoetnenneeorhoneChenkrroneluEcrmrMớ

'8OX +-:DESCBANG THE ACCURACY OF SCREENING TESS

To crayee i mart of asceenng pagan Minecesay founded ne conaany tach eng eo:

9 trea fo 0 sh “omg cece he es Tse cechng cece ch oe he

‘iyo testo eo ent cos ono teeny cess, UH eos hs Oe ss ay od ‘Posty emamty wha perenne a ch scene pata wen re ho el pose tbe pacey he perce {pe tof teh sleered peop we et nega oer wach saraey ue cy 0 I ope

‘Sma wt paciy ters sey ee Meaty perp corechy These acer ate Mee Ira reo pore te ho pce pave hs of nae vo et oh) a eng eae Per ope ofceone aici e cre ese ie eat nepene) For eno oo wah ey Deh 7

‘on 8 omc! wet hove & famenegotre foe of Sto 3 xe he toa do đt tế cor 2 free, ‘acter anne cra om reas eps one care

> a‹a

ta ere wow he ha unser porno oo tae ne sowered popactn vogoon 2 neo one oo sso oe tren tem veoe seen ot nega ne

“iy ot cone namo ete cance eer bq boca oft ne ch ond ơi

‘ten elon acecane nat kevect nee Sones hence, ow Ader spt he ts Imecare of oe cma ecvo te

‘anv pcre ‘Senoge of pcos wi pose os eh we Yum OP ne FP per (She efcbe xaya AB e nu bron

‘ae tow a te ur event ot

Sợ cendien s dich tơng te he meoc Soe oe conan bang scenes or Neto Alon PV icy nro eso „NG tắc cơn hs =metoet reeEe to ho bom otew Vanda bw ones ate hao (oe wo dere Be Roger Fag ‘Seow peonges er pines ome

"he ý ntes meses ot enng Geo092/ lì CICA 0 ow PY a a ey ance sweated o xeon rumba ot esas rae Pe he re ate cwacled wy Klong, Mowe EY

“ren chicos roy arc fo oars aay on realy enone wh sorry 8 aa 2

‘oe trọt at tet ond om tec wanxsar o ewan ny exe UY erconng woemer neu pte corer ieeeg a maf bam Re anne sees he Bs

Cư emooamToDotetnowatcacm Ầ

postive predic va in diferent sds depends

eavily on he aggressiveness ofthe followup ng

ployed forasusiios est Sates ten find more

sponse tos suspicion primary tt (7)- Using thismet-

lap, est tha has poor emiiiy and spe

icv vate of te soagy For example ster of

efoming mllpe se of biopsies ca all men with

‘ronneyes wood probably have aaher igh "ili

tne cancer Rye coo, in estan, becomes oe for

receiving the mor aecurate dagstc st, TRNB.A o>

sly sic qusdrants of he pst that yield

cloewhee inthe prose as result of be systematic

Sop

DIGITAL RECTAL EXAMINATION

“The digital rectal examination, in which he clinician stvemps to fee bnormalies inthe sie or shape ofthe prostae gan through the rec ia ime-honored test {or he early detection of prostate cancer despite very

‘soak agreement song published guidelines abou is

‘ale (100, The DRE simi insensitivity becaose of

an nhily to detect tumors deep witha the prostate land, Because larger tumors at easier to fee, DRE is

le to detect insignificant cancers (although this rik wil increase i suspicious DRE triggers a set of systematic biopsies in addition to a bop ofthe ss cáo ares), The detection of larger cancers also means chat a eluively high perceaage of DRE-deteted t0-

‘mor half or more) il ve sready spread beyond the confine of the prosaic capsle (139,279,271) Many lavesigats have been concerned aboot variation mong physicians in ther ability to detect cancers by DRE (271, expecially the possibility tha DREs per onmed by rimary care physicians may not be as dis ciminaing a urologists" exams, However, le empe- icslevience exis to address his concer (354)

‘Appendix C lists sade of primary DRE sreeing for prosate cancer, with bie descriptions of sy methods and resus Comparisons are dificult given dit- ferem patient popolations, diferent tresblds for eal- ing a DRE “suspicious,” and ifferem strategies of fl- lowup testing One study by Chodak snd eolleagues (79) rovdes the mest detailed presentation, and allows es

@©D cons me srtcnouse OF Powe Concer Serene W sey Men

"ARE ESMMATEO URGLNOOD RANOS FOR RESUS OF DISMAL RECIAL XAMINATON CHANGING TH 0005 OF SIGNIFICANT ROSTAE CANCER 60.5) OF DHFEREM PATHOLOGIC BE

Ueto Rete

<i creme na ow tegrated

imation ofthe likelihood of cancers with and witet

capsular penetration (able 3-1) fr each DRE est re

sul? Appendix B discusses the methods used opoace

thes etme No clinical aloft use of DRE slone

for the early detection of prostate cancer ae avalabe,

However, itera case con study (129 ora dei

‘on model (241 as suggested an inponant survival

‘ene for men eeened with DRE

PROSTATE-SPECIFIC ANTIGEN -Prosate-sperifie antigen is» sheoprtin produced

‘nthe prostate pland with probable ole ia the wasp

of semen, Because cancerous prove tse, gram for gram, prodoces preter quanitis of PSA than does no tal or bengnly enlarged tise, and because poste

‘cancer may inerest the likelBood tit PSA "leaks mo the peneral circulatory system, serum (lod) PSA lev-

‘lk appear tohave soe discriminating capaci For rs- tate cancer (99, 257) Psliminary evidence sogest prostate cancers need to be geste than min volume before hey cause an ncrease in srum PSA (49) “Thre PSA assayshave ben commonly sed liscl- Iya describe ithe lteatre (172) Hybxteh's Ta dem PSA assays detect PSA with monoclona! antibody,

Cương TroweroeesTo Orne Prosar concen ED

probes; these assays use radioactive antibodies and en-

_ymatc reactions to perform the measurement The Tan-

‘dem PSA tests c cunenly the ony atsys approved by

‘he U.S Food and Drug Administration (FDA) fr use in

conjunction with DRE as an aidin he detection of pros

tat cancer insnea overage $0.3 Abbot's IM PSA assay ses @ microparticle enzyme immunoassay technique

‘Yang's Pros-Check PSA assay ues a polylonal ani-

body probe to measure PSA (386) The levels of PSA measured by the Hybvtech and Abbon assays appear

oughly similar (190,355), while the polyclonal assay

ans values about 1-0 higher (148, 338) However,

investigators have recently raised concems about the

calibration ofthe Hybrtech and Abba assays (48,149,

226,266), wich ogeter dominate the PSA assay ma

‘et Clinicians need to know which ts their Iboratory

‘wes and to considera switch in assay inthe “iferen-

til dlagnoss” ofa changing PSA ina given patie

‘One potential ificuly with hs sevening esti that

‘actors ober than prostate cancer ca temporarily ele

‘ve PSA levels for severl week: acute inflammation

(f the prostate (prostatitis), acute urinary retention, a

diagnostic medial procedure called rigid cystoscopy,

“TRUS, TRNB, or prostate surgery (193, 262) recent

study has also found temporry elevation in PSA fol-

owing ejaculation (250) However, sever studies have

now documented that there is no clinically inporant

elevation in PSA values following routine DRE (95,

371), an imporant fading since physicians often per-

form DRE and PSA a the same visit

‘values (able 32) One study sed the 95 percentile of serum PSA among men without evidence of prostate cancer as the upper Boundary of the reference range

“-xeerefgsenooarobce trữ h N‡e0gvechoron ông rtzoelølerZh4 1) 2zegze\egxergirzzenrdf

© ors eo ercenvnes O Prose Cancer Scams Ee Ne

(260,26, wile he oder wed a sigh tees, bot methodological smitar approach define te ppt

limit (101 Another recent stay compared the perfor

‘mance of several PSA tet kits as prt f an neato

A sandardiation conference (329) ‘Appendix Diss published studies thats PSA asthe

primary srening tool to detect prose cancer (DRE

‘wed only to folowyp «suspicious PSA)7 Although

these sodie geerdly have a somewhat higher propor:

‘ionof subjects with a cance detected than do the sais

‘of primary DRE, these proportion are key unereti-

rmaes of tbe maximal ainable yiel since patents

‘were fle not biopsied uns followup DRE or TRUS vas also suspicious Using data tom the Calon and

‘Brawer studies, lelibod rats for Hybrtch PSA re

sul fiero cages were called as deserted in

appendix B and are provided in able 33 (44, 66,70) ‘Variation the use of PSA for srening have been

Proposed a improve the operating characterise of this

vest for prostate cancer (96, 182) These variations each

of which ns its own drawbacks inch: 1) A densi

(SAD), a method of comecting the rw PSA vale by

the volume ofthe posta, as measured by TRUS (32, 33,284); 2)aprediied ‘ume gaint which mensored PSA is compared 10 make PSA gPSA) base on gland vol

‘ecsions about proceeding iogpy (205% and 3) PSA loci, thera of change PSA overtime (3, 64)° Research curety underway may ea tet for more speciicrypes of PSA (36,37 106, 211,212.213)erath- er pes of biological substances (171, 298) that more recy identity men with prostate cance

‘One-Time Versus Repeoled FSA Screening Much lest ix known sbout the rea of repeated seeing wth PSA tan shout on-imesrening This

‘2p in our knowledge is significant since a Meare ‘rotate cancer sereeing benefit would man! klycov- periodic serenings, not one srenig pe iftime

‘The few sues that ar avalalesugest a decrease in the proportion of sreenes with cancer over repeated serecings (46 47, while the proporion of patients with cancer confined 10 the prose capac appears 10 in-

‘crete: 7 percent opposed 10 8 percent in ove ie (1,70, aod 87 perceat vers Số percent in anther se- res (46) Appendix E summarizes these srdies

Tế chi hovhy so cntergectrecn 2ynd eaecetrztcrce re reơccecxeezZerecanvcns 2zeag

KT na awưwnnnnnnnnnnnnnd

SA Screening Among Men with ‘expt as conferred by thee age) (239) and in one large

Symptoms of BPH

As noted cari, benign prostatic hyperplasia (BPH)

can ise PSA levels complicating PSA mearuement

Given the widespread prevalence of urinary symptoms

Inicatve of BPH amoog oler men, PSA screening for

‘prostate cancer among this lrg group may ye ile

‘weful information Men with symptoms of BPH do ot

sty, when controling fr age, men wit symptoms of prostatism actualy had lower chance of being found to hhavecancer trough DRE and PSA sreening (72)-Inad- dition, because BPH and prostate cancer share symp- tom nd the likelihood of elevated PSA levels, the pec-

‘cy of PSA deweriorates 10 50 1 79 percent smong

‘men with clinical evidence of BPH (173, 30) Baber- more, there appears to be a great degree of overlap

€ cínooteennesorhonrCacnternsokioormrMe

among men with localized Ginracapsulat) prostate can-

‘cer and BPH, frber limiting the vale of PSA testing

mong men with hese symptoms (3090

COMBINATION Alnough combination sereening with both DRE and OF DRE AND PSA

PSA may cumenty, be the most poplar sacgy of

segresiveofe-based eal detection of pros can-

cer among US wologiss, studies of he predicive val-

‘eof this strategy ae only just becoming avaiable for

lows populations DRE and PSA each dete some

cancers not denied bythe ober modaliy; therefore,

the yield of soreening program (ihe percentage of

screenees who ulimaely havea ance ontemed) can

‘be increased (1 roughly 4 percent) by combining bo

tess, In ado th sais of combination testing te

pore recently hve generally performed ast of y5-

vematie bogies if eer et suspicious, a well at

topes of suspicions lesion noted on followup TRUS;

thismore aggressive use of TRNB also contibues othe

higher yield sen in these ties However, hese more aggressive sraogis reel ia

forming bogies cn upto athird of al screenees: the

ditional cao detected mol be weighed aint the

cata ik of biopsy Furtbemre, these sade were

conducted among volutes, and some data suggest that

oluMeer may havea higher "prior probability" of pros-

fate cancer tan antcleced men in the community

on

“he newest sudies where DRE and PSA are per: formed inthe same men make itclear at PSA isabeer single ethan DREin erm of etcing ences nd of seecting cance stil confined within the prostatic cp: tule (28,72, 119,263,279), FOLLOWUP TESTING

Increasingly, followup strategies for a suspicious

‘DRE or PSA include bth TRUS and TRNB Moatinves-

‘igmors use TRUS to guide biopsies of reas deveined to besuspicious by DRE or TRUS Many ciicins now perform mul systematic (four to sx) bps of the Prostate (ina single procedure) in addition to bopses of ‘spicons area, since a patient with 2 normal TRUS may actully harbor cance 12033 percent ofthe time (depending on the PSA level) (157) Others base dec

‘ons about wheter to perform systematic topes on

‘aw PSA values or PSAD vale (29,99, 306) Akthough some investigators advocate simply following rea with

sd PSA elevation (Le inthe 4.110 10.0 ngage)

‘fe DRE and TRUS ae negative, when aggressively ranted ht group yield the highest percentage of i-

‘raapaular cancers, the el uret of screening

‘There is lo varsity in bow lncians allow men

‘who havea negative set of biopsies after 8 suspicious

SA test Some urologists recommend repeating the ys

‘emai biopsies atleast one (paricualy fora PSA

‘eater than 10 g/mL: others perorm followup PSA

Cưng3 EowaoeeoDmerFesusCacm €ŸQ

‘ess more fequenly than annually and rebiony fore

‘her persist elevations a arising PSA vale Often

then, a suspicious screning test, even if followed by a

negative biopsy, wil ad to beighened survellane for

ovate cancer and further ts and biopsies ine f-

ture On the other hand this ore intensive urvelince

{num inereaes the ya of screening to sme degree

‘Transtectal Utrasound

Because ofthe anatomy ofthe prostate gland isl,

‘TRUS has much beter senstviy for ences found a

certain pars ofthe prostate than fr ches (334) App:

lx tsar hat ose TRUS a primary mean for

carly prostate cancer detection In oe ofthese sues,

a demoasution projet ofthe American Cace Scie,

oo 14 peroent of men hada suspicious TRUS, an 15

erent hese men ad cancer lower predictive valve

‘han tudes of DRE or PSA alone (235) Inthe absence

of asuspicious DRE or clevaedPSA, the peice vài:

‘we in this series dropped to 54 percent (19,215) In 2

study based ina urologe practice where the prevalence of ence was especialy high (tection ate f 146 per-

cen), and where aboot half ofthe men were biopsied

‘based on esas of combined serening (DRE, PSA, and

TRUS), Cooner and associates found th fren had @

SA lea than 4m and anoneupiious rectal exam,

the yield of ulaasonopraphic cent Pat nanos way tbe overall ie ofthe testing screening was bout 2 per-

straegy oaly increased ffom 135 10 146 percent

trough the perfomance of TRUS in ation to DRE

and PSA (91)

Several studies provide more dict evidence about

‘etre senstvity and specify of TRUS than isaval- she fr DRE snd PSA Twostdis were sbletoestnate

‘he operating characteristics ofpeoperaive TRUS p= formed on men aeady scheduled for radical poste

tomy forcaneer or BPH, The sdy on men sched for prostatectomy for cancer showed a TRUS sensitivity of,

52 percent and a specificity of 68 percent (6), nd the idyof men with BPH showed a sensitivity of 20 per

ce (315) These eave low sensi estimates for TRUS ae amor eaon forthe inereasing tendency lo perform tematic biopsies for aspEious DRB o PSA resus, even if TRUS does ot india anything ssp lous, Fermore, thes and oer studies 337) soe set that TRUS teds o underestimate the siz of ean cers tht we detected, making it a problematic technology for Metfying men wth smal cancers who nay not need aggressive weameat Fly, evidence

bo suggests that BPH may also erode the ability of

‘TRUS to detet cancer (4) TRUS ine doesnot appear tops any ik fo pa

‘ents, lfiough does pose costs to phien or thế eat inser In 1992, Medicare reimbursements were

$89 fora dagnonic TRUS by all so S189 for 3

‘TRUS ued biopsy Tronarectal Needle Biopsy Moder transrectal eae biopsies CTRNBS) ae uso

ly done with ulrasound guidance wsing a needle rmoumed ina spring-loaded biopsy “gun” Biopsies can

te rected toward seas deemed suspicions by DRE or

“TRUS, cr performed systematically to sample he entire prostate fen six biopsies ar taken in a sextant pater ftom diferent pars ofthe prosate gland (326) TRNB {suncomfocable and can be complicated by infin or

‘eeding (89) Complications of biopsy include wiry

‘ret ifecton in 0.510 5 percent of patients and rose sis nan estizsted 0 pret (no deaths), despite rou: sine sta prophylsns (16,95, 108, 160) Some p-

‘ents abo experience bleeding ess than | percent) vith ery few one ou of 835 biopsies in one sty) requiring

‘eanstasion 1 108)

© oe so rca Or Prose Cance SemPuns hin Mon

‘TRNB soften considered he “gold sundaes for

the diagnosis of prostate cancer however, tis inees-

‘ng clear hate gold standard is "umished” to some

epee ners of the sensitivity of TRNB, ivestia-

tors from Washington University have found that when

‘mente fund to haveapersiseat mild elevation in PSA

(41099 api, repeated biopsies find a age number of caners presumably mised by previns biopsies Ia

‘one preliminary sepo, 25 percent of these men with one

_reviosly negative biopsy bad cancer, swell a 4 pe

‘eat wih 190 previously negative biopsies and 10 pr cent with thre previously negative biopsies (187) Al

hough many ofthese patents ad gina biopsies that

were directed by abaormal DRE oe TRUS resus instead

of moliple, systematic biosics, snalation modsing hat lio suggested systematic biopsies maybe relatively

Insensitive (103) Interns of specifiy, TRNB can dete “incident”

cancers of les han 05 ml in Yolume, which (as i-

cused in chapter 2) may likely pose mo eat tthe ps

‘iets heath, making the concepully equivalent 10

“fale potves." Ti rik ncreases more biopsies ae

performed, and particularly sith repeated syematic

biopsies Tes and colleagues rece estimated tht

‘he probably of nding an incemal cancer ona set of

sie bopsies was approxiatly 4 percent 338)

‘SCREENING THE MEDICARE POPULATION

‘Age hat a comple lect onthe resus of screening

fax prostate cancer Te price probability of cance in

ress wih age, bot the percentage of ogan-onfned

‘cancers decreases Farthemers th specifiy of PSA,

and probobly DRE as wel, deteriorates as more men in

the popolasca have greater amounts of BPH Richie and colleagues (279) present te net elect ofthese factors

‘sing da oe trae, si-cemer sy fsrenng:

* The deteriorating specifiy of he et with age = sued ina steply increasing namber of aden with saspicious esl on ether DRE or PSA that wuld

‘ener recommendation for biopy: 18 percent ages S010 59,28 percent at ages 601069, and 40 pee- cxntat ages 701079

1 The ring prevalence of cancer manned the me: ctv valoe relatively constant, 2 hat ance wa eee in 2,4, 2087 percent of these age BOP, = spective

Among men whose cancers were pttologially sage, te percentages (etinon no speci) by age groups were 74,76, that were organ confine and 0 percent

* his sad, for men ages 60 069, ang PSA in- creased the pereatage of men with a suspicions screening evaluation from 16 percent (with DRE lone 1028 perceot imerestngy, tae percentage of tins with patologicaly localize cancer did not decent withthe adton of PSA n Đi s 00p Formen ages 70.79, adding PSA to DRE increased the percentage of sspcious evaluations from 204 erent, with an ncreate inthe esing percentage

‘of eran-cofind cancers detected fom 45160 per- All of tes data sogest hata screening programs, especially those employing PSA as one screening technology: ae directed toward older populations, the numberof patient quiring more costly, invasive, and