The aim of the GHSG study was to adapt treatment in supradiaphragmatic clinical stage I-II patients according to their risk of occult infradiaphragmatic involvement. A retrospective multivariate analysis (logistic regression) was performed on pretherapeutic clinical characteristics of 391 laparotomized patients.13 Twenty one percent had subdiaphragmatic disease. Of the factors (clinical including nodal presentation and biologic) tested, four independently correlated with infradiaphrag-matic disease: left cervical involvement, mediastinal involvement, Karnofsky performance status less than 10, and histologic subtype (mixed cellularity or lym-phocytic depletion). The regression coefficients were then used to derive a quantitative estimate of the probability of infradiaphragmatic disease for individual patients. It ended in defining two groups of patients at low or high risk. The low-risk group was composed of patients with no left cervical involvement, mediastinal involvement, and Karnofsky index equal to 10 (19% of patients), or with no left cervical involvement, medi-astinal involvement, and lymphocyte predominant or nodular sclerosing histologic subtype (19% of patients), or with mediastinal involvement, Karnofsky index equal to 10 and lymphocyte predominant or nodular sclerosing histologic subtype (19% of patients). The high-risk group comprised of patients with left cervical involvement and mediastinal involvement (25% of patients), or with no mediastinal involvement and mixed cellularity or lymphocytic depletion histologic subtype (29% of patients). All other patients were considered with intermediate risk. The probability of infra-diaphragmatic disease was 8% in patients at low risk and more than 35% in patients at high risk.
Was this article helpful?