To Color Doppler Image the Prostate or Not That Is the Question prostatitis

					          Archie                 A. Alexander,                               MD

          To Color   Doppler     Image                                                                                                              the                Prostate                                     or Not:
          That  Is the Question’

I       N this          issue          of       Radiology,                   Newman                     et        tate     epithelium.             This      protease               is de-                 the        low     positive    predictive      values     of
          ab (1)              to support
                          present              the
                                                data                                                              tected       in the serum                with       one      of two                      18%-60%               for lesions      sampled     (1,25-28).
contention     that color Doppler          sonogra-                                                               methods:            the       monoclonal              Tandem               R             Part          of the problem         lies with the non-
phy (CDS)     increases     the information                                                                       and      Tandem            E (Hybritech,                San Diego,                       specificity                of lesions           imaged              with        TRUS,
obtained   from transrectal        ultrasono-                                                                     Calif)      or the        monoclonal-polyclonal                                IMX       especially       hypoechoic             ones. Only        a
graphic                 (TRUS)                biopsy              for        the        diagnosis                 (Abbott           Laboratories,             Abbott          Park,          Ill),         third     of them        are proved          to be cancer
of prostate                     cancer.                However,                       Kelly        et al          with      “normal”              being       0-4      ng/mL           (8,16-              with differential            possibilities       including
(2)       indicated                  that        the           increase                in posi-                   18).     Others         may       use    the       Pros-Check                            nodules       of benign         prostatic       hyperplasia,
tive        predictive                      value            (0.77)          achieved                             (Yang        Laboratories,               Bellevue,           Wash)                       atrophy,       infarcts,      and regions          of acute
was         not         sufficient                   to     warrant                   inclusion                   method,      which   has a slightly    lower                                             and        chronic             prostatitis             (29,30).           Further-
of this            technology                         in the            TRUS             evalua-                  range    of normal    (0-2.5  ng/mL),      but                                           more,   the underlying        echotexture      at the
tion        of prostate                       cancer             and           suggested                     it   sample     values  may be higher        than   those                                     site of cancer   may not be appreciably
would               detract              from             careful               gray-scale                        obtained      with the Hybritech      test. Al-                                          altered   to be distinguished          from the un-
analysis.                 The         one            feature             that          regularly                  though            the sensitivity              of the PSA level                     is   involved                 portions           of the prostate                     gland,
recurs     in discussions         about   prostate      can-                                                      clearly    better      than the digital      rectal   ex-                                which      is termed       an isoechoic      lesion.
cer is controversy.           So how did color-                                                                   amination         results,   problems      may arise                                     This lesion       is difficult    to detect,     and                                  its
coded      flow enter       into the diagnostic                                                                   with the nonspecificity              for PSA levels                                      appearance          depends      on the amount                                          of
evaluation       of the prostate        gland      in the                                                         (8,18-20).      This is particularly        true for                                     fibrosis           the       cancer         incites           or   its     degree
first place?                                                                                                      PSA levels        within    the range     of 4.1-10.0                                    of invasion    into adjacent      normal    gland.
     Everyone      is familiar      with the statistics                                                           ng/mL,             which may have increased                                 from         Yes, a small percentage        of lesions    may
that prostate        cancer     is now the leading                                                                benign            prostatic hyperplasia, acute                               and         even be hyperechoic,         which    means     that
type         of cancer                   in     men             and          is second                            chronic    prostatitis,           prostatic    intraepithe-                              cancer     may possess    any sonographic
on the             list of cancer-related                                       deaths             (3).           hal neoplasia,          infarcts,       and the like (19-                                appearance,                    thus        increasing               the non-
Most          are         aware               that          detection                   of    pros-               22).                                                                                     specificity        of TRUS (30). To improve                      le-
tate        cancer              can         be difficult,                      frequently                            To further        confuse          the issue,     prostate                            sion detection             and differentiation              with
relying              on       the        diagnostic                      triad           of the                   cancer        has been           found          in individuals                           TRUS,        investigators         have sought           addi-
digital            rectal            examination,                            prostate-spe-                        with normal            PSA levels      (20). These         may                           tional      features       to detect      the presence             of
cific       antigen                 (PSA)              level,          and           TRUS           (4-           be tumors          smaller     than    1 cm3, and some                                   cancer       such as loss of natural               bound-
8).       Once           detected,                    the       type            of     treatment                  of them        may be poorly          differentiated                                     aries, bulging            or contour        abnormalities,
given             the        patient             and            its outcome                       has             cancers       (21). In at least one study,               sev-                            and finally          increased       color-coded            flow
continued                     to stir           debate,                 with           some           sug-        eral patients         with normal          PSA levels                                    (2,3,29-31).
gesting               that          patients                 would               do      as    well               had clinically          marked      cancers        (20). To                                  The application             of color-coded             flow
with         no treatment                             (9-15).                                                     improve         the specificity       of the absolute                                    to the detection             of tumor        vessels     or neo-
     Most            physicians                        use an             abnormal                                PSA level, investigators                have developed                                   vascularity           is not new; increased              flow
finding              from       digital                  rectal           examination                             additional         parameters        such as down-                                       has        been          described            for      cancers             of the
and/or     an elevated        serum       PSA level as                                                            wardly       correcting        the absolute          PSA                                 liver (32,33),      kidney    (34,35),  breast      (36),
indicators      that suggest        the presence         of                                                       level to 3 ng/mL             for biopsy       to deter-                                  ovary   (37), rectum       (38), and prostate
prostate     cancer,    with neither         being      per-                                                      mine      the acceleration          of rise or calculate                                 gland   (2,31). This technology            displays
fect. Some       rely primarily         on the eleva-                                                             the PSA density    (19,21-22).    The latter                                             the mean       frequency       shifts of flowing
tion in PSA level, which              is a 34,000 mo-                                                             number  is derived    by dividing     the                                                red blood                 cells within                these        vessels           as
lecular    weight     serine     protease      liberated                                                          TRUS        prostatic           volume            (milliliters)            into          varying        shades     of the color spectrum
from both normal            and abnormal           pros-                                                          the serum            PSA level             in the patient.                 The           depending           on the direction          and speed
                                                                                                                  theory      behind         this calculation            is based                          of their flow (39,40).            Unfortunately,             the
                                                                                                                  on the contribution                 by benign          prostatic                         various       blood    flow speeds         detected
                                                                                                                  hyperplastic           tissue      to the PSA level,                                     within       the different       types    of tumor          yes-
Index         terms:                Prostate,         biopsy,      844.12985                                      since it yields          a 0.3 ng/mL            per gram         of                      sels may depend             on machine           sensitivity.
Prostate,            neoplasms,                  844.32         Prostate,
                                                            #{149}           US,                                  benign       prostatic         hyperplastic          tissue      el-                     Both operator-independent                     and opera-
844.12983,             844.12985
                                                                                                                  evation       compared            with the rise from                                     tor-dependent            parameters         such as gain,
                                                                                                                  cancer      that is 10 times             that amount.            The                     pulse-repetition           frequency,        and wall
Radiology                 1995; 195:11-13
                                                                                                                  cutoff     points       for initiation          of a TRUS-                               filter        settings          may        influence               detection                 of
                                                                                                                  guided       biopsy        vary from 0.1 to 0.14, de-                                    these          vessels          and        must        be understood
                                                                                                                  pending         on the study             (19,21-22).                                     by      the       operator            to    optimize               results
      1   From        the       Department                     of Radiology,                  Thomas                  Most investigators                 agree     that the vol-                           (39,40).
Jefferson      University    Hospital,     132 5 10th St.                                                         ume of the prostate                  gland      can be fairly                                  The development                           of tumor                 vessels          is
Philadelphia,        PA 19107-5244.       Received   January                                                      accurately            measured             with        TRUS          as long             often          loosely          termed              angiogenesis                    (41).
 1 1, 1995; accepted      January      20. Address    reprint
                                                                                                                  as a good           technique            is employed         (23,24).                    More           specifically,               tumor          vessels             may
requests      to the author.
      ,  RSNA,            1995
                                                                                                                  However,             detection           of cancer      with                             arise         de novo       (vasculogenesis),                                sprout
      See also           the article             by       Newman                   et al (pp          86-         TRUS-guided                 biopsy         is a different      mat-                      from          preexisting        vessels      adjacent                           to the
90) in this             issue.                                                                                    ter, with          many         investigators              pointing                to    organ           (angiogenesis),          or develop                             from

dilatation    of preexisting          vessels     supply-                            over, in microvessel          density     studies      of                                         radical       prostatectomy.                 Urol Clin North                     Am
ing the organ        (vascular       expansion).        Tu-                          prostatic   carcinoma,        Brawer      et al (48)                                              1990; 17:779-785.
                                                                                                                                                                                 12.   Walsh  PC.      Why                  make an early diagnosis
mor vessels      arising       de novo      or vasculo-                              supported      this concept        of neovascular-
                                                                                                                                                                                       of prostate           cancer?         J Urol 1992; 147:853-854.
genesis    are associated          with endothelial-                                 ity as a prerequisite        for tumor       progres-                                       13.   JohanssonJE,        Adami      HO, Andersson         SO,
lined vessels      lacking       a smooth        muscle                              sion. This group         of investigators         (48)                                            Bergstrom       R, Holmberg         L, Kruserno     UR.
coat      but surrounded                      by neoplastic              cells       also observed        a high degree        of correla-                                             High 10-year       survival     rate in patients     with
that may possess          low impedence             flow                             tion       with      the disease                stage,       which          may                   early untreated        prostate cancer. JAMA
                                                                                                                                                                                       1992; 267:2191-2196.
(41,42).  In addition,        high-speed         flows                               provide        prognostic       information.     Does
                                                                                                                                                                                 14.   Lu-Yao      GI, McLerran       D, Wasson        JH,
may be detected          if arteriovenous           shunts                           this mean         that 1.3% of patients        with                                               Wennberg      JF, Prostate Patient Outcomes
are present.      Another        feature     of the tu-                              neovascularity            that could      have been                                               Research Team.         An assessment   of radical
mor to be considered              is its size or vol-                                potentially         missed     with conventional                                                  prostatectomy      time trends, geographic
ume, with tumors            smaller      than 2 mm in                                gray-scale             analysis    are at increased risk                                          variation,          and      outcomes.            JAMA         1993;        269:
diameter     being     avascular        and those        at                          for distant             metastasis    and are dying of
                                                                                                                                                                                 15.   Adobfsson            J, Carstensen               J, Lowhagen               T.
least 1 cm3 possessing   vascularity  (42).                                          their disease?
                                                                                                                                                                                       Deferred            treatment           in clinically         localized
All of these factors will play a role in                                                 No one is advocating                             replacement                  of              prostatic          carcinoma.            BrJ Urol          1992; 69:183-
detection     depending         on                 the      color      unit          gray-scale             analysis             by color-coded                  flow,                 187.
used (39,40).                                                                        but, as with other       secondary   signs sug-                                             16.   Brawer    MK, Chetner       MP, Beatie J, Buchner
    The radiologic      detection                         of prostatic               gestive   of cancer,    such as the loss of nor-                                                  DM, Vessella      RI, Lange    PH.    Screening      for
                                                                                                                                                                                       prostatic  carcinoma      with prostate     specific
blood           flow      and,     in particular,             duplex                 mally   expected     boundaries    and contour
                                                                                                                                                                                       antigen.           J Urol     1992; 147:841-845.
Doppler       and color-coded                       flow,        has be-             bulges,           color-coded                 flow       may         represent              17.   Smith       F, Resnick           MI.         Prostate        specific            an-
come    interesting     with                    the goal        of im-               another     secondary       finding   to initiate  a                                              tigen:   clinical applications   and new devel-
proving            cancer         detection         and       lesion          dis-   biopsy    that otherwise         may not be per-                                                  opments.       Kidney 1993; 25(5):1-6.
crimination        (2,31). Early results  with                                       formed.     True,     it may not have the sensi-                                            18.   Colberg           JW, Smith           DS, Catalona              WJ.
                                                                                     tivity, specificity,      or positive   predictive                                                Prevalence             and pathological                  extent        of pros-
duplex      imaging       were less than optimal
                                                                                                                                                                                       tate     cancer       in men         with    prostate           specific
(43). Results       for the CDS detection      of                                    value   desired      by the purists,    but what
                                                                                                                                                                                       antigen           levels     of2.9      to 4.0 mg/ml.             J Urol
prostatic      blood flow were slightly     more                                     prostate           study           does?       Investigators                 who                  1993; 149:507-509.
encouraging         in the 619 men who under-                                        possess           the knowledge                     and        capability              to   19.   Bretton  PR, Evans                   WP,     BordenJD,                Castella-
went      imaging     in the study by Rifkin et                                      perform            CDS of the                 prostate           gland                            nos RD.             The use of prostate                    specific         anti-
al (31) and the 456 men in the study                                          by     should            do so. If further                  investigation                                gen      density           to improve         the sensitivity                of
                                                                                                                                                                                       prostate specific antigen in detecting      pros-
Kelly et al (2), since blood flow was                                                verifies          CDS-detected                   neovascularity
                                                                                                                                                                                       tate carcinoma.   Cancer  1994; 74:2991-2995.
easier          to image.                                                            that       correlates              with       a biologically               more             20.   SpencerJA,             Alexander             AA, Gomelba                L, Mat-
    Unfortunately,         CDS has its limita-                                       aggressive       cancer,                  then   CDS would      en-                               teucci      T, Goldberg              BB.         Clinical       and US
tions,    some of which        may be machine                                        able identification                       of a clinically impor-                                  findings           in prostate          cancer       patients          with
related.     First, peripheral    gland  flow has                                    tant       subgroup                of patients.                 S                                 normal        prostate-specific                  antigen        levels.          Ra-
been     described     as sparse   with pericap-                                                                                                                                       diology       1993; 189:389-393.
                                                                                                                                                                                 21.   Pollack       HM, Resnick     MI.                   Prostate-specific
sular       and         periurethral            flow,      which         may         References                                                                                        antigen and screening         for prostate cancer:
be related     to the wall filter settings           used                             1 . Newman                 JS, Bree RL, Rubin                 JM.       Prostate                 much ado about something?             (editorial).
(31,39,40).    Newman        et al (1) noted        more                                    cancer:    diagnosis   with color Doppler                                                  Radiology      1993; 189:353-356.
                                                                                            sonography        with histologic correlation   of                                         Benson M, Whang         IS, Olsson     CA, McMa-
prostatic    flow, especially       within      the pe-                                                                                                                          22.
                                                                                            each biopsy site. Radiology       1995; 195:86-90.                                         hon DJ, Cooner WH.            The use of prostate-
ripheral    gland    than initially      described
                                                                                      2.    Kelly 1MG, Lees WR, Rickards                              D.        Prostate               specific  antigen density to enhance             the
with units      that allowed      for lower       levels                                    cancer         andthe role of color Doppler           US.                                  predictive       value of intermediate        levels of
for the pulse-repetition                          frequency              and                Radiology        1993; 189:153-156.                                                        serum       PSA. J Urot 1992; 147:817-821.
wall filter settings.        Second,    all three in-                                 3.    Boring      CC, Squires       TS, Tong T.       Cancer                               23.   Littrup      PJ, Kane PA, Williams         CR, et at.
vestigators      who used CDS found               an                                        statistics,    1993. CA CancerJ           Clin 1993; 43:7-                                 Determination         of prostate     volume     with
                                                                                            26.                                                                                        transrectal       US for cancer    screening.      I. Corn-
overlap     between        increased     flow identi-
                                                                                      4.    Oesterbing       IF.    Prostate     specific  antigen:    a                               parison with prostate-specific      antigen                                  as-
fled with cancer        and prostatitis        (acute                                       critical assessment          of the most useful tu-                                        says. Radiology  1991; 178:537-542.
and chronic).       Third,      not every    cancer                                         mor marker          for adenocarcinoma          of the                               24.   Littrup   PJ, Williams                  CR, Egglin           TK, Kane
demonstrated                     sufficient       flow       to be de-                      prostate.         J Urol 1992; 145:907-921.                                                RA.     Determination                    of prostate          volume
tected           with      CDS,        and,     according           to                5.    Catabona      WI, Richie JP, Ahmann                             FR, et al.                 with transrectal                 US for cancer screening.                           II.
Kelly           et al (2), only           one additional                 can-               A mubticenter      evaluation  of PSA                           and digi-                  Accuracy            of in vitro      and in vivo tech-
                                                                                            tab rectal        examination              (DRE) for early de-                             niques.    Radiology     1991; 179:49-53.
cer was discovered           with CDS that was
                                                                                            tection               cancer
                                                                                                           of prostate     in 6,371 volun-                                       25.   Rifkin    MD, Choi H. Endorectal            prostate
not seen with standard              gray-scale         imag-                                teers (abstr). J Urol 1993; 149(suppl):412A.                                               ultrasound:     implications    of the small pe-
ing. This prompted            them      to suggest         that                       6.    Labrie  F, Dupont    A, Suburu    E, et al.   Se-                                          ripherally    placed hypoechoic        lesion. Radi-
the merits      for CDS were lacking                and                                     rum prostate-specific                     antigen        as pre-                           obogy 1988; 166:619-622.
that it might       even distract        the sonogra-                                       screening            test     for prostate          cancer.   J Urol                 26.   Lee F, Torp-Pedersen                        5, Littrup          PJ, et al.
pher from careful          observation          (2). This                                   1992; 147:846-852.                                                                         Hypoechoic                       of the prostate:
                                                                                                                                                                                                                  lesions                   clinical
                                                                                      7.    Lee        F, Littrup         PJ, Torp-Pedersen                 ST. et ab.                 relevance            of tumor   size, digital  rectal ex-
was not the case for Rifkin et al (31) or
                                                                                            Prostate         cancer         comparison            of transrectal                       arnination            and prostate-specific     antigen.
Newman         et al (1), and here lies the key                                             US and          digital       rectal     examination      for                              Radiology        1989;           170:29-32.
point     to be made      regarding         the role of                                     screening.            Radiology           1988; 168:389-394.                         27.   Ragde        H, Bagely            CMJr,     Aldape              HC,     Blasko
CDS.                                                                                  8.    Cooper       WH,            Mosley       BR, Rutherford       CT, et                       JC_ Prostate cancer                      screening    with high-
    Together,      the authors       identified         13                                  al.    Prostate             cancer      detection    in a clinical                         resolution  transrectal                   ultrasound.     J En-
                                                                                            urobogical           practice          by ultrasonography,                                 dourol 1989; 3:115-123.
cases of prostate        cancer      with CDS that
                                                                                            digital   rectal examination,            and prostate-                               28.   Ajzen SA, Goldenberg    SL, Allen                               GJ,
might            not have          been       detected,         and       even              specific    antigen.     J Urol 1990; 143:1146-1154.                                       Cooperberg   PL, Chan   NH, Jones                                EC.             Pal-
Kelly et al identified         one case that was                                      9.    Chodak       GW.      Questioning           the value of                                   pable prostate nodules:     comparison                                  of US
not seen with conventional               gray-scale                                         screening       for prostate      cancer      in asymptom-                                 and digital guidance    for fine-needle                                 aspira-
analysis,      accounting      for 1.3% of all pa-                                          atic men. Urology            1993; 42:116-118.                                             tion. Radiology   1989; 171:521-523.
tients    studied      (2,31). Now,    consider     the                              10.    ElderJS,      Jewett   IJ, Walsh      PC.       Radical                              29.   Starney       TA.           Diagnosis            of prostate           cancer:
                                                                                            perineal         prostatectomy           for clinical stage B                              a personal            view (editorial).              J Urol 1992;            147:
suggestion                made         by several          investiga-
                                                                                            carcinoma               of the prostate.    J Urol 1982; 127:                              830-832.
tors that tumor     vascularization        corre-                                           704-706.                                                                             30.   Shipohara             K, Wheeler            TM,      Scardino              PJ.
lates with its potential         for rapid  growth                                   11.    Middleton               RG,    Larsen         RH.       Selection         of               The appearance       of prostate                      cancer          on
and distant   metastasis        (44-47).   More-                                            patients         with stage B prostate                    cancer       for                 transrectal  ultrasonography:                            correlation               of

12        Radiology
       #{149}                                                                                                                                                                                                                                      April 1995
      imaging          and pathological                examinations.                   37.   Fleischer       AC, Rodgers           WH,      Rao    BK, et at.          44.   Folkman J. Tumor            angiogenesis:         thera-
      Urol 1989; 142:76-82.                                                                  Assessment     of ovarian tumor vascularity                                     peutic implications.        N EnglJ     Med       1971;
31.   Rifkin MD, Sudakoff                    CS, Alexander   AA.                             with transvaginal      color Doppler sonogra-                                   285:1181-1     186.
      Prostate:        techniques,           results, and potential                          phy. J Ultrasound      Med 1991; 10:563-568.                              45.   Folkman     J, Cole P. Zimmerman       S. Tumor
      applications   of color Doppler     US scanning.                                 38.   Alexander        AA, Liu JB, PalazzoJP,                  et al.                 behavior    in isolated  perfused   organs: in
      Radiology   1993; 186:509-513.                                                         Endorectal  color and duplex imaging of                                         vitro growth and metastasis       of biopsy ma-
32.   Numata   K, Tanaka     K, Mitsui   K, et al.                                           the normal rectal wall and rectal masses.                                       terial in rabbit thyroid    and canine    intesti-
      Flow characteristics    of hepatic   tumors    at                                      Ultrasound   Med 1994; 13:509-515.                                              nal segment.     Ann Surg 1966; 164:491-502.
      color Doppler               sonography:              correlation                 39.   Winkler P, Helmke K. Major pitfalls in                                    46.   Weidner      N, Semple     JP, Welch        WR,    Folk-
      with       arteriographic            findings.         AIR     1993;      160:         Doppler investigations      with particular   ref-                              man J. Tumor angiogenesis                 and metasta-
      515-521.                                                                               erence to the cerebral vascular system. I.                                      sis: correlation     in invasive breast carci-
33.   Dodd        GD III, Miller           WJ, Baron           RL, et al.                    Sources of error, resulting     pitfalls and mea-                               noma. N EnglJ Med 1991; 324:1-8.
      Detection   of malignant                      tumors in end-                           sures to prevent     errors.  Pediatr  Radiol                             47.   Weidner       N, Folkman      J, Pozza F, et ab. Tu-
      stage of cirrhotic livers:                    efficacy of sonogra-                     1990; 20:219-228.                                                               mor angiogenesis:         a new significant       and
      phy      as a screening             technique.          AJR      1992;           40.   Winkler    P, Helmke      K, Mahl M.     Major                     pit-         independent        prognostic      indicator   in early
      159:727-733.                                                                           falls in Doppler investigations.      II. Low                     flow          stage breast carcinoma.          J Natl Cancer      Inst
34.   Taylor        KJW,       Ramos       I, Carter        D, et al.          Cor-          velocities and colour Doppler                     applications.                 1992; 84:1875-1877.
      relation of Doppler US tumor signals with                                              Pediatr Radiol 1990; 20:304-310.                                          48.   Brawer      MK, Deering       RE, Brown       M, Preston
      neovascular   morphologic features. Radiol-                                      41.   Blood CH, Zetter      BR.     Tumor      interactions                           SD, Bigler SA.       Predictors       of pathologic
      ogy 1988; 166:57-62.                                                                   with the vasculature:       angiogenesis         and tu-                        stage in prostatic      carcinoma:       the robe of
35.   Ramos   IM, Taylor KJW,                       Kier    R, et al.          Tu-           mor metastasis.     Biochim      Biophys      Acta                              neovascubarity.      Cancer      1994; 73:678-687.
      mor vascular             signals      in renal         masses:         detec-          1990; 1032:89-118.
      tion with         Doppler           US. Radiology              1988;      168:   42.   Folkman    J, Cotran R. Relation of vascular
      633-637.                                                                               proliferation        to tumour growth.                Int Rev
36.   Cosgrove           DO,      Kedar      RP, BamberJC,                et al.             Exp Pathol         1976; 16:207-248.
      Breast       diseases:       color Doppler               US in differ-           43.   Fornage      BD.        Transrectal          duplex      sonogra-
      ential      diagnosis.        Radiology              1993;    189:99-104.              phy of prostatic              carcinoma:       preliminary
                                                                                             experience         (abstr).      Radiology       1989;     173(P):

Volume           195        Number
                         #{149}                 I                                                                                                                                                               Radiology                  13