274 Abstracts/Lung Cancer 10 (1993) 266-286

Medicine, National Cancer Center Hospital, Tsukiji 5-1-J. Chuo-ku. Tolryo 104. Cancer lV93;71:93-8.

Background sod Methods. Between 1978 and 1990, 51 cases of per&did effusion secoodary to luog cancer were treated at the National Cancer Center Hospital by creating P pericardial window, using the subxipboid approach, that was connected to P water-sealed drainage system. Results. Most patients had advanced disease, such as distaot m&s&is (76 %), pleural effusioo (88 96). and clinical Stage N2 orN3 disase(981). Forty-fivepatieotshadcardiac tampon&, and six hadoosympto~attributabletopericPrdinl effusion. Cardiactamponade was the initial maoifestation of lung caocer in only 3 patients; it was a late manifestation in 48. Of those specimens that were examined cytologically, 92% had positive tindings. The interval fromcreation of the pericardisl window until removal of the drainage tube ranged from 4-135 days (median. 11 days). The interval was significantly longer in patients who previously had received thoracic radiation therapy (P < 0.05). The overall median survival was 80 days, and the l-year survival rate was 10.5%. Postmortem examination showed that constrictive heart failure caused by pericardial lesions was the major contributory cause of death in 32% of patients. Using multivariate analysis, factors indicating P poor prognosis were: (1) the interval from the diagnosis of lung cancer to pericardial effusion development (P = 0.005) sod (2) the absence of prior surgery (P = 0.007). Conclusions. The creation of per&dial window effectively treated pericardial efisioo io 85 I of cases. However, themleof ietrapericsrdial instillationofaotisanceror sclerosing agents was unclear in this retrospective analysis.

Seamfragmentofcytokeratinsuhunit 19meawredbyCYFRAZl- 1 immunoradiometric assay as P marker of lung cancer Pujol I-L. Greuier I, Daures J-P, Daver A. Pujol H. Michel F-B. Cliniqwdes MaladiesRespiwoires. Vniversire&Monfpellier. Hopiral Aiguelongue, 34059 Montpellier Ceder. Cancer Res 1993;53:61-6.

Cytokeratio 19 is a subunit of cytokemtin intermediate filament expressedinsimpleepilhelisandtheirmplign.Therefore, it is expressed by respiratory epithelium cells and has beea detected in lung czmcer specimens. An immuoomdiometric assay was used to detect a fragment of the cytokeratie 19, referred to as CYFRA 21-1, in the serum of 165 patients with histologically proved lung cancer (128 noo- small cell and 37 small cell lung cancers). This prospective shldy was conducted to evaluate the reliability of this immuooradiometric assay and to identify the relationship between serum CYFRA 21-1 and different features of lung cancer including prognosis. The minimal detectable coocentration detected by this a.wy was 0.06 ng/ml. The reliability of the immunomdiometric assay was demonstrated by the linear relationship between CYFRA 21-1 measurement and dilution of theserum, thereproducibility ofthedosagein intmassay and interassay, and the high sensitivity ofthe method in discriminating low CYFRA 21- 1 concentrations. Using a threshold of 3.6 @ml, sensitivity and specificity were 0.52 and 0.87, respectively. Tbe sensitivity of the marker was highest in squamous cell carcinoma and lowest in small cell carcinoma. In non-small cell lung cancer patients, the marker varied significantly according to both stage of the disease (Kmskal-Wallis, 13.7; P < 0.005) and performance status (Kmskal-Wallis. 9.16; P < 0.05) inasmuch as a high s+rum CYFRA 21-1 level was associated with advanced stages, media&al lymph nodes, and poor performance status. Consequently, the marker was significantly lower in patients who were operated upon when compared with unresectable ooes. Lung cancerpatients withserumCYFRA 21-l over3.6nglmlproved to have a significantly shorter overall survival than those with a normal serum level (log mok, P = 0.007; Wilcoxon, P = 0.001). The negative prognostic effect of CYFRA 21-l was highly significant in squamous cell carcinomas whereas it was nonsignificant for the other histologies. In Cox’s model analysis, performance status, stage grouping, and CYFRA 21-1 were the only significant determinants of survival. This study supports the use of the serum fragment of cytokeratin subunit 19 CYFRA 21-1 as an independent prognostic marker of squamous cell carcinoma of the lung.

Quality of life assgunent in individuals with lung cancer: Testing the lung cancer symptom scale (LCSS) Hollen PJ. Gralla RI, Kris MG, Potaoovich LM. VrdwrsityofRockwr, 601 Elmwood Avenue, Rochester. NY 14642. Eur J Cancer Pat A Geo Top 1992;29:Suppl l:S51-8.

Thispsper~tsthecootinueddevelopmentnndmulti-institutioonl testing of PO instrument focusing oo measuring the physical sod fuoctional dimensions of quality of life. It empbasises evaluation of symptomsassociated with lung caoceraod their effect oo activity status. The Lung Cancer Symptom Scale (LCSS) is a disease- sod site-specific instrument which has both P patient and ao observer (health care pmfessional) form. Tbe patient scale required 8 min to administer and the observer scale 2 min. The readability index was second-grade level for the patient scale and ninth-grade level for the observer scale. Content validity revealed P mean of %% agreement for all major symptoms among 52experts surveyed (confidenceinterval = &i-99%, P = 0.05). 69 patients with non-small cell and 52 patients with small cell lung c~cerconfirmedthattbesympcomsmatchedtheirexperiences. lnten-ater reliability showed consistency for all s but one among 21 raters at eight iostitutions; that one was consistent for 20 of tbe 21 raters. Similar results were found oo B 9-month interval replication. Using the Kappa stntisticto~imnteextentofngreementforrepeptedinte~aterreliability, almost perfect agreement was obtained (mean coefficients. 0.95-0.98). Using the same rule of agreement as for Kappa (* one category) intmrater agreement was 95-10046 for all 21 raters. Past test re-test reliability indicated high patient reproducibility for 52 patieots (I > 0.75. P < 0.01 for all 6). We conclude that (1) the LCSS demonstrates good feasibility, reliability, and content validity, (2) high interrater reliabilityindicp~utilityinmultic~tretrjpls, and (3)continuedtesting for internal consistency. construct validity nod criterion-related validity is warranted.

Rognostic value of neuron-sperific enolpse in small cell lung cancer patients Sztunnowicz M, Roginska E, Roszkowski K. Kwiek S, Filipecki S, Rowinska-tiska E. Deportment of Internal Medicine, Inst. Tuberctdosis/PulmonaryDis., Plocka26.01-138 Warsaw. Lung Caecer (Irelaod) 1992;8:259-64.

Prognostic value of pretreatment serum neuron-specific enolase levels (s-NSE) and their changes during cytotoxic treatment was evaluated in 92 small cell lung cancer patients. Median survival time of the patients with pretreatment s-NSE exceeding 50 g/l was 7.3 months. significantly shorter @ = 0.05) thao survival of other patients. The negative influence of s-NSE exceeding 50 g/l on survival was also significant when only patients with metestases in two or more organs were taken into account. During treatmeot s-NSE showed oormalisation in both complete and partial responders, but also occasionally in nonresponders. Median survival time of the patients io whom s-NSE failed to normalize wps 7 months and it was significantly shorter than median survival time of the ptieots io whom s-NSE nonoalised (13.3 months) or was within normal limits all the time (15.3 months).

Epidermoid carcinoma of the lung in stage I: Factors of prognostic interest Carbajo M, Ortega FJ, Hemaodez M, Oodiviela R, Blaoco C, Buelta L et al. Departametuo de Anatomia Patologica. Hospital Univ. Marques de Vakkxilla, A&a. Vadecillas/n, 39008SanlMder. Histol. Histopathol. 1993;8:41-5

Factors affecting state I epidermoid caocer of the lung were studied in a series of 29 patiems treated only by surgery and followed up for ten years. A set of 13 variables with a possible influence on prognosis were investigated. The application of the Cox Univariate Analysis to the different variables showed the grade of cell differentiation and the mitotic index to be predictors. In the Cox Multivariate Analysis, the proportional regression equation revealed two independently significant variables @ < 0.01). which were the mitotic Index and Nuclear Area. Grouping patients on the basis of the prognostic variables indicated allows a better prediction for survival to be made for this series of patients.

Application of an algorithm for staging small-cell lung cancer can save one third of the initial evaluation costs Richardson GE, Venzon DJ, Phelps R, Edison M, Brown M. Frame IN et al. NC/-Navy Medical Oncology Branch, National Naval Medical Center, Eldg 8, Bethesda, MD 20889-5105. Arch Intern Med 1993;153:329-37.

Objective: Design of a cost-effective algorithm for staging disease