Immunoscintigraphy of Recurrences of Gynecologic Carcinomas

Jean-FrançoisChatal, Pierre Fumoleau, Jean-Claude Saccavini, Philippe Thedrez, Chantal Curtet, Alicia Bianco-Arco, Alain Chetanneau, Patrick Peltier, Mireille Kremer, and Yves Guillard

Unite Inserm U-211, UER de Médecine; Centre RenéGauducheau, Quai Moncousu Nantes; and Compagnie ORIS Industrie Gzfsur Yvette, France

Ina first, retrospectivestudy,15 patientswith knownovariancarcinomawereinjectedwith 13110C125 F(ab')2monoclonal (MAb).The sensitivityof immunoscintigraphybased on the numberof the tumorsiteswas 67% (12/18). In a second,prospectivestudy,29 patientswith gynecologiccarcinomawereinjectedwith 1311-OC125F(ab')@(24)or 13119@9 F(ab')@(5)MAbsaccordingto the histologictype. Basedon the numberof testedanatomic sites,sensitivitywas 72%andspecificity86%.Intwo patientsinjectedwith both‘311-OC-125 F(ab')@and1251-NSF(ab')@(nonspecificimmunoglobulin)1 and4 daysbeforetumorresection, tumor uptake of the specific antibody was 2.2 and 4.5 times greater than that of NS. Immunoscintigraphic results were complementary with those of ultrasonography and computedtomography.Finally,in one patientinjectedsuccessivelywith 1311-OC125F(ab')@ and1111n-DTPA-OC125F(ab')@,the recurrenttumorwasvisualizedwith bothradionuclides, with 1111nproviding better abdominal tumor contrast but causing much greater liver radioactivity than 1311.

J Nuci Med 28:1807—1819,1987

pithelial ovarian cancers still have poor prognosis, logic type (5). This antibody recognizes an antigen, CA with an overall 5-year survival rate of @30%(1). A 125, associated with ovarian carcinomas particularly of recent study concerning 770 patients reported a 5-year the serous type, and is shed into the circulation where survival rate of <10% for stages III and IV (FIGO it can be assayed by an immunoradiometric method classification), the most common stages at the time of (6, 7). Likewise, antibody 19-9 recognizes an antigen a first diagnosis (2). After initial surgery, treatment designated CA 19-9, associated with mucous carcino generally includes a chemotherapy regimen for ad mas (8), that is also shed into the circulation where it vanced stages with macroscopic residue and chemo can be assayed by the same method. therapy and/or abdominopelvic irradiation for early By serial biologic monitoring with both markers de stages without macroscopic residue. Following chemo pending on histologic type, recurrences can be detected therapy, second-look surgery is usually carried out to early in patients who previously reached a state of confirm possible complete remission (3). When corn complete remission (7). It is then important to be able plete remission is thus confirmed in patients initially to confirm recurrences by visualization, which is the with stages III and IV, a recurrence subsequently occurs role of immunoscintigraphy(IS), in order to provide in 25% of cases with poor prognosis (4). It would further effective surgical treatment. The aim of the therefore be important to be able to detect such recur present study was to specify the diagnostic value of IS rences early, before the appearance of clinical signs, in both by examining patients with previously known and the hope of improving therapeutic effectiveness. localizedtumor sites and, especially,in a prospective A rnonoclonal antibody (MAb) designated OC 125 view, by comparing the efficiency ofthe technique with has been produced after immunization by a serous that of conventionaldiagnosticmethods. ovarian cystadenocarcinorna, the most common histo METhODS Received Sept. 19, 1987; revision accepted June 11, 1987. For reprints contact: Jean-Francois Chatal, MD, U.21 1Inserm, UER de Médecine,1, rue Gaston Veil, 44035 Nantes Cedex 01, Two MAbs were usedin this study.—OC125 and 19-9— France. with the choice depending on the type of marker that had

Volume 28 •Number 12 •December1987 1807 significantlyincreasedserumconcentrationrelativeto normal initial 10-mm elliptical rotation. Then, with the patient in the values. MAb OC 125 was produced and made commercially same position, a second 40-mm rotation was performed after available'; its characteristics have been previously reported a 20% energy window was centered on the 1311photopeak. (5). F(ab')2 fragments were obtained after pepsin digestion, For each recording, 6-mm-thick transverse, sagittal, and cor and their purity was tested by SDS electrophoresis and high onal sections were then reconstructed using data filtering. The performance liquid chromatography (HPLC) on a TSK 3000 same reconstruction with thicker sections (12 mm) did not column. There were neither Fab' fragments nor intact anti change the interpretation of the images. The normal pattern body in the preparations. The F(ab')2 fragments were labeled ofan ECT section was characterized by patchy, heterogeneous with iodine-i 31 (1311) using the iodogen method, with a distribution with many foci more or less contrasted in relation specific activity of 3 mCi/mg. Immunoreactivity after ra to each other. Our interpretative criteria for considering dioiodination was tested by affinity chromatography using a whether a focus was pathologic required that this focus appear sandwichassay.CA 125antigenwaseluted through a sepha with the same contrast and localization features in at least rose column coupled to OC 125 MAb. The column was three superimposed sections and two tomographic planes. washed and the labeled MAb was then eluted through the Four to seven days after injection of the radioiodinated column. This sandwich assay was made possible by the repet antibody, planar scintiscans were recorded using an large field itive structure of the recognized epitope. The antigen-bound of-view gamma camera with a medium-energy collimator, fraction of the labeledantibody was eluted with a 3M am interfaced with a data processing system.@Anterior and lateral monium thiocyanate solution. The F(ab')2 fragments from views were recorded in order to cover the whole of the abdo 19-9 MAb were labeled with ‘@‘iusing the same method and men and pelvis. Before each scan, @mTcwas injected as in the with the same specific activity. The labeling efficiency ranged ECT study. from 80 to 90% with ‘31I-OC125 F(ab')@and from 90 to 95% Patients Studied with l3hI@l9@9F(ab')2. The immunoreactivity of MAb 19-9 From January 1983 to March 1986, 44 patients with gyn was tested in the same way as with MAb OC 125. The percentageofimmunoreactivitywascalculatedusingthe ratio ogic cancer underwent one or more immunoscintigraphic examinations. These patients can be divided into two groups of radioactivity of MAb eluted with NH4SCN to the sum of on the basis of the circumstances leading to . For MAb radioactivity not bound to antigen plus that eluted with 15 patients (Group I: mean age 58 yr, range 39 to 70 yr), NH4SCN. The percentages obtained were 47 ±4% with @I localization of the primary tumor or recurrence was deter OC 125 F(ab')2, and 80 ±5% with ‘@‘I-19-9F(ab')2. The mined by surgery, palpation of a mass, and/or ultrasonog immunoreactivity of MAb OC 125, the antibody most often usedin thisstudy,wasalsotestedby a cell-bindingassay.The raphy (US) or computed (CT) that showed a tumor mass prior to IS (Table 1). For the other 29 patients cell line used (OVCAR NIH)t came from a serous ovarian adenocarcinoma and expressed the CA 125 antigen (9). (Group II: mean age61 yr, range39 to 79 yr), ISwasthe first Briefly,3. l0@cells were distributed in a round-bottom 96- examination performed prospectively when a recurrence was suspected on the basis of significant and persistent elevation well plate. One hundred microliters of @‘I-OC125 F(ab')2 of the serum concentration of one of the two markers (CA were added to each well and left to incubate 3 hr at room 125 or CA l9-9)(Table 2). A total of 36 immunoscintigraphic temperature with continuous shaking. After centrifugation examinations were performed in these 29 patients, with two and three successive washes, cell radioactivity was counted. examinations in five patients (Nos. 1, 3, 5, 7, 13) and three in The results were expressed in percentage oftotal activity used Patient 4. After IS, US and/or CT was performedblindly in for each well. In these conditions, the percentage obtained was Group II patientswithout knowingthe findingsofIS. A total 46%, whereas it was 1.5% with a nonspecific immunoglobulin of 32 US examinations were performed with a ALOKA SSD labeled in the same conditions. 85 1real-time equipment and 27 examinations were performed Scintigraphic Technique on the third generation CGR CE 10000 Scanner with a scan An activity of 1.5-2 mCi (55.5—74MBq) (—0.6mg) for duration of 3.4 sec. There were fewer US and, especially, CT planar scintigraphy and 2.5-3.5 mCi (92.5—129.5 MBq) (—1 examinations than IS examinations since some patients con mg) for emission computed tomography (ECT) was injected sidered the IS protocol too difficult and demanding and sub into each patient after obtaining informed consent. Thyroid sequently refused US and/or CT. uptake of free ‘@Iwas blocked by oral administration of The localization diagnosis for recurrence (Group II) was Lugol's solution (100 mg/day) for 8 days beginning 2 days confirmed by surgery in 14 cases, by concordant results of US before injection ofthe radioantibody. Each radioantibody was and CT in six cases and by the course of the disease, with injected intravenously over a period of'—30mm in an infusion eventual appearance of a palpable mass in the location of of 100 ml ofsaline solution. No skin test was performed before antibody uptake, in five cases. Absence of recurrence was this injection. Three and four days after injection of the confirmed by surgery in three patients and suggested by the radioantibody, ECT studies were performed using a rotating coursein anotherpatientovermorethan a year'stime.Finally, single-head gamma camera with a medium-energy collima in seven cases the diagnosis at this writing still remains unde tor.* The abdomen and pelvis were evaluated in two scans 3 termined. In five of these seven cases, IS was positive, sug and 4 days after injection. Each recording first involved injec gestingrecurrence,whereasUSand/or CTwerenegative(Nos. tion oftechnetium-99m (@mTc)5 mCi (185 MBq), in different ib, 7b, 12, 13b, 20). In two cases (Nos. 13a, 2 1), all localization forms hydroxymethylene diphosphonate (HMDP), albumin, examinations, including IS, were negative despite elevated diethylenetriaminepentaaceticacid,or sulfurcolloid)toobtain serum CA 125 concentrations. In all these cases the patients anatomic landmarks and thus facilitate interpretation of im refusedsurgery,and the diagnosisremainedundetermined. munoscintigraphic images. This injection was followed by an Planar scintigraphy was performed in 18 cases (eight in

1808 Chatal,Fumoleau,Saccavinietal TheJournalof NuclearMedicine TABLE I Results ofIndudingSerum*Retrospective Study 15 Cases of Primaryor RecurrentOvarianCarcinoma

Patient Age CA 125 no.tomographyPSECT1(yr) U/mi Tumor site (size, cm)ImmunoscintigraphytUftrasonograph@Computed Abdomen7x3—++266 2,260 Pelvis>5+ND+347 38 6—+ND456 10,000 Pelvis6 x Abdomen>5+NDND560 323 Abdomen>5+ND—663 820 Pelvis>5+NDND770 ND Pelvis>5+ND+839 900 Pelvis4x3+++963 152 Abdomen6x3+++1055 157 >5++ND1150 1,370 Pelvis Pelvis>5——+ND1261 67 PeMs4x4+++Abdomen2.5x2.5———1363 292

carcinomatosis—+—1443 520 Diffuse 5++—NDAbdomen3x3++—NDUver<1——+ND1550 128 Pelvis10 x

Abdomen8x6+ND+.62 95

U/mI.tUpper limit of normal values: 40 scintigraphy.*PS = Planar ND = Not done.

Group I, ten in Group II) ECT in 40 cases(nine in Group I, Results ofthe prospective study are given in Table 2. 31 in Group II). In eight cases, patients were evaluated with Out of 24 immunoscintigraphic examinations per both scintigraphic methods. formed in 20 patients injected with ‘@‘i@125 F(ab')2 Three patients (Nos. 8 and 9 in Group I and No. 18 in whose diagnosis of recurrence was documented, 19 Group II) were injected, respectively,at 4-, 6- and 8-mo tumor sites were correctly visualized: six by planar intervalswithan identicalactivityof ‘31I-OC125F(ab')2and scintigraphy and 13 by ECT. In one case of abdominal 11‘In-DTPA-OC 125 F(ab'>@ to compare the scintigraphic pattern obtained with each .Everypatient who recurrence (No. 19), ECT was negative at Day 4, but had a knownrecurrenceof ovariancarcinoma,wasevaluated planar scintigraphy was positive at Day 7 after a 30- by ECT in the same conditions 3 days after injection. Patient mm acquisition time. 8 had been treated after the first immunoscintigraphy with a Figure 2 shows a case ofbilateral ovarian cancer. The high dose of 3,330 MBq of ‘@‘I()(125 F(ab')2. two tumors were the smallest visualized in this study, Finally, Patients 15 and 19 were injected simultaneously measuring 2 x 2.5 and 2.5 x 3 cm (No. la). For the with ‘31I-OC125 F(ab')@and ‘25I-NSF(ab')@(a nonspecific two patients simultaneously injected with the specific immunoglobulin ofthe same class as OC 125), respectively, 1 MAb ‘3'I-OC125 F(ab')2 and the nonspecific immu and 4 days before second-look surgery to study distribution in noglobulin ‘251-NSF(ab')2, respectively, 1 and 4 days the tumor and normal tissue specimens removed during the before the operation (No. 15 and 19), study of the operation. The results were expressedin % of injected dose distribution showed higher tumor uptake ofthe specific per gramof excisedtissue. MAb than of the nonspecific immunoglobulin (Fig. 3). The difference in uptake was greater at Day 4 (Fig. 3B) RESULTh than at Day 1 (Fig. 3A). This tumor uptake of the Results of the retrospective study are given in Table specific MAb was well correlated with the results of the 1. Among the 15 patients studied, a total of 18 tumor immunohistochemical study that in both cases showed sites was known before IS. Twelve sites were visualized: CA 125 antigen expression by over 50% of cells. In four by planar scintigraphy, six by ECT and two by normal tissues (liver, muscle, fat, skin), radioactive both methods. The smallest tumor was 3 cm in diam concentration of the specific MAb varied little from eter, but most tumors visualized were >5 cm in diam that of the nonspecific immunoglobulin. In the two eter. patients, tumor-to-liver, tumor-to-blood, and tumor-to Figure 1 shows the case of a pelvic recurrence clearly muscle ratios were, respectively, 2.45 and 2.56, 1.32 visualized by ECT, with a transverse section image and 1.08, and 5.64 and 4.76. The specificity indices, as corresponding to a CT image. The tumor measured 4 measured by the tumor-to-tissue ratios obtained with x 3 cm(No.8). 1311 divided by the same ratios obtained with 125!, ranged

Volume28 •Number12 •December1987 1809 TABLE 2 Results from 29 Prospectively Investigated Patients IS' Serum Means Patient 125 PSECTsite(s) of Tumorsize(cm) no.AgeU/mIla69Serousovanan1,0702x2.5carcinoma2+(yr)Primary carcinomaCA US@ CT diagnosist Surgery 3lb18,000(Pelvis)1 (Pelvis)2+1+1—2.5 x

+ND—UND279Serous ovarian carcinoma + x 5 2.53b200(Pelvis)1+1+NDEvolution4a58SerOUScarcinoma190(Pelvis) (PeMs)1 1+— 1+— 1+SurgeryConcordance4 3 x

ovarian node explored4b630(Pelvis)1carcinoma4,000—SurgeryAxillary 3 x 3 not

+—NDEvolution4c30.0005a65Serous

+1 + + x 5.5 ovarian Perftoneal seedlingsSb6703a65Serousovarian1,170—NDNDEvolutionPelVICcarcinoma95(Pelvis)1 ND1 NDConcordanceSurgery7

mass658SerOUS ovarian 2.57a72FallOpiantUbecarcinoma650—NDSurgery3 x

carcinoma60(Abdomen) + + + 4 x 4 Pelvis2xl7b1,480(Abdomen)1+——UND860Serous(Pelvis)1 1—I 1—1 1—SurgeryAbdomen

ovarian seedlings973Endometnumcarcinoma450——NDSurgeryPentoneal

31058Serous carcinoma2501 + (Pelvis)—NDSurgery4 x ovarian +—NDEvolution1carcinoma900(Pelvis)1 150Serous ovarian 61257Serous carcinoma2,6001 +1 +1 +Surgery5.5 x ovarian +—NDUND13a51Serousovanancarcinoma360(Pelvis)1 carcinoma215———UND13b260(Abdomen)1+——UND1476Serous

ovarian +Evolution1 carcinoma107(Pelvis)1 +—1 5 ovarian carcinoma + + 1674 55SerousSerOUSovarian 400(Abdomen)1 3+ Concordance>51764Serousovanancarcinoma3,300 —— (Abdomen)1 3+Surgery

mass1862Serouscarcinoma270(PeMs)1 +—1 +SurgeryPelvic benign ovarian 41947Serous carcinoma220(Abdomen)1 +1 +1 +Concordance4 x ovarian 52073Serousovariancarcinoma3601 + (Abdomen)1 ———SurgeryS x

+——UND2173Serouscarcinoma1381 + (Pelvis)1 ovarian carcinoma95———UND2260Serous ovarian apparent carcinoma132(PeMs)1 +—1 +SurgeryNo tumor

1 81 0 Chatal, Fumoleau, Saccavini et al TheJournalof Nudear Medicine TABLE2 (continued)

Serum Means Patient Age CA 125 PS ECT of Tumorsize (cm)23no. (yr) Primarycarcinoma U/mI sfte(s) lJ5@ CT diagnosist 74 SerousPentonealcarcinomaovarian 1,200 carcinosis24 1+ (Abdomen) 1+ — — Surgery 63 Serous1,400carcinomaovarian 2—Surgery1(Abdomen) 2+ — — 1—SerumCA+ (Pelvis)

19-9U/mI25

39 Mucinous3carcinomaovarian 135 Pelvis4 x Surgery26 (Pelvis) 1+ — — 66 Mucinous75carcinomaovarian 3(Liver)(Pelvis) 1+ 1+ 1+ Concordance Liver 3 x 1+27 1— 1+ 46 Mucinous45carcinomaovarian recurrence28 — — — Evolution No 80 Mucinous>2,400carcinomaovarian Concordance29 1+ (Pelvis) 1- — + 58 Endometrium35carcinoma apparenttumor. (Pelvis) 2+ ND — Surgery No

U/mI.tSee Table 1 ; Upper limit of normal value: 40 US.‘ISUND = Lfrdetermined; Concordance = Concordant results of CT and scintigraphy.from= Immunoscintigraphy;PS= Planar

andimages1.63 to 3.37 (mean 2.50). Figure 4 shows the ECT tebral node recurrence is well contrasted (Fig. 4A), sectionneousin the patient operatedon 4 days after simulta- its well-defineddevelopment on the coronal injection ofthe specific MAb and the nonspecific corresponds with that found by the surgeon and mdi immunoglobulin4C).formed (Fig. 3B). This ECT study was per- cated with metal clips (Fig. 3 wk before second-look surgery. The prever- In this prospective study with MAb 1311-OC125r

@ ..@,s.4 @-;•@@: •:

@ I@ \ .@, ,@. :@ @ / ‘@ i /

A B FIGURE I ECT imagein a patientwith pelvicrecurrenceof a serousovarianadenocarcinoma.A: ‘311-OC125 F(ab')@ECT image. A transversesection,showsan abnormalconcentrationof radioactivitybehindthe bladder(arrow).The outlinesof the bladder and the pelvic bones have been sketched in from a corresponding [@“Tc]HMDPsection. B: A CT scan shows a massbehindthe bladder(arrow)correspondingto the ECT imagefindingof Figure1A. The imagehas been inverted to correspondto the ECT image.

Volume28 •Number12 •December1987 1811 ‘L

A 1.

C

FIGURE 2 Bilateralovariancarcinoma ina patient previouslyoperated on for a malignantascitis ofunknown origin. A: The anterior 131lØt@125 F(ab')@ scan image shows two symmetrical abnormalfoci of radioactivityabove the bladder (b) (ax rows). B: A sonogramshows a pelvicmass (arrows)but failsto differentiatetwo distinct tumors. T = Tumor. C: A peMc transverse CT scan shows a left @lobatemass (arrow) wfthout obvious abnOrmalityon the nght alde. D: The surgical view confirms the photoscan finding, showing two tumors (arrows).

1812 Chatal,Fumoleau,Saccavinietal TheJournalof Nudear Medicine @ 500 Iu l31-OC125F(th@] I_P@I125-NSF(ab')2@j 400

300

% llYgm.10-5

0 11ij.F 111 A TUMOR BLOOD UVER MIJS@LE FAT SKIN 500 IRi131-CC125F(th@] 0 i 125-NSF(ab')2 400

300

% ltYgm.l0-5 200 FIGURE 3 Distribution of 1311-OC125 F(ab')@ 100 and a nonspecific immunoglobulin [@l-NS F(ab')@Jin two patients, 1 day (A)and 4 days (B)after injection. 0 Resufts are expressed in mean % of B TUMOR BLOOD UVER MUSCLE FAT SKIN injected activity per gram of tumor.

F(ab')2, eight tumor sites were not visualized (false involved. For the other patient (No. 22), a small pelvic negative results), including three abdominal recurrences focus had been confirmed by CT as a small retrovesical in the same patient (No. 16). In two cases, the recur mass. The patient's course may determine whether this rence was diffuse, taking the form of peritoneal seed case was a false positive or a recurrence signaled biolog lings (No. 5a and 8). In one case of primary fallopian ically but not apparent at the time of surgery. tube carcinoma, the tumor was small, 2 x 1 cm (No. In five patients injected with MAb ‘@‘I-l9-9F(ab')2, 7a).In another case,a 3 x 2.5cm pelvicrecurrencewas three tumor sites were correctly identified. A single close up against the urinary bladder. The patient had tumor site was not visualized that corresponded to a 3- only undergone planar scintigraphy at Day 4 (No. 6). cm liver metastasis (No. 26). Finally, for one patient Finally, in one case, recurrence was suggested by an (No. 29), two small, poorly contrasted pelvic foci that elevated serum CA 125 concentration of 670 U/ml corresponded to the interpretative criteria had been (No. 5b). Negative IS was not followed up by US or considered positive but were not confirmed surgically. CT. Ten months later a pelvic recurrence was palpable. Table 3 summarizes the overall diagnostic perform It is quite likely that this recurrence existed at the time ances ofIS. The results were considered to be a function of the immunoscintigraphic examination but was not of the number of anatomic sites tested and not of the ViSUalized even though it was already expressed biolog number of patients. To measure specificity, two ana ically. tomic sites were considered for each patient: the pelvis In two cases immunoscintigraphy was considered and the upper abdomen, including the liver and spleen. positive, whereas the surgeon found no apparent tumor When results are limited to cases where diagnosis was recurrence (false-positive results). For one of these pa confirmed by surgery, sensitivity was 76% (13/17) and tients (No. 17) there had in fact been an error in specificity 79% (15/19). interpretation as all the ECT criteria defined were not Table 4 shows the combined results of IS, US, and

Volume28 •Number 12 •December1987 I 813 CT. The diagnostic sensitivity of US was 47% (14/30). amination in the decision to perform third-look surgery. For CT it was 63%( 15/24). The 13 tumor sites visual Total resection of the recurrence led to a return to ized by IS and not by US corresponded to eight pelvic normal CA 125 values that had persisted until this and five abdominal localizations. In two cases, US writing. interpretation was hindered, once by the impossibility Finally, concerning the three patients explored suc ofkeeping the patient's bladder full during examination cessively by ‘@I-OC125 F(ab')2 and ‘‘‘In-DTPA-OC and another time by the patient's obesity (250 lb). The 125 F(ab')2, the recurrence in two cases was visualized @ eight tumor sites visualized by IS and not by CF were with ‘@‘Iand not with ‘‘In.One case concerned the pelvic (four sites) and abdominal (four sites) in origin. patient who had been treated with a high dose of 131!.. Figure 5 illustrates the case of Patient 2 with a OC-125 F(ab')2 after the first IS. In the other case, the moderate elevation of serum CA 125 concentration, recurrence was a metastasis of the lower part of the @ suggesting a recurrence which was not visualized by US right hepatic lobe. Using ‘‘In,excessive liver activity and CT. Positive pelvic ED' was the determining ex did not allow the metastasis to be clearly visualized, in

1814 Chatal,Fumoleau,Saccaviniet al The Journalof Nudear Medicine 3Sensitivity TABLE implies that the percentage of necrosis may have been andSpedficityTestedAnatomicBasedon Numberof high and that antibody accessibility to the tumor target SitesNo was limited by lower blood flow. oftestedana The major interest of the second, prospective stage of the study was to test the value of IS when it can be truly beneficial for the patient by confirming and local Clinicaltomicgroup SpecificityIsites TP FNt TN' Fl―Sensitivity izing a recurrence signaled only by an elevation in serum CA 125 or CA 19-9 concentration. This rise 67%(12/18)II18 12 6 / / usually occurs early in the course of the disease, before (24/28)=60 23 9 24 4 72%(23/32)86% the appearance of clinical signs, at a stage when further curative surgery is still possible. For the 29 patients in positive.t True negative.*FN = False this condition, the percentage of visualized recurrences negative.STN = True was 72% which was close to that of the retrospective FP = False positive. study. The analysis of false-negative results was similar to that ofthe retrospective study: small tumors, diffuse involvement and sometimes unexplained circum the absence of subtraction, when interpretation was stances in which the recurrence expressed CA 125 an made prospectively. Figure 6 illustrates the third case tigen, seemed well vascularized and yet was not visual in which an abdominal node recurrence was visualized ized. The specificity (86%) was satisfactory and could with both radionucides. Although the scintigraphic have been improved if an error in interpretation as a contrast to the tumor compared with abdominal back result of inexperience had been avoided (No. 17), and ground is better with ‘‘‘In,liver radioactivity is quite if the pelvic focus confirmed by CT and not by surgery high, whereas it is low with ‘@‘I. truly corresponded to a recurrence and is demonstrated by the patient's future course (No. 22). DISCUSSION Of these 29 patients in the prospective study, the diagnosisof recurrencewashistologicallyconfirmedin The initial retrospective part of this study showed 14 cases. If results are limited solely to surgically con that it was possible to visualize -‘@-70%of primary or firmed cases, comprising the gold standard, sensitivity secondary ovarian tumors. For patients who were not is comparable (76% versus 72%) and specificity slightly examined surgically, the diagnosis of recurrence was lower (79% versus 86%). considered highly probable on the basis of an associa In this study, the choice of antibody injected was tion of a significant elevation of serum CA 125 level determined by the histologic type ofthe primary tumor and a palpable and/or clearly visualized US and CT and by the nature of the antigen for which serum mass. Three of six false-negative results could be attrib concentration increased. Mainguenéet al. (8) have uted to small tumor size (a liver metastasis <1 cm in shown in an immunohistochemical study that 90% (27/ diameter and an abdominal tumor 2.5 cm in diameter) 30) of serous ovarian carcinomas, but no mucous car and in one of these cases to diffuse peritoneal involve cinomas, expressed CA 125 antigen. On the contrary, ment, a condition that in the absence of focalization 93% (13/14) of mucous carcinomas, and only 13% of sometimes makes it difficult to differentiate diffuse serous carcinomas, expressed CA 19-9 antigen. It is thus uptake of moderate intensity from simple abdominal clear that OC 125 and 19-9 antibodies are complemen vascular background. In the other three false-negative tary for ovarian carcinomas as a function of their his cases, the tumors were large—>'5 cm in diameter—that tologic type. The theoretic limiting factors for their immunoscintigraphic application is the shedding of an tigens into the circulation where they may form im TABLE4 munocomplexes with injected radiolabeled antibodies Combined Results of lmmunoscintigraphy, and thus limit the latter's accessibility to their tumor (Jtrasonography and Computed Tomography in Documented Tumor Sites target. This limitation is, however, not confirmed in practice since tumors have been visualized at very ele tomography vated concentrations of circulating antigens (10), whereas a limited percentage of tumors has been visu ImmunosdntigraphyUltrasonographyNegativePositivePositiveNegativeCOmputedPositive alized with antibodies such as 17-lA that recognize an 13 8 antigen not shed into the circulation (11). Negative 4 3 4 1 Moreover, for two patients, this study indicated that 9. TOTAL10 14 1611 15 uptake in a tumor tested with paired labeling is not related to simple trapping ofan immunoglobulin in the Resufts are to be read across. extravascular compartment ofthe tumor but a result of

Volume 28 •Number 12 •December1987 1815 EXTERNAL ARTERY

FIGURE 5 URETER Pelvicrecurrenceof an ovarianadenocarcinomain a pa tient with isolated elevation of serum tumor marker. A: 1311 OC 125 F(ab')@ECT image. A transverse section at the level of the pelvis shows an abnormal left lateral focus of radioactivity (arrow). B: A CT scan approximately at the same level as that of ECT falls to show any abnormal mass. C: An anatomic section, drawn by the surgeon from the surgicalview, shows the left-sided location of the TRANSVERSE SECTION resected recurrence, corresponding to the ECT finding of C Figure5A. specificity of the injected antibody for its antigenic liver uptake which hinders interpretation of the upper target. abdomen. However, tumor imaging ofthe pelvic region F(ab')2 fragments were preferred in the study since is markedly more contrasted than with ‘@‘I(Fig. 6). It their clearancein normal tissues is faster than that of is likely that ‘‘‘Inwill be the future radionuclide of intact antibodies, whereas clearance in tumors varies choice ifliver uptake can be significantly reduced. little with either form (12). The mode of radionuclide detection also deserves Other MABs have been used successfully to visualize discussion. Planar scintigraphy, with moderate doses of ovarian carcinomas, including antibodies directed ‘@‘Ior ‘‘‘In1to 2 mCi (37 to 74 MBq), allows recording against placental alkaline phosphatase (13) and an an of late images 5 to 7 days after injection, the period of tigen of human milk fat globule membranes (14—15). the best tumor-to-background contrast. However, it is It is not easy to compare the diagnostic results obtained difficult to interpret foci located close to normally ra with these antibodies since the selection of tested pa dioactiveorganssuch as the urinary bladder. ECT has tients varied greatly from one study to another. One the advantage of increasing contrast and providing future possibility would be to inject antibody “cock three-dimensional localization of a focus (17), in which tails―to take advantage of complementary specificities case it is easier to differentiate a pathologic focus from (16). adjacent vesical radioactivity. However, to satisfy the The choice of radionuclide is important and partly count statistic, a greater dose must be injected 3 to 3.5 conditions IS results. Iodine-l31, used in this study, has mCi (1 11to 129 MBq) than in planar scintigraphy, and radiophysical characteristics that are not very favorable prolonged acquisition times (40 mm of rotation in our to scintigraphic detection and dosimetry but that allow study) must be used. Moreover, interpretation of the exploration of the upper abdomen owing to lowliver different reconstructed sections requires experience and uptake. The radionuclide providing the best compro previously defined criteria. To benefit from the respec mise in associating energy, half-life, and availability is tive advantages of the two detection modes, our pro ‘I ‘In. Currently, its major drawback is a high level of spective study is being followed up by another associ

1816 Chatal,Fumoleau,Saccavinietal The Journalof NuclearMedicine A

FIGURE 6 Abdominal lymph node recurrence of an ovarian adenocarcinoma. A: [1―InJDTPA-OC125-F(ab')@ECT image. On this coronalsectionthe lumbarlymphnoderecurrenceis dearly visualized(arrow)belowandinwardfrom an intenseliver activity. B: 131l@ 125-F(ab')@ECT image recorded 8 mo before the image of Figure 6A. The lumbar lymph node recurrence is well defined (arrow) but the contrast compared to abdominal back ground is less than with “‘In.However, liver radioactivity is low in comparison with that of ‘@ln. ating both modes, in which an ECT examination is morphologic alterations, IS provides in addition infor performed 3 and 4 days after injection and is followed mation about relative oncologic specificity. Numerous at Day 7 by planar scintigraphy involving a global studies have reported the results of US in detecting posterior abdominal view with 30 mm acquisition time. primary or residual ovarian tumors after treatment (18- For four patients (Nos. 19, 23, 24, 28) in this study 21). Althoughthe criteriaof patientselection,which tested according to this protocol, and whose diagnosis vary from one study to another, make comparison of of recurrence was determined, planar scintigraphy findings difficult, simple addition of results gives an showed three recurrence sites (two pelvic, one abdomi overall 58% sensitivity (140/242) and 95% specificity nal) not clearly visualized on the ECT sections (Nos. (265/278). The most frequent causes of false negatives, 19, 24, 28). However, in one case (No. 24), ECT visu other than methodologic problems involving intestinal alized two abdominal recurrence foci which did not gas and a bladder that is not full, concern small nodules show up clearly in planar scintigraphy. < 1 cm in diameter, peritoneal studding and matted Finally, the clinical value of IS needs to be defined intestinal loops that are difficult to distinguish from relative to the other conventional diagnostic methods. recurrences. In comparison with these, which detect nonspecific Studies performed with CT are just as numerous and

Volume28 •Number12 •December1987 1817 offer the same variety with respect to the range of cases Drs. Alain Bécamand Odile Godin for their useful collabo (22-26). When exercisingthe samecautionin inter ration. preting the overall results ofthese different studies, 80% sensitivity (141/176) and 95% specificity (338/355) are REFERENCES found. The causes of false negatives include peritoneal studding, isodense masses or masses adhering to colonic 1. Cutler SJ, Myers MH, Green SB. Trends in survival rates of patients with cancer. N Engi J Med 1975; structures, masses against the pelvic side wall or para 293:122—124. aortic nodes. As with US, owing to a lack of oncologic 2. Einhorn N, Nilsson B, SjävallK. Factors influencing specificity, fibrotic lesions may cause false positives. survival in carcinoma of the ovary. Cancer 1985; There are fewer studies with magnetic resonance 20 19—2025. imaging (MRI) that have been introduced more re 3. Schwartz PE, Smith JP. Second-look operations in ovarian cancer. Am J Obstet Gynecol 1980; 138:1124— cently. The combined results of three studies (27-29) 1130. give 85% sensitivity (23/27). As with US and CT, MRI 4. GershensonDM, CopelandLI, Wharton JT, et al. lacks oncologic specificity and cannot distinguish in Prognosis of surgically determined complete re flammation or radiation changes from recurrent tumor sponders in advanced ovarian cancer. Cancer 1985; (27). 55:1129—1135. 5. Bast RC, Feeney M, Lazarus H, Ct al. Reactivity of a In order to define diagnostic strategy with the four monoclonal antibody with human ovarian carcinoma. diagnostic methods (IS, US, CT, and MRI) in the event iC/inInvest1981;68:1331—1337. of biologically suspected recurrence, it is advisable to 6. Bast RC, KIug TL, St John E, et al. A radioimmuno think in terms of the complementarity and not the assay using monoclonal antibody to monitor the competitiveness of these methods, as illustrated by the course of epithelial ovarian cancer. N Engi J Med 1983; 309:883—887. results of this study (Table 4). These results should 7. Ricolleau G, Chatal iF, Furnoleau P, et al. Radio admittedly be interpreted with due caution in compar immunoassay of the CA 125 antigen in ovarian car ing IS with US and CT since all patients did not undergo cinomas: advantages compared with CA 19-9 and all three examinations. It is thus not possible to provide CEA. TumourBiol 1984;5:151—159. statistical comparison of the respective diagnostic sen 8. MainguenéC, Aillet G, Kremer M, et al. Immunohis tochemical study ofovarian tumors using the OC 125 sitivities. However, for patients who underwent all three monoclonal antibody as a basis for potential in vivo examinations, in five cases (Nos. 2, 19, 23, 24, 25), IS and in vitro applications. J Nuci Med All Sd 1986; was the only positive examination in patients who had 30:19—22. negative US and CT. It is also appropriate to add four 9. Hamilton TC, Young RC, Lowe KG, et al. Charac ofthe seven cases ofundetermined diagnosis which had terization of a xenograft model of human ovarian carcinoma which produces ascitesand intraabdominal positive IS and negative US and CT (Nos. 7b, 13b, 20). carcinomatosis in mice. Cancer Res 1984; 44:5286— The short- and middle-term course of these patients 5290. will determine ifthese were also confirmed recurrences. 10. Rhodes BA, Burke DJ, Breslow K, et al. Effects of Identical situations in which IS was determinant for circulating antigen on antibody localization in vivo. localization ofa recurrence have also been reported for In: Burchiel SW, Rhodes BA, ed. Radioimmunoassay and radioimmunotherapy. New York: Elsevier Sci colorectal carcinomas (16,30). ence Publishing Co., 1983: 26—39. In fact, when an elevation in marker serum concen 11. Mach JP, Chatal JF, Lumbroso JD, et al. Tumor tration suggests a recurrence, it would be desirable to localization in patients by radiolabeled monoclonal employ the three methods (IS, US, and CT), thus corn antibodies against colon carcinoma. CancerRes 1983; pensating for the limitations of each and providing the 43:5593—5600. 12. Wahi RL, Parker CW, Philpott GW. Improved ra surgeon with the most precise evidence relative to the dioimaging and tumor localization with monoclonal different anatomic sites to guide his decision on second F(ab'>@.JNuclMed 1983;24:316—325. look surgery. 13. Epenetos AA, Carr D, Johnson PM, et al. Antibody guided radiolocalization of tumours in patients with testicular or ovarian cancer-using two radioiodinated NOTES monoclonal antibodies to placental alkaline phospha •Producedat the Dana FarberCancerInstitute,Boston(R. tase.BrfRadiol 1986;59:117—125. Bast, MD); Compagnie ORIS Industrie, 91 190Gifsur Yvette, 14. Pateisky N, Philipp K, Skodler WD, et a!. Radioim France. munodetection in patients with suspected ovarian can cer.JNuclMed 1985;26:1369—1376. t National Cancer Institute, Bethesda, MD (R.F. Ozols). 15. Granowska M, Shepherd J, Britton KE, et al. Ovarian *Gammatone 2-CGR-Sopha Medical, Les Ulis, France. cancer diagnosis using 123!monoclonal antibody in *(Searle Siemens) Cine 200-Intertechnique, Plaisir, France. comparison with surgical findings. Nuci Med Commun 1984; 5:485—499. ACKNOWLEDGMENTS 16. Chatal JF, Douillard JY, Saccavini JC, et a!. Clinical prospectivestudy with radioiodinated monoclonal an The authors thank Marie-Antoinettede Cusséfor her val tibodies directed against colorectalcancer. In: Baldwin uable secretarialassistancein preparingthis manuscript and RW, ByersVS. ecis.Monoclonal antibodies for cancer

1818 Chatal,Fumoleau,Saccavinietal TheJournalof NuclearMedicine detection and therapy. London: Academic Press, 1985: vanced ovarian cancer. Obstet Gyneacol 1983; 62:630— 159—180. 634. 17. Chatal JF. Detection des tumeurs par anticorps mon 24. JohnsonRI, BlackledgeG, EddlestonB,etal.Abdom oclonaux radiomarquésen tomoscintigraphie. Presse mo-pelvic computed tomography in the management Med 1983; 12:2361—2363. ofovarian carcinoma. 1983; 146:447—452. 18. Bowie JD. Ultrasound of gynecologic pelvic masses: 25. Mamtorca H, Isherwood I. Computed tomography in the indefinite uterus and other patterns associated with ovarian carcinoma: patterns ofdisease and limitations. diagnostic error. J Clin Ultrasound 1977; 5:323—328. ClinRadiol1982;33:165—171. 19. Khan 0, Cosgrove DO, Wiltshaw E, et al. Role of 26. Whitley N, BrennerD, FrancisA, et al. Use of the ultrasound in the management of ovarian carcinoma. computed tomographic whole body scanner to stage JR SocMed 1983; 76:821—827. and follow patients with advanced ovarian carcinoma. 20. PussellSi, Cosgrove DO, Hinton J, et al. Carcinoma InvestRadiol1981;16:479—486. of the ovary. Correlation of ultrasound with second 27. BiesJR, Ellis JH, KopeckyKK, et al. Assessmentof look laparotomy. Br J Obstet Gynaecol 1980; primary gynecologic malignancies: comparison of 87:1140—1144. O.15-T resistive MRI with CT. Am JRoentgenol 1984; 21. Wicks JD, Mettler FA, Hilgers 1W, Ct al. Correlation 143:1249—1257. of ultrasound and pathologic findings in patients with 28. Butler H, Bryan PJ, Lipuma JP, et al. Magnetic reso epitheial carcinoma of the ovary. J Clin Ultrasound nance imaging of the abnormal female pelvis. Am J 1984;12:397—402. Roenigenol1984;143:1259—1266. 22. Brenner DE, Shall MI, Jones HW, et al. Abdomino 29. Johnson IR, Symonds EM, Worthington BS, et al. pelvic computed tomography: evaluation in patients Imaging ovarian tumours by nuclear magnetic reso undergoing second-look laparotomy for ovarian car nance. Br J Obstet Gynaecol 1984;91:260—264. cinoma. Obstet Gyneacol 1985;65:715—719. 30. Begent RHJ, Keep PA, Searle F, et al. Radioimmu 23. Goldhirsch A, Triller JK, Greiner R, et al. Computed nolocalization and selection for surgery in recurrent tomography prior to second-look operation in ad . Br J Surg 1986; 73:64—67.

Volume28 •Number 12 •December1987 1819