Prognostic factors for subdiaphragmatic involvement in clinical stage I-II supradiaphragmatic Hodgkin's disease: a retrospective analysis of the GHSG.

Background: Staging laparotomy and splenectomy were routinely performed in patients with early-stage Hodgkin's disease (HD) qualifying for radiotherapy alone to determine the exact extent of disease. However, staging laparotomy is associated with a considerable number of side effects, warranting more sophisticated diagnostic procedures and new therapy strategies. We retrospectively analyzed patients undergoing staging laparotomy to identify pretherapy risk factors predicting the probability of abdominal disease and to define high-risk groups that might benefit from staging laparotomy and subsequent stage-adjusted treatment.

Patients And Methods: Between February 1988 and January 1993, 391 patients with CS I-II supradiaphragmatic Hodgkin's disease underwent staging laparotomy and splenectomy according to the treatment policy of the German Hodgkin's Lymphoma Study Group (GHSG) for early stages of Hodgkin's disease. Univariate and multivariate analysis of pretherapeutic clinical characteristics were performed in an attempt to predict staging laparotomy results and to identify risk groups.

Results: Of the 391 patients, 81 (21%) had subdiaphragmatic disease. Eighteen percent were upstaged to PS III and three percent to PS IV. By a multivariate model the following parameters were independent risk factors for positive surgical staging: left cervical involvement (P < 0.001), mediastinal involvement (P < 0.009), Karnofsky performance status (P < 0.004) and histology (P < 0.04). In our analysis gender (P < 0.08) and ESR (P < 0.06), often described as of high prognostic value, was not significant. The presence of systemic symptoms, number of involved areas and clinical stage were not associated with abdominal disease, as described in several former publications. To define high-risk groups, which comprise at least 15% of patients of the cohort and have a risk of subdiaphragmatic involvement of > 35%, combinations of only two or three of the predictive factors were analyzed. With respect to these criteria the following subgroups of patients were identified as having a high risk for subdiaphragmatic disease (> 35%): a) left cervical lymph node involvement and no mediastinal involvement (n = 98, observed risk 36%); b) no mediastinal involvement and MC/LD histology (n = 113, observed risk 40%).

Conclusions: We conclude that initial clinical characteristics are predictive for occult abdominal involvement in early clinical stages of Hodgkin's disease. The impact of these risk factors on future therapeutical strategies have to be evaluated.