Your Federal Quarterly Tax Payments are due April 15th Get Help Now >>

Editorials An Efficacy Lesson for Medical Radiologists from Our by xor56373



                                                        An                   Efficacy                                             Lesson                                               for Medical                                                        Radiologists
                                                                                                                                       from                                              Our Dental                                                        Colleagues’

T         HIS          issue       of Radiology  contains                                     an im-           only            one       asvmptomatic                            lesion              in 30,000                    about    l0     of that for medical     radiograph-
       portant                 article   by Dr. Zeichner                                      and              radiographic                         examinations                             of the          face        and      Ic examinations,        but it is a very high fig-
colleagues                   on the efficacy    of dental                                      radi-           jaw.        It seems                clear          that          nearly              all of these                  ure that        has increased         from       an estimated
ographv                for      detecting                asvmptomatic                              bone        rare        lesions               would             be      symptomatic                            when            $730      million      only   7 ‘ears        ago.    It is also                                           in-
tumors            of the            face and    jaw. The                           authors                     relatively                 small,               as they           often              are      associat-            teresting         to know     that     there      is an Amen-
conclude             that           dental   radiography                               is not                  ed with                 infectious                 or inflammatory                                 dis-            can Board                 of Oral               and Maxillofacial                               Radi-
efficacious       for this purpose,             and their                                                      ease   of the teeth     and                               gums.                                                    ologv  that               certifies               dental physicians                              with
careful     analysis       of the question            and clear-                                                   The authors     recruited                                      four          board-certi-                      special             expertise                  in     radiology.
lv written       report      constitute        an excellent                                                    fied        dental              radiologists                     to use              in an         analy-             Congratulations                           are due Dr. Zeichner
example       of efficacy        research        that is wor-                                                  sis of observer                        accuracy.                   The dental                        radi-         and colleagues                           for developing  an impor-
thy of reading           by all radiologists            and oth-                                               ologists    were                     shown       the               digitized                                       tant         paper         on       the         efficacy             of dental                 radi-
ers interested          in medical        utilization      ques-                                               radiographic                         images               in 97 cases                      with         bi-        ographv              for identifying                              occult           bony      le-
tions.                                                                                                         opsv-proved                         lesions               and          97 normal                   stud-           sions.          Certainly        the                  FDA          should             change
       Dns.      Zeichner,                    Ruttimann,                    and         Webber                 ies.      Five           levels           of confidence                          were        evalu-                its    recommendation                                    after       this         report.            In
of the National      Institute                                  of Dental                 Re-                  ated        and           receiver           operating                         characteristic                      medical                radiology                    we      have          many              impor-
search  have responded                                        to the Food                  and                 (ROC)              curves               obtained.                 The          analysis                            tant         diagnostic     and therapeutic     questions
Drug Administration’s                                         (FDA) plan                  to                   showed                  that        expert              dental            radiologists                        do   that         should     be subjected    to the same vig-
publish     a recommendation,              “based                                             on               not have                 a high         accuracy       for finding                                                 orous       scientific                    review-screening                                    for         os-
available     evidence,”        that  radiography                                                       of     asvmptomatic                        lesions.      Even      with       the                              least      teoporosis,            use               of “baseline”                          diagnostic
the face be performed              on “patients                                               seek-            rigorous                 conditions            (a 50      false-positive                                           and          nuclear            medicine                    tests         in     patients
ing routine       dental    care”    for “proper                                                diag-          rate),           the      true-positive                         rate          was          only                    with          cancer,             various                types          of      stress        tests
nosis.          treatment                 planning,                   and         prognosis                    about            80      with                multiple              (14-21      views)     in-                      in cardiology.                       laboratory                     screening                  for
of a variety                   of tumors.”                    The         authors              not             traoral            radiographs                      and           only    about      609k                          hvperparathyroidism,                                   and a large number
only       questioned                   this recommendation                                              but   with            panoramic                     views.             In the              usual           prac-         of interventional                               procedures.        to name     a
also      conducted                   a model     efficacy                         study                       tice setting,                     the true-positive                                 rate would                     few. Determination                                  of prevalence,       diag-
that proved                    it to be           incorrect.                 To        make              an    most    likely                   be only     50’,                      with           a similar                    nostic           accuracy,                    economic                  cost,        and       out-
appropriate                    scientific               judgment,                      they                    false-positive                          rate.                                                                      come           (including                      complications)                         are      neces-
evaluated                the        prevalence                      of disease,                 ana-                Using               current                typical           charges                   for    dental          sarv    for each efficacy                                  study.  This                  article
lvzed         diagnostic                  accuracy,                  estimated                     the         radiographs.                        the         authors                estimated                   that            evaluates     each topic                                 in a masterful                     way,
cost of          identifying                     lesions,      and     examined                                the       radiographic                           cost      of finding                       each                   with          one        exception.                      Treatment                   outcome                    is
potential             problems                    of radiation-induced                                         true-positive                      malignant                     lesion                would     be                not evaluated                        (the           variety             of bony              lesions
tumors.                                                                                                        more        than                $8 million,                  and      the             cost for the                 is large.              and   I could                     find      no        satisfactory
      To determine                       prevalence                       of bony              tu-             identification                      of each                 true-positive                    be-                   outcome                 data   in the                    literature),               and        a lim-
mors,         thes’          analyzed                 more           than         30 million                   nign            lesion            would             be      more              than           $400,000.             ited         analysis             is used                to estimate                  radiation-
insurance     records   and                              determined    that                                    The         total         cost          of diagnosis                      would               be high-             induced                 tumors.                However,                   this       is a minor
there are approximately                                     100 cases of bony                                  er because                      their         economic                   estimates    did                          criticism               for       this         study,            in which                 preva-
tumors    of the face and                                jaw each year per                                     not include                       the        important                   cost of surgical                          lence          of disease                     is very            low.          Few       studies
million           citizens               in     the      United              States,               of          follow-up                      for both              true-         and           false-posi-                       in     the       medical                 radiology                  literature                meet
which           less         than        five         cases         are      malignant.                        tive lesions.        The                        number        of surgical       or                                 the standards       of this report.
Any        practicing                 radiologist                     would   agree                            medical       follow-up                            examinations           would                                       There    is another     area in the                                             practice               of
that      these       bony             tumors                 are     not a signifi-                           be       yen,          high          since          the authors                        estimated                   dental            radiography                         that     presents         an               even
cant      health               risk-when                       was        the      last        time            that    there       are nearly        30,000   false-positive                                                      bigger            question                    for     efficacy        analysis:                   Is
you       saw      a lesion               of the face                  or       jaw       that                 lesions       for every       correctly      identified          ma-                                               there          an’        need            for “routine”                          radiographs
was       not      metastatic?                In fact,                for        the      second               lignant        lesion    due     to the combination                of a                                            to be          obtained                  in     the       annual                or semiannu-
part      of their              study,           the       authors                could            find        low prevalence             rate of malignant               lesions                                                 al visit            to the         dentist                in the             absence            of
                                                                                                               and         a low              sensitivity                 (true-positive                          rate).          symptoms                   or      clinical                findings?                 We       now
                                                                                                               The’             also      estimated                      the      number                    of radia-             know            there           is no reason                        for screening
                                                                                                               tion-induced                         lesions               and      suggested                        that          dental           radiographs                      to find      tumors,                      but is
      From the Department
      I                                of Radiology.       Mal-                                                the number                         of induced                     malignant                       lesions          there          any reason                      for radiographs                            to find
linckrodt    Institute       of Radiology.       Washington                                                    max’ be the same                                 as those identified      by di-                                   any asymptomatic                                 abnormalities?                          I hope
University     School       of Medicine,        510 South                                                      agnostic  screening                                 of a large population                                          Dr. Zeichner     and                            colleagues      are                    pursuing
Kingshighway           Boulevard.       St. Louis.    MO 63110.                                                (i.e.,          very       few).                                                                                   this         larger        question.
Received     November           21, 1986; accepted        Novem-
                                                                                                                      The         report            also         has       several                 interesting
ber 26. Address         reprint     requests      to the author.
                                                                                                               bits       of information                            new           to this             medical                                                                         RONALD               C.       EVENS,            M.D.
     c RSNA,    1987
    See the article      by Zeichner         et al. (pp. 691-                                                  radiologist.                      Dental             radiographic                            examina-                                                                       Director              of the        Institute
695) in this issue.                                                                                            tions            cost approximately                                    $2.5      billion                an-                                 Mallinckrodf                        institute             of Radiology
                                                                                                               nually              in the United                          States,             or $10             per                       Washington               University                       School           of Medicine
Radiology               1987; 162:873                                                                          year            for every                person.                This          cost          is only                                                                                    St. Louis,              Missouri


To top