EDITORIAL ImmunoscintigraphyForLungCancerDetection: RealityTesting

IMAGING TESTS FOR LUNG (Type Fo23C5) for lung cancer detec cost of screening programs depends CANCER DETECTION tion (6). This investigation is a pilot on the intrinsic cost of the test itself The development of new imaging study rather than an detailed test of and the added cost and morbidity of modalities will continue to play a cen the actual screening utility of immu false-positive tests that lead to other ti'a.lrole in the development of better noscintigraphy for lung cancer detec unnecessary procedures. The low tests for cancer screening and staging. tion. Biggi et al. report that 57 of 63 specificity of immunoscintigraphy Radiologic studies enjoy a special ad patients with confirmed lung cancer (0.45) in the small control group used vantageover many other medical tests and 6 of 11 controls with other non by Biggi et al. is quite disturbing be because all-importantlocalization cc lung cancerchest diseaseshad positive cause it implies lack oftumor specific curs in concert with detection. There uptake of the isotope. The high test ity of the preparation and! is no better illustration ofthe practical sensitivity (0.90) is promising but not or low signal-to-noise of importance of joint localization and necessarily what can be expected if an in the tumor. The use of immunoscin detection than in the heavy reliance actual screening program were con tigraphy in its present form would be on for the detection ducted. The patient population on likely to lead to high added cost and of non-palpable breastcancers. which the data are drawn are likely to morbidity from false-positive tests. In the case oflung cancer, effective be quite different from actual screen Although we have knowledge of the screening tests remain an important ing populations where the detection actual false-positive rate that would public health challenge because mci of minimal disease is at a premium. occur ifimmunoscintigraphy wereap dence and mortality remain high, and In fact, Biggi and colleagues report plied to normal ratherthan non-can large segments of the population con that false-negativestended to occur in cer chest disease controls as used by tinue to smoke cigarettes. The grow lesions < 2 cm diameter, and in those Biggi et al., it still appears that a large ing health risk from lung cancer is in the perihilar region. The inclusion fraction of false-positives would be a emphasized by new statistics showing of six small-cell cancers in the study considerable risk. In chest radiogra lung cancer superceding breast cancer group suggeststhat the high sensitivity phy, it is known that the false-positive as the most common cause of cancer may partlybe due to advanced disease rateisintherangeof0%—lO%per death among women (1). Chest ra in some cases. We do not have any conventional radiographic examina diography with (2) or without (3) sup precise knowledge of the stage of the tion for lung cancer (7). plementary sputum cytology can de lung cancer lesions reported by Biggi In its presentform, immunoscintig tect lung cancer at an earlier stage et al., nor do we know the relative raphy is unlikely to be a practical, or than controls, but screening with this performance ofconventional chest ra robustscreeningmethod for earlylung conventional modality is not thought diography in this group. It is likely cancerdetection. The sensitivity offers to reduce overall lung cancer mortal that conventional radiographywould little, if any, advantage over chest ity (4). It is clear that the search for have a good chance of detecting the and the false-positive rate an effective new lung cancer screening small peripheral lung cancers missed may be very high. Few subjects will test should look beyond conventional by immunoscintigraphy. It is unfor bewillingto takethe additionaltime radiography and sputum cytology. tunate that immunoscintigraphyjoins (and cost) involved in a return hospi conventional radiography in exhibit tal visit for imaging the after radio IMMUNOSCINTIGRAPHY ing rather low sensitivity to perihilar nucide injection. Even if immuno lung cancer lesions. became inexpensive and Radiolabeled monoclonal antibody convenient, there seems to be little to to lung cancer tumor-associated anti recommend it over radiography. REALITY TESTING gen (5) provides an attractive new Although it is possible that im imaging vehicle for cancer detection. The utility ofa cancer screeningtest provement in the radionucide anti In this issue, Biggi and colleagues de depends ultimately on its ability to body preparation may increase the scribe the use of immunoscintigraphy detect early disease and reduce mor signal-to-noise in tumor, it is hard to with ‘‘‘In-labeledanti-CEA Mab tality in the screened population by be very optimistic about the potential early treatment intervention. Conven ofimmunoscintigraphy in lung cancer

ReCOIVOdJUI.15, 1991;acceptedJul. 17, 1991. tional radiographicscreening for lung detection. A much simpler conven For reprints contact: Reginald Greene, MD, The cancer can accomplish the first crite tional form of imaging may actually Department of ,Harvard Medical School and MassachusettsGeneralHospital,Boston,MA rion for utility but it has not yet been do the job better. Detection of lung 02114. shown to meet the second aim. The cancer with immunoscintigraphy is

Immunoscintigraphyfor LungCancerDetection•Greene 2069 analogous to picking up a straightpin Massachusetts General Hospital Sloan-Kettering study in New York. Chest 1984;86:44—53. with a front-end loader. Much more Boston, Massachusetts 4. Fontana RS, Sanderson DR. Woolner LB. Tay than the pin will be scooped up, and br WF, Miller WE, Muhm JR. Lung cancer a great deal of effort will be required screening: the Mayo program. I Occup Med to cull out the pin. It is easierto stoop 1986;28:746—50. REFERENCES 5. Goldenberg DM, Kim FH, Deland FH, et al. over and pick up the pin in the con Radioimmunodetectionof cancer with radio ventional way. The quest for a more active to carcino.embryonic antigen. 1. SilverbergE, Lubera J. Cancer statisticsCA Cancer Res l980;40:2984—2992. effective test for early lung cancer de l988;38:5—22. 6. Biggi A, Buccheri G, Ferrigno D, et al. Detec tection should look beyond both con 2. Fontana RS. Screening for lung cancer. In: tion of suspected lung cancer by scintigraphy ventional radiographyand immuno Miller AB, ed. Screeningfor cancer. New York: with ‘‘‘In-anti-carcinoembryonic antigen AcademicPress;1985:377—395. monoclonalantibodies(TypeF023C5).I Nucl scintigraphy as it exists today. 3. Melamed MR. Flehinger N, Zaman MB, Hee Med 199l;32:2064—2068. ReginaldGreene lan RT, Perchick WA, Martinin N. Screening 7. Eddy DM. Screeningfor lung cancer.Ann In for early lung cancer. Results ofthe Memorial ternMed 1989;111:232—237. Harvard Medical School

Erratum In the September1991issueofthe Journal,the captionsfor Figures1and 2 in the article,“AnAnalysisofCerebralBlood Flow in Acute Closed-Head Injury Using Technetium-99m-HMPAO SPECT and Computed ,―by Steven N. Roper et al, were placed incorrectly. The corrected captions and figuresare reprinted below.

@ 1.1. 45@2@ 0 3'E$.'@ ____@ @1N @ECT

@ RT.@LT. @63

FIGURE 1. Imaging stud ies from a 45-yr-old male who was thrown from a horse. XCT (A) shows left occiptial contusion and SPECT (B)shows decreased CBFin bothoccipitallobes.

A.S. 1@ %

H@@R @LA @EDC?R

FIGURE 2. Imaging stud @ ies from a 19-yr-old male RT Lt. who suffered a non-pene trating gunshot wound to the left temporalare. XCT (A) shows a left frontotem poral contusion. SPECT(B) showsdecreasedCBF in the left frontotemporoparietal, right frontal, and right cc 4 cipital areas.

ADDENDUM

Please note the following change for the article “Reorientationof the Left Ventricular Long-Axis on Myocardial Transaxial Tomograms by a Linear Fitting Method―by He et al, which appeared in the September issue of the Journal (pages 1794—1900).On page 1796,line 9, the statement: (using y = ax + b, where a was the slope ofthe fitted line and b the intercepton the y-axisofthe fittedstraightline in the transferredcoordinate)shouldbe changedto: (usingy = bx + a, where b was the slopeof the fitted line and a was the intercepton the y-axisof the fitted straightline in the transferred coordinate).

2070 The Journal of •Vol. 32 •No. 11 •November 1991