Abstracts/Lung Cancer 15 (19%) 139-157 149

negative based on the presence or absence of increased FDG uptake in the lung and/or in the mediastinum. All 82 patients with lung cancer had increased FDG uptake in the lungs, whereas only 12 of 25 patients with nonmalignant diseases had increased FDG uptake. Sixteen lung cancer patients with mediastinal metast&es had increased FDG uptake in the mediastinum, of whom three had no lympbadenopathy on computed tomographyofthechest. Sixteenlungcancerpatientswithoutmediastinal nodal involvement had no FDG uptake in the mediastinum. Seven of these patients had lymphadenopathy on computed tomography. FDG- PET imaging is 100% accurate in predicting mediastinal involvement in patients with lung cancer. It is 100% sensitive and 52% specific in predicting the malignant nature of a chest radiographic abnormality.

Bronchioloalveolar carcinoma: domputed tomography findings Trigaux J-P, Genevois PA, Goncette L, Gouat F, Schumaker A, Weynants P. Department of Radiology, Hopital de Mont-Godinne. B- 5530 Yvoir. Eur Respir J 1996;9: 1 l-6.

The aim of the present study was to investigate the appearance of bronchioloalveolar lung carcinoma on computed tomography (CT) scans, and to determine the frequency of signs suggestive of this diagnosis. CT features of 42 cases with pathologically proven bronchioloalveolarcarcinomawereretrospectivelyanalysedforpattem, size, location and secondary signs suggesting the diagnosis. Bronchioloalvwlar carcinoma had one of the following patterns: solitary nodule or mass (16), lobar consolidation (IO), multilobar consolidations (13) and diffuse nodules (3). The 16 solitary nodules or masses ranged in size from 2.0 to 9.4 cm (mean f SD 3.75 f 1.7 cm). Eleven of the 16 nodules or masses were peripheral and five were central: Eight of the 16 tumours had pleural tags, seven had spiculated margins, and three had bubblelike lucencies. The consolidations were peripherally distributed in 13 out of 23 casea; cystic airspaces were observed in 19 out of 23 consolidations; bulging of interlobar fissures in 8 out of 23; and the angiogram sign in 7 out of 23. In conclusion, computed tomography findings of bronchioloalveolar carcinoma have a widespectrum, showing typically a peripheral nodule or consolidation. Computed tomography has a role in the diagnosis of nodular localized versus other forms, with subsequent therapeutic and prognostic implications.

Hilar lymphnode metastasis -of bronchogenic carcinoma; evaluation with ultrafast incremental dynamic CT (IDCT) Shimoyama K, Murata K, Takahashi M, Nitta T, Mishina A, Matsuo Y et al. Jpn J Clin Radio1 1996;41:41-51.

Eighty patients with bronchogenic carcinoma were studied with ultrafast IDCT using a 3mm collimation. Nineteen contiguous images encompassing pulmonary hilum were obtained in a single breath-hold with all injection of contrast material at the rateof 3ml/sec. Subsequent to the preliminary analysis of the peribronchovascular interstitium in 22 patients with normal hilum, the presence or absence of hilar lymphnode m&stases was judged in 151 hilar lymphnodes of 80 patients on the basis of convex interface of the interstitium into adjacent lung. The interfacesbetweenthenormal interstitiumandadjant lung demonstrated concave or linear appearances in 97.2%. Sensitivity, specificity, and accuracyofultrafast IDCTin thediagnosisofhilar lymphnodemetastasis were 91.4%, 87.996, and 88.7%, respectively.

Nl of lung cancer evaluated by thin-section dynamic CT Tsushima J, Tokuno E, Otsuji H, Uchida H, Maeda M, Yoshiya K et al. Jpn J Clin Radio1 1996;41:53-7.

We evaluated N 1 for correct staging of lung cancer by CT. Short and

long axis of hilar lymph nodes (HLN) of 40 lung cancer were measured on thin- section dynamic CT. All HLN wereproved by histopathological study after R2 resection. 114 HLN were detected by CT in 38 cases. True positive of HLN was 111, and false positive was three. There was statistically significant difference in size between metastic HLN (short axis 10.8mm (6.2 and long axis 14.6m.m + 6.6) and non-metastatic HLN (shortaxis6.4mm f 2.8 and longaxis 8.8mm f 3.8). ROC curve analysis for HLN showed that the optimal size criteria was 8mm in the short axis (specificity 7396, sensitivity 73%. and over all accuracy 73 96).

Nuclear ~53 overexpression is an independent prognostic parameter in node-negative non-small cell lung carcinoma Dalquen P, Sauter G, Torhorst J, Schultheiss E, Jordan P, Lehmann S. Institute of Pathology, University of Basel, Schonbeinstr~se 40, 4003 Basel. I Path01 1996; 178:53-8.

The prognosis of operated patients with non-small cell lung cancer (NSCLC) is poor despite thorough pre-operative staging. An improved preselection is needed of patients likely to profit from surgery. This study was undertaken to evaluate the prognostic significance of nuclear ~53 overexpression in a cohort of 247 surgically treated patients with NSCLC. It showed that the prevalence of immunohistochemically detectable ~53 overexpression varied between different tumour types. ~53 overexpression was equally frequent in large cell carcinoma (53 per cent) and in squamous cell carcinoma(54per cent), but significantly less frequent inadenocarcinoma(34percent; P = 0.009). ~53 overexpression was particularly rare in bronchioloalveolarcarcinoma (positlvity in 1 of 17 cases). These variations may reflect aetiological differences between the histological subtypes. ~53 overexpression was also associated with high tumour grade (P = 0.0157) and the presence of lymph node metastasis (P = 0.0259), but not with advanced tumourstage. Survival analysis showedno difference in clinical outcome between p53-positive andp53-negativetumourswithin 101 node-positivetumours. Incontrast, survival time was significantly better in p.53-negative tumours than in p53-positive tumours within thegroupof 113 node-negative tumours (P = 0.032). Stepwise regressionanalysis showed that ~53 overexpression is an independent prognostic factor in node-negative NSCLC.

Targeting of small-cell lung cancer using the anti-CD2 ganglioside monoclonal antibody 3F8: A pilot trial Grant SC, Kostakoglu L, Kris MG, Yeh SDJ, Larson SM, Finn RD et al. lhoracic Oncology Service. DivSolid Tumor Oncol, Depr Medicine, Memorial Sloan-Kettering Cancer Cent. 1275 York Avenue. New York, NY 10021. Eur J Nucl Med 1996;23: 145-9.

The present study evaluated the ability of the anti-GD2 ganglioside monoclonal antibody 3F8 to target tumor sites in patients with small-cell lung cancer (SCLC). Of 12 patients entered into the trial, ten received intravenous 3F8 labeled with 2 or 10 mCi iodine-131. The first five patients had recurrent or progressive disease after chemotherapy. Subsequent patients were studied before starting chemotherapy. Radionuclide scans were performed on days 1, 2, and 3 post-infusion and once between day 5 and day 7, Four patients underwent single- photcn emission tomography (SPET) imaging. Radionuclide scans demonstrated localization to all known sites of disease, other than small brain metastases in one patient, SPET/CT scan fusion images confirmed precise localization. No signiiicant toxicity was observed. Mean serum half-life was 64.2 h. Analysis of specimens from one patient who died of unrelated causes 6 days post-infusion continned the scan results. The present study demonstrates that 3F8 targets SCLC sites in patients. Further studies of anti-GD2 antibodies with higher doses of antibody and radionuclide are warranted to evaluate their role in SCLC.