Federico Bozzetti1, Giuliano Bonfanti1, Luigi Mariani2, Rosalba Miceli2, and Salvatore Andreola3. 1 Division of Surgery of the Gastrointestinal Tract, Istituto ...
Gastric Cancer (1999) 2: 74–75 74
F. Bozzetti et al.: Second malignancy after surgery for EGC 1999 by International and Japanese Gastric Cancer Associations
Short communication Risk of second malignancies after surgical treatment for early gastric cancer Federico Bozzetti1, Giuliano Bonfanti1, Luigi Mariani2, Rosalba Miceli2, and Salvatore Andreola3 1
Division of Surgery of the Gastrointestinal Tract, Istituto Nazionale per lo Studio e la Cura dei Tumori, Via Venezian 1, 20133 Milan, Italy Division of Medical Statistics and Biometry, Istituto Nazionale per lo Studio e la Cura dei Tumori, Milan, Italy 3 Division of Pathology, Istituto Nazionale per lo Studio e la Cura dei Tumori, Milan, Italy 2
Abstract: We evaluated the follow-up of 105 patients who received surgical treatment at our Institute for early gastric cancer (EGC). Median follow-up was 71 months. Operative mortality was 5.7% and the 5-year survival rate was 82.8%, approximately that of an age- and sex- matched Italian population. Overall mortality was assessed with regard to mortality from gastric cancer or other causes. Ten second primary malignancies were identified, and the standardized mortality ratio (1.12; 90% confidence interval [CI] 0.77–1.64) and standardized morbidity ratio (1.50; 90% CI 0.89–2.50) were determined using data from the Tumor Registry of Lombardy. To our knowledge, this is the biggest Western series to measure the risk EGC patients face of developing a second primary tumor. In conclusion, oncologists who follow-up patients surgically treated for EGC, a disease that is much less frequent in Western countries than in Eastern ones, should be aware of the possible occurrence of extragastric primary tumors so that they may plan the appropriate diagnostic investigation when necessary. Key words: early gastric cancer, multiple malignancies
Introduction The aim of this study was to survey, in patients surgically treated for early gastric cancer (EGC), the occurrence of second primary tumors, an occurrence which seems to be particularly frequent in the West but which has so far been investigated only in a very small series of patients [1,2] and without specific reference to National Tumor Registry standardized morbidity and mortality ratios.
Offprint requests to: F. Bozzetti Received for publication on March 16, 1998; accepted on Dec. 1, 1998 * This study was supported in part by a qrant of The Associazione Italiana Ricerca Contro it Cancro
Subjects and methods We evaluated 105 patients who were surgically treated for EGC at the Istituto Nazionale Tumori in Milan from 1973 to 1990. Median follow-up was 71 months. No patient was lost to follow-up. To describe survival in this case series, we determined the overall survival curve using the Kaplan-Meier method [3]. An expected survival curve based on the annual death rates in the Italian population according to sex and age (Italian Institute of Statistics; ISTAT) was also computed, using the program developed by Hakulinen and Abeywickrama [4]. Cumulative mortality from either gastric cancer or other causes was outlined according to the competing risk approach [5]. To have a summary measure for comparing the observed and expected overall survival rates, the standardized mortality ratio (SMR) was assessed, using an indirect method, and the corresponding 90% confidence interval was determined as described by Breslow and Day [6]. In our study, based on our prior hypothesis, our goal was to distinguish those tumors whose risk of occurrence was higher, rather than changed in general, below or above the population level. Consequently, we used 90% confidence intervals for standardized morbidity ratios, so that a lower limit greater than 1 corresponded to a significant one-sided test at the 95% significance level. The SMR was the number of deaths observed in the cohort divided by the expected number of deaths. The expected number of deaths was assessed by applying Italian population rates to the cohort sex and age structure. In order to study the occurrence of second primaries, we calculated standardized morbidity ratios and their confidence intervals using the same procedure as that mentioned above. We based the adopted population rates for all tumors (after subtracting the rates for gastric cancer), or specific for the malignancies observed in this cohort, on information collected by the Tumor
F. Bozzetti et al.: Second malignancy after surgery for EGC
Registry of Lombardy. For the years 1973–1982 and 1983–1990 the rates adopted were those published in Italian reports [7,8].
Results and discussion Of the 105 patients followed up to March 1993, 81 are still alive; 6 are dead from postoperative complications, 5 from recurrence of gastric cancer, and 13 from other causes. The 5-year survival rate was 82.8%. The standardized mortality ratio was 1.12 (90% confidence interval [CI], 0.77–1.64). Overall, there were ten second primary malignancies (two breast carcinomas, three lung carcinomas, two bladder carcinomas, one esophageal carcinoma, one malignant melanoma, one non-Hodgkin lymphoma). Three of these were synchronous (two lung cancers and one esophageal cancer), and the rest metachronous. The standardized morbidity ratio was 1.50 (90% CI, 0.89–2.50). Our values for the standardized global morbidity ratio, although not statistically significant, are in accordance with previous findings, and, despite the limited sample of patients investigated, represent, we believe, the largest analysis reported in Western countries. Nevertheless, we estimated [9] that four times more patients (matched with ours for age and sex) would have been required in order to achieve statistical significance for a 50% increase in morbidity with an 80% power.
75
In conclusion, our study revealed a higher incidence of second primaries than that found in the general population. One practical implication of this finding is the need for oncologists to recognize that these patients are at risk for a second primary so that they may plan appropriate diagnostic investigations when necessary.
References 1. Green PHR, O’Toole KM, Weinberg LM, Goldfarb JP. Early gastric cancer. Gastroenterology 1981;81:247–56. 2. Lehnert T, Sternberg SS, Sprossmann M, DeCosse JJ. Early gastric cancer. Am J Surg 1989;157:202–7. 3. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457–81. 4. Hakulinen T, Abeywickrama KH. A computer program package for relative survival analysis. Comput Methods Programs Biomed 1985;19:197–207. 5. Marubini E, Valsecchi MG. Analyzing survival data from clinical trials and observational studies. Chichester: John Wiley and Sons; 1995: chapter 10: Competing risks 331–61. 6. Breslow NE, Day NE. Statistical methods in cancer research, Volume 2. IARC Scientific Publications No. 82, Lyon: International Agency for Research on Cancer Design and analysis of cohort studies, 1987:67–71. 7. Incidenza dei tumori e cause di morte in Lombardia, Regione Lombardia, Settore Sanità e Igiene, Notizie Sanità, 10:1985. 8. La Lega Italiana Per La Lotta Contro i Tumori, Associazione Italiana di Epidemiologia. In: Zanetti R, Crosignani P, editors. Il cancro in Italia. I dati di incidenza dei Registri Tumori 1983–1987. Turin, 1992. 9. Cesana BM, Marubini E. Dimensione del campione nella ricerca biomedica. Milan: Collana monografie del Centro Zambon, Università degli Studi di Milano, 1987:48–55.
